Report No. 32711-SV El Salvador Investment Climate Assessment (In Two Volumes) Volume II: Detailed Findings and Recommendations July 12, 2005 Development Economics Research Group (DECRG) and Finance, Private Sector and Infrastructure Latin America and the Caribbean Region (LCSFR) Document of the World Bank GMP Good Manufacturing Practice GMP Good Manufacturing Practice GPD Gross Domestic Product HACCP HazardAnalysis Critical Control Point HACCP HazardAnalysis Critical Control Point IAAC Interamerican Accreditation Cooperation ICRG International Country Risk Guide ICs Investment Climate Survey ICT Informationand Communication Technologies IEC International Electrotechnical Commission ILAC International Laboratory Accreditation Cooperation IMF International Monetary Fund INCONTEC National StandardsBody of Colombia INSAFOW Government Training Institute I S 0 International Organization for Standardization IUDOP Instituto Universitario de Opini6n Publica MINEC Ministerio de Economia MME Mercado Mayorista de Energia Electrica MSMES Micro, Small andMediumEnterprises NAFTA North American Free Trade Agreement OECD Organization for Economic Cooperation and Development PAHO Pan-American HealthOrganization PPP Plan PueblaPanama R&D ResearchandDevelopment SGR Sociedadesde Garantias Reciprocas SIEPAC Central American Electric Interconnection System SIGET Superintendenceof Electricity and Telecommunications SME Small andMediumEnterprise TFP Total Factor Productivity UNDP UnitedNations Development Program USAID USAgency for InternationalDevelopment WBI World Bank Institute Vice President: Pamela Cox Country Director: Jane Armitage Sector Director: Makhtar Diop Sector Leader: Manuel Sevilla Sector Manager: Susan Goldmark Investment Climate Unit Manager: Axel Peuker ResearchManager: Luis Serven Task Manager: Pablo Fajnzylber TABLEOF CONTENTS: ACKNOWLEDGEMENTS ....................................................................................................... 13 CHAPTER 1: INTRODUCTION ........................................................................................... 14 BACKGROUND ............................................................................................................................ 14 WHAT I S THEINVESTMENT CLIMATEANDWHY I T MATTERS .................................................. 16 ANOVERVIEW OFELSALVADOR'S GENERALECONOMIC 18 CLIMATE SURVEY ................................................................................ TRENDS............................................ THE2004 INVESTMENT 26 MAININVESTMENT CLIMATE CONSTRAINTS: THEPERSPECTIVEOFTHEFIRMS......................... 27 IMPACT OF THE INVESTMENT CLIMATE: ECONOMETRIC ESTIMATES .......................................... 31 WITHIN-COUNTRYSIMULATIONS .............................................................................................. 34 INTERNATIONAL SIMULATIONS .................................................................................................. 37 THEPOTENTIAL BENEFITS THE FREETRADEAGREEMENT OF WITH THE US ............................. 42 ANNEX1.1:TABLES ................................................................................................................... 47 ANNEX 1.2: TECHNICALAPPENDIX ............................................................................................ 53 CHAPTER 2: GOVERNANCEAND INSECURITY .......................................................... 58 CORRUPTION.............................................................................................................................. 58 POLICY RECOMMENDATIONS ON CORRUPTION .......................................................................... 69 CRIME ........................................................................................................................................ 70 POLICY RECOMMENDATIONS ON CRIME .................................................................................... 84 CONTRACT ENFORCEMENT THE JUDICIARY AND ........................................................................ 85 POLICY RECOMMENDATIONS ON CONTRACT ENFORCEMENT THE JUDICIARY AND ..................103 BUSINESS-GOVERNMENT RELATIONS...................................................................................... 104 POLICY RECOMMENDATIONS ON BUSINESS-GOVERNMENT RELATIONS .................................. 123 ANNEX2.1: ECONOMETRIC ESTIMATES................................................................................... 126 CHAPTER 3: INFRASTRUCTURE ...................................................................................... 133 ELECTRICITY............................................................................................................................ 133 POLICY RECOMMENDATIONS ON ELECTRICITY ........................................................................ 146 TELECOMMUNICATIONS ........................................................................................................... 148 POLICY RECOMMENDATIONS ON TELECOMMUNICATIONS ....................................................... 157 TRANSPORT .............................................................................................................................. 159 TRANSPORT: POLICY RECOMMENDATIONS .............................................................................. 170 ANNEX3.1: ECONOMETRIC ESTIMATES ................................................................................... 172 CHAPTER4: ACCESS TO FINANCE .................................................................................. 178 THESTATE OFTHE SALVADORANFINANCIALSYSTEM............................................................ 179 ACCESS COST OF CREDIT ................................................................................................. AND 184 ACCESS CONSTRAINTS REPORTED BY FIRMS ........................................................................... 189 SOURCES OF FINANCE .............................................................................................................. 192 POLICY RECOMMENDATIONS ON FINANCE ............................................................................... 195 ANNEX4.1: TABLES ................................................................................................................. 199 ANNEX 4.2: ECONOMETRICESTIMATES ................................................................................. 200 CHAPTER 5: SKILLS, QUALITY AND TECHNOLOGY .............................................. 202 EDUCATION LABOR AND SKILLS .............................................................................................. 202 POLICY RECOMMENDATIONS ONEDUCATION AND LABOR SKILLS .......................................... 208 QUALITY .................................................................................................................................. 209 POLICY RECOMMENDATIONS ONQUALITY .............................................................................. 213 TECHNOLOGY .......................................................................................................................... 214 POLICY RECOMMENDATIONS ONTECHNOLOGY ....................................................................... 221 ANNEX5.1: ECONOMETRIC ESTIMATES ................................................................................... 224 ANNEX5.2: MAINCOMPONENTSOFA TYPICALNATIONAL QUALITY SYSTEM ........................ 228 REFERENCES .......................................................................................................................... 230 5 LIST OF BOXES: BOX 1: What i s an Investment Climate Assessment?.................................................................. 15 68 BOX 3: E-government inChile .................................................................................................. BOX 2: Korea's Move to E-Procurement..................................................................................... 156 BOX 4: The Great Transformation o fAmerican Secondary Education..................................... 203 LIST OF TABLES: 47 Table A1-2:InvestmentClimate Indicators for El Salvador. by Sector ....................................... Table A1.1: Investment Climate Indicators for El Salvador. by Firm Size.................................. 48 Table A1.3: InvestmentClimate (IC) Indicators for El Salvador, by Province............................ 49 50 Table A1.5: IC indicators for Large Firms(four countries) ......................................................... Table A1.4: I C indicators inEl Salvador, Guatemala, Honduras and Nicaragua......................... 51 Table A1.6: I C indicators for MSMEs(four countries)................................................................ 52 Table Al.7: The Impact o fthe Investment Climate on Salvadoran Plants................................... 55 Table A1-8: The Impact of the IC on Salvadoran and Central American Plants.......................... Table Al.9: The Impact of the I C on the TFP o f Salvadoran and Central American Plants........56 57 Table 2.1: Transparency International CorruptionPerception Index -El Salvador..................... 61 Table 2.2: Criminal offensesbetween June 2001 and May 2003 ................................................. 74 Table 2.3: Crime, victimization and insecurity rates, by province ............................................... 75 Table 2.4: Cases received by the Attorney General, January-November 2003 ............................ Table 2.5: Violent crimes committed in2000 and percentage inwhich firearms were used.......76 78 Table 2.6: Pre-Payment,Payment at the Time o f Sale, and Credit .............................................. 89 Table 2.8: Business Association Membershipand DisputeResolution ....................................... Table 2.7: ExpectedEffects of A Hypothetical 10% Price Increaseon Product Demand ...........90 92 Table 2.9: Firmswith Overdue Payments That Have FiledCourt Cases, by Size ....................... Table 2.11: Time, Costs andNumber o f Proceduresto Enforce a Debt Contract inCourt..........94 Table 2.10: Statistical Tests for Differences inCourt Use between Small and Large Firms.......93 99 Table 2.12: Attorney and Court Fees inUSD and as Percent o f GNIper capita........................ 107 Table 2.14: Costs of Business Registration ................................................................................ Table 2.13: Tests for Differences inTime to Registerbetween Small and Large Firms ...........101 108 Table 2.15: Statistical Tests for Differences inUse o f Professional Agents for Renewal o f Table 2.16: LandPurchaseand Construction Activities............................................................. Business Licenses andor Permits: A Comparison across Countries and Types o f Firms.. 111 Table 2.18: Tax Penalties and Proportion o f Sales Declared to the Tax Authorities .................112 Table 2.20: Percentageof Senior Management Time Spent Dealingwith Regulations.............115 120 Table A2.1: Determinants of Bribe Payments inEl Salvador .................................................... Table A2.2: Determinants o f BribePayments inEl Salvador and neighboring countries .........129 Table A2.3: Correlates o f expenditures insecurity, crime victimization and crime losses........130 131 Table A2.4: Determinants o f Court Use inCentral America...................................................... 132 Table 3.1:Numbero fusers, sales and market shares o f distribution companies....................... 136 Table 3.2: Average Electricity Spot Market Prices (US$/kWh)................................................. 140 Table 3.3: Average Electricity Sales Price (US$/kWh) (I) .......................................................... Table 3.4: Electricity Prices in Central America for SelectedUsers (US$/kWh) (I) ..................141 142 6 143 Table 3.6. Fixed LineRates inCentral America, 2003 (US$).................................................... Table 3.5: Electricity Reliability Measures ................................................................................ 152 Table 3.7. Incidence, lengthand impact oftelephone cuts (2002) ............................................. 152 Table 3.9: Maritime Transport Costs to the U.S.,for selectedgarments (% o f export value) ...161 Table 3.8: Container Movement inCentral American Ports, 2002-2003 .................................. 162 Table A3.1: Correlates o fproduction stoppages and monetary losses due to power outages....164 Table 3.10: Condition o fthe Priority RoadNetwork, 2002 ..................................................... Table A3.2: Correlates o fusingthe internet and email to reach clients or suppliers.................174 Table A3.3: Correlates o f constraints and losses associatedwith Transport Problems..............175 Table A3.4: Correlates o f constraints and losses associatedwith Customs Problems ...............176 Table A4.1: Major or severe obstacles related to finance, by exporter status and ownership....177 199 199 Table A4.3: Share of FirmsDemanding and Obtaining a loan, by Industry.............................. Table A4.2: Share of Firms Demanding and Obtaining a loan, by Region................................ 199 Table A4.4 -Determinants o f Supply and Demand for Loans, and interest rates ...................... 201 Table A5.1: Correlates o fthe demand for skilledlabor and provision of training to workers...222 Table 5.1: Technology Extension Services in Selected Countries ............................................ 225 Table A5.2: Correlates o f quality certification andquality control............................................ 226 Table A5.3: Correlates o f innovation, R&D employment anduse o f computers and computer- controlled production equipment........................................................................................ 227 LIST OF FIGURES: Figure 1.1:Annual Growth inGross Domestic Product............................................................... 19 Figure 1.2: Growth inper capita GDP, 1960-2000....................................................................... 19 Figure 1.3:Growth accounting by decade, 1960-2000................................................................. 20 Figure 1.4:Explanations ofchanges ingrowth rates, 1986-90 to 1996-99.................................. Figure 1.5: Inflows and Stocks ofFDIas % of GDP.................................................................... 21 23 Figure 1.6: Stock ofForeignDirect Investment by Sector, 1997and 2004 (June) ...................... 24 Figure 1-7:Inflows o fForeign Direct Investmentby Sector, 1997-2003..................................... 24 Figure 1.8: Main Sourcesof Growth assumingprogressto top 25% ofLAC and the world.......25 Figure 1.9: Sample composition by Sector and Province............................................................. 26 Figure 1.10:Map ofEl Salvador................................................................................................... 27 Figure 1.11: The Perspective of Salvadoran firms - Investment Climate Constraints ratedas Figure 1.12: The Perspective o f Salvadoran firms - Investment Climate Constraints rated as "major" or "very severe" by more than 25% o f surveyed firms inEl Salvador ...................28 Figure 1.13: The Perspective of Salvadoran firms - InvestmentClimate Constraints rated as "major" or "very severe" by between 10% and 25% o f surveyedfirms inEl Salvador.......29 "major" or "very severe" by less than 10% of surveyedfirms inEl Salvador ..................... 30 Figure 1.14: Percentageo f sales lost due to quantifiable constraints inthe areas ofInfrastructure and Governance .................................................................................................................... 31 Figure 1.15:Labor productivity growth ofthe average Salvadoran firmunder optimistic counterfactual assumptions (improvements to 50th, 75thand 90thpercentiles) ..................... 35 Figure 1.16: Labor productivity growth o f the average Salvadoran firm assuming investment climate improvements to the country's 75thpercentile, by size range.................................. 35 Figure 1.17: Labor productivity growth o f the average Salvadoran firm by sector, assuming investmentclimate improvements to the 75thpercentile ofthe corresponding size range ...37 7 Figure 1.18: Labor productivity growth ofthe average Salvadoran firmby province. assuming investmentclimate improvements to the 75`h percentile ofthe corresponding size range...37 Figure 1.19: Labor productivity advantagederivedfrom investment climate conditions facedby the average Salvadoran firm. comparedwith average conditions inGuatemala. Honduras and Nicaragua ....................................................................................................................... 38 Figure 1.20: Labor productivity advantagederivedfrom investmentclimate conditions. comparing El Salvador's 75thpercentiles with 75thpercentiles o f Guatemala. Honduras and Nicaragua .............................................................................................................................. 39 Figure 1.21:Labor productivity differences derived from investmentclimate conditions. by size (comparison with average firm of the same size inGuatemala. Honduras or Nicaragua) ...40 Figure 1.22: Labor productivity differences with Guatemala derivedfrom investmentclimate conditions. by size (comparison with average firm o fthe same size) .................................. 40 Figure 1.23:Labor productivity differences derivedfrom investment climate conditions. by sector (comparison with average firm o f the same sector inGuatemala. Honduras or Nicaragua) ............................................................................................................................. 41 Figure 1-24:Labor productivity differences with Guatemala derivedfrom investmentclimate conditions. by sector (comparison with average firm o fthe same sector)............................ 41 Figure 1.25: Firms' Awareness ofTrade Treaties and Expectations about their Effects (El Figure 1.26: Firms' Expectations about CAFTA Effects on Exports and FDI.by Sector ...........45 Salvador. Guatemala. Honduras andNicaragua) .................................................................. 46 Figure 1.27: Major and Very SevereConstraints to Maximizing Benefits of CAFTA (El Salvador 46 Figure2.1: International corruption rankings of Central American countries.............................. and average for Guatemala. Honduras and Nicaragua) ........................................................ 58 Figure 2.2: Governance Indicators for Central America. 2002 (Kaufmann et. al.) ...................... Figure2.3: Governance Indicators for ElSalvador and LatinAmerica (Kauffman et a1. 2003) 60 Figure2.4: Governance Indicators for 1996 and 2002 (Kauffman et al.).................................... 61 Figure 2.5: Percentage o f Firms Reporting Low Governance (EOS 2003) ................................... ..59 62 Figure 2.6: Governance percentile ranks among 102 countries (EOS 2003) ............................... 62 Figure 2.7: Inconsistency andunpredictability o fregulations. government inefficiency and 63 Figure2.8: Bribes "to get things done" and to win public contracts ............................................ bribery................................................................................................................................... 63 Figure2.9: Percentage of firms that paidor were offeredbribesto obtain government services (subsetof firms that requestedthe corresponding services) ................................................. 64 Figure2.10: Bribesto obtain government services. international comparison............................. 65 Figure2.11:Bribesto obtain government services. by size ......................................................... 65 Figure2.12: Bribesto obtain government services. by province ................................................. Figure2.13: Bribesto obtain government services. by sector...................................................... 66 66 Figure 2.16: Homicide rates per 100.000 populationfor selected Latin American countries ......70 Figure2.15: Homicide rates per 100.000 population. by gender and WHO region..................... Figure2.14: Bribesto obtain government services. by export status........................................... 67 71 Figure2.17: Householdvictimizationrates (12 months before survey)....................................... 72 72 Figure 2.19: Crime victimization rates recorded inopinion polls ................................................ Figure2.18: Homicide rates inEl Salvador. 1969-2003 .............................................................. Figure 2.20: Kidnappings inEl Salvador between 1997 and August 2004 .................................. 73 73 Figure 2.21:Percentage o f Crimes reportedto the police inCentral America. Colombia and Mexico (2004) ....................................................................................................................... 77 8 Figure 2.23: Crime Losses and security costs (% of sales) - International Comparison............79 Figure 2.22: Victimized firms and firms constrained by crime........................................ 79 Figure 2.24: Victimized Firmsand firms constrainedby crime, by Province (YO).................-80 Figure 2.25: Crime Losses and security costs by Province (% of sales).............................. 80 Figure2.26: Victimized Firmsand firms constrainedby crime, by sector (%) ....................... Figure 2.27: Crime Lossesand security costs by sector (% of sales) ................................ 80 -80 Figure 2.28: Victimized Firmsand firms constrainedby crime, by size (YO)........................ 81 Figure 2.30: Crime Reporting by victimized firms, by country....................................... Figure 2.29: Crime Losses and security costs constrainedby crime. by size (%)..................-81 81 Figure 2.31:Crime Reporting by victimized firms, by size........................................... Figure 2.32: Crimes by Youth Gangs among victimized firms, by country.......................... 81 82 Figure 2.33: Crimes by Youth Gangs among victimized firms, by country., ........................ 82 83 Figure 2.35: Crimes solved by the police, by size..................................................... Figure 2.34: Crimes solved by the police. by country................................................... -83 Figure 2.36: FirmsWhich Do Not Believe that the Courts Will UpholdTheir Contractual and Property Rights ina Commercial Dispute............................................................................ 86 Figure 2.37: FirmsWhich PerceiveProblems inCourt Efficiency and Contract Execution to be Major or Very Severe Obstaclesto Doing Business............................................................. 86 Figure 2.38: Firmsthat View Court Efficiency Problems as Major or Very Severe Obstacles. by Size........................................................................................................................................ 87 Figure 2.39: FirmsWhich View Court Efficiency Problems as Major or Very Severe Obstacles, Court users Vs.Whole sample.............................................................................................. 87 Figure 2.40: FirmsWhich Perceive Problems inCourt Efficiency and Contract Execution to be 88 Figure 2.41: Share of Payment DisputesBrought to Court: by FirmOwnership......................... Major or Very Severe Obstaclesto Doing Business, by Province ....................................... 93 Figure 2.42: Share o f Firms Which FiledCourt Cases andNone of These Were Resolved in Their Favor, by Firm Size..................................................................................................... 95 Figure 2.43: Share of FirmsWhich FiledCourt Cases andNone of These Were Resolved in Their Favor: by FirmOwnership.......................................................................................... 96 Figure2.44: Average Number of Weeks Necessary to Resolve a DisputedPayment inCourt. 2003-2004 ............................................................................................................................. 97 Figure 2.45: Durationof a Payment DisputeinDays: A Comparison of ICs and Doing Business Figure 2.46: Durationof Contract Enforcement inCourt: by Stage of the Judicial Process...... 100 Data....................................................................................................................................... 98 Figure 2.47: Minimumand Maximum Durationof Contract Enforcement inCourt, by country ............................................................................................................................................. 100 Figure 2.48: Index of Judicial Complexity -A Cross-country Comparison ............................. 105 Figure 2.50: Time Needed for DifferentRegistration Procedures: A Comparison.................... Figure2.49: Number of Days to Register a Firm:A Comparison of DifferentData.................102 106 Figure 2.5 1: Time for Registration versus Time for Re-registration and Renewal o fLicenses and Permits ................................................................................................................................ 108 Figure2.52: Time Neededfor DifferentTypes ofRenewal ofLicenses and Permits: A Comparison......................................................................................................................... 110 Figure 2.53: Hire of Professional Agents inBusiness Registration and License and Permit Renewals............................................................................................................................. 110 9 Figure 2.54: Costs o f Registration and Licensepermit Renewal by Country: Percent of Total Annual FirmSales............................................................................................................... 112 Figure2.55: Time to Comply with RequirementsandPermits to PurchaseLandand Undertake Construction Activities ....................................................................................................... 113 Figure 2.56: Costs o f Compliance with Requirements and Permits to PurchaseLand and Undertake Construction Activities: by Country ................................................................. 114 Figure2 57: Shareof Firms Which Would ChangetheNumber ofTheir PermanentEmployees . ifNoLaborRestrictionsWereinPlace.............................................................................. 117 Figure 2.58: Number o f Days Spent inInspections and Mandatory Meetingswith Government Officials over the Year Preceding the ICs.......................................................................... 118 Figure 2.59: Number o f Days Spent inInspections and Mandatory Meetings with Government Officials over the Year Preceding the ICs: By Type o f Agency ........................................ 118 Figure2.60: Salvadoran FirmsWhich Consider Regulations Major or Very Severe Problems (%) ............................................................................................................................................. 121 Figure2.61: Central American Considering Regulations Major or Very Severe Problems (%) 122 Figure 2.62: Salvadoran FirmsConsidering Aspects of the Tax Administration Major or Very Severe Problems (%) ........................................................................................................... 123 Figure 2.63: Central American Firms Considering Aspects o f the Tax Administration Major or 123 Figure 3.1:Installed Electricity Generation Capacity, kW per 100 pop..................................... Very Severe Problems (%).................................................................................................. 134 Figure 3.2: Electricity consumption, kWhper capita ................................................................. 134 137 Figure 3.4: Production stoppages and monetary losses due to Power Outages.......................... Figure 3.3: Distributionand Transmission Losses as Percentage o f Output .............................. 144 Figure 3-5: Losses Dueto Power Outages............................................................... 145 Figure 3.6: Days for new Electricity Connection...................................................... 145 Figure 3.7: Losses Dueto Power Outages by size...................................................... 146 146 Figure3.9: Days to Install aNew Electricity Connection, by Province............................ Figure 3.8: Days to Install aNew Electricity........................................................... 146 Figure 3.10: Figure 3.8: Losses Dueto Power Outages, by sector.................................. 146 Figure 3.11:Firmswith generator, by country (%).................................................... 147 147 Figure 3.13: Mobile and FixedTelephone Lines inEl Salvador, 1998-2003............................. Figure 3.12: Firmswith generator, by Size.,............................................................ Figure 3.14. FixedPhone Lines per 100pop inLatinAmerican countries (2003) ...................149 . Figure 3.15. Mobile Phone Lines per 100pop Figure 3.16. FixedPhone Lines per 100pop. And GDP per capita, 1998 and 2003..................150 . in LatinAmerican countries (2003).................149 150 Figure 3.17. International Long Distance Tariffs inCentral America, 2003 (US$ per min. to the Figure 3.18. National Long Distance Tariffs inCentral America, 2003 (US$ per min.)...........151 U.S.) .................................................................................................................................... 151 Figure 3.19. Internet Hosts per 10,000 pop. and personal computers per 100 pop. (2003)........153 154 Figure 3.21: Internet and email use, by country (%)................................................... Figure 3.20. Internet Hosts per 1000 pop. and GDP p/c in2003............................................... 154 Figure 3.22: Internet and email use, by size(%) ........................................................ 154 Figure 3.23: Internet Dial-up Prices per month, Latin America (2004) ............................. 155 Figure 3.24: Vehicle Traffic at Central American border crossings, 1999................................. 159 Figure 3.25: Mainports usedby importers (Investment Climate Survey data).......................... 160 Figure 3.26: Mainports used by exporters (InvestmentClimate Survey data) .......................... 160 10 Figure3.27: Maritime transport costs to the U.S.from selectedports as % of export value .....161 Figure3.28: Shipping cost, Miami (20 f. cont.). ........................................................ 163 Figure 3.29: Shippingcost, San Francisco (20 f. cont.). ............................................... 163 Figure3.30: Territorial andpopulation roaddensity, by country............................................... 163 Figure3.31: Transport Constrained firms (%).......................................................... 166 Figure3.32: Transport Constrained firms ifCAFTA (%)............................................. 166 Figure 3.33: Input delivery delays (days lost) .......................................................... 167 Figure 3.34: Transport losses (% of sales) .............................................................. 167 Figure3.35: FirmsConstrained by customs (%)....................................................... 169 Figure3.36: FirmsConstrained by customs ifCAFTA (%).......................................... 169 Figure3.37: Custom delays for Exports (investmentclimate data)............................................ 168 Figure 3.38: Custom delays for Imports (investment climate data)............................................ 169 Figure 4.1: Financial Market Sophistication, by country (0 to 7 Scale) ..................................... 180 181 Figure4.3: Overhead Cost as a Share ofTotal Assets, average 1999-2001............................... Figure 4.2: Active and PassiveUSDInterest Rates, 1998-2003 ............................................... 183 Figure4.4: Firmsreporting major or severe obstacles relatedto finance, by size ..................... 184 Figure 4.5: Share of Firmswith Loans (%), International Comparison ..................................... 185 Figure 4.6: Share o f firms demanding and obtaining loans, by size (%of firms) ....................... 186 Figure4.7: Firmsdemanding and obtaining loans, by Export and Ownership Status (% of firms) Figure 4.8: Interest Rates for loans inUSD granted to large firms after 1999, by country ........187 ............................................................................................................................................. Figure 4.9: Interest Rates for loans inUSDgranted after 1999, by size and exporter status..... 188 189 Figure4.10: Constraints preventing firms from applyingto a loan............................................ 190 Figure4.11:Average value of collateral (as % ofloan), by country.......................................... 191 Figure 4.12: Sources of collateral (as % o f loan), by size .El Salvador...................................... 191 Figure4.13:Main Sourcesof finance for working capital, international comparison (%) ........192 Figure 4.14: Main Sources of finance for working capital, by size (%) ..................................... 193 193 Figure4.16: MainSources of finance for investmentcapital, by size (%)................................. Figure 4.15: Main Sources o f finance for investmentcapital, by country (%)........................... 194 Figure4.17: Main Sourcesof finance for working capital, by financial status (%)...................194 Figure 4.18: Main Sources o f finance for investmentcapital, by financial status (%)...............195 Figure5.1: Average years of schooling in 1960 and 2000 -International comparison.............202 Figure5.2: Educational distribution in2000 -International comparison .................................. Figure5.4: Firmsthat provide internal and external training to their workers, by country (%).204 Figure 5.3: Share o f Skilledworkers by firm size(%) ................................................................ 203 204 Figure5.5: Labor Market Rigidity Indexesfrom Doing Business Database.............................. 205 Figure 5.6: Minimumwages and minimumtotal labor cost (US$) ............................................ 206 206 Figure5.8: Share of firms with some presenceofworkers' unions (%) .................................... Figure 5.7 :Labor Regulations as constraints to business operations (% o f firms) .................... 207 Figure 5.9: Share o fpermanent workers intotal employment (%)............................................. 207 Figure 5.10: Time to fill vacancies of skilled and unskilledworkers (weeks) ........................... 210 Figure5.12: Numberofnational standardsinCentral America, by country.............................. Figure 5.11: Current quality system inEl Salvador (planned certificationindotted lines) .......208 211 Figure5.13: Numberof accredited laboratories inCentral America, by country ...................... 211 Figure 5.14: Firms with I S 0 9001:2000 quality certification, by country (%) .......................... 212 Figure 5.15: Firmswith I S 0 quality certification, by size (%) .................................................. 212 11 Figure 5.16: Firmswith I S 0 quality certification. by exporting status e!)............................... 213 Figure 5.17: Innovatingfirms by type o f innovationand by country......................................... 214 Figure5.18: Main source of Technological Innovations. by size (%)........................................ 215 Figure 5.19: Self-reported innovations by exporter status (% o f firms) ..................................... Figure5.21:Mainsource ofTechnological Innovations. by size (%)........................................ Figure 5.20: Main sources of technological innovations -International Comparison (%)..........216 216 Figure 5.22: Technology licensing inEl Salvador. by firm characteristics (% o f firms) ..........217 Figure 5.23: Less frequent forms o f technology acquisition .by country 218 Figure5.24: Imports ofcapital goods. by country (% oftotal imports) ..................................... (% of firms) ..............217 219 Figure5-25:R&D Expenditures (YOof GDP). by country (average 1980-1995) ....................... Figure 5.26: : R&D Expenditures per worker (US$). by country (average 1980-1995) ............219 Figure5.27: Patentsinthe US.during 1996-2000.by country (patents per millionpopulation per 219 FigureA5.1: InstitutionalSetting of a Typical National Quality System................................... year) .................................................................................................................................... 220 229 12 ACKNOWLEDGEMENTS This report was prepared by a team led by Pablo Fajnzylber and comprising Stefka Slavova (Contract Enforcement and Business-Government Relations), Jose Miguel Cruz (Crime), Francesca Recanatini and Juanita Riano (Corruption), Eloy Vidal and Juan Manuel Galarza Tohen (Telecommunications), Fernando Lecaros (Electricity), Andres Pizarro and Ximena Clark (Transport), Juan Miguel Crivelli (Access to Finance), Diane Thompson (Quality) and Isabel Sanchez-Garcia (Technology). Leonid Koriukin providedvaluable assistance with the descriptive analysis o f the data, Alvaro Escribano and Heisnam Singh performed total factor productivity estimates, Manoela Francisco provided useful research assistance for the chapter on Access to Finance, Adolfo Rouillon providedvaluable insights into the aspects o f the Salvadoran legal system that affect the firms' access to Finance, Waleed Malik provided useful information on the hnctioning o f the Salvadoran Judicial system and the ongoing Judicial Reform Project, and Bernice van Bronkhorst provided useful comments and suggestions on the crime section. Additional valuable inputswere received from Marialisa Motta, Jordan Schwartz, Marianne Fay, Aquiles Almansi, Manuel Sevilla and August0 de la Torre. MarialisaMotta, inparticular, helped setting up the team for this project and provided valuable guidance at the concept stage. Special thanks are also due to Natalia Angelucci and Eric Palladini for their editorial support. Peer reviewers are IrinaAstrakhan (ECSPF) and Caroline Freund (DECRG). 13 CHAPTER 1: INTRODUCTION BACKGROUND 1. Improvementsin the investment climate could contributeto promote sustained growth and poverty reductionthrough higher levels of investmentand productivity growth. After averaging 5.9% per year in 1990-1995, annual GDP growth was 3.2% in the second half o f the 1990s and 2% in 2000-2003. This slowdown took place in the context o f deteriorating terms o f trade, severe earthquakes in 2001 and the slowdown o f the U.S. economy, and despite the country's impressive economic transformation and prudent macroeconomic policies. Positive future growth prospects are to a great extent associated with the achievement o f higher rates o f investment and productivity growth, which in turn depend on the risks and expected returns faced by potential investors. Those risks and returns are intrinsically associated with the policy, institutional, and behavioral environment in which investment decisions are made. This environment is what we call the "investment climate." In the context o f this report, measures to improve the investment climate are grouped into four areas: (1) policies directed at improving governance and reducing insecurity, (2) increasing the quantity and quality o f available physical infrastructure, (3) strengthening the banking system and improving the conditions o f access to financial services, (4) promoting the development o f the national quality system and the firms' access to skilledworkers and state-of-the-art technologies. 2. Objectiveand rationalof the report.Work to improve the investment climate is recognized as a key pillar o f World Bank Group work to promote economic growth and poverty alleviation in developing countries.2 The main focus of Investment Climate Assessments (ICAs) is on microeconomic and structural dimensions o f a nation's business environment, viewed in an international perspective (See Box 1). To this end, ICAs look indetail at factors constraining the effective functioning o f product markets, financial and non-financial factor markets, and infrastructure services, including inparticular weaknesses in an economy's legal, regulatory and institutional framework. ICAs also provide the tools and analytical framework to identify reform priorities in a country's investment climate, by linking constraints to firm-level costs and productivity. The main objective o f this report i s to develop a better understanding o f the investment climate constraints that limit the growth and competitiveness o f Salvadoran firms. In particular, the report seeks to measure ina standardized way the investment climate conditions in ' These four areas are contemplated in the incoming Saca Administration's priorities, which include strengthening the realsector, buildingsocial andphysicalinfrastructure, institutingpublicsector and institutionalreforms,building a stronger financial systemto reducevulnerabilities inthe context o f dollarization, and implementingCAFTA.Fitch Ratings,09/08/04. "The centralchallengeinreapinggreater benefitsfrom globalization lies in improvingthe investment climate--that is, in providing sound regulation of industry, includingthe promotion o f competition; in overcoming bureaucratic delay and inefficiency;infighting corruption; and inimprovingthe quality of infrastructure. While the investment climate is clearly important for large, formal sector firms, it is just as important -- if not more so--for small and medium enterprises (SMEs), the informal sector, agricultural productivity, and the generation of off-farm employment. For these reasons, the investment climate itself is a key issue for poverty reduction." Nicholas Stem, ChiefEconomist, March22,2001. 14 El Salvador, to provide comparisons of those conditions with those prevailinginother countries and regions, and to identify the features o f the investment climate that matter most for competitiveness and growth. While we recognize that issues related to macroeconomic and political stability are crucial for constituting a good investment climate, the focus o f this report i s on microeconomic issues. Indeed, the importance o f political economy and macroeconomic issues i s now well understood, and they have been addressed at length in other World Bank reports. BOX 1:What is an InvestmentClimateAssessment? Investmentclimate assessments systematicallyanalyze the conditions for private investment and enterprise growth in a country, drawing on the experience of local firms to pinpoint the areas where reform is most needed to improve the private sector's productivity and competitiveness. By providing a practical foundation for policy recommendations and involving local partners throughout the process, the assessments are designed to give greater impetus to policy reforms that can speed the private sector's growth. Producedby the World Bank Group in close partnership with apublic or private institution ineach country, the investment climate assessments are based on a survey of private enterprises designed to capture firms' experience in a range of areas - financing, governance, regulation, tax policy, labor relations, conflict resolution, infrastructure services, technology, and training, among others. All these are areas where difficulties can add substantiallyto the costs of doingbusiness.The survey attempts to quantify firms' costs relatedto the investment climatebottlenecks. Using a standardmethodology, the assessmentthen compares the survey findings with those in similar countries to evaluate how the country's private sector is competing. The findings of the survey, combinedwith relevant information from other sources, provide a practical basis for identifying the most important areas for reform aimed at improving the investment climate. The findings and policy recommendations emerging from the assessments are discussed extensively with the private sector and other stakeholders in the country. This broad dissemination o f the findings is aimed at engaging not only policymakersbut also business leaders, investors, nongovernmental organizations, and the donor communityin shaping the nationalprivate sector development strategy, forging consensus on the priorities for reform of the investment climate, and laying the groundwork for concrete responses to the problems identified. Source:WorldBank(2003), Improvingthe InvestmentClimate inBangladesh. 3. Structure of the report. The present chapter provides conceptual background on what i s understood by the investment climate and why it i s thought to matter for economic growth, it then reviews El Salvador's general economic trends, and summarizes the main characteristics and the general findings o f the econometric analysis performed on the basis o f the firm level survey that was executed for this report. The second through fifth chapters focus on the main micro components o f the investment climate: governance, physical infrastructure, access to finance and technology. Eachchapter benchmarks the corresponding component o f El Salvador's investment climate using standardized quantitative indicators, and performs sub-national comparisons across firms o f different sizes, and across the sectors and regions covered in the See World Bank (2004a), "El Salvador Country EconomicMemorandum" (CEM); World Bank (2004b), "Country ProcurementAssessment Report" (CPAR); andWorld Bank (2004c), "El Salvador PovertyAssessment." 15 survey. Inaddition, each chapter contains an analysis o f the main policy issues that are relevant for the improvement o f investmentclimate conditions in the corresponding area, and formulates policy recommendations derived both from the survey findings and from other analytical work performed by the World Bank inEl Salvador. WHAT I S THEINVESTMENTCLIMATEANDWHY I T MATTERS 4. The quantity and quality of investment flowing into any specific region or country dependupon the risks and the expected returnsfaced by potentialinvestors. Those risks and returns are in turn intrinsically associated with the policy, institutional, and behavioral environment in which investment decisions are made. This environment i s what policymakers and multilateral organizations have recently denominated the "investment climate". 4 With this approach the investment climate obviously encompasses a wide range o f factors, including public policies and private sector characteristics. For analytical purposes, it i s useful to group these various aspects into three broad headings: Macro or country-levelissues: stabilityand openness. This group includes mainly national public policies that affect the country's degree o f political and economic stability, and the extent to which the country i s integrated in the global economy. Thus, this heading encompasses political economy issues, macroeconomic, fiscal, monetary, and exchange rates policies, as well as the national policy towards foreign investment and trade. Good governance and strong institutions.The main issue here is the efficiency ofthe legal and regulatory framework, as reflected inthe quality o f the services provided or regulatedby the government and the burden that firms face incomplyingwith regulations inareas such as entry and exit, labor relations, financing, taxation, the environment, and other legitimate public interests. This area also encompasses the efficiency o f the judicial system, the extent o f corruption and, more generally, the prevalence o f the rule o f law. Physical, financial and technological infrastructure. This broad area includes the quality and quantity o f available physical infrastructure, including power, water, transport and telecommunications, the development o f the country's financial system as reflected on firms' access to financial services at reasonable costs and conditions, and the extent to which firms have access to skilledworkers and state-of-the-art technologies. The quality of investment matters as much as its quantity. Recent research has found little correlation, at least in the short run, between countries' levels o f investment and their rates o f economic growthn5Thus, our emphasis goes beyond the promotion o f a larger quantity o f investment, to encompass also the conditions that are conducive to investments o f a better quality. By this we mean that a good investment climate i s one where the level o f investment i s higher, but also where investments do pay off in terms o f increasing the productivity and For details see Stem (2002), World Bank (2002a), and Dollar et al. (2003a). See Easterly (1999). Loayza et a1 (2002) also find that increases in investment rates tend to follow rather than to preceedhigher rates o f economic growth. 16 competitiveness o f firms and generating sustained growth. The importance o f the investment climate for promoting long run growth i s illustrated by the findings o f the vast empirical literature on the determinants o f growth. Spurred by the endogenous growth theories o f Romer (1986) and Lucas (1988), this literature has produced a number o f empirical results that are fairly robust and provide macro evidence about the importance o f the investmentclimate. 6. Macro stability and growth. It is now well accepted that macroeconomic stability and the avoidance o f financial and balance o f payments crisis have an impact on the long-run performance o f countries. Using cross-country data, Fischer (1993) documented the detrimental growth effects o f high rates o f inflation. As it i s well known, inflation i s affected by exogenous shocks but also by problems o f monetary management. Regarding fiscal policies, Easterly and Rebelo (1993) show that countries with high government consumption exhibit lower rates o f economic growth. Although some types o f government expenditures are socially productive, developing countries with very high government spending usually have bloated and inefficient bureaucracies, which tend to impose heavy burdens on the private sector, and assume roles that are most appropriate for the latter. 7. International Integration. A number o f studies, most recently Frankel and Romer (1999) and Dollar and Kraay (2001), have found find that openness to trade and direct foreign investment accelerate growth. These findings confirm the predictions o f the new growth models, which stress the importance o f increased competition and market size for diffusing technological innovations and exploiting economies o f scale. Moreover, greater trade openness allows firms to exploit their areas o f comparative advantage, it reduces the incentives for rent-seeking unproductive activities, and mitigates anti-competitive practices o f domestic firms [Lederman (1996)]. 8. Governance. Measures o f governance, such as the strength o f property rights, rule o f law, and the level o f corruption are also well correlated with growth [Kaufmann, Kraay, and Zoido- Lobat6n (1999); Knack and Keefer (1995); Mauro (1995)l. These studies typically use data generated from surveys o f private businesses, and reflect the extent to which investors and/or firms perceive problems with harassment, corruption, and inefficient regulation. These governance measures, however, are based on subjective assessments o f large entrepreneurs. A contribution o f this report, inthis respect, will be that o f documenting the quality o f governance in El Salvador, using objective indicators, collected from a representative sample of small, medium and large businesses. 9. Infrastructure. One can expect the stock o fphysical infrastructure to affect growth by either acting as an additional input into the production function - together with non-infrastructure capital - or by raising the productivity o f other factors o f production and increasing the profitability o f non-infrastructure investments. The empirical cross-country evidence suggests that investments ininfrastructure have a significant positive effect on long-term growth [Easterly and Rebelo (1993); Esfahani and Ramirez (2002); Calder6n and ServCn (2003b)l. In addition, infrastructure has been found to have a significant contribution to output, with rates o f return that are sometimes found to be larger than those o f non-infrastructure capital. [Gramlich (1994), Calder6n and ServCn (2003a)l. 17 10. FinancialDevelopment.Well functioning financial systems ameliorate the problems created by information and transaction costs and help allocate resources across space and time. Not surprisingly, a large number of studies have found a robust relationship between the level o f financial development and long-run growth [Levine, Loayza and Beck (2000)l. Financial development affects capital accumulation and technological innovation through at least five channels: by facilitating risk management, by reducing the costs o f acquiring information about new investment opportunities, by simplifying corporate control over managers, by mobilizing savings and by facilitating exchanges and thus promoting specialization and innovation [Levine (1997)l 11.Skills and technology. Differences inper capita income are due to a larger extent to gaps in TFP growth than to differences in factor accumulation [Hall and Jones (1999), Parente and Prescott (2000)l. Productivity growth i s inturn intrinsically related to the availability o f a skilled workforce and to the extent to which countries invest inresearch and development (R&D), either to generate new technologies, or to absorb innovations generated by others [De Ferranti et al. (2003), Lederman and Maloney (2003)l. 12. Problems of macro indicators and the use of microeconomic survey data. One weakness that i s shared by most o f the above studies linking growth to investment climate variables i s that they are based on macro-indicators o fpolicy and investment climate that are usually quite crude. Although they help illustrate the importance o f many variables for explaining differences in growth rates across countries, they are o f little help when the objective i s to identify what specifically needs to be done to create a better investment climate in a given country. For instance, aggregate indexes cannot, by definition, provide information on investment climate differences that exist at the sub-national level, nor help understand their effects on firm performance. This type o f analysis i s crucial for formulating priorities in the context o f policies directed at improving the investment climate. For these reasons, inorder to make the investment climate approach to development operational, it i s necessary to complement the aggregate indicators used on the empirical growth literature with the collection o f microeconomic survey data. The analysis o f this type o f data allows to estimate the links that exist between the different areas o f the investment climate and the levels o f productivity and export orientation o f local firms. That i s the approach adopted inthis report, which makes use o f firm level data collected in El Salvador during2003 and 2004, through a survey of465 private manufacturing enterprises.6 ANOVERVIEW OF ELSALVADOR'S GENERAL ECONOMIC TRENDS 13. With a per capitaincome of US$2,200in 2003 (atlas methodology), ElSalvador ranksin the upper part of the income range for lower middle income countries. Lower middle income countries are defined to have per capita incomes between $736 and $2,935. It i s worth noting, however, that El Salvador i s well below the average per capita income o f $3,260 found in 2003 among Latin American and Caribbean countries. El Salvador has a population o f around 6.5 million inhabitants, growing at an annual rate o f 1.7%, which i s about twice the world The survey also covered a sample of 47 hotels, which are analyzed in a different report: Wayne and Croes (2004), "Impacts o f Investment Climate Constraints on Accommodation Sectors in Ecuador, Guatemala, Honduras and Nicaragua." 18 average inthe same income range, Inaddition, it i s estimated that about two million Salvadorans live abroad, o f which about one million inthe U.S..In2003, remittances were equivalent to 67% o f goods exports, and almost 15% of GDP. In2001 the income o f Salvadorans living inthe U.S. -about13.3US$billion-was roughlyequivalenttothecountry's GDP.7Largemigrationflows started during the 12-year civil war from which El Salvador emerged in 1991. However, migration has continued steadily inthe 1990s as well as in the present decade, when the country was hit by earthquakes, severe droughts and the decline o f coffee prices. ' Figure1.1: AnnualGrowthinGrossDomesticProduct Year Source: Investment Climate Survey 14. Following the economic collapse of the civil war period, El Salvador experienced high rates of growth during the early 1990s, and a decelerationin growth after 1995 (see Figure 1.1). After averaging 5.9% per year in 1990-1995, annual GDP growth was 3.2% inthe second half o f the decade and 2% in 2000-2003. This slowdown in growth rates took place despite the country's impressive economic transformation and prudent macroeconomic policies. Indeed, during the past decade successive administrations undertook significant stabilization and modernization efforts, encompassing important measures in the areas o f trade liberalization, financial sector strengthening, re-privatization o f the financial sector and other state enterprises, comprehensive tax reform, pension reform and improvements in the environment for private investment.` Figure1.2:Growthinper capita GDP, 1960-2000 3 46 te: 1961-1970 1971-1980 1981-1990 1991-2000 `.El Salvador BLAC (weighted average)1 Source: World Development Indicators. '*World See Yang (2003) and EIU(2004). Bank (2004a), Country Economic Memorandum, p 1. 19 15. Despite the deceleration in GDP growthduringthe second halfof the decade,per capita GDP growth during the 1990s was faster in El Salvador than in the average Latin American and Caribbean country (see Figure 1.2). As a result, in 2003 El Salvador's per capita GDP was 31% higher than in 1990. During this period, economic growth coupled with increases in social sector spending-particularly inthe second half o f the 1990s- contributed to substantial reductions in poverty and significant improvements in basic socio-economic indicators. Between 1991 and 2002 the share o f the population whose income falls below the poverty line declined by 27 percentage points, from 64.4% to 37.2%, while extreme poverty rates fell from 31.2% to 15.4% (16 points). There have also been considerable improvements in a broad range o f non-monetary indicators o f well being, including enrollment rates in basic education, infant mortality, chronic malnutrition, and access to safe water, among others. Figure 1.3: Growth accounting by decade, 1960-2000 5.6 4.6 --2.6 u 3 . 0 1961-1970 1971-1980 1981-1990 1991-2000 IElGDP WLabor OCapital OTFP 1 Source: World Development Indicators. 16. A standard growth accounting decomposition of GDP growth yields strongly negative rates of total factor productivity (TFP) growth in both the 1970s and the 1980s, and only slightly positive TFP growth during the 1990s (Figure 1.3). To a great extent, El Salvador's negative performance in the area o f productivity growth can be attributed to the economic collapse and the destruction caused by the civil war. The latter can be illustrated by the fact that at the end o f the war, in 1991, per capita GDP was 28% lower than in 1978. Therefore, between the beginning and the end o f the conflict El Salvador experienced a rate o f per capita GDP growth o f minus 2.5% per year. To provide some perspective on the dimensions of these losses, duringthe "last decade" of the 1980sthe average LatinAmerican country exhibited a rate o f per capita GDP growth o f minus 0.8% per year (GDP-weighted average o f 25 countries), which amounts to a contraction o f less than 8% over a ten year period. Thus, one i s ledto conclude that the negative rates o f TFP growth experienced by El Salvador duringthe 1970sand 1980s- with the results o f the 1970s being driven by the latter years o f the decade - do not reflect "technical regress in the sense o f literal forgetting o f technology", but rather they mirror the large disruptions in economic activity caused by the armed conflict. loMoreover, during the period o f the civil war TFP growth is probably underestimated because o f the failure of economic time World Bank (2004b), Poverty Assessment: StrengtheningSocial Policy, pp, 5-12 loSee Barro (1999) for a comment on the negative rates o f TFP growth observed inmany Latin American countries. 20 series to capture the destruction o f productive assets - including not only the loss o f physical capital as a result o f the war but also that o f the skilledworkers that migrated duringthis period. Figure 1.4: Explanations of changes in growth rates, 1986-90 to 1996-99 Sources -3.5 -2.5 -0.5 0.5 Contrib. to changes In plc GDP Growth -1.5 1.5 2.5 3.5 I @I 1986-90to 1991-95 0 1991-95to 1996-99 I Source: Loayza et al. (2004). 17. Empiricalcross-country analysis of growth determinants fail to explain the high rates of economicgrowth of the early 1990s and the decelerationof the second half of the decade. The results of this complementary empirical exercise aimed at identifying and quantifying the main determinants o f the country's growth performance during the 1990s are summarized in Figure 3." El Salvador's growth in per capita GDP was 3.3% faster in 1991-1995 than during the second half o f the 1980s. However, using our econometric model o f the determinants o f cross-country growth we are able to predict only a 1.1% increase in per capita GDP growth between those two periods. This predicted improvement - which i s more than 2 percentage points below the actual increase in growth rates - results from a combination o f the positive effects o f cyclical reversion (increase o f 0.3% in growth), better stabilization policies (0.3% increase), progress in structural reforms (I increase) and deteriorating external conditions .3% (responsible for a reduction o f 0.7% in growth). In the second half o f the decade, per capita growth was 2.8% slower, a reduction that, once again, our model fails to predict, as it would forecast no change ingrowth during the period. Indeed, while continued progress inthe fields o f stabilization and structural reform policies should have led to an increase o f 1.8% in per capita GDP growth rates, this was expected to be just compensated by the continued deterioration o f external conditions, as well as by the consequences o f cyclical reversion and transitional convergence. All inall, El Salvador grew at faster than expected rates during the early 1990s and at lower than expectedrates duringthe second half o f the decade. 18. A likely explanation for the differencesbetween observed and predictedgrowthratesis related to the construction and consumption booms that took place in the early 1990s, spurred by reconstruction efforts and represseddemand followingthe end of the civilwar. These factors - which are not captured by the cross-country model above - appear to have been I'The figures are taken from Loayza, Fajnzylber and Calderon, 2002, "Economic Growth in Latin America and the Caribbean: Stylized Facts, Explanations and Forecasts", World Bank. The authors perform a detailed discussion of economic growth in Latin America and the Caribbean, including estimations of the determinants of growth in each country. 21 important during the early 1990s, and their winding down contributed significantly to the reduction in growth rates observed during the second part o f the decade. Indeed, the sectors o f construction and services experienced the largest rates o f growth during 1991-1995, but declined sharply during the remainder o f the decade. While agricultural growth and growth in non- construction industrial activities also declined after 1995, the reductionin aggregate growth rates can be attributed almost entirely to the fall in the growth rates o f the services and construction sectors. Average growth inmanufacturing value added, for instance, was 5.7% per year in 1990- 1995, 4.8% in the second half o f the decade and 3.6% in 2000-2003. Moreover, it i s estimated that if services and construction growth rates hadbeen maintained after 1995 GDP growth would have averaged about 5.8% per year.'* consumption booms driven by represseddemand - growth in construction activities and in 19. In addition to factors related to the end of the war - reconstruction investments, the service sector was amplified by a considerable increase in workers' remittances. The latter increased from an average o f 2.6% o f GDP during the 1 9 8 0 to~ ~more than 11% during the 1990s. Arguably, the increase inremittances was also a factor inpromoting growing inflationary pressures, a reduction in the relative prices o f export goods with respect to non-tradables, a worsening trade deficit which reached 15.4% o f GDP in 1995 and monetary disequilibria that ultimately led to a tighter monetary stance on the part o f the Central Bank. This latter factor arguably played an important role in stopping the construction and consumption boom o f the early 1990s, as interest rates increased from an average o f 4% in 1990-1995 to 12% in 1996- 2000. While the objective o f sterilizing the foreign exchange inflows associated with remittances playedan important role inshaping monetary policy inthe secondhalf o f the decade, the contractionary measures adopted by the Central Bank were also motivated by the desire to contain a lending boom that followed the re-privatization and liberalization o f the financial sector, and which caused a sharp increase in private sector indebtedness and a deterioration in loan portfolio quality. 20. The 1990s also witnessed importantchangesinthe composition of Salvadorane~ports.'~ The share o f maquila exports -basically apparel products - increased from 12% to 59% o f total exports between 1990 and 2002, while traditional export products (coffee, sugar, shrimp) fell from 45% in 1990 to 5% in 2002. Coffee, in particular, suffered the consequences o f lower international prices and the devastation caused by the 2001 earthquakes, which together account for a decline in the share o f that product from 20% to 3% between 1995 and 2003. Shrimp production also declined considerably, with exports revenues dropping from $39 million in 1996 to $11 million in 2003, mainly because o f pollution problems and the damages caused by the 2001 earthquakes. l4 Finally, non-traditional exports, which consist mainly o f non-maquila manufacturing products, fell from 43% to 36% o f total exports between 1990 and 2002. Since maquila exports were the most dynamic during recent years, and given that their entirety i s destined to the U.S., the share o f the latter market increased from 37% to 67% o f total Salvadoran exports in the 1990-2002 period. Non-traditional exports, on the other hand, became even more concentrated in the Central American market, which in 2002 accounted for 67% o f '*World Bank (2004a), Country Economic Memorandum,pp. 14. l3 This paragraphandthe next rely heavily on Lawrence (2004), p.4. l4EIU(2004), p. 27. 22 the foreign sales ofthose products, and 25% oftotal exports. As aresult, exports to markets other than the US. and Central America were only 7% o f total exports in 2002, down from 44% in 1990. Figure 1.5: Inflows and Stocks of FDI as % of GDP '"et Inflow 2000-2002 BFDI Stock in 2001 1 77 Source: WorldDevelopment Indicators. 21. SalvaLaran exports of apparel and textile products increased significantly dur..ig the late 1990s. In2002, exports o f those products represented86% of total Salvadoran exports to the U.S., with the remainder concentrated inother manufacturing sectors (10.8%) and only about 3% originating in traditional agro-export sectors (coffee, sugar and marine products). l5As shown by Buitelaar (2000), between 1993 and 1998 El Salvador, Honduras and Nicaragua significantly increased their shares in the U.S. apparel import market, growing at rates that are comparable to those exhibited by Mexican apparel exporters, which in this period benefited from a large devaluation of the Peso and the preferences introduced by NAFTA. Thus, while Mexico's implicit tariff for apparel products fell from 6.4% in 1993 to 0.9% in 1998, and its market share increased from 4.3% to 13.5%, Salvadoran exporters of garments paid average tariffs o f 10.7% in 1993 and 9% in 1998but still were ableto triple their market share from 0.8% to 2.4% ofU.S. apparel imports.l6After 2000, additional preferences have been granted to El Salvador and other Central American countries in the context o f the Caribbean Basin Trade Preferences Act. By 2002, 63% o f Salvadoran clothing exports to the US.entered duty free, while the remaining paid an average tariff o f 12.5% - resulting inan implicit average tariff of 4.6%. However, inorder to enter the U.S.market duty free, Salvadoran exporters need to comply with strict rules o f origin that increase their production and administrative costs, and they still face quantitative non-tariff restrictions. Not surprisingly, El Salvador's market share in U.S. apparel imports has only increased slightly, to 2.7% in 2002. As for non-apparel products, their average tariff rate in the U.S. was only 0.9%, suggesting supply side constraints as opposed to barriers to market access as the likely explanation for the country's relatively small penetration in the US. market. Competitiveness problems are also suggested by the relatively high protection rates enjoyed in the local market by agricultural products (e.g. 26.9% for dairy products), processed foods (e.g. IsLawrence (2004), p, 5. Among Central American countries, the share o f clothing in exports to the U.S.is highest inEl Salvador, followedby Honduras (77%), Nicaragua (64%), Guatemala(60%) andCosta Rica (23%). l6The combined market share of Honduran and Nicaraguan apparel exporters increased from 1.54% in 1993 to 4.27% in 1998, with Nicaraguan exporters holding a 0.47% market share in the latter year, but paying an average implicit tariff o f 14.7% in the latter year, compared to 7.1% for their Honduran counterparts. Buitelaar (2000), p. 137. 23 15.7% for prepared foodstuffs) and labor intensive manufactures such as textiles and footwear, whose tariffs reached respectively 17.8% and 14.9% in 2002. Thus, despite the fact that El Salvador's trade liberalization efforts during the 1990s were very significant - the average tariff declined from 21.9% in 1989 to 7.4% in 2002 - several important sectors o f the Salvadoran economy still benefit from considerable tariff protection, which i s likely to provide negative incentives for productivity- increasing investments. Figure 1.6: Stock of Foreign Direct Investment by Sector, 1997 and 2004 (June) Other Other - Maauila 5% Industry 41% 41% 47% 1997 Stock June 2004 Stock Source: CentralBank of El Salvador. Figure 1.7: Inflows of Foreign Direct Investment by Sector, 1997-2003 Other Other 6% dustry M ; ; $ a n I 22% &25% ~ fl &ServicesTrade ~ ~ ~ 30% 1997-2000 2000-2003 Source: Central Bank of El Salvador. 22. In the area of foreign direct investment (FDI), El Salvador has increasingly adopted a liberal regulatory regime, culminating with the Investment Law of 1999. Bilateral investmenttreaties have been signedwith 23 countries, including a 1999 treaty with the U.S. The latter treaty establishes the right to invest in terms no less favorable than those granted to domestic or third country investors, and allows for the free transfer o f capital, profits and royalties. Despite this liberal FDI regime, El Salvador has a small stock o f FDI compared to neighboring countries: in 2001 that stock represented 17% o f GDP, while FDI inflows averaged only 1.5% o f GDP in 2000-2002 (figure 5). While these figures are similar to those o f neighboring Guatemala, they are much lower than inHonduras, Costa Rica and Nicaragua. There is, however, evidence o f an increasing trend in FDI flows into El Salvador. Indeed, FDIaveraged 1.6% o f GDP during the 1999-2003 period, compared to an average o f 0.5% o f GDP in 1991- 1997. The year o f 1998 was atypical, as foreign direct investments jumped to about 9% o f GDP, driven by the privatization of state enterprises in the electricity and telecom sectors, which 24 resultedininvestmentsof $850 millioninthat year only. As seen infigure 6, foreign investments ininfrastructure are now responsible for 47% ofthe total stock ofFDI, compared to less than 1% in 1997. However, growth inFDIhas not beenrestrictedto the power and telecom sector, which represented only 25% o f FDI inflows over the 2000-2003 period (figure 7). In fact, the non- infrastructure FDI stock grew at an annual rate o f 19% between 1997 and 2003, with the FDI stocks o f maquilas and the financial sector exhibiting the highest rates o f growth - respectively 29% and 31% per year between 1997 and 2003. It i s worth noting, however, that within manufacturing the non-maquila sector i s still responsible for a higher share o f both the stock and the flow o f FDI(figures 1.6 and 1.7). 23. Looking ahead, it i s usefulto consider the implications of the above mentioned cross- country growth analysis, in terms of pointing out the policy areas in which El Salvador faces the largest challenges and opp~rtunities.'~ With that objective, figure 5 reports the estimated increases in growth rates that could be achieved if, for each o f the determinants o f growth included in our econometric model, El Salvador were to move either to the 75% percentile o f the Latin America and Caribbean region or to the world's 75% percentile. In the first scenario, it is expectedthat ElSalvador would achieve an annual rate of per capita growth of 4.5%, while in the second more optimistic scenario the expected rate o f growth would be 6.3%. This exercise suggests that El Salvador's greatest challenges and opportunities for achieving higher growth are in the areas o f education, infrastructure and trade openness. Indeed, it is estimated that the payoff from reaching the regional best-practices in each o f those three areas would be an increase in growth rates with respect to the 1990s of, respectively, 1.1%0, 0.8% and 0.4%. Moreover, achieving world best-practices in those areas i s expected to provide respectively 1.5%, 1.4% and 0.7% o f additional annual growth. In the areas o f financial depth and volatility, El Salvador i s already at the level o f LAC best practices, but moving to the World's 75'h percentiles i s expected to lead to increases inp/c GDP growth o f respectively 0.4% and 0.2%. Figure 1.8: Main Sources of Growth assumingprogressto top 25% of LAC andthe world -1.0% 00% 10% 20% 30% 40% 50% 60% 70% Contrib. to changes in plc GDP Growth IW75th oercentile of LAC 075th Dercentileof the world 1 Source: Loayza et al. (2004). l7See Loayza, Fajnzylber and Calderon (2004). 25 THE2004INVESTMENT CLIMATE SURVEY 24. The present El Salvador Investment Climate Assessment i s based on a survey of 465 manufacturing firms performedbetween November of 2003 and May of 2004. It is difficult to measure and understandthe issues contemplated inthe investment climate concept only on the basis o f the macro indicators. Thus, the World Bank's conceptual emphasis on the importance o f the investmentclimate for achieving sustained growth has been accompanied by a practical focus on the collection o f microeconomic data on the investment climate, by means o f surveys o f private enterprises. The present El Salvador Investment Climate Assessment i s based on a survey o f 465 manufacturing firms, performed between November o f 2003 and May o f 2004. About 60% o f the sampled firms were from San Salvador (the capital city), 16% from the province o f L a Libertad, 8% from Santa Ana, 6% from San Miguel, 3% from Ahuachapan, and the remaining 7% from eight other smaller provinces (see figures 1.9 and 1.10). l8 Figure1.9: Sample compositionby Sector and Province IllApparel Achuachapan 4.1% HFoodslBev 7'3%13,20/07.7% mSanta Ana % OChemicals 5. 0Rubber&PI. 5.7%QLa Libertad W Nowmet.Min 0San Salvador ElTextiles .5% w Metals WSan Miguel 60.4% 10.3% Sector 0Footwear Province Elother rgns Source: Investment Climate Survey 25.The sample was drawn randomly from 9 sub-sectors, including those of food, beverages, chemicals, apparel, textiles, shoes, rubber and plastics, metallic products and non-metallic minerals.The sectors covered inthe survey are responsible for 86% o f total output and 90% o f employment in the manufacturing sector, which in 2003 responded for 24% o f the country's GDP and the majority o f export revenue^.'^ The sample frame was drawn from a list o f firms provided by the National Statistical Institute (Digestic), and it is representative at the level o f economic sectors, with sample sizes chosen to ensure that binomial variables can be estimated with a 90% confidence and a sampling error o f about 9%. After determining the sample size at the level o f sectors, the sample was stratified by province and size ranges. The final sample encompasses microenterprises (21% o f the sample), small (29%), medium (26%) and large firms (24%).20 About 9% o f the sample i s composed o f maquilas and another 37% i s ''The survey also covered 35 hotels, in order to gain insights on the tourism sector. However, this data has not yet been used inthis report, given the lack o f sufficient country comparators. l9Maquila exports were responsible for 60% o f total exports, while traditional agro-exports represented only 5% of export revenues. The remaining so-called "non-traditional exports" are known to be mostly constituted by manufacturingproducts. See Lawrence (2004). 2o We define micro enterprises as those having less than 10 employees, small firms as those with 10 to 25 employees, medium firms as those with 26 to 75 employees, and large firms as those with more than 75 employees. 26 made up o f non-maquila exporters. The data was collected by means o f face to face interviews of company management, performed by trained enumerators. The survey was executed by the private firm CID Gallup, under the supervision o f the World Bank. 26.The questionnaire used in the survey was based on the standard core investment climate survey instrumentthat the World Bankis applyingin other countries of the region, as well as in other regions of the world. Inparticular, similar surveys were performed in 2003 in Nicaragua, Honduras and Guatemala, and another is about to be launched in Costa Rica. Whenever possible, the core survey attempts to evaluate conditions in quantitative terms of dollars and days. Supplementing and framing these quantitative questions are well-tested qualitative questions (such as those from the World Business Environment Survey). It is important to cross-check the findings derived from both types o f questions, as rankings of problems based on subjective indicators may differ from those constructed on the basis o f objective ones. Figure 1-10:Mapof ElSalvador G U A T E M A L A H O N D l l R C I S MAIN INVESTMENT CLIMATE CONSTRAINTS: THE PERSPECTIVEOF THE FIRMS 27. In order to evaluate which are the most important investment climate constraints affecting Salvadoran firms, this reports uses both subjective and objective indicators.The emphasis, however, is placed on the second type of information. Subjective indicators are based on the perceptions o f the surveyed firms regarding the key factors that constrain their development. This approach provides valuable information on the priorities that entrepreneurs would adopt if faced with the task o f designing policies to improve the investment climate. However, we argue that for policy-making purposes the emphasis should be placed on the analysis o f objective and quantitative indicators regarding the costs associated with various investment climate problems. Indeed, the perceptions of the entrepreneurs may be biased by 27 recent events reported in the media, and they may also reflect their specific cultural and socioeconomic background. For instance, managers o f firms that concentrate on local as opposed to national or international markets may lack the necessary benchmarks to judge the severity o f the problems existing intheir cities or provinces, and compare them to national or international best practices. In addition, by providing quantitative estimates o f the costs associated with investmentclimate problems, objective indicators may be o f greater help to policy makers when defining priorities for addressing the various investment climate obstacles encountered in the country. 28. To construct subjective indicatorsof the maininvestment climate constraints existing in El Salvador, the surveyed firms were asked to judge the severity of twenty potential investment climate problems in a five-point scale. Figures 1.11 to 1.13 report the percentage o f manufacturing firms that evaluated the corresponding constraints as "major" or "very severe". Figure 1-11shows the top constraints, defined as those that were mentioned as important by more than 25% o f the surveyed firms, while Figure 1.11 and 1.13 include the constraints that were mentioned respectively by between 10% and 25% o f the firms, and by less than 10% o f the respondents. Each figure displays also the average percentage o f firms in Guatemala, Honduras and Nicaragua that rated the corresponding constraint as "major" or "very severe". Figure 1.11: The Perspectiveof Salvadoranfirms Investment ClimateConstraints ratedas "major" or - "very severe" by more than 25% of surveyedfirms inElSalvador Crime &Violence Anti-competitive practices Corruption Access to financing Policy Uncertainty Cost of financing Macroeconomic Instability 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 @ElSalvador HAvg. Gua.-Hon.-Nic.1 Source: Investment Climate Survey 29. Salvadoran entrepreneurs report much less investment climate constraints to their operations than do their counterparts from Guatemala, Honduras and Nicaragua. On average the percentage o f firms inthose three neighboring countries that rate investmentclimate constraints as "major" or "very severe" i s 77% higher than in El Salvador. At least in the subjective perspective o f the firms, the incidence of constraints to business operations i s less frequent in Salvador for all areas o f the investment climate. As an illustration, in comparison to the average Guatemalan, Honduranor Nicaraguan firm, Salvadoran entrepreneurs are about half as likely to report important constraints in the areas of corruption, cost o f financing, telecommunications and macroeconomic instability. 30. Together with macroeconomic instability, two of the four main microeconomic components of the investment climate are mentioned in the group of top constraints: governance and insecurity,and access to finance (Figure 1.11). The top three constraints are 28 from the area o f governance and insecurity: they are crime and violence (mentioned by 49% o f Salvadoran firms), anti-competitive or informal practices (45%) and corruption (35%). Without diminishing the importance o f the concerns expressed by Salvadoran firms, it is worth noting that the fraction of firms that report those three areas as sources o f major or severe constraints i s on average 74% higher in Guatemala, 31% higher in Honduras and 22% higher in Nicaragua. The second group o f investment climate issues revealed by the survey as sources o f severe constraints involves access and cost o f financing, which are rated as important constraints by respectively 31% and 28% o f Salvadoran firms. Once again, the fraction o f firms rating these areas as sources o f major or severe constraints i s about 90% higher in Honduras and Nicaragua, and 32% higher in Guatemala. Finally, also among the top constraints mentioned by Salvadoran firms we find the issues o f macroeconomic instability and policy uncertainty, which are rated as important by respectively 28% and 26% o f the firms. This i s about half the fraction o f firms that rate those problems as important in the three above mentioned neighboring countries. Moreover, Salvadoran firms are also half as likely as their counterparts from Guatemala, Honduras and Nicaragua to rate issues related to tax policy, trade policy and customs as major or very severe constraints. Nevertheless, the relatively large importance given by firms to problems associated to policy uncertainty and macroeconomic instability suggests that despite the country's achievements in those areas, considerable concerns still remain among firms about the stability o f the macro-regime, and the possibility o f policy changes.2' Figure 1.12: The Perspectiveof Salvadoran firms Investment ClimateConstraints ratedas "major" or - "very severe" by between 10% and 25% of surveyed firms in ElSalvador Tax rates Electricity Workers Skills and Education Court efficiency Tax Administration Access to Land Customs regulations 0 0 100 200 300 400 500 600 700 IUEl Salvador mAvg. Gua.-Hon.-Nic.I Source: InvestmentClimate Survey 31. The other two main microeconomic components of the investment climate - infrastructureand skills and technology - were mentioned less frequently and would thus appear to be less important in El Salvador, at least based on these qualitative questions. In the area o f infrastructure, problems are considered major or very severe by only 21.5% o f the firms in the case o f electricity, 11% for access to land, 9.5% for transportation and only 6% for telecommunications. The fraction o f firms that rate these infrastructure areas as sources o f major As explained above, however, the focus o f this report is on the microeconomic dimensions o f the investment climate. Other World Bank reports (2004a, 2004b) have recently analyzed in detail the various issues that relate to El Salvador challengesinthe field o f macroeconomic stability. 29 or severe constraints i s 54% higher in the three above mentioned neighboring countries. As for the skills and education of available workers, they are considered to be major problems by 20% o f Salvadoran firms, while labor regulations are seen as an important constraints by only 4% o f the surveyedfirms. Itis interesting to note that Honduran and Guatemalan firms are more likely than Salvadoran ones to be concerned about the skills o f their workers, and they are about four times more likely to rate labor regulations as important constraints. In contrast Nicaraguan entrepreneurs exhibit smaller concerns with their workers' skills, and have a relatively low preoccupation with labor regulations. Inthis respect, a small concern with the qualifications o f the workforce or with the rigidities associated with labor regulations may reveal either a good supply of educated workers and flexible labor regulations or, alternatively, a small demand for skills and low levels o f enforcement o f labor regulations. We will return to this issue inthe sixth chapter. Figure 1.13: The Perspectiveof Salvadoranfirms InvestmentClimate Constraints rated as "major" or - "very severe" by less than 10% of surveyedfirms in ElSalvador Transportation Business Licensing, Permits Property Rights Telecommunications . Labor regulations Trade regulations 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 119El Salvador Avg. Gua.-Hon.-Nic. 1 Source: Investment Climate Survey 32. The nature of the main investment climate constraints, at least in the perception of the firms, varies considerably by firm characteristics: MSMEs that sell only in the domestic market, have domestic ownership and report low levels of capacity utilization are more likely to report more severe investment climate constraints. In the case o f finance, for instance, the percentage o f entrepreneurs that report major or very severe constraints i s twice as large among MSMEsthan among large firms, about 40% larger among non-exporters compared to exporters, and four times higher for domestic than for foreign owned firms. Not surprisingly, firms with low capacity utilization express smaller concerns with issues related to finance. However, they are 30% more likely to report constraints in the areas o f governance and infrastructure, and 60% more likely to report labor-related constraints. Similarly higher rates o f investment climate constraints are also found inthose areas among MSMEs, domestically owned firms and non-exporters. As mentioned before, these differences may reflect either different actual constraints or distinct firms' perspectives on their gravity, driven for instance by the use o f different national or international benchmarks to determine the importance o f the corresponding problems. As for sectoral and regional differences, the companies surveyed in the sector o f rubber and plastic products and those from the province o f L a Libertad report the least severe 30 investment climate constraints, while the largest incidences o f constraints are found among footwear manufacturers and inthe province o f San Miguel. Figure 1.14: Percentageof saleslost due to quantifiableconstraints inthe areas of Infrastructure and Governance 7 I Crime Electricity Interruptions 3.07 Total losses:9.9% of sales in ESA vs. 13.3% Transport Problems for neighbors (averagefor all Payments to get things done Security costs 4 2 5 J -7.- --- -7 0.00 1.00 2.00 3.00 4.00 5.00 BE1Salvador .Average Gua.-Hon.-Nic. Source: Investment Climate Survey 33. Qualitative questions have the value of revealing the subjective perceptions of Salvadoran firms with regard to the rankings of their problems, but it i s useful to complement them with indicators of the objective shortcomings of the country's investment climate. For instance, the problems in the area o f infrastructure are not among the ones mentioned most frequently by the surveyed firms as major or very severe. However, as seen in Figure 1.14, failures in basic infrastructure services, including water, phones, electricity and transportation are responsible for losses that add up to 3% o f total sales for the average Salvadoran firm covered in the survey. Figure 1.14 also illustrates the actual losses that may lie behindthe perceptionthat corruption and crime are important constraints inEl Salvador: indeed, expenditures in bribes "to get things done" consume 1.7% o f the revenues of the average firm and losses due to crime prevention and victimization amount to 5.1% o f sales, on average. All in all, the above mentioned problems in infrastructure and governance are responsible for an increase inthe costs o f doing business inEl Salvador that i s equivalent to 9.9% o f the sales o f the average firm. Although those losses are below the corresponding figure o f 13.3% that was found on average for Guatemala, Honduras and Nicaragua, they are still sufficiently sizable to warrant a closer look at the investment climate problems faced by Salvadoran firms. IMPACT OF THE INVESTMENTCLIMATE: ECONOMETRIC ESTIMATES 34. In order to objectively quantify the impact of the Investment Climate on the performance of Salvadoran firms, we performed an empirical investigation of the impact of the investment climate on several measures of firm performance: labor productivity, total factor productivity (TFP), employment growth, profits, exports, wages and the employment share of skilled workers. To measure the investment climate conditions faced by firms we 31 employed a group o f four indicators, one for each o f the areas of the investment climate emphasized in this report: governance and insecurity, infrastructure, access to finance and technology. Each o f those indexes was constructed as the first principal component o f a group o f four or five indicators measured at the firm level for the corresponding investment climate area.22The impact o f those indicators on firm performance was estimated holding constant basic firm characteristics, includingplant location (province), industry affiliation, firm size, firm age, the average level o f education o f the workforce, beingpublicly listed, having externally audited financial statements, the share o f imported inputs, and being an exporter or a multinational firm (including foreign owned firms and Salvadoran firms with factories abroad). The detailed statistical results on the impact o f the investmentclimate on the performance o f Salvadoran firms are reported in table A.1 o f the Technical Appendix.23 As explained in the appendix, the estimation was performed in such a way that endogeneity concerns have been minimized and the estimated coefficients can be interpreted as reflecting causality from the investment climate to firmperformance. 35. We also estimated the relationship between the investment climate and firm performanceby poolingthe data for ElSalvador with data from similar surveys conducted by the World Bankin 2003 in Guatemala,Honduras and Nicaragua.The comparison o f the Salvadoran investment climate with that encountered in those other three Central American countries i s particularly relevant inthe present context, given that Salvadorans firms will soon be competing with their counterparts from Nicaragua, Honduras and Guatemala to maximize the attraction o f Foreign Direct Investment in the context o f CAFTA, and increase their market shares in a more accessible U.S.market. The results are reported in Table A.2 o f the Technical Appendix. In the case o f the determinants of labor productivity, they are qualitatively very similar to those obtained usingdata from Salvadoran firms only (table A.1). There are, however, differences inthe relative magnitudes o f the effects o f each investment climate indicator, as well as differences in the determinants o f the other firm performance variables considered in the analysis. Overall, the results of these econometric exercises allow us to determine which are the areas o f the investment climate in which Salvadoran firms have a competitive edge, which are the ones inwhich they are at a disadvantage with respect to their Central American competitors, and what would be the potential gains to be derived from closing those gaps. 22 The construction o f the investment climate indexes is as follows. The governance index is the negative o f the first principal component of the percentage o f sales lost due to informal payments (bribes), the percentage o f sales that is typically not declared for tax purposes, the percentage o f sales lost due to criminal activities, a dummy variable for firms that do not believe that in the event o f a commercial conflict the judiciary would defend their contractual and property rights, and the number o f government inspections reported by firms, normalized by employment. The infrastructure index is the first principal component o f the percentage o f sales lost due to power outages (which enters with a negative weight), the percentage o f water obtained from the public grid, the percentages o f firms using the intemet and email to communicate with clients or suppliers inthe state, sector and size range o f the firm, and the losses reportedby firms as a result o f problems inthe transportation system. The finance index is the first principal component o f dummy variables for firms with access to an overdraft facility or line o f credit, firms that use bank resources for their working capital needs or to finance investments, and firms that use supplier credit to finance their investments. Finally, the technology index i s the first principal component o f dummies for I S 0 certification, R&D activities, the provision o f training to workers, the use o f computer controlled production equipment, and the share ofworkers that use computers on thejob. 23 The details o f the methodology adopted for estimating total factor productivity are also described inthe Technical Appendix. Our approach i s inmany respects similar to the one proposedby Escribano and Guasch (2004). 32 36. The four investment climate indexes are found to be positively related to labor productivityin boththe sample of Salvadoranfirms and in the one that also includes firms from Guatemala, Honduras and Nicaragua. In both samples the technology index has the largest quantitative effect on productivity, but the ranking o f the other effects depends on whether the analysis i s restricted to Salvadoran plants or the extended sample i s used instead. Indeed, the index that measures governance and insecurity has the second largest coefficient in the smaller sample, but has the smallest coefficient of the four when the estimation is performed with the extended sample. Other qualitative findings o f our econometric analysis can be summarized as follows. 1. Governance and insecurity.Among Salvadoran firms, better conditions in this area lead to higher levels o f total factor productivity and employment growth, a higher probability o f reporting profits, and higher employment shares o f skilled workers. When firms from the four countries are pooled together, improvements in governance indicators are also found to lead to higher wages and higher probabilities o f exporting. 2. Infrastructure. In the pooled sample, improvements in this index are related to higher values o f all the indicators o f firm performance hereby employed. When the analysis is restricted to Salvadorans firms, however, significant effects are found only on labor productivity, exports and wages. 3. Finance. Access to finance i s positively related to higher rates o f growth and profitability (among Salvadoran firms), and a larger probability o f exporting (extended sample). When all four countries are pooled together, we find that better access to finance is related to higher probabilities o f exporting, but also lower average wages and lower employment shares o f skilled workers - these negatives effects are also present although they are non-significant in the Salvadoran sample. 4. Technology. Inboth samples, firms with better technology indexes exhibit higher levels o f wages and skilled labor shares, and they are more likely to export. In addition, those firms also have higher TFP (Salvadoran sample) and are more likely to report profits (extended sample). 37. In order to quantify the potentialbenefits of improvingthe investment climate and to help determinethe areas that could be prioritizedto achieve greater improvements in labor productivity, we conducted a series of counterfactual simulations. These simulations are aimed at answering the following questions. If considerable improvements were made in all the areas of the investment climate, what would be the magnitude o f the benefits ripped by the average Salvadoran firm in terms o f labor productivity? If all the firms were to face a substantially improved investment climate, which would be the provinces, sectors and size ranges that would benefit the most? What are the areas o f the investment climate inwhich policy reform could generate larger benefits, both for the average Salvadoran firm, and for specific subgroups o f firms (defined by location, sector o f activity or size)? The answers to those questions depend on the specific hypothesis that are chosen to describe counterfactual changes in the investment climate or, in other words, on the benchmarks that the Salvadoran firms are 33 assumed to reach in hypothetical scenarios. In addition, the magnitudes o f the effects o f those counterfactual changes depend on the econometric estimates that are employed to link the investment climate with firm performance. 38. We performed both within-country and international simulations. Within country counterfactualsassume improvements in all areas of the investment climate and calculate the benefits that could potentially be derived from those improvements. International simulations serve to estimate the share of the productivity differences found across countries that are due to investment climate conditions. Inthe within-country analysis we use national benchmarks derived from best practices that are already found among the best performing Salvadoran firms. We analyze the consequences on labor productivity o f reaching those benchmarks on the basis o f econometric estimates derived from the survey o f Salvadoran plants performed for this report (table A.1). In the case o f the international simulations, the investment climate conditions faced by Salvadoran firms are compared to those o f their counterparts from Guatemala, Honduras and Nicaragua, and the impact o f those differences on labor productivity i s calculated usingthe econometric estimates obtained with the pooled sample o f firms from the four countries (table A.2). These simulations are intendedto provide a general idea o f the potential benefits o f improving different components o f the investment climate. They are, however, subject to the caveats that apply to the econometric estimates on which they are based, namely the fact that the data i s for the manufacturing sector only, and the possible existence o f measurement error in both the indexes that represent the quality of the investment climate and inthe variables that represent firm performance. WITHIN-COUNTRY SIMULATIONS24 39. In our first counterfactual exercise we use the 50fh, 75fh and 90th percentiles of our investment climate indicators as best practices that are considered attainable by Salvadoran firms provided that appropriate policies are put in place. In these optimistic assumptions we impute the value o f those benchmarks to all the firms that report investment climate indicators below the corresponding best practice. Under this counterfactual assumption, productivity remains unchanged for the firms that report indicators above the benchmarks. For the rest o f the firms, the labor productivity effects o f the reforms i s given by the difference between observed investment climate indicators and the corresponding benchmarks, multiplied by the corresponding coefficient from table A.l (column 1). Figure 1.15 reports the average effects o f reforms in each o f the four investment climate areas hereby considered and each o f the three alternative assumptions, as well as the total effect o f investment climate improvements on average labor productivity. We also report the average increase in labor productivity that would take place if all the firms with observed productivity below the 50th, 75thor 90thpercentile were to reach those benchmarks. 40. Reducing within-country differences in the investment climate conditions faced by Salvadoran firms would go a long way in increasing their productivity to the levels achieved by the nationaltop-performers.The largestproductivitygains are associated with improvements in the area of technology. As shown in Figure 1.15, if all the firms below the national 90thpercentile o f labor productivity were to reach that level o f performance, the average 24A similar simulation approach is applied to Chinese cities inDollar et al. (2003~). 34 increase in productivity would be 131%. As suggested by our counterfactual simulations, more than 90% o f that increase (119%) could be obtained through improvements in the investment climate, assuming that all firms below the 90thpercentile o f our investment climate indicators would reach that benchmark. Similarly, upgrading the investment climate conditions faced by firms below the 75th percentile or below the median would lead to average increases in labor productivity o f respectively 75% and 29%. Once again, those improvements correspond to respectively 90% and 74% o f the average increase inproductivity that would take place if all firms below the 75'h and 50thpercentiles o f labor productivity were to reach those levels. In the two more optimistic assumptions, the largest productivity effects are associated with improvements in the area of technology, which leads to gains that are about twice as large as those estimated for improvements in the areas o f governance and insecurity, infrastructure and finance. In the least optimistic scenario - improvements to the median - the gains associated with the area o f governance are slightly larger than those coming from that o f technology, and both are about twice as large as those derived from finance and infrastructure. Figure 1.15: Laborproductivity growth of the averageSalvadoran firm under optimisticcounterfactual assumptions(improvementsto 50th,75'hand 90thpercentiles) finance , infrastructure I I governance I I technology Th I I ~ I. .- total IC effects Observed Labor % Productivity 0.0% 20.0% 40.0% 60.0% 80.0% 100.0% 120.0% 140.0% l m e d i a n Bd75th perc. 0 9 0 t h perc. Source: Investment Climate Survey Figure 1.16: Laborproductivitygrowth of the average Salvadoran firm assuminginvestmentclimate improvementsto the country's 75thpercentile,by size range micro (111,5%) small (95.3%) medium (65.8%) large (33.2%) Bfinance Clinfrastructure Dgovernance Dtechnology 1 Source: Investment Climate Survey 35 41. Since smaller firms are farther away from national best practices in the various investment climate areas, they stand to gain more from the reduction in within-country differences in investment climate conditions. Progress in the areas of infrastructure and technology are relatively more important for smaller firms. Thus, micro and small and mediumenterprises (MSMEs)have muchmore to gain interms o f labor productivity growth as a result o f improvements in the investment climate (Figure 1.16). Thus, it i s estimated that in the counterfactual assumption o f generalized upgrades to the country's 75th percentiles o f the four investment climate indicators, large firms would experience a 32.4% labor productivity increase, compared to respectively 112%, 95% and 65% for micro, small and medium firms, respectively. The sources o f those gains vary with firm size, with technology and infrastructure responding for a larger share o f the total productivity gain estimated for smaller firms (e.g. 62% for microenterprises) and progress in governance and finance being responsible for a larger share o f the total productivity effects for large firms (61%). 42. All the indicators used to compute the indexes of finance and technology and most of those used for governance and infrastructureincrease monotonicallywith firm size (table 1.1). The larger the firm, the more likely it is to use bank finance and supplier credit. As an example, almost 70% o f large firms report having an overdraft account and using banks to finance their investments, compared to respectively 16% and 28% for microenterprises. Similarly, the provision o f training to workers, the use o f computers and automated production equipment, and the presence o f research and development activities and I S 0 certifications are all much more common in larger firms. In the case o f infrastructure, the losses due to power outages, the number o f interruptions in water services, and the difhsion o f information and communication technologies also tend to increase with firm size, but the losses associated with the transport system are higher among larger firms. Finally, with regard to governance and insecurity, we find that crime losses are larger among smaller firms, and so i s the distrust o f the judicial system and the percentage o f sales that go undeclared for tax purposes. Smaller firms also face a heavier regulatory burden, when measured by the number o f inspections by government officials, normalized by employment size. Bribes, however, represent a higher percentage o f sales for medium firms, followed by micro and small enterprises. 43. The sectors of apparel and non-metallic mineralproducts, and the smaller provinces of El Salvador, as well as those of Santa Ana and Ahuachapan stand to reap the largest gains from improvements in the investment climate. The firms located in La Libertad as well as those that produce chemical products are closest to the national investment climate best practices. The differences across sectors and regions interms o f the potential productivity gains derived from improvements in the investment climate are considerable, even if one takes into account the different size distribution o f firms (figures 1.17 and 1.18). Indeed, assuming that each firm reaches the best practice o f its corresponding size range (the 75th percentile o f the country in that size range), the average labor productivity gain inthe sectors o f apparel and non- metalic mineral products i s close to 70%, compared to about 50% among manufacturers of chemical and rubber and plastic products. Productivity increases o f the same order o f magnitude are found respectively for the firms located in Santa Ana, Ahuchapan and other small provinces (70% gain), and those operating in L a Libertad (gain o f about 50%). Averages by sector and province for the components o f the various investment climates indexes are reported in tables 1.2 and 1.3. The largest differences in potential productivity gains across different sectors and 36 provinces arise in the area o f technology, followed by governance in the case o f sectors, and infrastructure for that o f provinces. Figure 1.17: Laborproductivity growth of the average Salvadoran firm by sector,assuminginvestment climate improvementsto the 75'bpercentileof the correspondingsize range 119.7% I H n n n 21.3% 20.4% - n 12.1% 19.8% H H 18.6% H 15.3% 14.1% 15.6% 12.7% 11.6% T Mineral Prod. Apparel Textiles Metal Prod. Food & Bev. Rubber& Chemicals (72.9%) (69.9%) (61.4%) (56.8%) (56.8%) Plastics (48.8%) (53.1%) Bfinance infrastructure Ogovernance 0technology Source: InvestmentClimateSurvey Figure 1.18: Labor productivity growthof the average Salvadoranfirm by province, assuminginvestment climateimprovementsto the 75'hpercentileof the correspondingsize range Santa Ana Other Ahuachapan San Salvador San Miguel La Libertad (76.3%) (71.4%) (67.2%) (61.2%) (55.5%) (49.3%) Dlfinance infrastructure 0governance 0technology i Source: InvestmentClimate Survey INTERNATIONAL SIMULATIONS 44. In the context of the Free Trade Agreement negotiated between the United States and Central America, it may be useful to use El Salvador's neighboring countries as benchmarks to evaluate the areas of the investment climate that pose greater competitive challenges to Salvadoran firms. For that purpose, we now report the results o f a second group o f counterfactual simulations, based on econometric estimates obtained with a pooled sample o f 37 firms surveyed inEl Salvador, Guatemala, Honduras and Nicaragua (table A.2). The basic idea i s that o f calculating the differences in investment climate indexes found between firms from El Salvador and their counterparts from those three countries, and estimating the impact o f those differences on labor productivity. Figure 1.19 reports the effects on labor productivity o f the differences between average Salvadoran investment climate indexes and their equivalents in the three above mentioned neighboring countries. A similar comparison is reported inFigure 1.20, this time usingdifferences inbest practices - measured as the 75thpercentile of each country's indexes - instead o f average values to calculate the labor productivity advantages or disadvantages o f Salvadoran firms. Figure 1.19: Labor productivity advantagederived from investmentclimate conditionsfaced by the average Salvadoranfirm, comparedwith average conditionsin Guatemala, Honduras and Nicaragua 34.6% Guat., Hond. & Nic. Nicaragua h 18 0% Honduras 1 , * 7 0% 11.6% 1Eifinance minfrastructure Ooovernance Otechnolow .total (InvestmentClimate) mobserved Labor Productivity1 Source: Investment Climate Survey 45. On average Salvadoran firms report substantially better investment climate indicators than their counterparts from Honduras and Nicaragua, but the differences with Guatemalan firms are much smaller, and the latter are ahead in the area of technology. The average Salvadoran firm has considerable productivity advantages with respect to Nicaraguan and Honduran firms, derived from better indicators in all four areas o f the investment climate. However, the average Guatemalan firm is ahead of its Salvadoran counterpart in the area o f technology, which more than compensates the effects o f the better infrastructure, governance and finance indicators reported by Salvadoran firms. Thus, the latter are estimated to have a small investment-climate-related productivity disadvantage with respect to their Guatemalan counterparts (Figure 1.19). In contrast, as a result o f better investment climate indexes Salvadoran manufacturers are estimatedto have labor productivity advantages o f respectively 26% and 49% with respect to Honduran and Nicaraguan firms, which explains 85% and 67% o f the observed productivity differences. Similarly, when the comparison i s made between hypothetical firms in that operate at the level o f the 75`h percentile o f each country's investment climate indexes (Figure 1.20), we find that Salvadoran firms have considerable productivity advantages driven by relatively better investment climate conditions. As before, the 38 productivity advantages are largest with respect to Nicaragua (89%), followed by Honduras (48%) and Guatemala (7%). All in all, this analysis suggests that the main challenge faced by Salvadoran firms i s that o f catching up with respect to their Guatemalan competitors in the area o f technology. In particular, as seen in table 1.4, Guatemalan firms appear to be ahead o f their Salvadoran counterparts in terms o f the provision o f training to workers and the realization o f research and development activities. Figure 1.20: Labor productivityadvantage derived from investmentclimateconditions,comparingEl Salvador's 75thpercentiles with 75'hpercentilesof Guatemala,Honduras and Nicaragua Guat., Hond. & Nic. 88.7% Nicaragua Honduras g.8% -7.3 h Guatem 10.1% [wfinance infrastructure ngovernance 0technology total (Investment climate)w ObservedLabor Productivity 1 Source: InvestmentClimate Survey 46. Large Salvadoranfirms enjoy bigger productivityadvantages due to investment climate conditions than do micro, small and mediumenterprises (MSMEs). When the comparison is made with Guatemala, only large firms enjoy a productivity edge as a result o f better investment climate conditions (Figure 1.21). Similarly, while Salvadoran firms o f all sizes enjoy better investment climate conditions than their counterparts from Honduras and Nicaragua, the corresponding productivity advantage tends to increase with firm size. Looking in more detail at the comparison with Guatemalan firms (Figure 1.22), we find that Salvadoran MSMEs have a sizable disadvantage in the area o f technology, as well as a slight disadvantage in infrastructure. Inaddition, Salvadoranmicroenterprises also face worse conditions interms ofaccess to finance. Detailed cross-country comparisons o f the sub-components o f our investmentclimate indicators for large firms and MSMEs are reported intables 1.5 and 1.6 respectively. 39 Figure 1.21: Labor productivity differencesderivedfrom investmentclimate conditions,by size (comparison with average firm of the same size in Guatemala, Hondurasor Nicaragua) 36.3% 2 7 . 3 % 27.0% 23.5% U large medium small m icro -6 8 % U -15 0% -13 4 % =Guatemala =Honduras I N i c a r a g u a I G u a t . . Hond. 8 Ntc. Source: InvestmentClimate Survey Figure 1.22: Laborproductivity differenceswith Guatemala derivedfrom investmentclimate conditions,by size (comparisonwith averagefirm of the same size) 15.3% U -14 4% -18.8% IEIfinance Binfrastructure Oaovernance Dtechnoloav .total (Investment Climate) 1 Source: Investment Climate Survey 47. In all but one manufacturing sector, investment climate conditions are estimated to put Salvadoran firms at a disadvantage with respect to competitors from at least one of its neighbors. As seen in Figure 1.23, the largest challenges are associated with competition from Guatemalan firms, which in five out o f seven sectors enjoy better investment climate conditions than their Salvadoran counterparts. The latter are also at a disadvantage with regard to Honduran manufacturers o f textiles, apparel and rubber and plastic products, and with respect to Nicaraguan producers o f non-metalic mineral products. A more detailed analysis o f the sources o f sectoral productivity differences between Guatemalan and Salvadoran firms (Figure 1.24) indicates that the main challenges are to be found inthe area o f technology, followed by those o f infrastructure and finance. 40 Figure1.23: Labor productivity differencesderivedfrom investmentclimate conditions,by sector (comparisonwith average firm of the same sector in Guatemala, Honduras or Nicaragua) 71.9% InGuatemala =Honduras lNicaragua I G u a t . , Hond. B Nic. 1 Source: Investment Climate Survey Figure 1.24: Labor productivitydifferenceswith Guatemala derivedfrom investmentclimate conditions,by sector (comparisonwith average firm of the same sector) 16.2% II3 finance Dinfrastructure Ogovernance Otechnology .total (Investment Climate)1 Source: Investment Climate Survey 48. Overall, our intra-national and international simulations of the potential gains in manufacturingproductivityderivedfrom improvements in the investmentclimate suggest the followingconclusions. 1. There are important potential productivity gains associated with improvements in all the four areas o f the investmentclimate considered inthis report. 41 2. The largest potential productivity gains from generalizing national investment climate best practices are obtained inthe area o f technology. 3. Micro, small and medium enterprises stand to benefit the most from the generalization to all firms o f national investmentclimate best practices, particularly in the areasof infrastructure andtechnology. 4. The sectors o f apparel and non-metallic mineral products and the provinces o f Santa Ana and Ahuachapan - among other small provinces - are the ones that stand to gain the most from improvements inthe investment climate. 5. El Salvador i s well ahead o f Honduras and Nicaragua in all areas o f the investment climate, but trails behindGuatemala inthe areas o f infrastructure and, to a much larger extent, inthat o f technology. 6. Compared to large firms, micro, small and medium enterprises enjoy smaller advantages with respect to their counterparts from Honduras and Nicaragua, and have larger gaps with respect to Guatemalan firms. 7. In all but one manufacturing sector (metal products), investment climate conditions are estimated to put Salvadoran firms at a disadvantage with respect to competitors from at least one o f its neighbors (Guatemala, inmost cases). 49. The above findings suggest that policies aimed at increasing productivity by means of improvements in the investment climate should cover the four areas that are addressed in this report. The next chapters will analyze inmore detail the results of the investment climate survey in each o f those four areas, and combine them with other sources o f information in an attempt to provide a diagnosis o f the main policy challenges. At the end o f each chapter, we will provide a series o f specific policy recommendations that could be considered in the context o f efforts to improve the Salvadoran investment climate. THEPOTENTIALBENEFITSOFTHEFREE TRADE AGREEMENTWITH THE U.S. 50. Free trade agreement negotiations between the United States, El Salvador, Guatemala, Costa Rica, Honduras and Nicaragua began formally in January 2003. Nine rounds o f negotiations were completed over the course o f the year, culminating in the signing o f the agreement by four o f those four Central American nations on December 17, 2003. Costa Rica signed onto the agreement a month later, on January 26,2004. The Central American Free Trade Agreement (CAFTA) i s a second-generation trade agreement, similar to NAFTA and the U.S.-Chile FTA. In other words, it does not cover free trade o f goods only, but rather stipulates free trade for all sectors, including services. It opens market access by locking into low tariffs with clearly defined rules of origin. Relevant market sectors include agriculture, manufacturing and services. In addition to free trade o f goods and services, the agreement contains provisions regarding non-market access issues, such as labor and environmental regulations, intellectual property rights, and dispute settlement. The adoption o f CAFTA generates significant benefits 42 for Central American countries, such as making access to the U.S. market more permanent, thereby creating an irrevocable anchor to free trade. Under the current trade regime, the Caribbean Basin Initiative (CBI), these nations enjoy nearly equal access to U.S. markets, but due to the unilateral nature o f the agreement, long term investors cannot be confident that preferential trade status will undoubtedly continue. CAFTA also grants exporters and investors a fkller set o f protection and dispute resolution rules. Furthermore, the agreement acts as a catalyst to reform in the regionby instilling an obligation on the C A countries to upgrade their existing rules and institutions regarding investment and trade (such as improving protection for intellectual property rights and labor and environmental regulations). These changes are expected to improve the business climate and thus lead to higher foreign and domestic investment. Yet, before CAFTA can be implemented, it faces a final daunting obstacle -ratificationby the U.S. congress. 51. CAFTA is likely to generate bothstatic and dynamic gains for the Salvadoran economy, providedthat complementary policies are adopted to facilitate the adjustment process and attract foreign direct investment (FDI). In theory, static gains from trade liberalization are potentially associated with trade liberalization in general and with those types o f agreement in particular. Those gains result from increased levels o f specialization according to comparative advantage, as factors o f production move towards the sectors in which the country's relative productioncosts are lowest, and returns are highestafter the trade policy reform. Inthe short run, such an adjustment process necessarily leads to the emergence o f winners and losers, as workers and entrepreneurs involved in the production o f goods whose relative prices fall after trade liberalization face the challenge o f transiting towards the sectors in which the country has comparative advantage (and relative prices increase after trade liberalization). The extent to which this reshuffling o f resources takes place determines the size o f the static gains derived from trade liberalization. During the process o f adjustment, specific policies may be needed to help those workers and entrepreneurs that are negatively affected by changes in relative prices adjust efficiently to the new incentives that are brought about by trade liberalization. As put by Jaramillo and Lederman (2004), the workers that will bear the costs o f adjustment by having to change their economic activities need to be compensated appropriately for their efforts, so that the necessary adjustment takes place, and the static gains from trade are materialized. Besides policies directed at facilitating the adjustments o f those that appear as losers, at least in the short term, an efficient adjustment process also requires improvements in the overall investment climate, so that factors o f production can move towards the sectors inwhich their productivity is highest, without having to incur in additional costs associated, for instance, with excessive regulations, rent seeking activities, security problems, deficiencies in infrastructure, limited access to finance, or lack o f adequately trained workers. Similarly, a good investment climate i s indispensable for the materialization o f the dynamic gains from trade that are potentially associated with the adoption o f foreign technologies and the increase in foreign direct in~estment.~' 52. Mexico's experience with NAFTA highlightsthe importanceof complementarypolicies in order to maximize the benefits of regional free trade agreements. A recent study on the effects o f NAFTA on Mexico's economic performance found that Mexico's global exports would have been 25% lower and FDI about 40% less without NAFTA. In addition, NAFTA 25 See Jaramillo and Lederman (2004). 43 contributed to cut in half the time required for Mexican manufacturers to a adopt US. technological innovations. In more general terms, however, the study found that while NAFTA helped achieve moderate declines in poverty, and helped generate more and better jobs for Mexican workers, its effects were not sufficient to ensure economic convergence among North American regions and countries. To some extent this can be attributed to some limitations o f NAFTA's design upon which future trade agreements could improve - particularly with regard to rules o f origin, as well as anti-dumping and countervailing duties. More importantly, however, NAFTA's experience suggests that significant policy and institutional reforms are necessary to realize all the potential benefits o f an FTA, encompassing improvements in the overall investment climate, in areas such as governance, education and technology. In addition, to maximize the gains from trade liberalization, and minimize the possible costs from trade diversion, the Mexican experience indicates that regional trade integration should ideally be accompanied by unilateral, bilateral and multilateral actions on other trade fronts. Indeed, Mexico's impressive gains in export market share reflected, to a large extent, its unilateral trade liberalization since the 1980s. Moreover, higher initial tariffs may not have significant benefits in terms of helping negotiate better market access to industrialized countries, and can in fact have the negative effect of reducing the credibility associated with the country's willingness to implementfurther trade reforms.26 53. Empirical cross-country evidence also suggests that complementary policies related to improvements in the investment climate are essentialto maximize the benefits of free trade agreements (FTA) in terms of higher levels of income and economic growth. A recent paper by Bolaky and Freund shows that in countries characterized by high levels o f labor and entry regulations increased trade i s not necessarily associated with higher levels o f income and growth, and infact trade can actually be harmful inthose conditions. Excessive regulations may not only impede the relocation o f resources according to comparative advantage but also, in a worst case scenario, lead to increased production o f goods in which the country does not have comparative advantage, with a consequent reduction in welfare. As put by the authors, "not only do bad institutions lower growth, but they also prevent trade from generating growth." 27 26For a summary of research findings on the consequences o f NAFTA on the Mexican economy, and an analysis of the lessonsfor LatinAmerica andthe Caribbean, see Lederman et al. (2003). 27Bolaky and Caroline Freund (2004), p. 4. 44 Figure 1.25: Firms' Awareness of Trade Treaties and Expectations about their Effects (El Salvador, Guatemala, Honduras and Nicaragua) 96.3% Aware of CBI Benefited from Aware of CAFTA w c t s direct Expects Positive Comparison bstussn 2888, 1998 (top-botton order) 60. Data recently collected by the World Economic Forum (WEF) places the quality of governance in El Salvador above that of the average Central American country, but suggests that the country's world rankingis lowest with regardto bribery in the judiciary. Moreover, this survey suggeststhat ElSalvador i s very close to itsneighborsinterms of the incidence of state capture, illegal political financing and undue influence. In 2003 about 7743 enterprises were interviewed in 102 countries as part o f a WEF study on business environment - the Executive Opinion Survey (EOS).37 The firms interviewed provide an assessment o f governance along several key sub-components - from diversion o f public goods to bribery in taxation. On most governance dimensions, El Salvador displays better quality o f governance than the average Central American country (Figure 2.5). However, the differences are much smaller in the sub-components associated with state capture, illegal political financing and undue influence, for which the answers o f Salvadoran firms are quite close to those o f firms from other Central American countries. Moreover, when comparing the world percentile rank o f El Salvador with that o f its neighbors (Figure 2.6), we find that the sub-component inwhich the 36The Economist Intelligence Unit, "Country Monitor 16 August 2004 Main Report." 37For more information on the Executive Opinion Survey (EOS) see: http://www.weforum.org/site/homepublic.nsf/Content/Global+Competitiveness+Programme. 61 country performs worse and i s thus closest to its neighbors i s that o f bribery in the judiciary, in which El Salvador i s inthe 33rd percentile o f the group o f 102 countries covered inthis survey - compared with an average ranking o f 59 for the other sub-components. Relatively low rankings (close to the 43'd percentile) are also obtained in the above mentioned areas o f state capture, illegal political financing and undue influence. These results suggest that besides the problems associated with administrative corruption on which the investment climate survey provides detailed information (see below) - e.g. bribes in permits or government procurement - El Salvador also faces considerable challenges associated with corruption in the judiciary and the use o f political power for private gain. Figure 2.5: Percentage of Firms Reporting Low Governance (EOS 2003) Source: World EconomicForum (EOS 2003). Figure 2.6: Governance percentile ranks among 102 countries (EOS 2003) loo r Central Am erica Source: World EconomicForum(EOS 2003). 61. The InvestmentClimate Surveys suggest that corruption is lower in El Salvador than in neighboring Guatemala, Honduras and Nicaragua, but also that important challenges remain with regard to the need to improve the quality of governance in El Salvador. The cross-country data presentedabove suggests that El Salvador faces important challenges interms o f reducing corruption and more generally increasing the quality o f governance to levels that are closer to those found in industrialized countries and even among the larger Latin American and East Asian countries. Still, those challenges appear to be smaller than in neighboring Central American countries. The I C surveys performed in the region offer a similar, though more in- 62 depth, picture o f the reality faced by private firms, based on interviews performed with 465 manufacturing firms in El Salvador and more than 1,300 firms in Nicaragua, Honduras and Guatemala. Figure 2.7: Inconsistency and unpredictability of regulations,governmentinefficiencyand bribery 75.6 El Salvador Nicaragua Honduras Guatemala I b#jGovernment considered Inefficient Interpretations of regulations inconsistent and unpredictable Bribes "to get Things Done" considered common I Source: Investment Climate Survey 62. Salvadoran firms are less likely to describe the government as inefficient, to state that government regulations are interpretedin an inconsistent and unpredictableway, and to report that bribes "to get things done" are a common practicein their sector. Only 30% of the firms surveyed in El Salvador believe that the government is inefficient in its provision of public services, compared to almost 50% o f the firms in Honduras and Nicaragua, and 76% in Guatemala (Figure 2.6). Similarly, Salvadoran firms have more confidence in public officials than do their counterparts from other Central American countries: only 35% o f the firms state that public officials do not interpret government regulations in a consistent and predictable way, compared to about 45% in Honduras and Nicaragua, and 71% in Guatemala. This suggests a relatively lower degree o f discretionality on the part o f Salvadoran public officials, so it i s not surprising to find that the percentage o f firms that report that bribery is a common practice in their sector of activity is also lower in Salvador: 45%, compared to 50% inNicaragua, 56% in Guatemala and 63% in Honduras. Still, it i s important to emphasize that while lower than in neighboring countries, administrative corruption seems to affect, directly or indirectly, almost half o f the firms surveyed inEl Salvador. Figure 2.8: Bribes"to get things done" and to win public contracts 9.0 Source: InvestmentClimate Survey 63. The resources spent by private firms to bribe public officials are significant but generally smaller in El Salvador than in other countries for which comparable data is 63 available. The I C survey suggests that Salvadoran firms spend on average 1.7% of their revenues on bribes "to get things done". The same or very similar figures are found inNicaragua (1.7%), Honduras (1.9%) and China (1A%), while higher values are obtained for Bangladesh (2.5%), Ecuador (2.6%) and Guatemala (3.2%). Corruption related to public procurement is also a serious challenge, as highlightedinthe Country Procurement Assessment Report (World Bank, 2004). Indeed, the survey data indicates that in order to secure contracts with the government, Salvadoran managers pay bribes equivalent to 3.7% o f the contract on average. This i s slightly less than what was reported by firms in Nicaragua and Guatemala (bribes o f 4.4% o f public contracts) and about the same than in Honduras. In contrast, firms surveyed in China and Bangladesh report procurement bribes equivalent to respectively 2.1% and 4.3% o f government contracts. Figure2.9: Percentageof firms that paidor were offered bribesto obtain governmentservices (subset of firms that requestedthe correspondingservices) Import License (123 firms) 1 6.5% ConstructionPermit (47 firms) I 6.4% Sanitary Registration (133 firms) 3.0% Electrical Connection(75 firms) 2.7% Renew Health Minishy Permits (99 f m s ) 2.0% Renew Operating License at Alcaldia (21 1firms) 1.4% lntemal Revenue Service Inspection(244 firms) Mainline Telephone Connection (179 firms) Social Security Institute Inspection(179 firms) 0.6% Source: Investment Climate Survey 64. A more detailedanalysis of administrativecorruptionshows that in El Salvador import and construction permitsare the procedures most affectedby corruption.About 6.5% of the firms that applied to those two government permits were suggested the possibility o f making informal payments to speed up the procedures (Figure 2.9). Other permits or public services for which bribes were reported by Salvadoran firms are sanitary registrations (3%), electrical connections (2.7%) and permits issued by the Health Ministry (1.4%). Finally, bribes were reported by less than 1% o f the firms that dealt with phone installations and inspections by the tax authority or the social security institute. 64 Figure2.10: Bribesto obtain governmentservices, internationalcomparison Some Bribe (26 perm Iic. insp) I6 1 ImportLicense Construction Permit Sanitary Registry Electrical Connection 21 7 Health Ministry Inspection Operating License (Alcaldia) lntemal Revenue Service Insp Mainline Telephone Connection Social Secunty Institute Insp Source: InvestmentClimate Survey 65. While the incidence of bribes to obtain government permits and public services warrants justifiable concerns on the part of Salvadoran policy makers, the frequency of those informal payments is much smaller in El Salvador than in the three neighboring countries for which comparable data is available. Administrative corruption appears to be a smaller problem in El Salvador than in neighboring Guatemala, Honduras and Nicaragua. Indeed, as seen in Figure 2.10, when asked about the presence o f corruption in the procedures needed to obtain 26 permits or public services, only 5.2% o f Salvadoran firms report having been offered to pay bribes to expedite at least one o f them (Figure 2.10), compared to an average o f about 15% inthe three other countries. Moreover, for all the permits and services for which there i s evidence that some administrative corruption exists in El Salvador, the incidence o f bribe payments was found to be much higher in at least one o f - if not all - the three above mentioned countries. As an example, bribes are reportedby about 10% o f the Honduras firms that requested import permits, 16% o f the Guatemalan firms that applied to construction permits and almost 22% o f the Nicaraguan firms that requested electrical connections. Figure2.11: Bribesto obtain government services, by size Some Bnbe (26 permAic.linsp.) ProcurementBribe Bribes (% of sales) Consmction P m i t I S 2 lmpon License HeahhMinistryInspection ElectricalConnection Sanitary Regintry lntemal RevenueServiceImp. SocialSecurityInstituteInsp. Mainline TelephoneConnecnon OperatingLicense(Alcaldia) IPJMicro PJSmall UMedium .Large ~ Source: InvestmentClimate Survey 65 Figure 2.12: Bribes to obtain government services, by province Some Bnbe (26 permllic imsp ) Procurement Bribe Bribes (%ofsales) hport Lcense 13 3 HealthMinistry hspection Sanitary Registry Operating bcense (Alcaldia) hternal RevenueService hsp Construction Permit ElectricalConnection Social Security Institute hsp OSanSalvador 01.Libcrtsd OReii ofthecountry Source:InvestmentClimate Survey 66. Within El Salvador, small and especially medium-sized firms tend to be more affected by administrative corruption and spend a larger share of their revenues in bribes. The share of sales spent in bribes varies from 0.9% for large firms to 2.3% for medium firms, with micro and small firms spending respectively 1.5% and 2% (Figure 2.1 1). In the case o f public procurement, the average share o f public contracts that i s reportedly paid as a bribe by medium firms i s 5.4%, compared to between 2.6% and 3.8% in other size ranges. Similarly, 10% o f all medium firms report having been offered a bribe to obtain at least one permit or public service, compared to about 3% for micro and small firms and 4.5% for large companies. The bribes offered to large firms, however, are restricted to import licenses, while at least some medium firms report having been offered to pay a bribe to obtain each o f the permits and services listedin Figure 2.11. Figure 2.13: Bribes to obtain government services, by sector 5.7 11 1 1 . 9 4 1 3 s ' OBnbes"to get Thines Done"(%sales) ProcurementBnbe 1 Source:InvestmentClimateSurvey 67,Administrative corruption appears to be more frequent in San Salvador and La Libertad than in the rest of the country. The incidence of corruption also varies with the firms' sector o f economic activity and with their export status. As shown in Figure 2.12, procurement bribes appear to be more costly in the capital city - 4.4% o f contracts, compared to 66 at most 3% in other provinces - the average fraction o f sales that i s spent on bribes i s higher in L a Libertad (2.1%) than in the rest o f the country (1.6 percent). Similarly, for most o f the agencies that are associated with some degree o f administrative corruption, the fraction o f firms affected in L a Libertad i s higher than in San Salvador and the rest o f the country, the only exceptions being construction permits, electrical connections and social security inspections, for which the only cases o f corruption are reported in San Salvador. As for the variation across sectors, we find that manufacturers o f textile products are the ones that spend the most on bribes to get things done, but procurement bribes are most frequent in the sectors o f metal products, chemicals and apparel (Figure 2.13). Finally, there i s a significant difference in the incidence o f corruption among exporters and non-exporters, as well as between maquilas and non-maquila exporters (Figure 2.14). Indeed, firms oriented exclusively to the domestic market spend about 1.9% o f their sales on bribes, compared to 1.2% for maquilas and 1.5% for other exporters. Not surprisingly, maquilas are less exposed to corruption inpublic procurement but they report much more frequent informal payments with regard to import permits. As for the bribes to obtain construction permits, they are reported only by non-exporters. 68. The econometric analysis of the empiricaldeterminantsof the practiceof payingbribes "to get things done" yields both expected and unexpected findings. The main non- surprising result is that this type of corruption affects to a larger extent the firms with a greater need of public services and those that operate in more heavily regulated sectors or locations. Both among Salvadoran firms and in a sample that also encompasses companies surveyed in Guatemala, Honduras and Nicaragua, we find that the firms that had more interactions with government officials or providers o f public services are more likely to report that administrative corruption i s common in their sectors, and they also report that bribes consume a larger fraction o f their sales. This result provides support to the hypothesis that bribe payments are the result o f bargaining processes between firms and corrupt public officials, in which the firms that need to deal with the public sector to a greater extent are in a particularly weak negotiating position and as a result pay more and higher bribes.38 Figure2.14: Bribesto obtaingovernmentservices, by export status Some Bribe (26 perm./lic./insp.) Procurement Bribe 1 1 I " ConstructionPermit h p o r t Lcense "on-exporters mExporters (non-maq ) OMaquilas ' Source: InvestmentClimate Survey 38Our econometricanalysisfollows Svenson (2003) andis describedinthe appendixto this chapter. 67 BOX 2: Korea's Move to E-Procurement In 1997 the Korean govemment began reforming its notoriously complicated, nontransparent, corrupt public procurement system, introducing e-procurement to exploit the country's well-developed information and communications infrastmcture. Through extensive business process reengineeringand informationstrategy planning, the Public Procurement Service -the agency formerly responsible for buyinggovernment goods and services has been transformed into a one-stop information center. E-procurement has generated numerous benefits including: enhanced transparency andpublic trust -by reducingcontacts between officials and suppliersandby sharinginformationbetween govemment agencies and the public - and increased managerial efficiency - by achieving economies of scale in procurement,with an estimated $2.5 billion a year insavings from the $26 million investment. Before e-procurement was introduced, legal and regulatory requirements resulted in complicated procurement procedures. Companies seeking govemment business had to scan newspapers, magazines, and government gazettes to identify opportunities, then register separately with each govemment agency. Moreover, procuring agencies did not share information. Instead, each maintained its own register of suppliers. This time-consuming, costly application process discouragedsuppliers, and prearrangedbiddingwas not uncommon- contributing to corruption. The process also limited economies o f scale. Although a few ministries had computerizedprocurementsystems when reformbegan in 1997, sophisticationwas low. And becausethese standalone systems were incompatible,they couldnot handle large volumes of procurement. Meaningful change began only when former PresidentDae Jung Kim initiated comprehensive e-govemment reform, vowing in his 1998 inaugural speech to have state-of-the-art systems in place by the end of his term in 2002. A PresidentialE-GovemmentCommitteewas createdto oversee e-govemmentreform, which involved 11 major projects: the National Pension System, Intemal Tax Service, Integrated Local Administration System, National Education Information System, Financial Management Information System, Human Resource Management system, Electronic Data Interchange, e-Signature System, Government wide Information System, Government for Citizens (known as G4C), and e-Procurement. E-procurementreformhas reliedon carefulsequencingandpacing. After reformwas approved, the Ministry of Planningand Budgetconductedan innovativebusiness process reengineeringto translateits work processesinto an online system. The ministry also established a project managementunit that undertook an informationstrategy planningexercise, developingactionplansto automate the recommendedprocesses. The exercise emphasizedthe importanceof automatingthe most commonpurchasesfirst and o f actively involvingusers inthe system's design, to ensure its smooth and swift acceptance.The e-procurement systemgives govemment agenciesand suppliers a single pointof contactfor registrationandinformation. The systemhas also expandedthe selectiono fproductsand standardizedtheir classifications. Moreover, it involvessimpler documents and faster payments.The system is linked to more than30 procurement-relatedextemalagencies-includingsupplier certificationagencies, financial clearing institutes, the Ministry o f IntemalAffairs, the Ministryof Finance,and e-guaranteeand e-payment systems. Encryption technologyanddigitalcertificationensure secure transactionsandprotect online information. 1 Source: WorldBank (2004e), Korea's Move to E-Procurement. 69. W e also find that firms that have low confidence in the judiciary and those that receive information on government regulations from industry associations are more likely to pay bribes. The first result i s far from unexpected, as firms that believe that their rightswould not be adequately protected by the judiciary are understandably less likely to refuse offers from corrupt officials. The second result is more surprising, as one would expect that the firms' bargaining position could be strengthened - and not weakened, as the results seem to suggest - by their access to information providedby chambers o f industry or commerce. One possible explanation i s that the information collected by industry associations i s positively related to the firms' awareness o f corrupt practices in their sectors - which i s our actual empirical proxy for the incidence o f corruption -but not necessarily with their direct involvement inthose practices. 68 70. Our econometric estimates also indicatethat larger companies are morelikely to report administrativecorruption in their sectors, but smaller firms report larger bribe payments as a percentage of their sales. Firms with larger capital stocks report a lower prevalence o f bribe payments, and so do more profitable Salvadoran firms. The negative relationship between profits and bribe payments i s specific to El Salvador, as the opposite finding - profitable firms paying more bribes - was obtained for the other three countries. On the other hand, in all four countries it appears that larger firms are more targeted by corrupt public officials, but the value o f bribes does not increase inproportion with firm size. Moreover, for a given employment size corruption appears to affect to a lower extent the more capital intensive firms, possibly because they have greater means to defend themselves from corrupt officials, so that the latter fear a larger probability o f getting caught when dealing with those firms. POLICYRECOMMENDATIONS ON CORRUPTION 71. While administrative corruption appears to be a lesser obstacle for the private sector than in neighboringGuatemala, Honduras and Nicaragua, corruption in El Salvador also has a significant effect on the costs of doingbusinessin the country, and may thus affect the country's attractiveness to foreign investors. Import licenses, construction permits and transactions that involve public procurement are especially costly and burdensome for firms inEl Salvador. Thus, it would be important to increase transparency in the affected public agencies, clarifying the requirements (in term o f costs and time) to complete the corresponding transactions. Moreover, to increase the accountability o f civil servants working in public agencies, the government should also consider the introduction o f a system o f score cards (compiled by business people) for these services, whose results should be published regularly in local newspapers. 72. Managersdealing with the public sector and procurementpay an additionalcost due to corruption and limitedinformationavailable. As emphasizedinthe CPAR, it is important to simplify the procurement process and make it more transparent. INthis area, El Salvador can benefit from the experience o f other developing countries that have successfilly modernized their procurement procedures (see box 2 on Korea's experience). It will also help to establish internal and external auditing mechanisms in key public agencies dealing with public contracts. These objectives could be achieved with the creation o f a public source o f information available on the internet3' The geographical heterogeneity o f corruption and poor governance should also be addressed. Businesses in a few departments report a much higher cost o f corruption than in other parts o f the country. In response to this challenge, it would be important to strengthen accountability and transparency o f public institutions, especially in L a Liberdad and San Salvador. 73. Our econometric results suggest that collective action on the part of the business community - for instance through industry associations - may also be a good strategy to reduce the incidence of administrative corruption. Indeed, the finding that corruption is more frequent among firms with more dealings with the public sector i s consistent with bribe payments being the result o f negotiations between firms and corrupt officials. This, in turn, warrants the conclusion that actions that strengthen the bargaining position o f individual firms 39An example is the systemcreated by the Governmentof Peru:see www.inei.nob.ue. 69 could help reduce corruption. As argued by Svenson (2003) in a similar context, examples o f such initiatives, involving collective action by the business community, include the collection and dissemination o f corrupt practices, and the provision o f information on service standards, guidelines and norms o f major service providers. 74. The finding that a greater distrustin the judicial system is linked to a larger incidence of administrativecorruption impliesthe need for measures aimed at improvingthe honesty and efficiency of the judiciary. A perception of generalized impunity can increase the aggressiveness o f corrupt officials, and reduce the incentives o f firms to refuse to pay bribes.As we will argue in the last section o f this chapter, the lack o f transparency and irregularities in court proceedings affect the quality and efficiency o f the court system, which can reduce its effectiveness inpreventing and prosecuting ~orruption.~' CRIME 75. El Salvador exhibitsvery high homicide rates, even by Latin American standards. With the exception o f Colombia, Latin America is now mostly free o f the political violence and ethnic conflicts that pervade sub-Saharan Africa. Yet, according to the World Health Organization the low and middle income countries o f the Americas still exhibit the highest rates o f intentional homicide o f the world: 27.5 homicides per 100,000 population in2000, compared to 22.2 for the African region, 14.8 for low and middle income European countries and less than 10 for high income countries and other regions o f the developing world (see Figure 2.15).41 In 2003 El Salvador had 36.5 homicides per 100,000 inhabitants. Thus, even within the highly violent Latin American region El Salvador ranks near the top in terms o f its rates o f intentional homicide, trailing only Colombia and Honduras (Figure2.16). Figure2.15: Homicideratesper 100,000 population,by gender and WHO region IOAll BMales OFemales , , 51.0 Source: Word Health Organization(Krug et al. 2002) 40See Lothian (2003) for a detailed analysis o f corruption within the judiciary in El Salvador. 41See Krug et al. (2002), Ayres (1998) and Buvinic et al. (2002). 70 76. International victimization surveys also place El Salvador among the most insecure countries in Latin America. Whereas the homicide rate is the most common indicator for measuring levels o f criminal violence, it only refers to violence directed towards the physical integrity of individuals, and is inmany cases unrelated to the extent o f criminal violence against property, motivated by economic reasons.42A more direct measure o f the latter type o f violence i s the survey done by the Latinobarometro Corporation, which measures the percentage o f households that report having been the victim o f a criminal act duringthe twelve months prior to being interviewed. While the distribution o f countries by household victimization rates does not follow the same pattern apparent for homicides, El Salvador is also near the top o f the group o f 17 Latin American countries covered by Latinobarometro, ranking second in 1996 and fourth in 2003, behind only Mexico, Venezuela and Argentina (Figure 2.17). Moreover, a more recent survey conducted in early 2004 with the support o f the United States Agency for International Development (USAID) also finds higher rates o f victimization in El Salvador than in other Central American countries.43 Figure2.16: Homicideratesper 100,000 populationfor selected LatinAmerican countries 64 0 Sources: PAHO (2002); Carri6n (2004); Rico (2004); GESO (2004); Acero (2004); Krug et al. (2002). 77. Violent crime is not a new phenomenon in El Salvador, neither can its current level be attributed entirely to the armed conflict that the country endured for over a decade.44 Statistics from the Pan-American Health Organization (PAHO) along with anthropological and sociological studies o f violence in El Salvador show that the country was known for high homicide rates in years prior to the eightie~.~'For obvious reasons, the political conflict at the end o f the seventies and the civil war during the eighties brought the homicide rate to much higher levels. By 1980, that rate had reached the unprecedented level o f 400 deaths per 100,000 inhabitants, but homicides decreased in the mid 1980s when the civil war entered a stagnant phase (see Figure 2.18). In 1987, for instance, there were about 60 homicides per 100,000 42 According to the National Civil Police, around 70 per cent o f homicides can be attributed to social violence and 20 percent to economic violence o f a criminal nature (the rest of the homicides have not been classified). See MinistryofGovernance at: www.gobemacion,gob,sv/NR 43 See Cordova and Cruz (2004). In this survey households were asked about criminal acts o f which they were victims over a period of four months. El Salvador's victimization rate was estimated to be 17.1, behind Mexico's 17.4, but above the rates for other Central American countries, which ranged from 12.8 to 15.2. 44 Cruz and Gonzalez, 1997; Cruz, 2003a 45 Alvarenga (1996), Gonzalez (1997). For example, between the late 1960s and the early 1970s the rate o f intentional homicide in El Salvador was close to 30 deaths per 100,000 inhabitants. 71 population. However, while the levels of conflict-related violence stayed at high but relatively stable levels until the end o f the conflict, the first post-war years witnessed a considerable reached 139per 100,000 inhabitant^.^^ increase in homicide rates: according to a pioneering IDB study, by 1996 the homicide rate had Figure 2.17: Householdvictimization rates (12 months before survey) Cn Source: Latinobarbmetro Corporation. 78. Since the late 1990s, however, El Salvador has experienced a clear downward trend in homicide rates, as well as considerable reductions in household victimization rates. Homicide rates averaged about 42 per 100,000 in 1999-2000 and 36 per 100,000 in 2002-2203. The results of nation-wide victimization surveys done by Universidad Centroamericana since 1993 also suggest a declining trend in economically motivated violence. Indeed, in 1994 about 34% of households reported that they had been victims of violent acts over the four months preceding the survey. By 2001, the victimization rate had reached 16% and, aside from an isolated rebound in 2002, the predominant tendency of crime against property has continued to be a negative one. Figure 2.18: Homicide rates in El Salvador, 1969-2003 5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 011 1 Y 1 9 1 Y 1 9 1 Y 1 Y 1 Y 1 Y 1 Y 1 Y 1 Y 1 Y 1 Y 13 1 Y LU io 6 9 7 1 7 3 7 5 7 7 7 9 8 1 8 3 8 5 8 7 8 9 9 1 9 3 9 5 9 7 9 9 0 1 0 3 I Sources: Adapted from McElhinny and Seligson (2001), with data from Fiscalia General de El Salvador and Cruz, Trigueros and Gonzalez (1999). 46See Cruz et.al., 1997 and Cruz, Trigueros and Gonziilez (2000). 72 Figure 2.19: Crime victimization rates recordedin opinionpolls 0% I I l:: 0%1993 1995 1997 1999 2001 2003 Year Source: constructedusing IUDOP opinion polls for several years. 79. Data from the National Civil Police also reveal drastic reductions in rates of economically motivated crimes. It is also possible to identify a general downward trend in statistics that measure the prevalence o f organized crime. According to National Civil Police data gathered by the UnitedNations Development Program (UNDP), bank robberies and commercial vehicle robberies peaked in 1998, with respectively 35 and 1,258 cases reported to the police. However, the incidence o f these crimes diminished considerably in recent years: in 2000 there were only 10 bank robberies and the number o f commercial vehicles robberies fell to 501 inthat year and to 127 in 2003. Another clear indicator o f the downward trend in violence related to organized crime i s the recent reduction inthe number o f abductions reported nationwide. As seen in Figure 2.20, kidnappings almost doubled between 1997 and 2000, reaching 114 in the latter year (Figure 2.20). However, the incidence o f this type o f crime diminished drastically in the following years, with only 7 cases reported in 2003. This reduction i s arguably the result o f a clear and decisive policy followed by the Salvadoran authorities, particularly after revelations o f alleged police involvement inorganized crime rings (CIDAI, 2000).47 Figure 2.20: Kidnappingsin ElSalvador between 1997 and August 200 I 1 4 0 1 2 0 1 0 0 8 0 6 0 4 9 4 0 2 0 019 9 7 1 9 9 8 1 9 9 9 2 0 0 0 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 ( A r o f A u g u r l ) Source: National Civil Police. 47Since that time the government of El Salvador set inmotiona purgingprocess inthe police force and set down a policy aimedparticularlyat fightingorganizedcrime. 73 80. The latestpolice statistics show continuedsuccess in the fight against organizedcrime, a reductionin the number of economically-motivatedcrimes, but less tangible results in the combat of crime against life and physicalintegrity.A National Civil Police report on rates o f criminal activity from 2001 to 2003 shows that the authorities' efforts to combat economically- motivated crime and organized crime has yielded more tangible results than those aimed at the crimes against life and physical integrity. Table 2.2 compares the number o f crimes reported to the police during the period from mid 2002 to mid 2003 with the preceding 12 month period. These statistics confirm that the activities o f organized crime have been curtailed, as illustrated by considerable reductions in the number o f abductions, car thefts, robberies and commercial vehicle robberies. The main exception to this trend i s the slight increase in the number o f car robberies. Less significant reductions have also been achieved in the number o f extortions and assaults, while crimes against life and physical integrity have recently increased - rape and homicides increased respectively by 10% and 4% over the above mentioned period. Moreover, during the first eight months o f 2004 the number o f homicides reported to the National Civil Police increasedby 16% compared to the same period in2003. Table 2.2: Criminaloffensesbetween June 2001andMay 2003 Crime Jun 01- May 02 Jun 02-May03 Difference Kidnappings 25 18 - 2 8 % Car thefts 3,363 2,525 --- 25 25 % Robberies 6,550 4,948 % Commercial vehicle robberies 551 176 68 % Extortion 364 329 Theft 12,308 11,110 -- 10% 10% Assaults 5,248 4,547 - 1 3 % Car Robberies 1,753 1,786 1.8 % Rape 781 858 10 % Homicides 2,098 2,183 4 % Source: National Civil Police (2003) 81. The high levels of crime and insecurityrecordedin the early postwar period during the mid 1990s and their subsequent decline can be largely explained respectively by the dismantling of the old police forces and the gradual consolidation of the National Civil Police. One o f the fundamental political agreements o f the 1992 Peace Accords that ended El Salvador's civil war was the dismantling o f the old security institutions linkedto the military and the creation o f a new law enforcement system under civilian ~ommand.~' While the dismantling o f the old police forces occurred relatively rapidly, the process o f establishing the new civil police was neither fast nor exempt from vicissitudes. In practice, it took the new police force several years to completely assume its law enforcement functions nationwide. Duringthe interim period there was an enormous security void in vast areas o f the country, where there was no governmental institution capable o f effectively performing law enforcement duties (Cmz, 2004). Given the large numbers of firearms in civilian hands, the prevailing militaristic culture after a decade o f armed conflict, and the strong legacy o f personal disputes and resentments caused by 48The agreement encompassed the dissolution o f the National Police, the National Guard and the Treasury Police ("Policia de Hacienda") and called for military troops to be withdrawn from law-enforcement duties. It also contemplated the creation o f a completely new police force called the National Civil Police, 40% o f which was made up o f former army and guerilla combatants. See Costa (1999). 74 the war, this temporary institutional vacuum in the area o f law enforcement was associated with an impressive escalation in the levels o f violence. By the late 199Os, however, the increasing consolidation o f the National Civil Police began to be reflected in significant reductions incrime rates. Table 2.3: Crime,victimization and insecurity rates, by province Car robbery (a) Commercial Bank Homicide Victimization Insecurity vehicle robbery (b) with firearm (a) rate (%) (c) Perception (d) robbery (a) Ahuachapan 3.2 4.4 0 20.8 6.5 6.2 CabaAas 2.0 1.6 1 34.3 14.5 6.0 Chalatenango 5.6 3.0 0 14.0 14.6 6.0 Cuscatlan 5.4 2.5 0 29.0 10.4 5.9 L a Libertad 26.8 9.8 1 27.4 14.9 6.5 L a Paz 12.3 5.0 1 31.6 18.5 6.4 L a Uni6n 1.O --- 0 29.8 12.2 6.6 M o r a z h 0.6 _ _ _ 0 13.0 7.4 5.8 San Miguel 7.5 0.4 0 26.4 18.8 6.8 San Salvador 65.9 6.7 13 23.6 20.7 6.9 San Vicente 6.8 4.2 1 22.6 10.8 5.6 Santa Ana 8.7 3.2 2 31.6 15.2 6.5 Sonsonate 6.1 4.2 0 31.6 15.5 6.5 Usulutin 2.1 1.O 0 20.9 9.4 6.1 (a) Rate per 100,000 population in 2000. (b) Absolute number in 2000. (c) measured in 2001. (d) measured in 2001 on a 1 to I O scale, where 10 i s the highest level o f insecurity. Source: UNDP (2002), T[JDOP (2001, 2002). 82. While the highest incidence of robberiesis found in the provinces of San Salvador and La Libertad, homicide rates are higher in the western provinces of Santa Ana and Sonsonate, and in the central provinces of La Paz and Cabaiias. Even after normalizing by the population living in each province, the rates of car and commercial vehicle robberies are muchhigher in San Salvador and La Libertad (Table 2.3). Moreover, in2000 almost 70% o f all bank robberies took place in San Salvador, which also reported the highest percentage o f victimized households and the highest perception o f insecurity (in 2001). However, homicides are more frequent (relative to the population) in the provinces o f Santa Ana, Sonsonate, L a Paz and Cabaiias. These provinces also share a moderate level o f perceived insecurity but, with the only exception o f L a Paz, they are characterized by relatively low rates o f victimization and economically motivated crimes. It thus appears that there are quite different regional pattems for what Moser and Winton (2002) have denominated economic and social ~iolence.~' 83.There is a widespread perception that juvenile gangs are responsible for a large fractionof the crimes committedin ElSalvador.However, there is limitedhard evidence to 49 In their work on violence in Central America, Moser and Winton (2002) propose a typology that covers three major types o fviolence: political-institutional, economic and social. The first type o f violence -practically absent in ElSalvador - is defined as the perpetrationof violent acts motivatedby a conscious or unconscious desire to obtain or maintain political power, in contrast to economic and social violence, where violent acts are motivated respectively by the desire o f economic gain or economic power, and by the desire to gain social benefit or to obtain or maintain social power. For an analysis o f the differtent regional patterns of economic and social violence in El Salvador see Cruz, Trigueros and Gonzalez (1999). 75 support this perception. Many studies have pointedout that gang problems are one of the most characteristic expressions o f criminal violence in Central America. According to conservative estimates there are 20,000 youth enrolled injuvenile gangs inEl Salvador, and the National Civil Police holds those organizations responsible for a good part o f the violence that occurs in the country. Indeed, a 2003 survey o f gang members found that 80% had committed thefts or robberies, almost 50% had participated in drug-trafficking and a small fraction had also committed murders (Carranza, 2003). Moreover, other studies have found that juvenile gangs are a breeding ground for organized crime, with older members eventually joining professional networks engaged in more complex criminal activities with a greater impact on the economic climate (e.g. kidnappings, business hold-ups and commercial-vehicle theft^).^' However, in spite o f their undeniable contribution to the climate o f insecurity that exists in El Salvador, independent studies suggest that the role o f gangs in explaining the incidence o f violent crime may have been blown out o f proportion. A report by the Legal Medicine Institute, for instance, shows that out o f 2,388 homicides registered in 2003, only 192 (8%) were known to have been committed by juvenile gang-members. Even though this may underestimate the number o f murders associated to gang violence - due to the lack o f reliable information in many o f the cases - it i s estimated that the homicides that are attributable to gangs do not exceed 18% of the total number o f violent deaths inthe ~ountry.~' Table 2.4: Cases received by the Attorney General, January-November 2003 Category Number Global Percentage of Percentage cases filed Initiated (cases filed) 94,912 100% Cases taken to the courts 39,178 41.28% 100% Cases in preliminary hearing 4,615 4.86% 11.78% Cases taken to trial 1,729 1.82% 4.41% Source: FESPAD (2004). 84. One of the main factors underlying the high levels of violence encountered in El Salvador is the presence of weak law enforcement institutions,which is illustratedby low levels of prosecution efficiency and effectiveness. According to a working paper prepared by the World Bank, UNDP, UNICEF and PAHO (2004), the results of a large number of studies suggest that the violence in El Salvador i s driven by factors associated with income inequality, social exclusion, limited educational and employment opportunities, cultural patterns that propitiate and reproduce violence and the proliferation o f firearms. However, the factor that i s arguably most closely related to the types o f economically motivated crimes that more directly affect the Salvadoran private sector i s related to weaknesses in the institutions that are in charge o f providing security and justice. In this respect, the available statistics reveal that despite the considerable progress observed since the mid 1990s, great challenges remain in order to achieve acceptable standards o f efficiency and effectiveness in the area o f law enforcement. Indeed, a UNDP study estimated that between 1998 and 2001 the efficiency o f the office o f the Attorney General - measured as the ratio o f the number o f cases filed to the number o f reported cases - increased from 18% to 35% (UNDP, 2002). However, the effectiveness o f that office - measured by the percentage o f the cases inwhich the judge initiatedthe phase o f instruction as a response to a requisition by the Attorney General - decreased from 59% to 53% from 1999 to 2001. ''Interview 50Santacmz and Concha-Eastman(2001), Carranza(2003). with Dr.FabioMolina, directorof the statistics unit o f the "Instituto de MedicinaLegal El Salvador". 76 Moreover, other studies reveal that from 1998 to 2003 the fraction o f the homicide cases for which any arrests were made increased from 6% to 42%.52 Nevertheless, only a fraction o f the arrested suspects are normally prosecuted, about 41% o f the cases that are filed are actually taken to the courts, about 5% reach the phase o f preliminary hearing ands less than 2% are actually taken to trial and can thus lead to actual convictions (see Table 2.4). Figure2.21: Percentageof Crimes reportedto the police in CentralAmerica, Colombia and Mexico (2004) 51`4 49.1 A 7 5 Source: Cordova and Cruz (2004) andLatin America Public Opinion Projectdatabase. 85. The ability of law enforcement agencies to curtail crime and violence is hampered by low levels of confidence in those institutions, which are reflected in the fact that a small fraction of crimes are reported to the police. Victimization surveys indicate that less than a third o f the households that were victims o f criminal acts reported those crimes to the police, which is less than in most Central American countries (Figure 2.21).53 Reporting rates vary widely by type o f crime. According to a 2001 victimization survey funded by El Salvador's Ministry of Governance, the crimes more frequently reported to the police are vehicle theft (80%) and homicides (61.6%). On the other hand, the least reported crimes are armed robbery (17.2 percent) and assaults (18.7 percent). It is interesting to note that even very serious acts o f violence such as homicides and robberies using firearms exhibit relatively low police reporting rates. This i s probably due, according to various opinion polls, to the low credibility o f law enforcement institutions among Salvadoran citizens. Indeed, when asked about the reasons for not reporting crimes to the police, 47% o f Salvadorans indicate that reporting the crime would be "useless" and 20% indicate that it would be dangerous to do 86. Another important factor for explaining the high levels of violent crime in El Salvador i s the proliferationof firearms in the population.A study conducted by IUDOP and FESPAD found that between 2000 and 2002 there were approximately 450,000 firearms in civilian hands inEl Salvador, with only 40% beinglegally registered. The proliferationof firearms results from '*See the studies performedrespectivelyby the NationalCouncil for Public Security and the Foundation for the Study of the Application of Law -CNSP (1998) and FESPAD (2004). 53 See Cordova and Cruz (2004). Similar reporting rates are encountered by Seligson, Cruz and Cordova (2000), and IUDOP (2003). A study by the Foundation for the Application o f Law (Aguilar, 2002) finds that the number of crimes reported to the National Civil Police rose from almost 17 thousand in 1996 to 55,676 in 1998, and has remained around that level since. 54 See Cordova and Cruz (2004). Other reasons that are mentioned the absence o fproofs (18%) and the fact that the crime was not serious (1OW). 77 a combination o f factors, including the copious arsenals left in civilian hands after the civil war - despite the disarmament plans established in the Peace Agreements - and the associated development o f a large international illegal trade in weapons. This trade was initially based on the arsenals that resulted from the armed conflicts in El Salvador, Nicaragua and Guatemala. In addition, the liberalization o f the market for weapons led to an impressive growth in imports and sales of firearms: between 1994 and 1999 El Salvador was the seventh largest importer o f handguns and pistols from the United States, after Mexico and before South Africa." Thus, almost 70% o f the homicides and about 38% o f the total number o f crimes recordedby the police are committed with firearms (see Table 2.5). Moreover, according to police reports in about half o f the cases the weapons usedby criminals are legally registered. Table 2.5: Violent crimes committedin2000 and percentageinwhich firearmswere used Crime Numberof crimes Percentage o ftotal Crimescommitted No. of crimes with firearm (%) SexualAggression 913 1.a 6.8 Attempted sexual aggression 254 0.4 3.7 Threats 7,194 13.8 29.4 Homicide 2,099 4.1 68.6 Attemptedhomicide 351 0.6 67.0 Robbery 11,051 21.2 68.6 Theft 18,385 35.3 3.3 Assaults 7,824 15.0 13.4 Kidnappings 545 1.o 56.7 Damages 2,700 5.2 5.1 FamilyViolence 528 1.o 0.0 Other crimes 188 0.4 85.1 Totals 52,024 100 26.3 Source: UNDP (2003a). 87. The survey data suggests that the impact of crime and violence on the activities of private Salvadoran firms i s smaller than in Guatemala but larger than in Nicaragua and Honduras. Almost 41% o f the surveyed Salvadoran firms report having beenvictims o f criminal acts in the year preceding the survey and 49% indicate that crime and violence are "major or very severe" constraints. As seen inthe previous chapter, the problems associated with crime and violence are the most frequently mentioned by Salvadoran firms as representing important constraints to their operations and growth. Comparing El Salvador with neighboring countries, we find that the percentage o f firms that describe crime as a major constraint (49%) is higher than in Nicaragua (40%) but lower than in Honduras (62%) and Guatemala (81%). However, objective indicators o f the impact o f crime on the activities o f private firms suggest that while the problems faced by Salvadoran companies are smaller than those encountered by their counterparts from Guatemala, they are larger than in both Nicaragua and Honduras. Indeed in these two countries the I C surveys suggest victimization rates o f respectively 26% and 31%, 55 Godnick, Muggah and Waszink (2002), using data from the U.S. Department of Commerce. The Salvadoran legislation that govems the firearms trade has been reformed and improved over the last few years, and it is relativelyadvanced in intemational terms. However, it does not guaranty an effective controlof civilian arm usage. For instance, the current legislationdoes not restrict the number o f firearms a citizenmay have and it allows for the use of weapons o f up to 0.458 inch caliber, which encompasses assault rifles such as such as the M-16 or AK-47 (providedthat they are impededto functionautomatically). 78 compared to 41% in El Salvador and 51% in Guatemala (Figure 2.22). Moreover, the share o f firms that report expenditures in security i s also highest in Guatemala (78%), followed by El Salvador (72%), Honduras (65%) and Nicaragua (55%). The average Salvadoran firm looses about 0.95% o f its sales as a result o f criminal acts and spends about 4.25% inprivate security in order to prevent crime. As seen in Figure 2.23, these figures are lower than those reported by Guatemalan firms, but larger than those gathered inboth Nicaragua and Honduras. Figure2.22: Victimizedfirms and Figure2.23: Crime Losses and security costs firms constrainedby crime (YOof sales) InternationalComparison - 5 6 6 1 Elvictimized (%) Wconstrained by crime (%) Osome sec. Cost Ocrime losses Wsecunty costs Osec. cost (if any) Source: InvestmentClimate Survey Source: InvestmentClimate Survey 88. Within El Salvador, the probabilityof a firm being a victim of crimes is higher in the provinces of San Salvador and La Libertad, and in the sectors of food and beverages and chemical products. The finding that crimes against private firms are more frequent in San Salvador and L a Libertad i s consistent with the general crime statistics presented above. It i s worth noting, however, that the percentage o f firms that mention crime and violence as major constraints is even higher in other regions o f the country, notably Santa Ana (Figure 2.24). We also find that the losses associated with crime are lowest inthe provinces o f and San Miguel and Ahuachapan and that firms located in San Miguel and Santa Ana report smaller security expenditures (Figure 2.25). Across sectors, the producers o f food and beverages and chemical products appear to be the most affected by crime, both in terms o f the percentage that report having been directly affected, and in terms o f the perceived constraints related to crime (Figure 2.26). These higher levels o f crime victimization for the sector o f chemical products (including pharmaceutical) and the food and beverage industry are probably associated to their more dense product-distribution chains. Indeed, firms from those sectors generally deliver merchandise to points-of-sale located all across the country, and products are thus more subject to being robbed while on transit. Less clear patterns are apparent with regard to the monetary costs associated with crime losses and security expenditures, which are highest respectively inthe sectors o f non metalic mineral products and metal products. 79 Figure2.24: Victimized Firms and firms Figure 2.25: Crime Losses and security costs I constrained by crime, by Province(%) I by Province(YOof sales) e3 12.1 n mvictimzed f i r m (Oh BCOnStrUnedbycrim(%) Osomesec cost (Oh locrime losses ~secuntvcosts nsec costs (itany) 1 Source: Investment ClimateSurvey Source:Investment Climate Survey Figure 2.26: VictimizedFirms and firms Figure 2.27: Crime Losses and security costs constrainedby crime, by sector (%) by sector (% of sales) 9.1 ovictimzed f i r m (Oh mconstraid by crime(%) osomsec cost Ocrime losses Osecurity costs Usec costs (if any) Source: Investment Climate Survey Source:Investment Climate Survey 89. Smaller firms are less likely to be victims of criminalactivities but their average losses as a result of crime are higher than for mediumand large firms. The percentage o f firms that report having been victims o f crime increases from around 30% among micro and small firms to more than 50% for large firms (Figure 2.28). Similarly, the fraction that report any security costs increases from 39% for micro firms to 93% for large companies. Despite these large differences, the average loss resulting from crime is, as a fraction o f sales, lowest among large firms. Moreover, among the firms that report at least some security expenditures - presumably those that face a larger risk o f victimization - the share o f those costs in sales i s much larger for smaller firms (Figure 2.29).56 56The data confirms that the fraction of victimized firms is higher among those that report at least some security expenditures,with causalityarguably runningfrom higher crimerisksto higher security expenditures. 80 Figure2.28: VictimizedFirms and firms Figure2.29: Crime Losses and security costs constrainedby crime, by size (YO) by size (YOof sales) 86 93 8.4 Oany sec. Cost (%) Ocnme losses Qsecunty costs Osec costs (fany) Source:Investment Climate Survey Source:InvestmentClimateSurvey 90. Salvadoran and Guatemalan firms are more likely to report crimes to the police than are firms from Honduras and Nicaragua. Reporting rates are higher among larger firms. Salvadoran and Guatemalan firms that have suffered from criminal acts report about 70% o f the corresponding crimes, compared to 53% and 56% for their counterparts from respectively Nicaragua and Honduras (Figure 2.30). The higher reporting rates by Salvadoran firms contradict the findings from household victimization surveys (see Figure 2.21), and suggest a different pattern for the specific types o f crimes that affect private firms - as opposed to the general population. Since reporting rates tend to be higher when there i s greater likelihood o f catching the perpetrators and for crimes which cause larger losses to the victims, the above result could reflect either a greater degree o f confidence in the police in El Salvador and Guatemala, or the fact that the average losses derived from crime are larger in the first two countries.57Within El Salvador, reporting rates are increasing inthe size of the firm: from 56% for micro-enterprises to 80% for large firms (Figure 2.3 1). Figure2.30: Crime Reporting by Figure2.31: Crime Reporting by victimized firms, by country victimized firms, by size Oreport rate (avg.) manyreports (%affirms) Oreport rate(avg ) ~aanyreporla(%offirms) Source:Investment Climate Survey Source: InvestmentClimateSurvey 57The data confirms that crime losses are larger among firms that report crimes to the police: this i s the case in the four Central American countries for which we have data. Within El Salvador, the only exception are micro- enterprises, a result that could reflect lower confidence inthe police among those firms. 81 91.The share of crimes committed by gangs is similar in El Salvador, Honduras and Nicaragua,and much higher in Guatemala.In El Salvador, gang-related crime affects to a much greater extent micro and small firms. About one third of the Salvadoran firms that report having being the victims o f criminal acts indicate that at least some o f them were committed by youth gangs. On average, gangs are responsible for 27% o f the crimes committed against the surveyed Salvadoran firms, compared to 67% inGuatemala and respectively 28% and 25% inHonduras and Nicaragua (Figure 2.32). Within El Salvador, gangs affect to a much larger extent the activities o f smaller firms: while 46% and 37% o f the crimes committed against micro and small firms are attributed to youth gangs, the corresponding figures for medium and large companies are 19% and 18% respectively. Similarly, for the firms that report having been the victim o f criminal acts, one in two microenterprises indicate that youth gangs were responsible for some o f those crimes, compared to one in four large firms(Figure2.33). Across firms located in different regions, those from San Salvador, San Miguel and la Libertad report the highest shares o f crimes committed by gangs (respectively 33%, 22% and 19y0). Figure2.32: Crimes by Youth Gangs Figure2.33: Crimes by Youth Gangs amongvictimized firms, by country amongvictimized firms, by size mgang share (avg ) manygang crime (%of firms) Ogang share (avg ) many gang crime (Oh of firms) Source: InvestmentClimateSurvey Source: Investment Climate Survey 92. Salvadoran firms indicatethat only 12% of the crimes that they reportto the police are eventually solved. The percentage o f crimes that are reportedly solved is higher among micro enterprises. The effectiveness o f the police in solving crimes reported by private firms i s slightly higher inEl Salvador compared to Guatemala and Honduras, where respectively 10% and 6% o f the crimes reported are solved (Figure 2.34). Nicaraguan firms report a much higher percentage o f crimes solved by the police (30%). The survey data also suggests that the type o f victimization that affects small businesses - including crimes committed by youth gangs - i s more susceptible o f being investigated and solved by the police. Thus, across Salvadoran firms o f different sizes, the highest shares o f solved crimes are obtained for microenterprises: 23% compared to 9% for small and medium firms, and 12% for large companies. Moreover, in all four countries the percentage o f crimes that are reportedly solved i s slightly higher among the firms that are affected by gang activity. Within El Salvador, this i s clearly the case for micro and small firms, but not among mediumand large companies (Figure 2.35). 82 Figure 2.34: Crimes solved by Figure2.35: Crimes solved by the police,by country the police, by size 36 34 Ucrimes solved 1Dcrimes solved if some cnmes bv aanas " n e s solved 1 Source: InvestmentClimate Survey Source:InvestmentClimate Survey 93. Compared to other firms of the same size, sector and province, Salvadoran maquilas are less likely to report security expenditures and to be victims of criminal acts. Similar but quantitatively smaller differentials are found in other Central American countries. The probability o f spending on private security i s 18% smaller for Salvadoran maquilas than for non- maquila firms o f the same size, sector and province (see appendix table A2.3). Similarly, maquilas have a 23% lower probability o f being the victim o f a criminal act in El Salvador. These differentials are larger than those found in other Central American countries, as revealed by the analysis using data from Salvadoran firms pooled with firms from Nicaragua, Honduras and Guatemala: in the latter case, the above differentials are respectively 14% and 13%. In addition, at least in the pooled sample, the losses o f victimized firms as a percentage o f sales are 1.3% larger for maquilas - the estimate i s positive but not significant in the Salvadoran sample, possibly due to its smaller size. 94. Our econometric estimates also suggest that the risk of victimization is lower for firms that are located in regions and sectors with a higher share of companies that invest in private security. This effect i s found only among Salvadoran firms. Its magnitude i s such that a 50% increase in the share o f firms that invest in security reduces in 18% the probability o f victimization o f firms o f the same size range, province and sector. Estimates reported in the appendix also confirm several o f the descriptive results previously reported: larger firms are more likely to report security costs and to be victimized but they loose a smaller percentage o f sales when victimized; firms that have been victimized are more likely to have security expenditures; the highest victimization risks are obtained for the sectors o f food and beverages and chemical products; the sectors that in El Salvador are more likely to report security expenditures are those o f Food and Beverages and Metal Products; and the sector inwhich losses due to victimization are largest i s that o f non-metalic mineralproducts. 83 POLICY RECOMMENDATIONS ONCRIME 95. It i s important to improve the collection of information on the impact of crime on private sector activities.Despitethe large number o f studies available on the issue o f crime and violence in El Salvador (see Moser and Winton, 2002), little information has been collected on the impact of crime on the private sector. The ICs is the first survey to systematically collect information on the impact o f crime on private firms. In this respect, the government's General Direction o f Statistics and Census (DIGESTYC) and the business surveys o f organizations like FUSADES could including questions on crime victimization at the firm 1evel.It would also be very important that business associations such as the chamber o f commerce develop databases on the frequency and importance o f crimes against their members. Finally, it i s fundamental to improve the information systems o f the institutions in charge o f public security and justice, particularly those o f the Attorney General's Office and the courts, in order to evaluate not only the general trends o f crime but also the effectiveness o f those institutions indeterring crime. 96. A security policy aimed at improving El Salvador's investment climate should encompass measures directed at curtailing the incidence of both property crimes and crimes against the integrity of individuals. Moreover, such a policy should not necessarily have its main focus on the problems created by youth gangs. As seen in this section, the efforts to reduce the incidence o f economically motivated crimes have been more successful than those aimed at reducing violence against individuals - e.g. homicides and rapes. However, high rates o f "social violence" also have detrimental effects on economic activity, as they increase the perception o f insecurity and lead to large unproductive expenditures in security and in the treatment and compensation o f victims. Thus, El Salvador's security policy should encompass actions directed at reducing both economic and social violence. As for the role o f gangs in explaining the high rates o f crime and violence, the I C s showed that gang members are responsible for less than one third o f the crimes committed against private businesses. Thus, the sole concentration o f police and judicial resources in fighting gang-related crime may not be totally warranted, and the fight against gang-related crime should be put in the context o f a comprehensive vision o f the violence problem inEl Salvador. 97. A comprehensive national security policy should also incorporate measures aimed at strengthening justice sector institutions, creating educational programs to promote the peaceful resolution of conflicts, and broadening education and employment opportunities for at-risk groups. Legal and illegal weapons constitute one o f the most common precipitants o f violence in El Salvador, where there are almost half a million weapons in civilian hands, and firearms are used in more than 75% o f homicides and robberies. Recent reforms approved by Congress in February 2005 should help curtail the proliferation o f firearms, thus contributing to the reduction o f violent crime. Efforts must also be made to increase the efficiency and effectiveness o f the police and the judicial systems, so that they act as more effective deterrents o f criminal activities. As for the role o f educational programs, a good part o f the social violence associated with homicides i s the product o f a system o f norms and values that emphasize the use o f force to resolve conflicts. While this culture o f violence has existed inthe country for decades, it can be contained through peaceful conflict resolution programs. Finally, inthe case o f at-risk youth, it i s very important to make every effort to keep young people in the educational system 84 for as long as possible, and to provide employment opportunities for at a greater risk o f joining criminal gangs. 98. The active participation of the private sector in formulating and supporting the country's policies against crime is fundamental. The participation o f the private sector inthe design, support and execution o f policies to face the problems o f crime and violence has been very limited. The main exceptions have been a campaign to collect arms at the end o f the 1990s, the support given to the UNDP program "Sociedad din Violencia" (society without violence), and the participation o f the judicial teams o f the National Association o f Private Enterprises in some high-profile criminal cases. However, the business sector can play a much more important role, not only by continuing its support o f campaigns to collect arms and by producing information on the criminal cases that affect the members o f business associations, but also by directly participating in educational and job training programs aimed at at-risk youth, as well as inother localandnationalcrime preventioninitiatives. CONTRACTENFORCEMENT THE JUDICIARYAND 99. Business disputes are inherent in commercial transactions. In a complex economy, involving many buyers and sellers, and complex goods and transactions, firms necessarily undertake risks in dealing with new clients and suppliers, and face potential disputes with them. Without efficient courts and the expectation that courts will uphold their contractual rights and obligations, firms will be less willing to deal with new clients and suppliers, and fewer transactions will take place. 100. The functions of enforcing contracts and reducing the uncertainty faced by firms when engaging in commercial transactions is not restrictedto the courts. These include both formal and informal mechanisms o f enforcement, such as social networks, business associations, or public information channels to decide on whom to do business with and under what conditions. A t a general level, contract enforcement mechanisms could be divided into personal relationships based on trust; self-enforcement through repeated transactions; third-party enforcement based on reputation; 5! rivate enforcement; administrative government intervention, and court enforcement (litigation) . While each could potentially serve the same purpose, some o f these mechanisms are inefficient since they exclude new clients (for instance commercially transacting only with known, repeated customers), or structuring transactions with a view to avoid business disputes, such as requiring pre-payment. Others, such as business associations, may sometimes not work well and be rarely used. '*See Hendley and Murre11 (2002) for a classification and analysis o f the use o f these mechanisms in supporting contractual agreements among Romanian firms. 85 Trust in the Judiciary 101. El Salvador does reasonably well - relative to other Central and South American countries- in terms of private firms' confidence in the judicial system. Still, a quarter o f respondents (26%) to the ICs disagree fully, or at least in the majority o f cases, that the judicial system will uphold their contractual and property rights in the case o f business disputes. As depicted in Figure 2.36, El Salvador does substantially better on this "trust in the judiciary" measure than Guatemala, Nicaragua, Honduras and Ecuador, but worse than Brazil and Peru. Furthermore, only about 20% o f interviewed firms agree entirely or in most cases that the judiciary would uphold their contractual and property rights in business disputes. Therefore, more than half o f the ICs respondents in El Salvador seem to harbor doubts about the usefulness o f the courts in resolving their disputes with other business parties, and express a tendency to only weakly agree or disagree with the validity o f the statement about their confidence in the courts. Figure 2.36: Firms Which Do Not Believethat the Courts Will Uphold Their Contractualand Property Rightsina CommercialDispute 0 I O 20 30 40 50 60 Source: World Bank InvestmentClimate Surveys Figure2.37: FirmsWhich PerceiveProblemsin Court Efficiencyand ContractExecutionto be Major or Very Severe Obstacles to Doing Business 0 I I O I S 1 0 2 5 1 0 15 4 0 p " C c 0 l Source: World Bank vestment Climate Surveys 102. Salvadoran firms also report less severe problems associated with such aspects of the contract enforcement regime as court efficiencykonflict resolution and contract executiodproperty rights. In line with the previous paragraph, we find that the proportion o f 86 firms that consider the problems relatedto court efficiency and contract execution to be major or very severe obstacles to their operations i s smaller inEl Salvador than inother Central American countries, Ecuador and Brazil (Figure 2.37). Thus, for instance, issues related to court efficiency are seen as important constraints by 16% o f the firms surveyed in El Salvador, compared to almost 30% on average inGuatemala, Nicaragua and Honduras. 103. The fraction of firms which report major concerns with court efficiency is largest among medium-sized companies and among those that report having used the judicial system to resolve commercialdisputes. As seen in figure 2.38, in El Salvador medium-sized firm are the most likely to report severe constraints related to problems in court efficiency. This i s to some extent different from what i s found in other Central American countries, where concerns with court efficiency increase monotonically with firm size and are largest for medium and large firms. However, both in El Salvador and in the other Central American countries, concerns with court efficiency are driven by the extent o f the actual use o f the courts to solve commercial disputes,which i s generally increasing in size, but i s largest for medium sized firms inElSalvador. Indeed, comparing the whole sample with the sample ofcourt userss9,we observe that concerns with court efficiency are larger among the former (see Figure 2.39). Figure2.38: Firmsthat View Court EfficiencyProblemsas Major or Very Severe Obstr :les, by Size -T- El Salvador I I I II 3 0 99 I I 1 I I I I I Guatemala 37 B5 I i I t 50 I <" I "carague =" 0 I O 20 30 40 50 60 psrosnt IPB Micro I PMedium Cl Large Small I Source:WorldBank InvestmentClimate Surveys Figure2.39: FirmsWhich View Court EfficiencyProblemsas Major or Very Severe Obstacles, Court users Vs. Whole sample Source:World Bank InvestmentClimateSurveys 59 Court users - as defined later on - are those firms, which have filed some of their payment-relateddisputes with the courts. 87 104. There are also some regionaldifferences within El Salvador in terms of the gravity of the problems with court efficiency and contract execution. A larger proportion of firms in the provinces of San Miguel, La Libertad and Ahuachapan consider court efficiency to be a major or very severe problem. Problems related to contract execution are reported most frequently as major or very severe inthe province o f San Miguel (see Figure2.40). Figure2.40: FirmsWhich PerceiveProblemsin Court Efficiency and ContractExecutionto be Major or Very Severe Obstacles to DoingBusiness, by Province Le Libsflad Ahuashapan Snn Salvador santa Ana Other 0 5 10 15 20 25 PCIESIII 1ElProps- rights and ~ ~ 0 ~ ~ 1 1of0contra~fsICOURefficiency and conflict rerolutmn 1 1 j Source: World Bank InvestmentClimate Surveys How Do Salvadoran Firms Avoid Disputes? 105. Among the interviewed Salvadoran firms close to a quarter, i.e. 23%, report requiring pre-payment from their customers. The average size o f these pre-paid sales for those firms which do require pre-payment is 30% o f their annual sales. Therefore, almost a quarter o f Salvadoran firms require from the buyerso f their products to pay inadvance, and such pre-paid sales constitute slightly less than one-third o f their annual sales for this subset o f firms. Among all interviewed firms, the average share o f pre-paid sales in total annual sales i s approximately 7%. This compares favorably with other countries in the region, in the sense that pre-payment - which i s an indicator o f inefficient contract enforcement - i s less frequent in El Salvador than in other countries for which we have comparable data (see Panel A o f Table 2.6). 6o 106. A large share of Salvadoran firms - 81 % - report that they require from their customers to pay on the spot (at the time of the sale) for their goods and services. This is similar to the same indicator for Peru and Guatemala, but substantially higher than in Brazil and Ecuador (58% and 69% respectively). Such sales constitute about half o f the annual sales volume o f these firms (49%). Among all interviewedfirms, the share o f sales paid at the time o f sale i s 40% (see Panel C o f Table 2.6). The ICs question on the use o f credit in firm sales asks firms what proportion o f their sales are paid by customers in advance (Le. before the delivery o f the good), at the time o f the sale, or on credit. Previous research has found that when countries' contract enforcement in court works well, firms would be more willing to extend credit to their customers as well as to engage intransactions with new, less known ones. They would generally rely less on transacting with familiar, trusted customers. When contract enforcement is poor, 6oThe only exception is Guatemala, which has indicatorsthat are similar to El Salvador's. 88 firms would reduce contractual risk by investing in building long-term relationships with customers and suppliers based on trust, and would shy away from trading with less known parties or extending trade credit, even if it were on more beneficial terms.61 This relational contracting results inloss o f economic value. Table 2.6: Pre-Payment, Payment at the Time of Sale, and Credit Panel A ...paid in advance Share offirms which YeqUirepre-paidsa[es ("/.ofall Pre-paid sales of all by clients pre-payment sales) acrossfirms which sales) across allfirms requirepre-payment Nicaragua 39% 40% 16% Honduras 31% 42% 13% Peru 34% 30% 10% Brazil 35% 17% 6% Ecuador 28% 28% 8% ElSalvador 23yo 30% 7yo Guatemala 19% 34% 6% Panel B ...soldon credit to Share offirms which sell on Percentage sold on credit Percentage sold on credit clients credit acrossPrms which do sell across allfirms on credit Nicaragua 65% 50% 32% Honduras 65% 64% 42% Peru 69% 66% 46% Brazil 94% 84% 79% Ecuador 90% 76% 68% ElSalvador 81% 66% 53yo Guatemala 77% 67% 51% Panel C ...paid at the time of Share offirms whichget Percentage paid at the time Percentage paid at time of sale by clients payment at the time of sale of sale acrossfirms which sale across allfirms get paid at time of sale Nicaragua 88% 59% 52% Honduras 76% 60% 45% Peru 82% 50% 41% Brazil 58% 24% 14% Ecuador 69% 35% 24% ElSalvador 81% 49yo 40yo Guatemala 79% 53% 42% Source: World Bank InvestmentClimate Surveys 107. About one fifth of interviewed Salvadoran firms report that the do not sell their goods on credit; in 81% of the cases firms reportat least some sales on credit. This is higher than the share o f firms selling on credit in Nicaragua, Honduras, Guatemala and Peru, but lower than the same in Brazil (94%) and Ecuador (90Y0).For those firms which do sell on credit to their customers the share o f credit sales is about 66% o f their annual sales (see Panel B o f Table 6' See Johnson, McMillan and Woodruff (2002) for five transition economies, Fafchamps (1996) on Ghana, McMillan and Woodruff (1999) on Vietnam. 89 2.6). Across all interviewed Salvadoran firms the share o f sales sold on credit i s 53% on average. Thus, pre-payment and payment on the spot - at the time o f the exchange - are quite common, accounting for almost half (47%) o f the annual sales volume o f all interviewed firms. It is likely that pre-payment and payment at the time o f the sale are used as mechanisms to prevent fbture disputes over delayed paymenthon-payment, and is a signal that public enforcement o f contracts i s probably seen as inadequate inensuringa fast and reliable resolution o f payment disputes. 108. Firms also avoid disputes with clients or suppliers by relying on known, existing partners and foregoing transactions with new partners, even when these are on better terms. For example, less than one quarter o f interviewed Salvadoran firms expect to lose their customers entirely to competition should they raise the price o f their product by lo%, assuming that their main competitors leave prices unchanged. This reflects firms preferringto contract with known, trusted clients, thereby reducing contractual risk. However, this also implies loss o f economic value as opportunities for transacting on better terms are lost, which also deters entry o f new, less-known firms. Importantly, unwillingness to switch to new suppliers, when the existing one raises prices, could also be interpreted as an indicator o f inadequate contract enforcement and dispute resolution inthe courts. Table 2.7 shows some comparative results, and indicates that El Salvador does similarly to the other Central American countries and Ecuador, butworse than Brazil, where a higherproportiono ffirms (31YO)anticipate losing their customers ifthey were to raise prices. Inaddition, the largest shares of firms expecting their customers to still continue purchasing at the higher price the same quantities as before are in Nicaragua and Honduras (19%), suggesting that firms there are particularly risk-averse to transacting with new firms, even if it were on better terms. Country FirmsExpectingto Existing Customerswill Continue Existing Customerswill Continue Lose Customers Purchasingthe Same or Somewhat Purchasingthe Same Quantities Entirely (% of firms) Lower Quantitiesas Before(% of firms) as Before(%of firms) ElSalvador 24% 35% 9yo Guatemala 25% 41% 10% Nicaragua 23% 45% 19% Honduras 21% 44% 19% Peru NA NA NA Brazil 31% 33% 6% Ecuador 22% 41% 11% 109. Salvadoran firms report a relatively large share of their sales to private customers being paid with a delay. Despite strategies to avoid disputes with clients and suppliers by requiring pre-payment, only 11% o f the firms that deal with private clients report no payment delays. The average share o f delayed payments is about 33% o f total annual sales (and 37% if we exclude firms which do not experience payment delays). Consistent with the information above that a larger proportion o f Nicaraguan and Honduran firms require prepayment and a smaller proportion o f the same allow credit, we find that a larger fraction o f respondents in these two countries report not experiencing payment delays (19% in Honduras and 20% inNicaragua). In contrast, firms in El Salvador and Guatemala which are more likely to extend trade credit and less likely to require prepayment, are also more likely to experience payment delays with their 90 private customers. Only 11% o f Salvadoran firms and 8% o f Guatemalan firms report zero delays. Accordingly, the size o f delayed payments is similar inEl Salvador and Guatemala (33% and 34% o f total annual sales respectively), and lower in Nicaragua and Honduras (25% and 28% respectively). This might possibly indicate a somewhat higher trust in the judiciary or in contract enforcement institutions in El Salvador and Guatemala than inNicaragua and Honduras -whichappearstobetrueforElSalvadorontheanalysis ofperceptionsoftrustinthejudiciary, butnot so for Guatemala. 110. Salvadoran firms dealing with the Government report that approximately 35% of their total annual sales to the Government are paid with a delay. Of the 82 interviewed firms which sell to the Government just above one-third (34%) do not experience any delayed payments on the part o f the Government. This i s three times higher than the respective number for no delayed payments from private customers (11%). Therefore, the Government i s less likely to delay payments than private firms. Similar results hold for the other three Central American countries: 35% o f Guatemalan firms which deal with the Government report zero delays; the respective figures for Honduras and Nicaragua are again higher - at 42% and 51YOrespectively. The size o f the delayed Government payments is 44% in Guatemala, and 27% and 26% o f total annual sales inHonduras and Guatemala respectively.62 111. Those Salvadoran firms which sell goods on credit typically allow for more than one month (39 days) to their customers to execute payment. There is not much variation in this variable across the four Central American countries: Honduran and Guatemalan firms allow 38 and 35 days to their customers to pay, while firms inNicaragua allow somewhat less - at 28 days on average. 112. For those Salvadoran firms which experience payment delays in their sales on credit it takes about one and a half months on average (49 days) to resolve them. The ICs question does not indicate through what means the resolution was achieved - whether through private negotiation, third-party mechanisms such as business associations, Government, private enforcement, or through the courts. Compared to the rest o f Central America, Salvadoran firms take somewhat longer to resolve delayed payment disputes: the corresponding duration o f payment delays resolution i s 38 days in Guatemala, 33 days in Honduras and 26 days in Nicaragua63. TheRole of Business Associations in Dispute Resolution 113. Almost 60% of interviewed Salvadoran firms belong to a business association, of which 59% receive assistance from the association in resolving business disputes. Ingeneral, o f those firms which belong to a business association and receive dispute-resolution services from it, more than one third (37%) attach a high or critical importance to the dispute resolution 62Ifwe restrict our sample to only those firms with a non-zero share ofpayment delays (out of total annual sales), the means of the size of the delayed payments rise to 53% in El Salvador and Nicaragua, 47% inHonduras, and 67% inGuatemala. 63 Some respondents in Honduras (9), Guatemala (1) and Nicaragua (15) indicated that they need zero days to resolve their payment delays. We have removed these observations, assuming that it takes at least one day to resolve a payment dispute.Leaving these in, however, does not change the results substantially. 91 services received64.However, o f the same subset which belong to a business association and get dispute resolution services from it, more than a quarter (27%) attach zero or low value to the services received (see Table 2.8). Business associations can play the role o f a third-party contract enforcement mechanism. 114. Across Central America business association membership is highest in Honduras (71%) followed by El Salvador at 59% of interviewed firms. Business association membership is somewhat lower in Nicaragua (43%) and Guatemala (53%). Compared to other South American countries, these figures are similar to those in Brazil and Peru (74% and 47% respectively), but lower than Ecuador, where 97% o f interviewed firms report being members o f business associations. Among the countries in Table 2.8, Guatemala has the highest proportion o f interviewed firms which indicate that business associations help them resolve disputes; in contrast, Peru has the lowest - at 14% o f Peruvian firm-re~pondents~~.Despite the differences, it appears that in Central and South America firms use the business associations as an avenue for resolving commercial, labor and administrative disputes, and, on average, attach relatively high value to the quality o f the services provided. In contrast, a similar firm-level survey -- the 2002 Business Environment and Enterprise Performance Survey (BEEPS-2) -- conducted in 2002 in Central and Eastern Europe and the CIS countries reveals that for a subset o f eight South Eastern European countries on average only 4% believe that the business association to which they belong provides a major or critical value in resolving disputes, and an average o f 84% believe that the business association to which they belong provides no value or, at best, only negligible value inresolving business disputes66. Table 2.8: BusinessAssociation Membershi? and DisputeResolution Firmsbelongingto a Firms, which report thatFirms which report that Firms which report that business the association helps in the association's the association's Association, (share of all resolvingdisputes with assistancein dispute assistancein dispute firms) Government, workers, resolution is of zero or resolution is of high or andor otherfirms low importance crucial importance El Salvador 59% 59% 27% 37% Guatemala 53% 75% 26% 35% Nicaragua 43% 54% 13% 48% Honduras 71% 52% 27% 42% Peru 47% 14% NA NA Brazil 74% 58% 16% 56% Ecuador 97% 32% 22% 32% 64 If we consider the firms which do not get dispute-resolution services from the associations and chambers of commerce to which they belong, and assume that they assign zero importance to these services, then the share o f firms which attach high or critical importance declines to 22% in El Salvador and Honduras, and to 26% in Guatemala andNicaragua. 65 Unfortunately, the Peru ICs does not have the questions which determine the degree o f importance and value o f the services provided by business associations to their members. 66 Inboth the BEEPS-2 and ICs questionnaires the format ofthe question about belonging to business associations and the value o f services they provide are very similar. The value o f dispute resolution services is assessed on a 5- point rating scale (0 -no value to 4 - critical (crucial) value). 92 Dispute Resolution in Court 115. Salvadoran firms very seldom use the courts to resolve their overdue payments. Only about 2.0% o f cases o f overdue payments were brought to the courts in the two years preceding the ICs - but this number varies systematically across firms o f different size and ownership. More than three quarters o f Salvadoran firms which experience payment disputes (79%) do not bringthese disputes before the courts. Medium-sized and large firms are bringinga significantly higher share o f their payment disputes before the courts than small and micro firms, and so are foreign-owned firms compared to domestic firms (3.4% vs. 1.9%). Nevertheless, the results indicate that firms are primarily resorting to direct negotiation and alternative mechanisms for resolving payment disputes when such occur. This i s usually the case in other countries too - e.g., Hendley and Murre11 (2002) for Romania, Broadman et al. (2004) for eight South Eastern European countries. 116. The results on payment cases filed with the courts or tribunals for El Salvador generally apply to the other three CentralAmerican comparator countries. For instance, the percentage o f overdue payments brought to court over the preceding two years i s also 2.2% on average in Honduras, 2.3% in Nicaragua, and 1.3% in Guatemala. Similar shares o f firms experiencing overdue payments report filing court cases to resolve those: 21% in El Salvador, 20% in Honduras, 19% in Nicaragua and 17% in Guatemala. These shares, as shown in Table 2.9, vary across firms o f different size. Overall, micro firms file the least cases - less than 10% o f those micro firms which experience payment disputes file them with the court, followed by small firms -- 10% and 18% o f whom file such disputes with the courts. The heaviest users o f formal court process are the medium and large firms, with a larger fraction o f medium-sized firms filing disputes with the courts in El Salvador and Honduras, and vice versa - in Guatemala and Nicaragua. Micro Firms Small Firms Medium Firms Large Firms El Salvador 7.5% 16.4% 30.6% 24.2% Guatemala 9.8% 10.4% 16.2% 28.9% Nicaragua 4.9% 18.3% 35.4% 40.0% Honduras 4.4% 16.9% 36.2% 25.3% Figure2.41: Share of Payment DisputesBrought to Court: by Firm Ownersh P I . .... ... .. ... ., iF1DO m s m I E I PormlBn Source:World Bank InvestmentClimate Surveys 93 117. A larger share of foreign firms file their payment disputes with the courts than domestic firms do in three of the four Central American countries, includingEl Salvador. As shown in Figure2.41, the difference is most pronounced inNicaragua, where the fraction of foreign-owned firms filing payment disputes in court i s more than twice the fiaction o f domestic firms filing payment disputes in court Only in Honduras i s there no large difference between . thethe shares o f foreign-owned and domestic firms filing court cases. We have performed t-tests for differences in mean shares o f domestic and foreign-owned firms payment disputes filed in court, and found that the differences between displayed inFigure 2.41 are only significant for the case o f Nicaragua at the 5% level. For the other three countries the differences are not statistically significant. Similarly, t-tests for differences inmean percentage o f payment disputes filed with the courts reveal significant differences between domestic and foreign-owned firms only inNicaragua (1.9% vs. 4.9%), but innone o f the other three countries.67 118. Tests for differences in mean shares of firms filing payment disputes in court and mean percentage of payment disputes filed in court confirm that medium and large companies use the judicial system to a greater extent than micro and small firms. A summary o f these findings i s presentedinTable 2.10. The average percentage o f micro and small firms filing payment disputes i s generally lower than that among medium and large firms. We find significant differences in each of the four Central American countries for the former variable, but not always so for the latter. Source: World Bank Investment Climate Surveys Note 1: One-tailed t-tests are reported. P-values are shown inparentheses next to the t-statistics. Significant results are marked inbold. Note 2: OLS regressions on court use are conducted for the whole sample, controlling for firm size, ownership and other firm characteristics, and their results are presented in the Annex. 67 These results are not reported for considerations o f space, but are available upon request. 94 119. Among the surveyed Salvadoran firms which brought payment cases before the courts, 45% of respondents indicate that none of the cases they filed was resolved in their favor. On average, among all the firms with court cases, 36% o f these cases were resolved in their favor. Large firms had the largest share o f payment disputes resolved in their favor by the court - 49%, followed by medium-size firms with 39%. Incontrast, micro firms report that only 10% o f their court cases over payment disputes resultedin a decision in their favor. The same number for small firms is 23%. This suggests, perhaps, an existing bias on the part of the courts infavor of larger firms, and is potentially a source of distrust inthe judiciary. However, this is not borne out by the evidence in the section on Trust in the Judiciary, where medium and large firms were found to consider court efficiency and contract execution most problematic. Figure2.42: Share of FirmsWhich FiledCourtCases and Noneof These Were Resolvedin Their Favor,by Firm Size h l c a r a g u l 0 I O 1 0 1 0 1 0 S O 6 0 7 0 8 0 9 0 p I r E I n l [SmW h o l e S a m p l e I M i ~ DrS m~a l l O M n d i u m I L a r g rI Source: World Bank Investment ClimateSurveys 120. We compare our results for El Salvador on percentage of payment-related court cases decided in the plaintiff'sfavor to those of the other three CentralAmericancountries. We findthat -just like inEl Salvador - 45% o f Guatemalan firms with court cases over payment disputes did not have a case resolved intheir favor. The corresponding figures for Nicaragua and Honduras are even higher - 54% and 61% respectively. Since the dispute inthis case i s assumed to be over a delayed payment and the plaintiff i s the creditor, these numbers might mean an existing pro-debtor bias in the courts. Alternatively, these figures o f failure rate may be indicative o f a particular type o f creditor, who is less able to defend her position in court (micro or small firm, firms without a legal representative in court, etc.). Upon inspection o f the average percentage o f payment disputes resolved in the creditor's favor, we find that El Salvador ranks highest - with 36% o f court cases decided in the creditor's favor, followed by Guatemala with 29%, Honduras and Nicaragua with 24% each. Interms o f an existing bias in favor o f particular types o f firms, we find that unlike in El Salvador, where there appears to be a bias in favor o f medium and large firms, in the other three countries the bias i s usually in favor o f small and micro firms, and the differences are not particularly large. Interms o f ownership, we observe that in El Salvador, Honduras and Nicaragua domestic firms get a higher percentage of their court cases resolved in their favor, whereas in Guatemala the opposite i s true - Le., foreign-owned firms get more court judgments over payment disputes in their favor (44% vs. 39%). However, in terms o f shares of creditors who get judgments against them in all cases, a larger fraction o f foreign firms appear to get judgments against them in Honduras and Guatemala, whereas the 95 opposite i s true for El Salvador and Nicaragua. Figures 2.42 and 2.43 depict some o f these findings. Figure 2.43: Share of FirmsWhich Filed Court Cases and Noneof These Were Resolved in Their Favor: by Firm Ownership 0 I O 1 0 30 P O 50 60 1 0 I I D a m e n t i r I F o r c i g n - O w n e d 1 Source: World Ddnk Investment Climate Surveys 121. Among the Salvadoran firms which brought disputes over delayed payments before the courts, 36% of respondents report having at least some of these cases resolved outside the court system. On average, only 15% o fcases over disputedpayment, that enteredthe courts, were resolved outside the courts. There are significant differences in the average percentage o f payment disputes resolved outside the courts - e.g. large firms and medium firms report having 23% and 14% respectively o f their payment disputes that entered the courts as being ultimately resolved outside the courts. In contrast, micro firms have only 0.2% o f their court cases over payment disputes as being resolved outside the courts. This seems to indicate that medium-sized and large firms are better positioned to avail themselves o f alternative dispute resolution mechanisms such as mediation, resolution through the Chambers o f Commerce, or simply have more bargaining power to deal with their clients directly. 122. Compared to the rest of the Central American countries, El Salvador has the lowest rate of resolution of payment-related cases - initially filed with the courts - outside of the courts. For example, in Guatemala and Honduras respondents indicate that more than half o f the cases which they filed with the courts were ultimately decided outside o f the courts (53% and 52% respectively). This may indicate that court resolution i s slow, expensive and complex, and parties prefer to settle the case outside o f the court system. An alternative interpretation is that there exist good alternative dispute resolution mechanisms, which in some instances even the court judge might suggest that the parties to the case follow. Inother words, it is not necessarily inefficient from an economic point o f view to have a large fraction o f disputes resolved outside o f the formaljudicial system. 123. The ICs also gathers data on the duration of payment disputes in court. Of 64 interviewed Salvadoran firms which had payment disputes resolved in court over the past two years, 18 report that these cases are still pending. Among the other 46 firms the average time for the courts to deliver judgment was 18 weeks - or about 4 months. There are some pronounced differences among different types o f firms (see Figure 2.44). For instance, court cases are resolved faster when the party to the case is inthe maquila sector (10 weeks) vs. 96 19 weeks for non-maquila firms; i s foreign-owned (11 weeks vs. 18 weeks for private domestic firms); and belongs to the textile (4 weeks) or metals (8 weeks) industries. Also, micro firms report faster times to disposition by the courts (12 weeks on average). Case disposition times are longer than average in the footwear, non-metal manufacturing, chemical and apparel industries, as well as inL a Libertadand Other Regions (the residual regional category). Figure2.44: AverageNumber of WeeksNecessary to Resolve a DisputedPaymentinCourt, 2003-2004 30 I 25 20 1 5 10 5 0 1 Source: World Bank El Salvador Investment Climate Survey 124. We conduct a comparison of the ICs time to disposition of a payment-relatedcourt case - as reported by survey respondents - with the other available ICs data for Central America. El Salvador has the lowest times to disposition of a business case in the courts, followed by Nicaragua, Guatemala and Honduras. Figure 2.45 shows the average duration o f these identical overdue payment cases from filing in court until the judge delivers a decision for the four Central American countries. It also shows the duration o f a payment dispute in court in the country's largest city for each ofthe four Central American countries according to the World Bank Doing Business Database. While the ICs figures show that Honduras has the highest time to disposition, the Doing Business numbers reveal that Guatemala has the longest contract enforcement process in a first-instance court (including execution o f the judgment, i.e. attachment o f debtor's assets, holding o f an auction, if necessary, and actual payment o f the proceeds to the creditor). Since the I C s data on duration do not include time for execution o f the judgment, we would naturally expect that the DoingBusiness data would be higher than the ICs (which i s the case in El Salvador). The numbers for Nicaragua display less o f a discrepancy, while Honduras has a higher duration according to Doing Business in 2004 compared to 2003 (and twice what the ICs indicates). Nevertheless, these differences are not as dramatic as the one inGuatemala, where - according to the Doing Business data - contract enforcement incourt is the lengthiest inthe LatinAmerica Region, and indeed, inthe world, at 1460 days or 4 years6*. A potential source o f difference between the ICs and Doing Business data inthis particular example are also the numerous specific assumptions, which are made in the case considered by Doing Business Project. These include assumptions about the transaction and its value, the place, the parties (both commercial entities), etc. Furthermore, the data are based on a detailed written self-administered questionnaire provided by one o f the largest law firms in each country. Therefore, there is some degree o f subjective assessment of quantitative measures such as duration, costs, etc. although sufficient care is taken to double-check and verify all the Doing Business information. 97 Figure2.45: Durationof a PaymentDisputein Days:A Comparisonof ICs andDoingBusiness Data -7 H o n d ra1 Y 115 I 1 6 5 I '"I I I I I Source:WorldBank El Salvador Investment ClimateSurvey 125. Data on contract enforcement in court from the World Bank Doing Business in 2005 database reveal that resolving a payment dispute in a court when the value of the claim i s 500% of El Salvador's GNI per capita (about US$ll,OOO) is going to take 275 days on average (as shown in Figure 2.45). These data assume that the case i s heard in the country's largest commercial city (i.e. San Salvador) as well as other specific procedural assumptions. As already mentioned, the difference between the duration o f the Doing Business data and the ICs data can be attributed to a number o f reasons such as differences in respondents - the Doing Business data are collected through a detailed questionnaire answered by lawyers rather than companies; one law firm only fills out the Doing Business questionnaire as opposed to close to 500 firms interviewed for the ICs; the Doing Business questionnaire covers a standard commercial dispute,,butalso builds many assumptions into the case, which may not necessarily be true for the average payment dispute, as envisaged by respondents to the ICs; and finally, the ICs data cover the whole country, whereas the Doing Business data is representative o f the largest commercial center only. 126. Two other measures of contract enforcement in court indicate that court process is burdensome in El Salvador as the judge and the parties to the case have to go through 41 different procedural actions from the filing of the lawsuit until the moment the creditor receives his payment. The whole process would cost the plaintiff (creditor) 12.5% o f the debt amount in official fees, such as attorney fees, court fees, and enforcement (bailiff) fees. Given the assumed claim amount, total costs in fees would be US$1,375 - not an insignificant amount in a country with an average per capita income of US$2,200 per annum. While by number of procedures El Salvador ranks among the most burdensomejurisdictions inLatin America and the Caribbean (the L C R average i s 35 procedures), by official costs it has one o f the cheapest court processes inthe region (only Chile and Bolivia have a cheaper court process). 98 Table 2.11: Time Costs and Numberof Proceduresto Enforcea DebtContract inCourt Panama 45 355 37 Colombia 3 1 363 18.6 Haiti 35 368 25 Ecuador 41 388 15.3 Mexico 37 421 20 Peru 35 441 34.7 Venezuela 41 445 28.7 Source: World BankDoingBusinessdatabase. Note: Data are as of January 1,2004. 127. A regional comparison of the Doing Business data on contract enforcement shows that El Salvador's number of days to enforce a contract is close to that of Paraguay (285 days), but higher than in Nicaragua (155 days). However, it i s much faster than enforcement inthe courts of Guatemala (which is the longest inthe region and inthe world with 1459 days) and Honduras (545 days). Table 2.11 presents the three measures from Doing Business in 2005 on contract enforcement. 128. El Salvador does well compared to its neighbors and other South American countries when we break down the Doing Business data on time to enforce a contract in court into its three constituent stages: filing and service of the claim on the defendant; obtaining a judgment, including hearings and all procedures for evidence submission, opening and closing arguments, etc.; and enforcement (execution) of the judgment through court bailiffs or other enforcement agents. Figure 2.46 shows the duration o f the process in each o f these three stages - in El Salvador - as elsewhere - the first stage o f Filing and Service o f Process takes the shortest time to accomplish. The next two stages are similar in length- but not so in the other comparator countries. Enforcement takes the longest o f all three stages in Honduras and Peru; obtaining a judgment takes the longest in Guatemala, Brazil, Ecuador and Nicaragua, 99 129. Apart from these average DoingBusiness estimates of duration, we also have at our disposal Doing Business data on the minimum and maximum duration of the hypothetical contract enforcement case (as described earlier) in court. These data are important as they proxy the degree o f uncertainty a party to a court case faces upon initiation o f judicial proceedings. According to the data supplied by the Doing Business respondent lawyers, while the average duration o f the case is 275 days, the maximum it could take is 1,825 days - or 5 years. This i s the second longest maximum after Guatemala (1850 days), and i s followed by Brazil at 1460 days. Therefore, the degree o f uncertainty Salvadoran plaintiffs and defendants face i s rather high.Figure2.47 presents these data. Figure2.46: Durationof Contract Enforcementin Court: by Stage of the JudicialProcess Source:World Bank Doing BusinessDatabase Figure2.47: Minimum and MaximumDuration of ContractEnforcementin Court, by country I 8 0 0 I 6 0 0 1 4 0 0 1 2 0 0 5 I O 0 0 B O O 6 0 0 4 0 0 2 0 0 0 O u a f e m a l a E l S a l v a d o r B r s r i l H o n d u r a i P a r " E c u a d o r N 8 ~ a r n p u o ImBM i l j i m u m I M a x l m u mI Source: World Bank El Salvador Investment Climate Survey 130. The Doing Business data also allow us a closer look at the'monetary costs litigants face when going to court. Two o f these costs - as provided by the Doing Business survey respondents - give an estimate o f the average attorney and court fees. Table 2.12 summarizes these data in absolute terms, and as percentage o f each country's GNI per capita. El Salvador ranks in the middle o f the table according the absolute value o f applicable attorney fees (US$550), which represents one-quarter o f its income per capita. Attorney fees are highest - relative to income per capita - in Honduras (62%) and Peru (60%), and lowest in Guatemala at 20%. Interms of court fees, both El Salvador and Nicaragua allow for a free process - i.e. court fees are zero. Court fees are highest inPeruand Guatemala at 9% o f GNI per capita. 100 131. While the informationon attorney and court fees is interesting,and provides some indication to the costliness of the judicial process, more data is necessary to understand better the market for legal services. Attorney fees should be - in principle - determined by supply and demand in the market for legal services. High attorney fees may be indicative o f a scarcity o f quality legal counsel. On the other hand, low attorney fees may bejust a function o f a country's poor income. The way lawyers' fees are determined is also crucial: in some countries lawyers charge by the hour - which creates incentives on the lawyer's part to delay the case, other things being equal. In other countries the party and his legal representative are free to negotiate a lump-sum for the services provided. In still other countries, they can make the fees contingent upon the outcome o f the case. And in yet other places, there are mandatory rules which prescribe the amount o f attorney fees to be paid for each stage o f a certain court process. In El Salvador the most common way to remunerate attorneys is by using the contingent fee option - a percentage o f the judgment amount collected i s awarded to the attorney. Table 2.12: Attorney and Court Fees in USD and as Percentof GNIper capita bountry bttomey fees, US$ kourt fees, US$ bttorney fees, % oflcourt fees, % o f GNI~ GNIper capita per capita Peru 1290 200 60 9 Brazil 685 154 25 6 Honduras 602 40 62 4 ElSalvador 550 0 25 0 Ecuador 537 10 30 1 Guatemala 382 170 20 9 Nicaragua 238 0 33 0 132. Apart fromtime andcost of court enforcement of contracts,many authors6'pointto judicial formalism or complexity as being associated with court delays and judicial corruption. El Salvador's judicial complexity appears to be quite high - as given by the Doing Business data. The index ofjudicial complexity is a composite measure, covering such features o f thejudicial process such as the type o fjudge, the necessity o f legal representation, the presence o f written versus oral elements during the process, the regulation o f appeals, the regulation o f evidence, and i s built on a scale o f 0 to 100, with higher values standing for higher complexity. Figure 2.48 presents the index o f procedural complexity for the four Central American and three South American countries, which we compare across the ICs data as well. El Salvador ranks second after Guatemala in terms o f how complex and cumbersome its procedure is. 69See Djankov et al. (2003), Hendersonet al. (2004), Doing Business in 2004, Chapter 4. 101 Figure2.48: Index of Judicial Complexity -A Cross-Country Comparison O"afcm.la 61 S a l v a d o r P C , " N l C . , B 8 " @ E c u a d o r H o n d u r i i B 11.11 0 I O 2 0 3 0 4 0 S O E O 7 0 S O 9 0 I Source: World BankDoingBusinessDatabase Court Performance and Judicial Corruption 133. Anecdotal and additional available survey evidence also point to major problems in the Salvadorancourts and the public's trust in the judiciary. A report commissioned by the World Bank was prepared by the Institute o f Public Opinion o f the Central American University "Jose Simeon Canas" inEl Salvador. The study was aimed at evaluating public opinion about the courts and the judicial system through a survey which included questions in four main areas: Jurisdictional Matters; Knowledge o f Laws and Institutions; Confidence in the Courts; and Access to Justice. Confidence in the fairness o f the courts was low: 35% o f interviewed respondents answered that courts are never or seldom fair. Statements were given such as "Judges favor the rich more than the poor." Inparticular, it was found that respondents believed that economic power had the strongest influence on judges' decisions (73% o f respondents thought that economic power has such an influence). The corresponding figures for the influence o f the government and political parties onjudges' decisions are 36% and 39%. The survey found that 58.7% o f individuals had zero or little confidence in the courts, and 50% had zero or little confidence in the Justices o f the Peace. The figure for confidence in the Supreme Court o f Justice i s similar - 56% o f interviewed individuals had zero or little confidence init. 134. Althoughjudicial corruption is not the main theme of this chapter, many observers and legal scholars point to corruption in the judiciary as one of its main problems. The above report reveals that bothjudges and lawyers are perceived as corrupt: on a rating scale o f 1 to 10, where 1 stands for "zero corruption" and 10 for "very corrupt", 45% o f respondents assign values o f 7 to 10 for how corrupt judges are, and 55% o f respondents do so for how corrupt lawyers are. 73% o f surveyed individuals agree that one has to pay ajudge to win a court case. In line with these findings, 53% o f respondents found the judicial system having no or little honesty. As a measure o f the bias in judges' decisions, 72% o f surveyed individuals agreed strongly with the statement: " Judges favor the rich more than the poor". 135. Access to justice is determinedby informationabout laws, legal aid, and the costs of hiring a lawyer. Among the surveyed individuals, 91% had not used a court in the preceding year. 90% o f respondents agree that lawyers are expensive to hire, but less than half o f respondents (48%) report that they know where to find a lawyer without having to pay, i.e. can 102 access free legal aid. Lack o f information applies to other aspects o f the judicial process: for instance 73% o f interviewed individuals report that there exists insufficient informationabout the legal rightso f workers who have been fired without due cause. POLICYRECOMMENDATIONSCONTRACT ENFORCEMENT THE JUDICIARY ON AND 136. To improve court performancemeasures need to be taken to reduce case backlogs. One o f the symptoms o f slow justice i s the case backlog - and there is evidence that typical civil cases take 3 to 4 years from filing to disposition7' - something which was confirmed by the Doing Business data on maximum times for case disposition. A typical recommendation to address high caseload is to increase the number o f judges or bring in more computers and equipment - but studies have shown that this seldom works. Instead, the focus should be on making sure that court procedures are less cumbersome, case management works well and certain types o f cases - such as small claim cases or commercial cases can go to a specialized court, or be put on a track for mandatory mediation, for instance by industry association^^^. Alternative dispute resolution mechanisms are governed in El Salvador by the 2002 Law o f mediation, conciliation and arbitration. statistics - on numbers of cases filed per year, numbers pending, types of cases, etc. The 137. It is also important to improve court management and the collection of judicial current World Bank El Salvador Judicial Modernization Project (JMP) i s already addressing some o f these concerns. Improving court case management through computerization and assigning these tasks to court clerks rather than the judges has produced good results in some countries which have undertaken similar reforms in recent years (e.g. in some o f the courts in Slovakia72).According to World Bank (2003), each judge currently has to spend 70 percent o f hisher time on administrative tasks, thereby reducing their time `for their main task - case adjudication. Clearly, this could be an area to work - through the JMP, but also through other donors' support. 138. Cumbersome judicial processes should be simplified as they breed judicial corruption and create delays. As evidenced inthe Doing Business data, procedural complexity i s high, and even though procedures are usually instituted to ensure fairness and impartiality o f judgment, the empirical evidence suggests otherwise73. Simplification can apply to many elements o f the typical civil judicial process - through instituting more oral rather than written 70 See World Bank (2003), "El Salvador: Social and Institutional Context, Profile and Challenges o f the Judiciary and the Judicial Modemization Project". This diagnostic reports that in 2003 there were about 260,000 cases pending in El Salvador, o fwhich about 60% in the First Instance Courts, and 25% inthe Justice of the Peace Courts. It states that inrecent years there has been an increase inlabor, family, small claim and commercial cases filed with the civil courts. The report mentions that backlogs have grown as a result of insufficient capacity and inefficiencies, and are more pronounced at the first instance courts rather than at the appellate level and the Supreme Court o f Justice (SCJ). 7' This has proven very successful and reduced backlogs inArgentina. See World Bank (2001a), "Argentina: Legal and Judicial Sector Assessment", Legal Vice-presidency. 72 See World Bank (2001b), "Administration o f Justice and the Legal Profession in Slovakia", Poverty Reduction and Economic Management Unit, Europe and CentralAsia Region. 73 See Djankov et al. (2003). 103 procedures; through eliminating the need for plaintiffs to present legally motivated claims; through eliminating interlocutory appeals and motions, which defendants file just to delay payment; through improving the system o f notifications o f claim to the defendant and o f the judgment to both parties. 139. The Government of El Salvador should also address the perceived lack of independence and impartiality among judges, and interferenceby political and economic power in court decisions. For example, Lothian (2003) argues that some of the main problems o f the Salvadoranjudicial system are the political control over it (through judicial appointments to the SCJ), the centralizedcontrol o f the judiciary by the Supreme Court o f Justice, and the lack o f transparency and openness injudicial proceedings and decisions. She proposes several sets o f reforms, such as changes in the appointment procedure for Supreme Court Justices, the separation o f administrative and judicial functions - which was mentioned earlier as well - and the creation o f a career judiciary at the level o f the first-instance courts through competitive examinations for judicial appointments and changes to the curricula in law schools. Finally, she proposes increased transparency - for instance, at present judicial decisions are not published and disseminated ina timely manner. 140. Finally, reforms should also target information-sharing, which supports both execution of judgments as well as informalcontract enforcement. Public credit registries are often cited as an example o f information-sharing institutions, which can facilitate exchange by allowing a business to check the credit history o f another business, thus expanding transaction opportunities. Credit and asset registries can also be helpful in ensuring that the debtor's assets can be located and seized to pay the debt. Having ajudgment without the ability to collect on it is not much use. Therefore, efforts could be made to institute and improve information-sharing among market participants. BUSINESS-GOVERNMENTRELATIONS Business Registration 141. A heavier regulationof entry is associatedwith worse economic outcomes. A number o f recent empirical studies (see Djankov et al. (2002), Djankov et al. (2003), Glaeser and Shleifer (2002), Shleifer (2004)) have found that a heavier regulation o f business registration- measured in both the time required to register a firm and its cost - is associated with worse rather than better economic outcomes. The Doing Business Report (2004) argues that only two registration procedures - notification o f company existence, and a tax and social security registration - are sufficient for business regi~tration~~. As reported there, however, most countries impose more registration procedures such as labor registration (87% of Doing Business sample countries), administrative registration (76%), bank deposit (68%), notarization (62%), health benefits registration (62%), etc. 142. Accordingto ICs data, El Salvador has a faster registrationsystem than most other countries in the region. The El Salvador ICs data offer information about the number o f days and types o f registration procedures required by respondent manufacturing firms to register for l4See Chapter 2 in World Bank (2004), Doing Business in 2004: Understanding Regulation. 104 the first time.75A comparison among four Central American countries (El Salvador, Honduras, Nicaragua and Guatemala) reveals that - interms o f the ICs data - El Salvador ranks reasonably well in term o f days to register a business. Figure 2.49 shows that a typical firm in El Salvador takes 46 days to register compared to 29 days inNicaragua, 56 days in.Guatemala and 83 days in Honduras. Summingup the average number o f days requiredto go through six different types of registration procedures, El Salvador actually ranks better than Nicaragua with 74 days versus Nicaragua's 82, and better than Guatemala (251 days) and Honduras (214 days).76 Figure 2.49 also benchmarks the I C s data for these four countries against the Doing Business data. Somewhat surprisingly- inthe light o f the ICs data - the Doing Business database finds that the registration o f a limitedliability company in San Salvador takes 115 days on average (or close to four months), which i s the longest inCentral America.. The Doing Business and I C s data are not well correlated, particularly for the variable which is a summation o f the number o f days requiredto get six registrationpermits (the correlation is negative and highat -0.72). This could be attributed to Doing Business relating to only limited liability companies, and reflecting the answers o f only one law firm, which supplies the data on duration o f business registration. In contrast, the I C s covers more than 400 firms in each country, o f different legal status, and the respondents are the firm managers. The negative correlation is, nevertheless, puzzling. Figure2.49: Number ofDaysto Register a Firm:A Comparisonof DifferentData Nicaragua El Salvador Guatemala 25 I Honduras 0 50 IO0 I 5 0 200 250 300 days ONumber ofdays for registering a firm, Doing Business database .Total duration of registration (including obtaining all necessary licenses and permits), I C s BITotal average duration of registration (sum of average days needed to obtain six licenses and permits), ICs Source: World Bank Investment Climate Surveys and World Bank Doing Business Database 143. In El Salvador, the longest registration processes are related to health permits (30 days) and draft of constitution of the company (16 days). The other registration procedures take considerably less time - inscription in the Commercial Register and getting an operational license take 11 days each, whereas tax registration takes 4 days. Obtaining the necessary permits from the Ministry o f Agriculture and Agro-industry takes 8 days. Figure 2.50 shows similar patterns in the neighboring countries considered - tax registration i s relatively short, whereas 75These data pertain to firms registered in2000 or later. 76The six registration processes are draft of constitution of the firm, inscription o fthe firm inthe Public Registry, registration with the tax authority, operating license, registrationwith the Health Ministry, and environmental permits. 105 getting health permits is also a lengthy process in Honduras (alongside draft constitution o f the company). InNicaragua and Guatemala, agricultural permitstake a considerable amount o f time (approximately 25 and 30 days respectively). Overall, the figure 'depicts lengthier registration procedures in Honduras and Guatemala relative to El Salvador andNicaragua. 144. Micro and small Salvadoran firms spend less time than medium and large firms on the different steps of business registration.The same is also observed inthe three comparator countries. Table 2.13 reports simple t-tests for differences in mean number o f days to register a business for the first time, counting all the necessary permits and licenses to beginoperations. In each o f the four countries, micro and small firms report a significantly lower number o f days to register than medium and large firms do. We also find that in El Salvador a larger fraction o f medium and large firms hire outside assistance in getting registered (67%) compared to the fraction o f micro and small firms that get such assistance (50%). Similar results hold for Honduras, Nicaragua and Guatemala. Figure2.50: Time Neededfor DifferentRegistrationProcedures: A Comparison 60 50 40 6 30 20 I O 0 El Salvador Guatemala Honduras Nicaragua ,I Draft of constitution of the society Hlnscnption of the society inthe Public Registry Bl Agroindustry Ministry (MAG) I 106 Table2.13: Tests for Differencesin Time to Register betweenSmalland Large Firms Total Number of Days to Register a Firmfor the First Time, (mean) Wholesample Micro and small firms 41.5 Medium and large firms 87.0 t-statistic -4.0348 (0.0001) El Salvador Micro and small firms 33.9 Medium and large firms 77.3 t-statistic -1.9340 (0.0393) Honduras Micro and small firms 68.3 Mediumand large firms 103.4 t-statistic -1.5813 (0.0599) Nicaragua Micro and small firms 20.7 Medium and large firms 57.9 t-statistic -1.9517 (0.0415) Guatemala Micro and small firms 28.0 Medium and large firms 80.9 t-statistic -3.1664 (0.0024) Source: WorldBank Investment Climate Surveys Note 1: One-tailed t-tests are reported. P-values are shown in parentheses next to the t-statistics. Significant results at the 10% levelat a minimumare markedin bold. 145. The cost of registration in El Salvador - defined as the cost of hiring outside the ICs - is higher than that of Nicaragua but lower then in Guatemala and Honduras. specialized agents to assist with the various steps of businessregistration and measuredby Therefore, costs follow the pattern o f time needed to register. Table 2.14 presents a comparison o f costs o f hiring agents to assist with business registration in the four Central American countries and compares them with the number o f days to register a firm. One might expect that higher costs would be associated with faster registration - in other words, firms may be willing to pay more for getting registered faster. However, as found by the World Bank Doing Business report (2004), this i s not the case. Higher costs o f registration are found to be highly associated with lengthier regi~tration~~.Table 2.14 also shows the costs of business registration according to the Doing Business database. There i s a highnegative correlation between the costs measuredby the ICs and those measured by Doing Business (-0.72). For example, Nicaragua is at the two ends o f the ranking using the Doing Business official cost measure and the I C s hired-agents measure o f costs. The reasons for this could be due to both variables measuring different things: the I C s data measure what firms pay to hired professional agents to facilitate business registration; in contrast, Doing Business measures the official costs o f registration, such as paying various stamp duties, license fees, etc. under a specific set o f assumptions about the 77We need to stress that there is a caveat to this finding inthe sense that not all firms which reportthe average duration o f a first-time registration hired agents to assist them with the registration process or reportedthe costs o f hiring such agents. Nevertheless, in Guatemala and Honduras a large fraction o f firms which reported average number o f days to register also hiredagents and provide the costs o f so doing (93% and 78% respectively). InEl Salvador the corresponding fraction is 53%, and inNicaragua it is 30%. Therefore, comparing time to register and cost of hiring agents is more problematic inthe latter two countries. 107 firm7*.Inthis sense, the ICs data onregistration costs may not be entirely incompatible with the Doing Business data on costs o f registration. Country Cost of Registering a Firm Costs of Business Registration, Days to Register a Firm, ICs (payments to agents hired), (DoingBusiness data), data ICs data, %of annualsales % of GNIper capita Nicaragua 0.24 170.1 29 El Salvador 0.46 128.0 46 Guatemala 0.57 62.8 56 Honduras 1.21 72.9 83 Re-registration and Renewal of Licenses and Permits 146. Salvadoran firms take less time than their counterparts from Honduras and Guatemala but more than those from Nicaragua to obtain license and permit renewals. Among the subset o f firms registeredbefore 1999 the ICs asks whether in the year preceding the survey the firm applied for licenses andpermits necessary to continue operation. The survey then goes on to obtain information how long each specific license or permit took to get. Inaddition, a follow-up question asks how many days the renewal o f all types o f licenses and permits took. In terms o f the ranking o f the four Central American countries, we get results similar to those for first-time registration (Figure 2.51): overall re-registration and renewal duration is shortest in Nicaragua (14 days), followed by El Salvador (26 days) and Honduras and Guatemala (42 and 34 days respectively). Nt~aragua+;9 14 I i -1 I El Salvador 46 56 Guatemala - 34 0 10 20 30 40 5 0 60 70 80 90 days 0Total durationofregistration(including obtainingall necessarylicenserand permits) I durationofre-registrationandrenewaloflicensesandpermits(includingailnecessaryllcenrerandpermits) Total Source: WorldBank InvestmentClimate Surveys 147. In El Salvador renewals of health permits and environmental permits take the longest time (at 35 and 36 days each respectively).The same patternis also observed in two of the other three countries. For instance, in Guatemala, these permits take even longer to '*The assumedfirm is a limited-liability companysituatedinthe country's largest commercial center.For more on the assumptions, see Djankov et al. (2002). 108 renew (at 43 and 58 days respectively) and in Nicaragua health and environmental permits take considerably less to renew at 10 days each (Figure 2.52). Much like in El Salvador, environmental permit renewals take the longest in Honduras at 45 days. Altogether, environmental permit renewals take longer in El Salvador than inNicaragua (36 versus 10 days) but less than inGuatemala (58 days) and Honduras (45 days). Interms o fhealthpermit renewals, El Salvador trails bothNicaragua and Honduras, but does better than Guatemala. All other types o f document renewals in El Salvador take no more than one week at worst. Across the entire sample, renewal o f environmental permits takes the longest at 44 days, reflecting the above findings for each country, followed by renewals by the Agriculture7' and Health Ministries (27 days each). All other renewals take considerably less time - operational license renewals across the four countries takes about two weeks on average (13 days), and the other renewals take 10 or less days to obtain. 148. A significantly lower fraction of Salvadoran firms (26%) hire professional agents, lawyers and other intermediaries to assist them in getting their licenses and permits renewed compared to firms in Honduras (37%) and Guatemala (47%), but this fraction i s significantly higher than in Nicaragua (where only 16% of interviewed firms resort to hiring agents for license renewal). This indicator i s important as it can potentially signal in which countries obtaining license and permit renewals i s more cumbersome and difficult. These results also mirror the use o f professional help in registration - with the difference that the fraction o f firms hiring professional agents to get registered i s higher than that o f firms hiring agents to renew licenses and permits andor re-register. Figure 2.53 displays these findings. Table 2.15 presents t-tests for differences in mean fractions o f firms hiring agents for license renewal by country and firm size. By firm size, the tests indicate that inall the four countries and across the entire sample a significantly lower fraction o f micro and small firms use professional help compared to medium and large firms. These results indicate that insome o f the countries - e.g. Honduras and Guatemala - firms potentially face more burdensome renewal procedures, and are therefore more likely to use outside assistance in obtaining them. Furthermore, medium and large firms may also be subject to heavier license renewal regulations, which is why they also resort to the same strategy more often than micro and small firms do. 79Note that the Honduras ICs does not contain a category of Agriculture Ministry license renewals. 109 Figure2.52: Time Neededfor DifferentTypes of Renewalof Licenses and Permits:A Comparison 50 40 30 20 I O 0 El Salvador Guatemala Honduras Nicaragua Whole sample Operatinglicense Tax inspectorate OLicense of state suppliers OlmponLicense HExportLicense BHealth (Secretaria de Salud) Agriculture and Agro-industryMinistry (MAG) OEnvironment Permit Source:WorldBank InvestmentClimate Surveys Figure2.53: Hireof ProfessionalAgents in Business Registrationand Licenseand Permit Renewals Nicaragua El Salvador Guatemala Honduras I I I I I I 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 percent HPercentageoffirms hiring agentsto renew licensesandpermits @Percentageof firms hiring agentsto registerfor the first time Source:WorldBank InvestmentClimate Surveys 110 Table2.15: StatisticalTests for Differencesin Use of ProfessionalAgents for Renewalof BusinessLicenses and/or Permits: A Comparisonacross Countries and Types of Firms Firms using professional agents for renewal of licenses and/or permits, (%) By Country El Salvador 26.3 Honduras 47.3 t-statistic -6.2357 (0.0000) El Salvador 26.3 Guatemala 37.1 t-statistic -3.3310 (0.0005) El Salvador 26.3 Nicaragua 15.5 t-statistic 3.7727 (0.0001) Honduras 47.3 Guatemala 37.1 t-statistic 2.8602 (0.0022) Honduras 47.3 Nicaragua 15.5 t-statistic 9.9783 (0.0000) Guatemala 37.1 Nicaragua 15.5 t-statistic 7.0368 (0.0000) By Firm Size Wholesample Micro and small firms 22.3 Medium and large firms 44.2 t-statistic -9.1446 (0.0000) El Salvador Micro and small firms 17.8 Medium and large firms 33.9 t-statistic -3.8495 (0.0001) Honduras Micro and small firms 34.7 Medium and large firms 66.0 t-statistic -6.1628 (0.0000) Nicaragua Micro and small firms 12.7 Medium and large firms 23.5 t-statistic -2.2683 (0.0124) Guatemala Micro and small firms 26.5 Medium and large firms 50.6 t-statistic -5.0222 (0.0000) Source: World Bank InvestmentClimate Surveys Note 1: One-tailed t-tests are reported. P-values are shown in parentheses next to the t-statistics. Significant results at the 10% level at a minimum are markedinbold. 149. Like costs of first-time registration, costs of license/permit renewals in El Salvador are lower than those in Honduras and Guatemala, and about the same as in Nicaragua (Figure 2.54). As with time and costs o f business registration, we again find that higher costs do not imply a shorter time for renewals - on the contrary, the higher the costs o f license/permit 111 renewal by specially-appointed professional agents, the longer the time for such renewal (the correlation coefficient between time and cost o f license/permit renewal i s 0.91)80. Therefore, countries which regulate license and permit renewal more heavily, also impose higher costs (as defined by the ICs) o f obtaining such renewals. There is no trade-off between time and cost, as one might expect. Figure2.54: Costs of Registrationand LicenselPermit Renewalby Country: Percentof Total Annual Firm Sales I I I I I I I I El Salvador 0 46 I I I 0.00 0.20 0.40 0.60 0.80 1.oo 1.20 I.40 percent lmCost ofRegisteringa Firm mCostof Renewal of BusinessLicensesand/or Permits Source: WorldBank InvestmentClimate Surveys Land Transactions and Construction Permits 150. The regulation of land transactions and construction activities by firms is also an area of Government interventioninbusinessactivity. The ICs data reveal that inEl Salvador, 13% o f interviewed firms bought land, and 14% -- engaged in construction activities in the two years prior to the survey. These figures are similar to the other three Central American countries (Table 2.16). Country Percentage of Firms that Percentage of Firms that PurchasedLand Constructed a New Building Guatemala 8.8 14.6 El Salvador 12.9 13.6 Honduras 14.4 20.3 Nicaragua 14.9 21.6 _____~ Again, as with time and cost of first-timeregistration,this finding is subject to the caveat that not all firms which appliedfor some type of licenseor permit renewal, resortedto hiring agents for that renewalandreportedthe costs o f hiringagents. 112 151. Constructionpermits take longer to obtain but are cheaper in El Salvador than in the other three Central American countries. In contrast, land-related permits and registrations take less time, but are costlier than in the comparator countries. Thus, the time for obtaining construction permits i s longest in El Salvador (39 days), followed closely by Guatemala (33 days). The shortest time for obtaining a construction permit is in Nicaragua (20 days), while in Honduras it i s 26 days. In terms o f costs o f complying with construction regulations and obtaining the necessary permits, El Salvador registers the lowest costs o f construction-related permits (at 0.16% o f total annual sales). In contrast to time for obtaining construction permits, for their most recent land purchase Salvadoran firms take the least time to obtain the necessary permits and comply with land registration requirements (51 days). This measure is somewhat higher for the other three countries - it takes 54 days to get such land permits and approvals in Honduras, and 58 days in Guatemala and Nicaragua each (Figure 2.55). As shown in Figure 2.55, the time needed to obtain construction permits and all necessary requirements i s substantially shorter than the time needed for complying with land regulations. Unlike its costs for construction permits, El Salvador has the second highest costs of obtaining land-purchase-relatedpermits, at 0.48% o f annual sales (Figure 2.56). 70 1 __ - - - ~ .. 50 40 30 20 10 0 Guatemala Honduras El Salvador 1Tune to Aoquirs Permits for Land Purchases I to Asquirs Connrmctian Permits Time I Source: World Bank Investment Climate Surveys We do not show these t-test results for considerations of space. They are available upon request. 113 153. There appears to be a trade-off between time and cost of obtaining land permits and constructionpermits. Thus, we find a negative correlation between the cost o f obtaining land permits and the time for completion o f a land purchase transaction (correlation coefficient o f -0.78). Therefore, it appears that - at least for land transactions - higher costs are associated with speedier land registration and issuance of relevant permits. The same holds - albeit to a lesser degree - for construction permits and regulations. Time and costs o f obtaining construction permits and complying with construction regulations are negatively related in the four Central American countries, but the correlation i s weak (correlation coefficient o f -0.20). Finally, an interesting observation i s the fact that 22% (13 out o f 60) o f the subset o f Salvadoran firms which undertook construction activities, report that they constructed without the necessary permit.This variable is not available for the other three countries, so no comparison is possible, but it indicates that delays in obtaining construction permits may instigate construction without due permission. Figure 2.56 compares the two types o f regulatory costs - for construction and land transactions-by country. Figure 2.56: Costsof Compliancewith Requirements and Permitsto PurchaseLand andUndertake ConstructionActivities: by Country 0 7 01 - - __ _ _ _-- - I 0 60 0 50 0 40 0 3 0 0 20 0 I O 0 00 Honduras El Salvador Nicaragua Guatemala /&(Costof Land Acquisition and Registration DCort ofBuilding Permits AcquisitionI Source:WorldBank InvestmentClimate Surveys TaxRegulations 154. A relatively low fraction of interviewed Salvadoran firms (13%) report that they had to pay penalties and fines to the Directorate General of Internal Revenues in the year prior to the survey. However, the size of the tax fines is high compared to the other three Central American countries at 0.66% of total annual sales, which is higher than Guatemala (0.24%) and similar to Honduras (O.62%), but lower than Nicaragua (0.85%). InHonduras, the proportion of firms paying tax fines is similar to that of El Salvador (11%). In contrast, a significantlyg2higher fraction o f Guatemalan and Nicaraguan (24%) firms report having had to pay tax-related fines and penalties compared to firms in El Salvador and Honduras. Table 2.18 summarizes the data on tax fines and share o f sales reported to the tax authorities. Importantly, the average share of sales reported to the tax authorities is highest inGuatemala and El Salvador at 77% versus 66% and 68% in Nicaragua and Honduras respectively. The differences between the average shares o f reported sales are significantly higher in Guatemala and El Salvador *'These t-testresultsare not reportedfor brevity.They are availableuponrequest. 114 compared to Nicaragua and Honduras. Furthermore, the data indicate that countries with a higher size o f tax penalties report a lower proportion o f their sales to the tax authorities) - the correlation coefficient between the size of the tax penalty and the share of reported sales for tax purposes is (-0.69). Therefore, more tax penalties lower rather than raise the share of declared sales (tax revenues). 155. According to information by FUSADES (2003), tax regulations and tax rates are perceivedas one of the most problematicbusiness regulations by Salvadoranfirms. On the basis of a survey among Salvadoranprivate firms, the severity o f tax regulations and tax rates i s perceived to have worsened between 2001 and 2003 among exporters (from an average of 2.2 to 2Sg3in2003). Non-exporters also see tax regulations as most burdensome-the averagerating is 2.3 inboth 2001 and 2003. Country Share o f Firms Which Size of Tax Penalty, Percentage of Total Annual Sales Pay a Tax Penalty percent of total annual Declared to the Tax Authorities sales Nicaragua 23.9% 0.85 65.6 Guatemala 23.7% 0.24 77.0 ElSalvador 13.1% 0.66 76.9 Honduras 10.7% 0.62 67.7 156. In all four Central American countries medium and large firms report a higher proportion of their annual sales to the tax authorities, but the differences are only significant in Nicaragua, Honduras and Guatemala. In El Salvador there is no significant difference between reported sales by micro and small firms on the one hand, and medium and large firms on the other. In a similar fashion, foreign firms declare a somewhat larger fraction o f their sales to the tax authorities relative to domestic firms, but the results are statistically significant only in Nicaragua and Guatemala. Regarding the size o f the tax fines, a similar picture emerges: micro and small firms pay a higher fraction of their sales in tax fines than medium and large firms do, but the difference inmeans i s only significant for Honduras. In terms o f firm ownership, domestic firms pay a significantly higher fraction o f their sales in tax fines than foreign firms in all countries except for Honduras. A higher percentage of reported sales to tax authorities normally reflects a lower share o f the informal economy within a sector or industry.The above results suggest that inEl Salvador micro and small firms operating formally (informal firms are not covered by the survey) do not differ from their larger counterparts in terms o f reported sales, but do bear a higher burdeno f tax fines and penalties as a share o f their sales. 157. Smaller Salvadoran firms are very likely to hire outside assistance to comply with tax regulationsand the payment of monthly and annual taxes. A study by Martinez and de Vieytez (2002) basedon a survey, carried out by FUNDAPYME in2000, finds that among small and medium Salvadoran firms close to one third (32%) hire outside services to deal with government regulations. Among the major reasons cited for hiringoutside assistanceare taxation 83The severity of regulationis measured on a five-point rating scale, where 1=zero, 2=little, 3=medium, 4=high, 5=maximum obstacle. 115 procedures: for payment o f annual taxes (20%), monthly taxes (21%), rental taxes (24%), etc. The report also establishes that small firms are the ones which have most often required outside assistance for purposes o f dealing with the Government. Labor Regulations 158. Compared to firms from other Central American countries, Salvadoran firms appear to be less constrained by labor regulations. The ICs provides some informationabout the severity of employment regulation in the respondent countries. To gauge the constraints of labor laws and regulations a hypothetical question asks firms whether they would increase, decrease or leave unchanged the number o f their permanent employees ifsuch a change could be made without any restrictions (such as without seeking permission, making severance payments, etc.). Figure 2.57 depicts the findings o f this question. It shows that in El Salvador labor regulations appear the lightest compared to the other three countries as the lowest share o f firms (25%) report that they would increase employment ifrestrictions were liftedg4.They are also less likely to fire workers should regulations be eased - only 6.5% report an anticipated reduction in the number o f their permanent employees. In contrast, almost one third o f interviewed Guatemalan firms report that they would hire new workers under the scenario described; furthermore, almost 13% o f Guatemalan firms also expect to reduce their permanent employment. Therefore, close to one half o f interviewed Guatemalan firms (45%) anticipate changes to their workforce should labor regulations and restrictions be removed. Similarly, 31% o f Honduran firms expect to hire new workers and 8% -- to fire workers if labor regulations are lifted. InNicaragua, the difference between the share o f firms expecting to increase the number o f permanent employees i s similar to El Salvador (26%), and so i s the share o f firms which expect to fire workers (7%). Labor regulations affect to a greater extent micro and small firms. A significantly higher fraction o f micro and small firms report plans to increase employment compared to medium and large firms in El Salvador and Honduras, but not in Guatemala and Nicaragua. Moreover, foreign firms are less constrained than domestic ones. A significantly higher fraction of Salvadoran and Honduran domestic firms report plans to raise permanent workers compared to foreign-owned firms; again, the results are not statistically significant in Guatemala and Nicaragua. 159. Salvadoran firms cite prospects for an increase or decrease in sales and non-wage labor costs (such as social security costs) as the two most importantreasonsfor maintaining the level of their workforce above or below their establishment's current needs. Other labor regulations such as labor union pressures, laws and regulations regarding the dismissal o f workers, and difficult procedures at the Ministry o f Labor are deemed less important. The same ranking o f different types o f employment regulations are provided by firms in Guatemala and Nicaragua. However, in Honduras the most important reason given for maintaining the level o f employment above or below current business needs i s related to the laws and regulations regarding dismissal o f workers. Both micro and small firms rank sales prospects as the top 84This finding is also confirmedby FUSADES(2003), whereby the costs of dismissal to Salvadoran firms are quoted to be equal to the average for Latin America at 3 monthly salaries (Heckman and Pages (2000)). However, the same report lists social security costs (pension andhealth insurance) borne by the employer and employee at about 24% o f gross salaries, which is higher than the same measure for Latin America (18%). 116 reason for maintaining their current level o f employment above or below their actual needs, however a significant proportion o f small firms (43%) also view laws governing worker dismissal and labor costs as the top reason for keeping employment as it is. This proportion is much higher than for micro firms (16%), and higher than those for medium and large firms (32%) and 42% respectively. Therefore, when micro and small firms appear more constrainedby labor regulations, it may be in fact due to small firms' being subject to laws governing dismissal and to non-wage labor costs in a manner like larger firms. Incontrast, micro firms may be better able at avoiding some o f these regulations (for instance only 6% o f them see laws related to dismissal as the prime reason compared to 21% o f small firms). Figure2.57: Share of Firms Which Would Changethe Number of Their PermanentEmployeesifNo Labor RestrictionsWere in Place Guatemala 55 5~ Honduras 60 8 Nicaragua 6714 El Salvador I $8 8 0Percentageoffirms that wouldmcrease thenumberofworkers Source: World Bank Investment Climate Surveys Inspections and overall regulatory burden 160. As elsewhere in Central America, inspections by the Tax Authority are most burdensome in El Salvador (5 days per annum), followed by inspections by officials of the Ministry of Public Health (2 days per annum). The overall number of days spent in inspectionsper annum is lower in El Salvador than in Nicaragua and Honduras, and close to that of Guatemala. Firms often complain that inspections by different Government agencies and mandatory meetings with Government officials impose a serious burdenon their operations. As inthe earlier sections, utilizing the ICs data, we attempt to quantify the time and cost of such inspections and mandatory meetings for El Salvador and the three Central American comparator countries. An I C s question asks firms how many days in the previous year they spent in inspections and mandatory meetings with public officials such as representatives o f the Tax Authority, the Institute o f Social Security, the Fire and Safety officials, Public Health and Sanitation officials, etc. Figure 2.58 demonstrates that in all four countries firms devote the longest time to meetings and inspections performed by officials o f the tax authorities. In El Salvador the second longest period o f time spent in inspections i s with officials o f the Ministry o f Health (sanitary inspections), followed by the Salvadoran Institute o f Social Security (2 days each). On average, Salvadoran firms spend relatively less time with officials regarding Fire and Safety Regulations and with officials from the Police. Figure 2.59 shows that labor and social security, public health and environmental inspections take a similar amount o f time across the 117 four countries, on average. Municipal officials' inspections impose the highestburdens on firms inNicaragua and Honduras, but much less so inEl Salvador and Guatemala. Police inspections take less than half a day everywhere, apart from Nicaragua which averages 2 days per year. As in El Salvador, inspections related to fire and building standards impose the lowest burden across the four countries. Finally, summing up the days spent in inspections with officials from all the listed Government agencies, other than the Tax Authority, reveals that Nicaraguan finns bear the highest burden at about over 13 days per annum, followed by Honduras at close to 9 days per annum. Incontrast, firms' total time spent in inspections by the same agencies are substantially lower inEl Salvador (under 7 days) and Guatemala (over 5 days). Figure2.58: Number of Days Spent in Inspectionsand MandatoryMeetingswith Government Officials over the Year Precedingthe ICs 9 0 8 0 ' 7 0 6 0 ' A 5 0 4 0 3 0 2 0 1 0 0 0 Nicaragua El Salvador Guatemala Honduras 'mTax inspectorate .Labor and Salal Sccunty UFmand BulldugSafety UPublic Health .Police UEnnronmmtal BOfticials from the Municipality Source: World Bank InvestmentClimateSurveys Figure2.59: Number of Days Spent in Inspectionsand MandatoryMeetings with Government Officials over the Year Precedingthe ICs: By Type of Agency E O 7 0 5 0 4 0 30 20 I O 0 0 Tax Inapestoma bbor and Sofia1 F w andBuilding Public Hcsilh Pdiec Envcronmmlal Omelaln h m !he sesunty Safely Municipality jdNmapa .ElSalvador OGualmala OHanduraJI Source: World Bank InvestmentClimate Surveys 118 161. Salvadoranmedium and large firms do not spend significantly more time with the Tax Authority than micro and small firms, but they do spend a significantlyhigher amount of time in inspections by all other Government agencies taken together. We study whether there are differences across different firm groups to gauge whether certain types o f firms are more susceptible to lengthy inspections than others. If we restrict our attention only to the number o f days spent with the Tax Inspectorate officials, and conduct tests for differences in means by firm size and ownership,we establish that in El Salvador there are no significant differences between micro and small firms on the one hand, and medium and large firms on the other - eventhough the latter group do spend longer intax-related inspections. Incontrast, inthe other three countries medium and large firms - taken together - spend a significantly higher amount o f time with the tax authorities than micro and small firms do. Furthermore, when we do the same analysis by ownership, we findthat inall four countries foreign firms spendlonger with the tax inspectors on average -but the differences inaverage number o f days are only significant inHonduras and Nicaragua. Again, El Salvador (alongside Guatemala) does not seem to subject foreign-owned firms to substantially longer tax-related inspections than it does with domestically-owned firmsg5. We also conduct a test for differences in mean number o f days spent ininspections by all agencies other than the Tax Authority. Inall four countries, micro and small firms spend significantly less time on such inspections than mediumand large firms do. 162. Tax inspections appear costlier in terms of fines paid and goods confiscated by the visiting inspectors in Nicaragua and Guatemalarelativeto El Salvador and Honduras.The cost o f official inspections o f interviewed firms (such as the value o f fines paid or confiscated goods) i s not a very reliable measure due to a large proportion o f firms experiencing such inspections indicating that no fines were levied (or goods confiscated). However, even with this caveat in mind, some interesting findings emerge. For example, we only look at the value o f fineskonfiscated goods by visiting tax inspectors, and find that a larger fraction o f Nicaraguan and Guatemalan firms (20% and 16% respectively) did pay such fines. Incontrast, inEl Salvador and Honduras only 6% and 9% o f firms visited by officials o f the Tax Inspectorate ended up paying any tax-related fines. The average size o f the tax fines paid i s also higher in Nicaragua and Guatemala compared to El Salvador and Honduras. A higher fraction o f medium and large firms than micro and small f i r m s across the four countries end up paying tax inspection-related fines, with significant differences observed only in Nicaragua and Guatemala. In El Salvador there i s no significant difference in the proportions o f domestic and foreign firms paying fines, but in the other three countries a higher proportion of foreign firms end up paying, with Guatemala showing a significant difference. Finally, a significantly higher proportion o f exporting firms compared to non-exporting firms are fined by the tax inspectorates in El Salvador and Guatemala. 163. Salvadoranmanagers spend on average about 7.4% of their working time per year in meetings regardingdifferent types of government regulations.This is lower than the same measure for the other three Central American countries (11% in Guatemala, 8.6% in Honduras and 8.0% in Nicaragua). The I C s also gathers information about the time senior management spends in dealing with various Government regulations and agencies (such as tax authorities, customs, labor costs, licenses, permits, etc.). Table 2.20 presents this measure by firm size. The ''The results of these t-tests are not shown for brevity. They are available upon request. 119 larger a firm is, the higher the share o f senior management time spent dealing with regulations. This pattern is observedinthe other three comparator countries as well. Type ofFirm Mean Median observations Micro 5.3% I12.8% II No. of 99 Medium 9.2% I 5.0% I 111 Large II8.0% I 5.0% 1 109 Source:El Salvador ICs 164. Overall, the Salvadoran maquila firms are subjected to more inspections by different Government agencies other than the Tax Authority, and spend on average more time with regulators than non-maquila firms. However, they spend one average less time renewing licenses and permits to continue operations as well as obtaining construction and land permits. A higher proportion o f maquila firms in El Salvador are subject to paying tax fines, but the size of the fine is lower than for non-maquila firms. All in all, the ICs data suggest that the Salvadoran maquila firms benefit from a favorable treatment by the Government, but do end up spending more time in meetings with Government officials and in more days in inspections. Similar results are found for maquila firms inGuatemala, Honduras and Nicaragua. InHonduras, maquila firms are disadvantaged also by longer times for license renewals, construction permits, and landpermits. Perceptions of the Regulatory Burden 165. El Salvador does better than the other three Central American countries in ranking the severity of obstacle which different areas of government regulation represent,with the exception of the area of quality standards and certification,in which it fares worst. The ICs gathers information on perceptions as to how problematic different aspects o f government regulations areg6. Salvadoran firms see the areas o f quality standards and certification, and environmental regulations as most problematic for doing business with between 10 to 15% o f interviewed firms viewing them as a major or very severe obstacle. Safety and sanitary regulations, and procedures to register a new firm are perceived as a major or severe problem by 9% and 8% o f Salvadoran firms. In contrast, inspections, patent and trademark regulations and currency regulations are perceived as a major or very severe problem by around 6% or less o f interviewedfirms (Figure 2.60). We compare El Salvador with the other three Central American countries in Figure 2.61, and find that the fraction o f Salvadoran firms which consider each o f the seven areas o f business regulation to be a major or very severe constraint is lowest in the areas o f currency regulations, patent and trademark regulations, inspections (although close to those in Nicaragua and Honduras), and procedures to register a firm. This is in line with the 86 The question asks respondentswhether seven different types o f government regulationspose a problemfor the operation andor growth o ftheir business, and elicit answers on a scale of 0 to 4, inincrements of 1, whereby 0 indicates no perceivedproblem, 1-a minor problem,2 -amoderateproblem, 3 -a major problem, and4 - a very severeproblem. 120 findings o f relatively lower time and costs of different government regulatory activities in El Salvador compared to the other three countries. However, in the area o f quality standards and certification the fraction o f Salvadoran firms deeming this area a major or very severe obstacle i s the highest among the four countries. Finally, in the areas o f environmental regulations and safety and sanitary regulations El Salvador ranks better than Guatemala and Honduras, but worse than Nicaragua. Across firms with different characteristics, non-maquila firms inEl Salvador are found significantly more troubled by the area o f quality standards and certification than maquila firms are. The same applies to inspections - in fact none o f the interviewed Salvadoran maquila firms find inspections a major or very severe obstacle to doing business. This i s in contrast to the finding that maquilas spend longer with Government officials and being inspected. Exporters are also significantly less likely to consider inspections problematic compared to non-exporting firms. As in the other three countries, micro and small firms are significantly less likely to find environmental regulations problematic than are medium and large firms. In contrast, currency regulations are perceived as problematic by a significantly higher fraction o f micro and small firms. In the other areas o f business regulations, however, there are no major differences among firms o f different characteristics such as size, ownership, beinga maquila or an exporter. Figure2.60: SalvadoranFirmsWhich Consider RegulationsMajor or Very Severe Problems( O h ) Quality ntandards and c c r t i h a l i o n Environmental rcgulstianr Safety and sanitary regulstlonE Procedures to register a now en lerprire lnrpcctlonr Patent and trademark regulations Foreign exchangelcurrcnoy regulatlonr 0 2 4 6 8 10 12 I 4 percent Source: World Bank InvestmentClimate Surveys 121 Figure2.61: CentralAmerican ConsideringRegulationsMajor or Very Severe Problems(%) Patentand trademark regulations Inspections Foreign exchangelcurrency regulations Safer) and sanitwy regularions Environmental regulations Qualib standardsand ccnificauon Procedures to rugislcra ncu enterprise 0 5 10 I 5 20 2s percent IWEl Salvador mGuatemala OHonduras ONicaraguaj Source: World Bank Investment Climate Surveys 166. Salvadoran firms see the effectiveness of appealing tax-related decisions, the severity of tax penalties and frequent changes in tax rules and tax rates as the most cumbersome aspects of tax administration - a view which is shared by firms in Honduras, Nicaragua and Guatemala. However, these areas of tax administration are seen as much less problematic in El Salvador compared to the other three countries. In a manner analogous to the one discussed in the previous paragraph firms are asked to assess the operation o f the Tax Administration Authority on the basis o f their experience in'the twelve months prior to the survey. Seven different aspects o f tax administration are evaluated. The results for El Salvador are depicted in Figure2.62. Effectiveness o f tax appeal mechanisms are seen as one o f the more problematic areas o f tax administration, followed by the severity o f tax penalties and the frequency o f changes in tax rules and rates. Compared to the other three Central American countries, El Salvador does better in each o f the seven evaluated areas o f tax administration. In six o f these areas El Salvador does best, except for the area o f burden o f tax audits, where all four countries rank similarly. Inthe three most problematic areas o f tax administration -such as the existence o f effective mechanisms for tax appeals, the severity o f tax penalties and the frequency o f changes in tax rules and tax rates,- a substantially lower fraction o f Salvadoran firms see these as a major or very severe problem compared to firms in Guatemala, Honduras and Nicaragua. There are no significant differences among different types o f Salvadoran firms in terms o f their perceptions o f how problematic the frequent changes intax rules and tax rates are; however, maquila firms and foreign-owned firms see effectiveness o f tax appeals as less o f a problem than non-maquila and domestically-owned firms respectively. The difference between maquila and non-maquila firms is statistically significant. Finally, mediumand large firms, non- maquila firms and non-exporting firms see the area o f tax penalties as significantly more problematic than micro and small firms, maquila and exporting firms do. 122 Figure 2.62: SalvadoranFirmsConsideringAspects of the Tax Administration Major or Very Severe Problems(YO) Effcctitencrr ofappeals mcrhanirmr Severity o f penalties Frequency ofchanges in NICSand tax rates Burden of audits Ava,lability of information on tax requirements I Treatment by tax authorities Completiagifiling tax forms 0 2 4 6 8 10 12 14 percent Source: World BankInvestmentClimateSurveys Figure2.63: CentralAmerican FirmsConsideringAspects of the Tax AdministrationMajor or Very Severe Problems(YO) Effectivenessof appeals mechanism Severity of penalties Frequencyofchanges in mles and tax rates Burdenof audits Completmg/filing tax forms Availability of information on tax requirements Treatmentby tax authorities 0 5 I O I5 20 25 30 35 40 45 percent IRIIEl Salvador Guatemala 0 Honduras ONicaraguaI Source: World Bank InvestmentClimate Surveys POLICY RECOMMENDATIONS ON BUSINESS-GOVERNMENT RELATIONS usefulness of regulations - the so-called "grabbing hand" versus "the helping hand" 167. As mentioned at the outset, there are opposing theoretical views about the arguments. There is no doubt that due to externalities, market power and asymmetric information in the market some degree o f government regulation i s needed. For instance, governments need to ensure that the products firms produce are safe for consumers. Or, governments need to make sure that certain environmental standards are met before firms can begin operation. In a similar vein, governments regulate employment through labor laws to ensure that worker safety is not at risk. Despite all these arguments for regulation, recent 123 empirical evidence87suggests that modern governments regularly over-regulate business activity, and that such practices are detrimental rather than beneficial for economic activity and growth. We summarize some general policy recommendations for specific regulatory practices in El Salvador, giventhe above survey results as well as available data from alternative sources. 168. Efforts to reduce the time and costs of business registration should continue, particularly in terms of expanding the experience of the Oficina Nacional de Inversiones (ONI) to domestic firms, SMEs and firms across all regions of the country. In terms of business registration, time and costs inEl Salvador have already been reduced given the ongoing reform efforts - i.e. the establishment o f a one-stop-shop for business registration (ONI). Its establishment follows current best practice, which has been in the direction o f removing registration o f firms out o f the courts, reducing the involvement o f notaries, and making it an administrative process (see Doing Business in2004 for more on this). ON1has been instrumental over the past couple o f years inreducing the time for business registration o f foreign firms from about 100 days to 10 days on average88.It i s now trying to expand its services to domestic firms. In addition, efforts are being made to copy the ON1 experience to other agencies such as CONAMYPE, which cater to micro and small firms. Similar one-stop windows are being put in place there, thereby encouraging informal firms to register for the first time and become formal. 169. As evidenced by the data, El Salvador ranks well region-wide in terms of days for re-registration and license renewals, however some improvements can be made, particularly with respect to renewalof environmentaland health (sanitary permits). Inthe area o f construction and land permits there are some problems as well -although not as severe as in some of the other comparator countries. Labor regulations are not seen as that problematic either - although other available evidence8' suggests that labor laws are rigid and firms find it hardto hire and fire workers. 170. Tax reformsshould target simplificationof tax administrationprocedures and laws, the mechanismsfor appeals and reclamationof corporate taxes paid. Inspections by the Tax Authorities are reported to take the longest among inspections by various public agencies with regulatory functions. At the same time a relatively low fraction o f Salvadoran firms complain that the workings o f the tax administration are a major problem, and Salvadoran firms declare a relatively highshare o f their revenues for taxation purposes. Nevertheless, efforts can be made to further improve the interactions of firms with the Tax Authority and the Ministry of Finance regarding corporate taxation. In addition, firms complain about ineffective mechanisms to re- claim taxes paid, or appeal taxation decisions, as well as about the frequent changes o f tax laws and tax rates. Additional evidence, e.g. FUSADES (2003), supports these results and finds tax regulations to be perceived as particularly burdensome by Salvadoran firms. Therefore, some attention should be given to the above issues. "SeeDjankovetal.(2002)ontheregulationoffirmentry;Djankovetal.(2003)ontheregulationofcourtsand contract enforcement; Glaeser and Shleifer (2002) on the rise of the regulatorystate ,etc. "Seeforexample,ImplementationCompletionReport,ElSalvadorCompetitivenessEnhancementTechnical Assistance Project,World Bank (2004). 89See Doing Business in 2004, Chapter 3. El Salvador's index o f the flexibility o f hiring ofworkers (81 on a 0 to 100 scale, with highervalues signifying less flexibility) is among the worst in the Doing Business sample of 130 countries, and also equalto those ofMexico and Panama. 124 Finally, policy reforms should aim at creating a level playing field for all firms through targeted reforms of administrative simplification for MSMEs, firms different industrial sectors or geographic regions. The ICs data revealthat certaintypes of firms appear to be more disadvantaged than other firms - e.g. medium and large firms take longer to register, and are subjected to heavier regulations. The opposite usually holds for costs for different regulatory activities - micro and small firms seem to pay proportionately more than medium and large firms. While a more careful analysis i s needed to ascertain the effects o f type o f firm, ownership, and other firm characteristics on time and cost o f regulations, it i s clear that the playing field for Salvadoran firms i s far from level. Accordingly, actions could be taken to alleviate some o f these differences. 125 ANNEX2.1: ECONOMETRICESTIMATES To investigate the determinants of bribe payments across firms, we test empirically the validity of Svenson's (2003) framework, in which the likelihood of such payments depends positively on the firms' required dealings with the public sector and on its current and expected profits.Firmsmay needto deal to a greater extent with the public sector -and thus be ina weaker positionintheir negotiations with corrupt public officials-either becausethey need public services to a greater extent or because they operate in a more heavily regulated sector or location. In addition, more profitable firms may also have a weaker bargaining position, since corrupt officials can demand larger bribes, which those firms are better able to pay. Svenson also postulates that firms with a larger cost o f refusing to pay - measured by the difference between the profits that could be earned if the firm was to pay the requestedbribes and stay inbusiness, and those that could be earned in a different activity - are also positively related to the probability o f paying bribes. Even though we do not have adequate proxies for the alternative return on capital - associated with the sunk cost component o f the firm's technology - we postulate that a firm's willingness to refuse paying bribes should be positively affected by its access to information on government regulations and by its confidence in the honesty and efficiency o f thejudicial system. We proxy for the firms' dealings with the public sector using five alternative indicators, and also control for current profits, capital stocks, employment size, confidence in the judiciary and access to information on government regulations. The five proxies for government contacts are: a dummy for the firms that requested public services (electricity, telephone and water connections) during the two years preceding the survey, the number o f government inspections reported during the year preceding the survey, the number o f government permits obtained by the firm duringthe two years preceding the survey or as a result o f its initial creation, a dummy for the firms that sell to the government, and the first principal component o f the four above mentioned variables. As Svenson (2003), we control for current profits and proxy for expected future profits using capital stocks, Moreover, since "presumably smaller firms can more easily avoid detection by government authorities" (p. 214) we also control for employment size. To measure the availability o f information on government regulations, we include a dummy variable for firms that report being members o f industry associations that provide that type o f information. Finally, we include a dummy for firms that state that they do not believe that the judicial system would enforce their contractual and property rights in the event o f a commercial dispute. Our empirical estimated equation i s o f the following form: (A2.1) where Bribei i s either a dummy variable for firms that report that bribes "to get things done" are common intheir sector or the level o f reported bribes per worker, Sizei i s the log o f employment, Infoi and Judici are the above described dummies for firms that get information on government regulations from their industry associations and firms that do not trust the judicial system, Profti i s a vector that includes both current profits and the capital stock (to proxy for future profits), Govi i s one o f the five above mentioned proxies for the presence o f dealings with the public sector, and ~i i sa random error term. When the dependent variable i s the level o f bribes per 126 worker, we follow Svenson in rescaling profits and capital stocks by the level o f employment. We estimate equations (A2.1) using data from Salvadoran plants and pooled data from plants from El Salvador, Guatemala, Honduras and Nicaragua. In the second case we also include country dummies, as'well as an interactive between profits and a dummy for El Salvador (in order to test whether that variable does infact carry a different signinboth samples). The results as well as other estimation details are reported inTables A2.1 and A2.2, respectively. We also estimated an econometric model aimed at investigating, in a multivariate framework, the characteristics of the firms that are more likely to report security expenditures, as well as those associated with crime victimization and monetary losses relatedto criminalactivities. Our empirical estimatedequationis ofthe following form: where Crimei is either a dummy variable for firms that report expenditures inprivate security, a dummy for firms that indicate having been victims of criminal acts, or the percentage o f sales that were lost due to criminal activities (for victimized firms). We control for employment size (Sizei) and include a dummy variable for maquilas, as well as sector (Sectori) and province (Provj) dummies in all specifications. Finally, ExpOSurei i s either a dummy for firms that were victims of criminal acts (when the dependent variable is the dummy for security expenditures), or the percentage o f the firms o f the same size, sector and region that report security expenditures (when the dependent variables are either the dummy for crime victimization, or crime losses as a percentage o f sales). The results, with samples covering either just Salvadoran firms, or also firms from Nicaragua, Honduras and Guatemala are reported intable A2.3. For the third sub-section of this chapter, we estimate an econometric model aimed at identifying the characteristicsof the firms which use the courts as a vehicle for resolving their paymentdisputes, as well as other determinants of court use such as the time it takes to resolve a court case or the proceduralcomplexityof commerciallitigation. Our empirical estimated equation i s o f the following form: Court-Use, = P'Size, + P'Sector, +P30wnership, +P4Maquila, +P5Exporter, + +P6Age, P'BA-Used, + P8Private-Delayed, + j39Days-to-Resolve, + P'OCourt-Eflciencyj ++ p"Court-Feesj + 0 (A2.3) where Court-Use, is either a dummy variable for firms that report having filed court cases related to payment disputes inthe two years preceding the survey, or the percentage o f payment disputes which ended up going to court. We control for employment size (Size,), using dummy variables for micro, small, medium and large firms. We also include among the independent variables firm ownership dummies (Ownership$ for domestically-owned private firms, and for foreign-owned firms. Next, we include a dummy variable for exporter (Exporter,), and a dummy variable for maquilas (Maquila,,. In addition, sector (Sectori) dummies appear in all specifications. Among firm characteristics, we also include the log o f firm age (Agey). The variable BA-Used, is a dummy for whether or not business associations are used in resolving disputes with workers, the Govemment, or other firms. We would like to test whether using 127 business associations to resolve disputes i s a substitute or complement to court use. The variable Private-Delayedg measures the (positive) percentage o f annual sales paid by private clients with a delay. We would expect that a higher fraction o f delayed payments would raise court use. Finally, the variable Days-to Resolveg measures the number o f days needed on average to resolve an overdue paymentgo: A higher duration o f alternative dispute resolution mechanisms will be expected to increase court use. All the preceding variables are firm characteristics describing firm i in countryj. The regressions also include country-level variables related to the operation o f the courts and based on the World Bank Doing Business data. The variable Court Efficiencyj i s either the Doing Business estimate o f the time necessary to enforce a commercial contract in court; the Doing Business the index o f procedural complexity, which measures how burdensome the court procedures are; or the number o f procedural steps to enforce a commercial contract in court. The variable Court-Feesj i s the Doing Business estimate o f the mandatory court fees payable by plaintiffs in each country j. We would expect that more procedures, a more burdensome court process, a higher duration o f contract enforcement in court, and higher fees would reduce the use o f the courts. We utilize a pooled sample coveringg1 firms from El Salvador, Nicaragua, Honduras and Guatemala, and the main results are reported intable A2.4. 90Presumably by other means rather than court enforcement, although the question does not specify the exact means. 91Restrictingthe sample to firms from El Salvador only does not allow us to use the Doing Businessdata on court efficiency and fees. Furthermore, the sample size would be considerably reduced. 128 Table A2.1: Determinantsof BribePaymentsin ElSalvador Bribe Bribe Bribe Bribe Bribe Log(bribes Log(bribes Dummy(a) Dummy (a) Dummy(a) Dummy(a) Dummy(a) p/w)@) p/w)") Log (employment) 0.056** 0.055** 0.060** 0.065** 0.052* -0.756*** 0.118 (0.038) (0.040) (0.027) (0.016) (0 057) I (0.000) (0 732) I Gov.Regul.Info. from Ind.Assoc. 0.144** 0.135** 0.148** 0.149** 0.134"" 0.154 1.096 (0.014) (0.023) (0.011) (0.012) (0.024) (0.753) (0.310) Firmdistrusts Judicialsystem 0,083 0.079 0.077 0.078 0.085 1.271*** -0.561 (0.150) (0.166) (0 174) I (0.168) (0.142) (0,010) (0.613) Profitrate (ifpositive) -0.076 -0.021 -0.035 -0.040 -0.050 (0.676) (0.908) (0.848) (0.831) (0.784) Log (capital stock) -0.030* -0.031* -0.030" -0.031* -0.033* (0.083) (0.074) (0.082) (0.074) (0.061) Requestedpublic service 0.063* (0.059) Number of Inspections(list of 7) 0.023 (0.178) Number of Permits(list of 14) 0.006 (0.670) Firmsells to the govemment 0.169*** (0.005) Index of GovemmentContacts 0.043** 0.182 0.934** (0.047) (0.364) (0.014) Profitsperworker (US$ thousands) 0.031 0.089 (0.440) (0.195) Log (capitalperworker) -0.025 -0.083 Observations 428 428 428 428 428 113 348 Robustp values inparentheses. significant at 10%; * ** significant at 5%; *** significantat 1%. (a) Probit estimateswith robuststandard errors. (b) OLS estimateswith robust standard errors.(a) Tobit estimates. 129 Table A2.2: Determinantsof BribePayments in ElSalvador and neighboringcountries Bribe Bribe Bribe Bribe Bribe Bribe Log(bribesLog(bribes Dummy(a)Dummy(a)Dummy(a)Dummy(a)Dummy(a)Dummy(a)p/w) @) p/w)(') Log (employment) 0.037*** 0.032** 0.034** 0,041*** 0.027** 0.027** -0.462*** -0.157 (0.006) (0.017) (0.011) (0.002) (0.048) (0.048) (0.000) (0.256) Gov. Regul.Info. from Ind. Assoc. 0.071** 0.061" 0.068** 0.072** 0.056* 0.056* -0.120 0.356 (0.023) (0.053) (0.029) (0.020) (0.076) (0.077) (0.609) (0.399) Firm distrustsJudicial system 0.084*** 0.088*** 0.086*** 0.089*** 0.085*** 0.085*** -0.168 0.360 (0.001) (0,001) (0,001) (0,001) (0.001) (0,001) (0.386) (0.317) Profit rate (ifpositive) 0.210** 0.229** 0.225** 0.218** 0.211** 0.315*** (0.018) (0,010) (0.012) (0.015) (0,019) (0.003) Log (capital stock) -0.017" -0.017" -0.017* -0.015" -0.020** -0.021** (0.059) (0.051) (0.052) (0,090) (0.022) (0.020) Requestedpublic service 0.075*** (0.000) Numberof Inspections(list of 7) 0.032*** (0.000) Numberof Permits (list of 14) 0.023*** (0.004) Firm sells to the government 0.112*** (0.000) Index of Govemment Contacts 0.064*** 0.065*** 0.316*** 1.024*** (0.000) (0.000) (0.000) (0.000) Profitsper worker (US$thousands) 0.017*** 0.022** (0.000) (0.021) Log (capitalper worker) 0.027 -0.119 (0.705) (0.316) Profit rate ESA dummy * -0.391" (0 059) I Observations 1644 1644 1644 1644 Robustp values inparentheses. significant at 10%; * ** significantat 5%; *** 1644 1644 597 1351 significant at 1%. (a) Probit estimates with robust standarderrors. (b) OLS estimateswith robust standarderrors. (a) Tobit estimates. 130 Table A2.3: Correlatesof expendituresin security, crimevictimizationand crimelosses (plant level data from ElSalvador,Guatemala,Hondurasand Nicaragua) Private Private Crime Crime Crime Crime Crime Crime security security victim victim victim losses (% losses (% losses (% dummy (a) dummy (a)dummy (a)dummy (a) dummy (a)of sales) @) o f sales)(b) o f sales)(b) Employment(log) 0.154*** 0.174*** 0.102*** 0.069*** 0.071*** -0.779** -0.669*** -0.663*** (0.000) (0.000) (0.000) (0,000) (0.000) (0.029) (0,001) (0.001) Maquiladummy -0.184 -0.144** -0.229** -0.129*** -0.132*** 2.712 1.319 1.282 (0,109) (0.026) (0.011) (0,008) (0.006) (0.240) (0.134) (0.146) Crimevictim dummy 0.053 0.098*** (0.229) (0.000) Privatesecurity (a) -0.365** 0.070 0.138* -0.386 -0.102 0.282 (0 024) I (0.324) (0.072) (0.877) (0 946) I (0.866) ESAdum.*Priv. Sec. (a) -0.309** -1.730 (0.014) (0.490) Chemicals 0.119 0.170*** 0.238*** 0.118** 0.120** 0.412 -0.600 -0.590 (0,109) (0.000) (0.010) (0.023) (0.022) (0.751) (0.495) (0.504) Rubber& plastics 0.141 * 0.164*** 0.048 -0.002 -0.001 0.068 -1.460** -1.441** (0.058) (0.003) (0.649) (0 973) I (0.983) (0.953) (0.032) (0.035) Mineralprod. 0.000 0.112*** -0.058 0.068 0.051 3.572* 0.849 0.754 (0.994) (0,001) (0.521) (0.157) (0.293) (0.062) (0.411) (0.464) Textiles 0.041 0.059 -0.079 0.065 0.066 0.082 -0.712 -0.738 (0.748) (0.318) (0.555) (0.328) (0.323) (0.952) (0.437) (0.421) Metalprod. 0.089 0.145*** -0.056 0.008 0.003 0.521 -0.746 -0.839 (0.120) (0.000) (0.475) (0.870) (0.958) (0.719) (0.406) (0 338) I Food& bev. 0.100* 0.107*** 0.163** 0.163*** 0.160*** 0.276 -1.236** -1.264** (0.066) (0.002) (0.019) (0.000) (0.000) (0.785) (0.032) (0 027) I Observations 464 1766 464 1766 1766 181 601 601 Robustp values inparentheses. significant at 10%; * ** significant at 5%; *** significant at 1%. (a) Probit estimates with robust standarderrors (marginaleffectsreported) (b) OLS estimates with robuststandarderrors.Provincedummies includedinall regressions. 131 Table A2.4: Determinants of Court Use in Central America Court Use CourtUse Court Use Court Use CourtUse Percentage Percentage Dummy (a) Dummy(a) D ~ m m(a) Dummy(a) Dummy (a) of Cases y of Cases Filed@) Filed(b) Small 0.450*** 0.467*** 0.461""" 0.462*** -0.002 7.576** 7.936** (0.008) (0.007) (0.007) (0,009) (0.995) (0.023) (0.018) Medium 0.827*** 0.838*** 0.828*** 0.854*** 0.502** 15.368*** 15.589*** (0.000) (0.000) (OIO0O) (0.000) (0.023) (0.000) (0.000) Large 1.025*** 1.052*** 1.042*** 1.043*** 0.679*** 17.523*** 18.099*** (0.000) (0.000) (0.000) (0.000) (0.005) (0.000) (0.000) Foreign 0.263*** 0.247*** 0.259*** 0.232*** 0.226 7.116"" 6.683** (0.002) (0.000) (0.000) (0.000) (0.214) (0.035) (0.048) Exporter 0.064 0.080 0.071 0.085 -0.023 0.946 1.365 (0.535) (0.440) (0.491) (0.417) (0 856) I (0.693) (0.570) Maquila -0.217 -0.215 -0.227 -0.165 -0.373 -6.025 -6.054 (0.342) (0-338) (0.327) (0.480) (0.164) (0.267) (0.265) Log (age) 0.046 0.027 0.033 0.031 0.038 1.089 0.654 (0.397) (0.644) (0.560) (0.620) (0.558) (0.373) (0.591) Percent of PrivateDelayedPayments 0.005*** 0.005*** 0.005*** 0.005*** 0.002 0.106*** 0.108*** (0,009) (0.007) (0.007) (0.008) (0.390) (0.002) (0.002) Log (days to resolveapaymentcase) 0.178*** 0.178*** 0.174"** 0.187*** 0.211*** 3.113*** 3.115""" (0.000) (0.000) (0.000) (0.000) (0.000) (0.003) (0.003) ProceduralComplexity -0.165*** -0.387** (0.000) (0.O16) Log (Court Time) -0.138*** -3.112** (0.000) (0.013) Court Fees -0.002*** (0.001) Log (Number o f Court Procedures) -0.287""" (0.000) BA ServicesUsed 0.196* Observations 1066 1066 1066 Robustp values inparentheses. significantat 10%; * ** significantat 5%; ***1066 662 1066 1066 significant at 1%. (a) Probit estimateswith robust standarderrors. (b) ) Tobit estimates. Standarderrors are adjustedfor clusteringon countrywhen country-level independentvariables areusedinthe Probit regressions.Sector dummies are includedinall specifications. 132 CHAPTER3: INFRASTRUCTURE ELECTRICITY 171. Electricity supply is fundamental for efficientlyrunning private enterprises. Access to a reliable and reasonably priced electricity service i s a major consideration for potential investors seeking to take advantage o f the country's resources. Additionally, the power sector in itself i s a potential destination o f substantial foreign direct investment (FDI). Both o f these aspects will be analyzed inthis section in order to identify strengths, weaknesses and challenges facing the sector. 172. Between 1996 and 2000 the sector was restructured and a large number of Government-owned assets in generation and distribution were privatized. Until 1996 electricity service in El Salvador followed the pattern o f many countries in the region with a vertically integrated structure at the generation and transmission level owned by Government through a national power enterprise (Comision Ejecutiva Hidroelkctrica del Rio Lempa, also known as CEL). At the distribution level, supply was a private business under a 50-year concession contract that expired in 1986; at that time, the enterprise reverted to public ownership and was taken over by CEL. However, CEL maintained the administrative and operational structure o f the distribution enterprise. The power sector's restructuring was primarily motivated by the Government's need to mobilize private sector resources for investment, particularly in generation. The privatization process was also aimed at freeing capital that could be channeled towards the rehabilitation o f the social sectors, which faced important challenges after the country's civil war. Thus, in contrast to many other countries, efficiency increases were not urgently required: CEL was a well-organized company, and the distribution business had been kept inshape within the limitationsassociated incopingwith war. 173. The restructuring was executed according to what is now known as the standard model, with a competitive wholesale market, competition in generation, regulated transmission and distribution, and the introduction of the supply business. With the market (Mercado Mayorista de Energia Ele`ctrica - MME). On the supply side, actors in the exception o f some supplies from a few small generators, all electricity i s traded inthe wholesale market comprise generating companies (GenCos) and suppliers who trade energy imported from other countries in the region. On the demand side, market participants include distribution companies (Discos), some large consumers, and suppliers. Trade inthe wholesale market takes place through (a) bilateral contracts between GenCos/ suppliers and Discos/large consumers/suppliers, and (b) a spot market (Mercado Regulador del Sistema -MRS), which operates according to price bids. 174. The contract market -with freely agreed prices between participants- accounts for over 50% of all trades in the MME. Bid prices in the spot market determine (i)a marginal price which represents the market's spot price, (ii) the dispatch o f different productionresources according to whether their bidprice i s lower or equal to the spot price, and (iii) the remuneration, at the spot price, o f dispatched bidders, and charges, at the spot price, to those participants that extract energy from the spot market. Final demand is supplied as follows: (1) large consumers who participate in the MME receive their supplies from GenCos or from importing suppliers at 133 prices that correspond to contracts and an eventual remainder at the spot price; (2) consumers who have opted to enter into agreements with suppliers other than their local Disco, can negotiate and agree their purchase price; and (3) other consumers are suppliedby the Discos and are billed at a regulated price. In 2003, 4,311GWh were traded in the MME; the Discos extracted 3,895GWh (go%), large users extracted 294GWh (7%), and suppliers extracted 122GWh (3%). By type of commercial instrument, 2,218GWh (51%) were traded through contracts and the balance was traded through the spot market. In other markets, the contract market usually accounts for a greater proportion o f all trades (around 80%). Figure 3.1: Installed Electricity Generation Capacity, kW per 100 pop. 86.2 Source: World Development Indicators. 175. Interconnections with Guatemala and Honduras play an important role in the wholesale market. They allow power imports from these countries and, more broadly, from other Central American power systems (Panama, Costa Rica, Nicaragua, Honduras, Guatemala). Agreements among the countries have enabled the operation o f a joint regional market so that users in El Salvador can be supplied by contract by a generator outside the country, and generators in El Salvador can seek customers physically attached to other systems, as well as enter into agreements with other generators, e.g. for backup purposes. Currently, an interconnection with Mexico i s being built, which will hrther expand this market. In 2003, El Salvador's energy balance showed a net import o f 325GWh, equivalent to around 8% o f energy trades inthe MME. Figure 3.2: Electricity consumption, kWh per capita c 3000 2500 2000 01980 1500 m1990 1000 02001 500 0 Source: World Development Indicators. 134 176. El Salvador has committedto increasingits transmissioncapacity in the context of the regional SIEPAC agreements. Due to its modest extension, transmission in El Salvador can be achieved through a 115kV network. The only higher voltage (230kV) lines are those interconnecting the system with Guatemala and Honduras. This may soon change as required by commitments within the context o f the regional SIEPAC interconnection line. This transmission investment will close a loop from Guatemala to Honduras and will improve power transfers within the Salvadoran system. Given the uncertainty associated with assuring the supply/demand balance, backing ETESAL in order to count on support from the interconnection becomes a high-priority element o f sector policy. 177. Generation is divided between a state-owned company and several private thermal generators. Generators in the wholesale market include a government owned company (CEL) which operates hydro power plants (427MW available capacity), and several thermal power generators with private sector participation. The main private generators are Duke Energy, which acquired some o f CEL's thermal plants (289MW) during the privatization process, Nejapa Power, which also acquired thermal plants duringprivatization (144MW), and LaGeo, which i s jointly owned by the Government and a strategic partner (ENEL o f Italy) and operates and develops geothermal facilities (115MW). In addition, there are two large industrial cogenerators (Cementos El Salvador and Compaiiia Azucarera Salvadoreiia) with a total 50MW o f thermal power. Total installed capacity amounts to around 1,100MW and available capacity i s on the order o f 1,025MW. As seen inFigure 3.1, El Salvador's generation capacity i s lower inper apita terms thanthat o f most South American countries. However, El Salvador's per capita electricity consumption i s higher than that o f neighboring Guatemala, Honduras and Nicaragua (Figure 3.2). 178. There is a relatively high degree of concentration in distribution. There are five distribution companies (Discos) that supply final users: (i)CAESS: Compaiiia de Alumbrado Electrico de San Salvador ; (ii) AES-CLESA: Compaiiia de Luz Electrica de Santa Ana y Cia; (iii)DELSUR:Distribuidora deElectricidaddelSur; (iv) EEO:EmpresaElectricadeOriente; and (v) DEUSEM: DistribuidoraElectrica de Usulutan. Their respective sizes interms o f market share and number o f users (2003 data) are given in Table 3.1. It must be noted, however, that although CAESS and EEO were initially sold to ENERSAL (a subsidiary o f Electricidad de Caracas), AES has subsequently acquired control o f those two companies, together with DEUSEM. Consequently, AES now controls four distribution companies, which account for 79% o f all users and 73% o f retail energy sold. The fifth company, DELSUR, was sold to PPL (originally Pennsylvania Power and Light) and EMEL o f Chile (majority owned by PPL). Electricity sales o f Discos amounted to 3,770GWh in 2003, broken down into domestic (37%), small (12%), medium (5%) and large (42%) industrial and commercial users - the balance corresponds to public lighting and other users. This is a relatively robust structure, with most sales going towards larger users, which makes for easier meteringand billing. 135 Table3.1: Numberof users, sales and marketshares of distributioncompanies CAESS AES- DELSUR EEO DEUSEM Total CLESA Users 473 251 260 186 50 1,220 (thousands) Share of 39% 21% 21% 15% 4% 100% Users Sales 1,689 624 1,001 372 84 3,770 (GWh) Share of 45% 17% 27% 10% 2% 100% Sales Source: Siget. 179. The restructuring of the power sector also involved the appearance of intermediaries, as well as the creation a new transmission company, and a new company aimed at performing system dispatch and clearing-house functions. The restructuring allowed for the appearance o f suppliers ("comercializadores "), which broker transactions between users and generators. These companies are market participants who are allowed to buy energy from generators, from the spot market, or from foreign suppliers, and sell it to final customers. In 2003 there were eight registered suppliers in the Salvadoran wholesale market, only two o f which were active. Their clients include both large industrial and small residential and commercial users. The supply business allows for free contracting between parties, subject to the payment o f network charges to the distribution company that physically connects a user to the electricity system. Enabling connections between agents o f the wholesale market i s the responsibility o f the transmission company, ETESAL. This government-owned company (its major shareholder i s CEL) was also created as part o f the restructuring process. ETESAL i s remunerated by the generators that inject power into its network. This regulated income allows it to operate the existing transmission infrastructure and invest in new facilities. Finally, system dispatch and clearing-house functions in the wholesale market are performed by Unidad de Transacciones (UT), a private company whose members include generators, distributors, the transmission company, and user representatives. Suppliers are also expected to be represented in the UT in the near future. The UT also manages the Central American Wholesale Market (Mercado Ele`ctrico Regional -MER). The latter operates as a supra-national market that coordinates transactions between wholesale markets inindividual countries. 180. Establishing market rules and electricity pricing is the responsibility of the Superintendence of Electricity and Telecommunications (SIGET). The regulatory framework i s contained in the Electricity Law and its associated secondary legislation ("reglamento "). Electricity prices for consumers that do not participate inthe wholesale market are regulatedby SIGET. They comprise four components (generation, transmission, distribution, and supply). The generation component reflects an average spot price in the wholesale market and, starting in 2003, it could include contract prices in the wholesale market that have been negotiated through an open bidding process. The transmission component and the distribution component are regulated values that reflect the cost o f transmission charges, o f distribution assets, and their operation. The supply component reflects consumer-related costs (e.g. customer service, meter reading, and billing). The formulas for determining these components are established on a 5-year basis, with annual adjustments to reflect changes in their underlying parameters. 136 Figure3.3: Distributionand TransmissionLosses as Percentageof Output 25 20 15 10 5 0 Source: World Development Indicators. 181. Distribution and Transmission losses are lower than in neighboring Guatemala, Honduras and Nicaragua. In 2003, losses in the distribution network amounted to 11% o f energy injected into the DisCos' systems. This i s a very good indicator when compared to most utilities in the region (e.g. Panama, with 15.5% losses in distribution for 2003). Overall losses, including transmission, amounted to 12% o f energy injected into the system. Losses show a decreasing trend, e.g. a reduction o f around half a percentage point between 2002 and 2003. Moreover, the available international figures (for 2001) indicate that transmission and distribution losses are much lower in El Salvador than inneighboring Guatemala, Honduras and Nicaragua, and they are also lower than inthe Brazil, Colombia and Mexico. DisCos' estimates o f non-technical losses (i.e. theft, unmetered consumption and meter error) range between less than 1% in CAESS, the largest company, and 4.7% inthe smallest (DEUSEM). Although losses vary among DisCos, allowable losses (those that can be passed onto consumers) are limitedby SIGET in regulated prices, thereby providing an incentive to reduce them. Another common efficiency measure i s the number o f customers per employee. For the Salvadoran DisCos, this varies between 900 and 1000, which i s substantially above that o f many companies in similar settings, which do not exceed 300 customers/employee. These figures point towards well- managed distribution operations that reflect on their financial statements through consistent profits as well .as good credit ratings. 182. One of the main challenges faced by Salvadoran regulators i s that of promoting new investments in electricity generation. One o f the major current concerns o f the Government i s the need to maintain a balance between supply and demand o f electricity. In 2000, maximum demand was 760MW for an installed capacity o f 1,100MW; a 30% margin was considered adequate for supplying the load (although in a hydro-dependent system it is not necessarily the best measure). In 2004, demand i s expected to reach 820MW for a supply o f 930MW; reserve has reduced to an unacceptable 12% when the system is considered in isolation. International interconnections have provided a much-needed backup, but there i s a limit to their capacity. For 2005, the situation i s expected to worsen, with installed capacity barely covering maximum demand; in 2006 two new geothermal plants are expected to come on line and provide some respite. These developments are at odds with the expectations o f the restructuring effort; putting inplace the wholesale market was supposed to create a price signal for stimulating private sector investmentinthe sector, and ingeneration inparticular, which has not materialized. 137 183. One of the main reasons behind the absence of new investments in generation capacity is the fact that the current system entails a high degree of uncertainty regarding the expected returns of those investments. Investors' appetite for enteringthe power sector in developing countries has declined after a number o f fiascos, particularly in Latin America, as shown by a Bank survey conducted in 200292.Moreover, inthe specific case o f El Salvador, the government chose a market structure in which remuneration o f generators takes place exclusively on the basis o f the spot price for energy. Albeit theoretically correct (e.g. it was the model adopted in England and Wales, as well as in several other systems), this system only provides a return on investedcapital o f marginal producers when prices are high due to reduced margins or to a dry season that limits supplies from hydro resources. However, this only occurs sporadically and governments usually intervene when prices reach too high a level, e.g. by establishing price ceilings. El Salvador has not been an exception to this rule: in 2003 the Government introduced new regulations, one o f which established a cost-based ceiling for the spot market price in order to reduce its volatility, under anticompetitive practices detected by SIGET. Inthis context, investors are rightly wary o f runningthe risk o f sinkingtheir capital into enterprises with erratic and uncertain returns. 184. Various alternatives are being considered to promote investments in the sector, including Power Purchase Agreements and compulsory long term contracts. As an alternative to the current scheme, potential investors have sought to avoid market risk by suggesting that the Government enter into Power Purchase Agreements (PPAs) with them. The Government has radically opposed such contracts, as they would defeat the purpose o f the reform process by creating contingent liabilities for the public sector, which were supposed to have been eliminated. Past experience inEl Salvador, as well as in other countries, has revealed many flaws when entering into these types o f contracts, e.g. a PPA contracted between CEL and ElPaso was considered unsatisfactory due to a number o f conditions (minimumofftake, capacity payments) and it was ultimately bought out by CEL. Another alternative i s for investors to enter into long term contracts in the market, either with distribution companies or with large consumers, in order to cover their market risk. The existing contract market i s unsuitedfor this purpose, as its objective i s to smoothen the contracting companies' cash flow, rather than to assure supplies to the Discos. Indeed, contract prices are mostly tied to the spot price and contracts themselves run for around one year, as opposed to the 10-20 year span that investors would expect from a PPA. This explains the relatively weak participation o f the contract market within the wholesale market. In 2003 the Government allowed the price o f contracts to be introduced into tariffs as long as they are solicited through a public bidding process. With current contracts the measure remains ineffectual as long as they remain short term, but it is a step in the right direction by facilitating longer-term agreements. However, as long as the absence o f long-term contracts bears no consequences for the Discos and other net purchasers in the wholesale market, such arrangements are unlikely to materialize, whether or not the contract prices can be reflected inelectricity tariffs. 185. Another well substantiated government concern is the possibilitythat in the context of a price-biddingsystem a relativelyhigh degree of concentrationin generationcould lead to excessive price volatility. With only a handful o f generators, the spot market exhibits a high 92Private Power Investorsin Developing Countries. Survey 2002 - Preliminary Findings, Ranjit Lamech& Kazim Saeed 138 potential for exerting market power, particularly in a price-bidding market context. This has apparently contributed to a highvolatility o f prices in the spot market. The price-bidding market adopted in El Salvador differs from other countries in the region, which have adopted a cost- based approach whereby participants declare costs and availability. The price-bid approach has been criticized as beingprone to manipulation inthe presence o f market power, and it i s asserted that a cost-based system i s less likely to be subject to opportunistic pricing. This i s probably correct insofar as pricing under extreme conditions, when the demandsupply margin i s very low and outrageous price bids are likely to occur. However, cost pricing has also been criticized insofar as cost declarations are difficult to verify and require scrutinizing by the regulator. In addition, market power can still be exercised by spurious unavailability declarations on the part o f dominant producers, i.e. they can create an artificial scarcity. Inother words, inthe presence o f market concentration, the influence o f major actors will manifest itselfone way or another. 186. A relatedconcernis that the distributionbusinessis also highlyconcentrated,with a minimalparticipationof suppliers and final users in the wholesale market. The distribution business comprises a natural monopoly associated with physically conveying electricity to consumers through its network (the so-called `wires business'), and also -but not necessarily- the actual supply of electricity to consumers. The `wires business' is necessarily a regulated activity, whereas the supply business -which requires modest infrastructure investment, e.g. in metering facilities- i s potentially competitive. Market concentration in the `wires business' can be controlled through regulation, but the latter must then become more intrusive, as efficiency costs become more opaque, e.g. the regulator can no longer rely on instruments such as `yardstick regulation' or cost comparisons among firms. Because Discos traditionally operated the `wires business' together with a supply business to consumers in their geographical service area -indeed, the distribution companies were privatized with both businesses-, they are reluctant to allow some o f their consumers to `escape' to alternative suppliers. Breaking the Discos' hold on their consumers i s not easy and, because they are in effect `suppliers o f last resort', care must be taken to compensate for eventually stranding them with high-cost users by allowing other suppliers to `cream-skim' their markets. 139 Table 3.2: Average Electricity Spot Market Prices (US$/kWh) Month 1998 1999 2000 2001 2002 2003 Jan 5.85 6.13 8.70 6.41 6.77 7.51 Feb 7.06 5.79 9.18 6.64 7.08 7.89 Mar 5.92 6.19 10.67 6.68 5.61 7.86 Apr 5.49 6.15 17.37 7.25 6.49 7.83 May 7.32 6.58 7.44 7.05 6.91 7.03 Jun 5.98 7.68 6.53 7.08 5.31 6.02 Jul 5.20 6.49 5.81 7.36 6.39 7.20 Aug 4.67 5.72 6.40 6.99 7.00 7.45 Sep 4.66 6.14 6.48 5.45 6.66 6.55 Oct 5.23 5.69 5.89 5.83 6.74 6.85 Nov 5.80 6.73 6.05 6.3 1 7.20 6.62 Dec 6.65 7.44 5.96 6.98 7.28 7.53 TOTAL 5.93 6.36 8.30 6.66 6.61 7.25 Source: SIGET 187. While the operational performance of the wholesale market has been generally satisfactory, the government has recently taken measures to reduce price volatility. The wholesale market has operated satisfactorily in the sense that the mechanics o f dispatching the system and clearing transactions, and financial obligations between the participants have been observed. As for prices in the spot market, there have been considerable variations from one month to another, as reflected in Table 3.2. These variations reflect different factors such as dry/wet season supplies o f hydro energy, as well as unit availability. For a market susceptible to significant market power, average spot market prices appear to be fairly stable, which i s partly a consequence o f CEL's behavior in the market. As a Government enterprise, and although it i s constrained to show a profit, it has exerted its influence as the largest generator in the system by seeking to mitigate possible opportunistic pricing on the part o f other market participants. However, Table 3.2 does not capture the full divergence o f spot prices within shorter periods. During2003, within a 1-month period spot prices exhibited madmin ratios that ran from a low o f 103:43 in November to 165:0.25 in January. Price volatility in the spot market and the presence o f `price spikes' have been perceived as undesirable because o f their potential for disturbing regulated retail prices. Thus, as noted above, the Government moved in 2003 to establish a ceiling spot market price based on the cost o f diesel-based generation under anticompetitive practices detected by SIGET, and to modify the price o f extracting energy from the spot market. 188. Electricity services are subsidized to favor small domestic consumers with under 100kWh/month. The difference between the full tariff and a reference tariff established in November 1999 i s subsidized by 85% in these consumer categories. Overall, this i s equivalent to subsidizing around 57% and 43% o f the tariff for each group, respectively. In terms o f electricity sales, subsidized energy i s equivalent to around 10%o f kWh sales, and to around 28% o f total domestic kWh sales. Although the subsidy scheme does not discriminate between wealthy or poor consumers, the lOOkWh limit i s not unreasonable. In addition, due to the existence o f fixed charges for supply service and network use, the average price per kWh for subsidized consumers i s actually higher than that o f domestic users above 1OOkWhper month. 140 Table 3.3: Average Electricity Sales Price (US$/kWh)"' 1998 1999 2000 2001 2002 2003 CAESS 9.79 9.42 10.45 10.97 10.04 10.72 DELSUR 11.14 10.95 11.83 12.05 11.49 11.83 AES-CLESA 11.62 11.22 12.23 12.79 12.50 12.57 EEO 12.98 12.41 13.63 14.02 13.38 13.88 DEUSEM 14.82 13.26 14.10 15.43 14.41 14.25 Total 10.69 10.42 11.45 11.98 11.22 11.71 (-2.5%)2 (9.8%)2 (4.6%)2 (-6.3%)2 (4.3%)2 'Excluding subsidizedconsumers 'Percentage changewith respect to previous year Source: SIGET 189. Variations in spot market prices are reflected in regulated retail prices, with adjustments currently taking place on a biannual basis. The evolution o f retail electricity prices i s shown inTable 3.3. The price variations across Discos are due to market composition - CAESS, the largest Disco, has the lowest price because o f a greater proportion o f larger consumers - and to the fact that regulated network charges are higher for the smaller companies, which have a lower density market. Changes in the average sales price are due, as noted before, to adjustments in the parameters o f the regulated network and supply components, but principally to variations in the spot market price. For a small consumer, for instance, the generation component - which corresponds to an average o f spot prices inthe wholesale market - is 73% of the retail price, with the rest corresponding to network use (23%) and supply charges (4%). Initially, this component was adjusted on a quarterly basis. Because spot prices could vary significantly from one quarter to another, consumers were confronted with erratic bills. For example, during 2000 average retail prices (a) decreased by 3.3% in January, (b) increased by 33% in April, and (c) decreased by 33% in August. Starting January 2001, the energy component was readjusted monthly to avoid large swings, but in June 2003 adjustments reverted to a longer-term (6 month) interval. How to calculate the generation component of regulated retail prices i s a tricky problem in most systems. In general, the approach consists o f finding a suitable formula that conciliates the trade-off between price stability (i.e. dampening the vagaries o f the spot market) and ensuring that Discos and other suppliers can recoup their wholesale market costs without incurring excessive financial costs or reaping excessive short term profits (the objective being to achieve effective 'pass-through' conditions), but a consensus on how to attain this objective has yet to develop. 190. There is evidence that commercial and industrial users pay substantially lower electricity prices in ElSalvador than in other neighboring CentralAmerican countries. On the question o f whether prices o f electricity in El Salvador to final consumers are reasonable, a comparison in regional terms as shown in TabIe 3.4 provides some helpfil insight. Electricity prices in El Salvador (at least in the case o f the largest company, CAESS) appear to be the highest inthe region for domestic users, but they are among the lowest for large commercial and industrial users. Residential prices above the lOOkWh level are higher El Salvador than in Honduras, Nicaragua and Costa Rica as a consequence o f limited subsidies (as well as the absence o f cross-subsidization). Prices shown inTable 3.4 were calculated from tariff schedules 141 o f selected companies, for selected types o f users (prices for other companies may differ from those shown).93 Table 3.4: Electricity Prices in Central America for Selected Users (US$/kWh)(I) Guatemala ElSalvador Honduras Nicaragua Costa Rica Panama E.E.G.S.A C.A.E.S.S. E.N.E.E E.D.Ny S I C E ELEKTRANE RESIDENTIAL 100 kWh 8.56 7.54 5.54 8.78 6.20 12.56 250 kWh 8.49 11.44 7.61 8.98 7.19 12.39 1,000 kWh 14.16 10.91 10.46 13.09 10.14 12.74 COMMERCIAL 1,000 kWh (small) 16.68 10.90 12.55 13.30 11.61 9.55 15,000 kWh, 41 kW(medium) 15.24 11.00 12.55 11.86 9.80 11.53 50,000 kWh, 137 kW (large) 14.98 8.46 12.55 12.66 9.26 11.52 INDUSTRIAL' 15 000 kWh, 41 kW(medium) 15.17 8.46 12.55 11.81 9.80 11.53 50 000 kWh, 137 kW (large) 14.98 8.46 12.55 11.88 9.26 11.52 100000 kWh, 274 kW (large) 12.42 8.46 10.14 10.14 9.26 9.80 930 MWh, 2500kW(very lge.) 8.44 9.06 10.09 6.10 7.84 1,488 MWh,4MW 8.44 9.06 10.09 6.10 7.84 '30 June 2004 'Based on 50% LoadFactor Source:Calculationsoriginate inICE(Instituto Costarricensede Energia), basedupon information suppliedby powercompanies:E.E.G.S.A : EmpresaElectricade GuatemalaSA, C.A.E.S.S.: Compaiiiade Alumbrado Electric0de San Salvador, E.N.E.E: EmpresaNacionalde EnergiaElectrica, DISNORTE y DISSUR : Nicaragua, I.C.E, ELEKTRA NORESTES.A : Panama 191. Salvadoran regulators face important challenges in terms of achieving increases in the quality of electricity supply. Quality control includes three major areas: technical service quality, technical product quality and commercial service quality. Measurements o f technical service quality in electricity utilities comprise primarily frequency and duration o f power cuts. In El Salvador, SIGET is trying out a procedure for providing incentives to improve these indices, The usual approach i s to institute a reimbursement to consumers who are affected by indices that exceed a predetermined standard. During an initial trial period (which started in October 2002 and ended in June 2003) Discos implemented the required measurement procedures; during a transition period (ending December 2004) Discos are expected to take remedial measures to achieve the required standards, and starting January 2005, compensation to consumers will begin to take place when standards are not met. The reliability benchmarks set by SIGET are shown in Table 3.5, together with some representative values o f utilities from the U.S.and Panama, and actual values registered by CAESS in 2003.94It would appear that the 93 It is worth noting that other sources of information - notably OLADES - provide somewhat conflicting information on the average electricity prices charged inthe region, thus suggestingthe need for future more detailed analysis of this issue. 94 This table should be interpretedwith care, as the indices may vary accordingto definitions that are particular to each utility (typically certain utilities exclude catastrophic events such as hurricanes), and they depend on factors suchas loaddensity, weather patterns, andwhether undergrounddistribution predominates. 142 SIGET benchmarks should be tightened progressively and that Discos such as CAESS will require a significant effort to adjust to them. Thus, for instance, the actual number of power outages per consumer per year during 2003 for the largest Salvadoran distributor (CAESS) was about 3.5 times larger than SIGET's benchmark for 2005. The latter is in turn about twice as large as the actual number o f outages observed in Panama, and among the bottom quartile o f US.distributors. As for technical product quality indicators, they are measuredon a sample of measures to achieve the required standard^.^^ consumers indicated by SIGET, and Discos are constrained to take necessary corrective Finally, commercial service quality i s measured by SIGET through a number of indicators such as times to connect new users, to resolve billing disputes, to respond to requests for service etc. Although as o f December 2003 most o f the quality benchmarks set by SIGET had yet to be achieved, there has been definite recent improvements in overall indexes.Thus, for instance, the number o f transmission and distribution line outages decreased by 26% between 2002 and 2003. Table3.5: ElectricityReliability Measures SAIFI' SAID? CAID13 SIGET benchmark2005 (urban) 9 15 1.7 US average4 1.26 1.95 1.47 USbottom25%4 3.9 7.05 3.28 Panama,2003 (urban, actual) 3.3-6.9 4.2-8.1 1.27-1.18 CAESS 2003 (urban) 32 45 1.41 ' SAIFI (System Average lnterrupti?nFrequency Index) -the averagenumber of intenuptioils experiencedby customers per year. - SAIDI (SystemAverage InterruptionDurationIndex) -the averagenumber of interruptionhours experiencedby customers per year. CAlDl (Customer Average InterruptionDurationIndex)-the averagedurationof an intemption, equal to SAIDI divided by SAIFI. 'Richard E.Brownand MikeW. Marshall, ABB Consulting, Trmsmissiori undDisfrihiitim World,Dec 2001 192. Despite the challenges faced by Salvadoran regulators in terms of the quality of electricity supply, data from investment climate surveys places El Salvador ahead of Guatemala, Honduras and Nicaragua. As seen inFigure 3.4,48% o f the surveyed Salvadoran manufacturing firms reported that they had monetary losses as a result o f power outages, compared to about 63% for the three above mentioned neighboring countries. Moreover, 71% o f Salvadoran firms report at least one power outage, and 56% indicate that they had to stop production as a consequence: in comparison, as much as 91% o f Honduras firms had outages, and 77% o f Nicaraguan firms report production stoppages resulting from interruptions in electricity services. 95Technicalproductquality refersto parameters such as voltage control, flicker, andharmonics,whichhavebecome increasingly significant with the utilizationo f sensitiveelectronic equipment. 143 Figure3.4: Productionstoppages and monetary losses due to Power Outages 91 07 0atleastoneoutage stoppedproduction 0had monetary loss Source: Investment Climate Survey 193. Even though the quality of electricity supply appears to be better in El Salvador than in neighboring countries, the current status-quo is not at all satisfactory, with average monetary losses due to power outages consuming 1.3% of the firms' revenues. The survey data indicates that Salvadoran firms have an average o f 10 power outages per year, compared to 14 in Guatemala, and around 30 in Honduras and Nicaragua (Figure 3.5). Not surprisingly, the monetary losses reported by Salvadoran firms are smaller than in the other three countries: they represent 1.3% of sales for the average firm, compared to an average o f 3.5% for the other countries. New installations also take less time in El Salvador: 17 days, compared to 18 for Nicaragua, 33 inHonduras and 65 inGuatemala (Figure 3.6). Figure3.5: Losses Due to Power Outages Figure3.6: Days for new ElectricityConnection "umber ofpower outages m h s s e s as % of sales Source: Investment Climate Survey. Source: Investment Climate Survey. 194. Within El Salvador, large firms suffer to a smaller extent from the existing quality problems in electricity supply. Moreover, firms located outside o f San Salvador wait less for new electricity connections and manufacturers o f chemicals and non-metalic mineral products experience the lowest losses resulting from power outages. Firms with more than 75 employees report that power outages cause monetary losses equivalent to 0.7% o f their sales, while micro, small and medium firms report losses o f 1.4% on average (Figure 3.7). Similarly, micro and small firms take on average more than 23 days to obtain an electricity connection, while their larger counterparts obtain that service in less than 13 days (Figure 3.8). As for the differences that are apparent across firms from different provinces and sectors, we find that electricity 144 connections take longer in San Salvador, and losses due to outages are smallest among firms that produce chemicals and non-metalic mineralproducts(Figure 3.9 and 3.10). Figure 3.7: Losses Dueto Power Outages by size Figure 3.8: Days to Install a New Electricity Connection, by size 1.6 25 micro small medium large micro small medium large Source:InvestmentClimate Survey. Source: InvestmentClimate Survey 195. Firmsthat report owning or sharing a generator are less likelyto stop production as a result of outages, and have fewer monetary losses as a percentage of sales. About 29% o f the surveyed Salvadoran firms reported owning or sharing a generator. While this percentage is smaller than that found in Honduras (33%) and i s less than a half o f what is encountered in countries such as India and Bangladesh, the use o f private generators is more frequent than in neighboring Guatemala and Nicaragua (Figure 3.11). Not surprisingly, access to private generators increases sharply with firm size (Figure 3.12), and it i s associated with a lower likelihood o f stopping production as well as with lower losses as a result o f power outages (see appendix). Figure 3.9: Daysto Install a New Electricity Figure 3.10: Figure 3.8: LossesDue to Power Connection, by Province Outages, by sector 2.1 23 Source: InvestmentClimateSurvey. Source: InvestmentClimate Survey 145 Figure 3.11: Firms with generator, by country (YO) Figure 3.12: Firms with generator, by Size RRCl'15 Micro Small Medium Large Source: InvestmentClimate Survey. . Source: Investment Climate Survey 196. Maquilas are less likely to stop productionas a result of power outages, and so are smaller firms. Econometric estimates reported inthe appendix show that for a given number of outages and controlling for size, sector, province and the presence o f a generator, maquilas have a lower probability o f stopping production as a result o f electricity interruptions, but they do not behave differently from other firms in terms o f the monetary losses resulting from those interruptions. The results reported in the appendix also confirm that larger firms experience lower losses as a result o f outages, despite the fact o f being more likely to stop production in those events. Moreover, after controlling for the above mentioned firm characteristics, there are almost no significant differences in the consequences o f power outages across firms from different sectors and locations - the main exception beingthat producers o f non-metalic mineral products loose less, and that firms outside San Salvador are less likely to stop production when affected by an outage POLICY RECOMMENDATIONS ON ELECTRICITY 197. The government is currently considering alternative measures directed at promotingnew investments in electricity generation. The government's main concern is the question o f assuring hture supplies. If they do not develop and the Government i s forced to enter into PPA arrangements with new producers, guarantee their investments, or take market risk, this would be a major setback of the reformprocess. The highprofile that the Government has assigned to resolving this problem i s well justified. The studies that the Government i s conductingg6 for solving the problem center around (a) establishing compulsory long term contracts for suppliers and Discos to assure supplies, and/or (b) changing the current wholesale market scheme and implementing a cost-based approach in order to reduce the incidence o f market power, and/or (c) instituting a capacity charge to cover a proportion o f fixed costs and thereby provide an incentive for investing in new generation. These approaches have been analyzed indetail and have beenconsulted with the different stakeholders, and specific proposals for implementationare under consideration. 96The Ministry of the Economy has retained the services o f Mr. Manuel Dussan, a former regulator, to propose approaches and implement solutions to the problem.. 146 compulsory long term contracts - should take into account their effects on the 198. The analysis of measuresto promote investmentsin generation- e.g. the adoption of complementary objective of promotingmore competition in the sector. El Salvador has yet to develop a more dynamic supply function, thereby stimulating greater trade and enabling larger consumers to `escape' their local Disco if favorable conditions can be found elsewhere. Moreover, the wholesale market i s prone to being manipulated through market power and measures to increase competition are requiredto achieve the objectives o f the sector's reform. If compulsory long-term contracts were to be implemented, this would probably force suppliers to seek new generators. Moreover, the suppliers' compliance would be facilitated by the regulations adopted in 2003 for allowing contract prices to be incorporated into tariffs - i.e. allowing a pass- through o f generation costs. However, the Discos, who currently dominate the supply market, are likely to object to any measures that jeopardize their market, and in particular they would oppose to measures that would seek to stimulate competition in supply. Once committed to long-term purchases, a Disco would seek to guarantee its market by holding on to its customers thereby mitigating the commercial risk o f its contracts. Compulsory long-term contracts could therefore run contrary to focusing Discos on their `wires business' and opening up the supply market. Moreover, even though long term contracts would clearly create an opportunity for new investments in generation, the corresponding decisions would also depend on other factors, such as securing fuel supplies and overall guarantees for private sector investment.97 199. The merits of instituting a capacity charge are also controversial, while the alternative of switching to a cost-basedwholesale market does not guaranteethat dominant producers will not continue to exert undue influence. Instituting a capacity charge would provide an incentive for new generation by providing a degree o f income assurance to its investor, who would no longer be entirely dependent on the spot market. Such a charge i s likely to replicate the existing cold reserve remunerationg8(interms o f $/kW), which i s translated to an equivalent energy charge (i.e. $/kWh), and added to transactions in the wholesale market, thereby levyingthe charge from all consumers as it is passed-through from the wholesale market to retail sales. However, in systems where a general capacity charge has been implemented, its concept has been the subject o f substantial criticism, and in some cases it hasn't had the desired effect o f attracting new investment. As for the alternative o f switching to a cost-based wholesale market, it is worth emphasizing that in that scenario dominant producers could continue to exert undue influence through alternative means such as spurious availability declarations. Moreover, while this alternative would have the merit of aligning the Salvadoran structure with that o f the other Central American countries participating inthe regional market, it is worth noting that the exceptional character o f the Salvadoran market has not been an obstacle to trades with countries within the region inthe framework o fthe MER. 200. When deciding on major modifications to the sector's structure, the government should take into consideration a number of strategic questions related to the regional developments of the Central American market. In particular, the question o f urgency regarding the need for additional investment in generation should be tempered by the real 97It must also be noted that El Salvador does not appear to present particular comparative advantages that would justify locatinggenerationfacilities inthe country with a view to exporttowards other markets. 98As a consequence of Duke Energy's intentionto dismantle gas turbines that were rarely dispatchedand weren't earningtheir way, the Governmentimplementeda `coldreserve remuneration' for units on extended standby 147 possibility o f securing supplies from neighboring systems. If the Salvadoran power system can depend on imports in the near term, the question lies in orienting institutional reforms towards adapting to the ultimate market structure that could emerge in a regional context. Thus, even though establishing compulsory contracts or a capacity charge may be ways to ensure future supplies, they may become quickly outdated ifthe regional market expands rapidly. 201. From a strategic point of view, it is important to reinforce transmission in the context of El Salvador's SIEPAC commitments, and to make an effort to avoid short-lived modificationsto the sector's structure. Inthe event that the reforms that are being considered by the Government do not yield the desired results, an alternative plan would have to be formulated. In this context, insurance to maximize the eventual support available from the regional market can be obtained by providing ETESALwith the necessary resources (through its regulated income) for reinforcing transmission in the context o f El Salvador's SIEPAC commitments. Moreover, the reforms under consideration, together with those already adopted in 2003, could give an impression o f successive `patching' o f the underlying model, thereby projecting an image o f regulatory instability. Thus, consideration should be given to adopting reforms that are likely to be relatively long-lived. 202. The government should also take specific measures directed at increasing ` competitionin the wholesale market. Facilitating the operation o f suppliers within the market i s clearly needed to mitigate market power and to encourage fbrther trade with generators inthe regional market. Complaints by suppliers regarding bureaucratic obstacles put up by Discos to discourage what they perceive as `losing customers' should be addressed by streamlining procedures for accessing the wholesale market, as well as making sure that Discos are getting full compensation for the use o ftheir facilities. TELECOMMUNICATIONS 203. The 1998 reforms of El Salvador's telecommunications sector can be credited for an impressive growthin phone density. The new Telecoms Law and its regulations were designed to promote private sector investment and competition. The incumbent state company (ANTEL) was sold to a consortium o f French and local investors and the sector was opened to competition. Spectrum auctions were carried out to attract new investors to the mobile telephone market; and a new regulator was created (SIGET). The reforms produced significant improvements in efficiency, competition, and private investment in the telecommunication sector. As a result o f the reforms the telecom sector has exploded. As shown in Figure 3.13, fixed lines grew from 386,700 units in 1998, to 752,600 in 2003, while mobile telephones grew from 137,100 units in 1998, to 1,149,800 in2003. 148 Figure 3.13: Mobile and Fixed TelephoneLinesin ElSalvador, 1998-2003 1,149,790 -4 137,l I998 I999 2000 2001 2002 2003 l+Fixed +Mobile 1 Source Investment Climate Survey 204. In 2003, the density of fixed and mobile phones was larger in El Salvador than in most Central American countries. According to the World Economic Forumg9, El Salvador was ranked fifth among all Latin American countries in growth o f main lines in operation from 1999 to 2002. Not surprisingly, the standing o f El Salvador's telecom sector with respect to other countries o f the region has improved significantly, especially when the comparisons are made with neighboring Central American countries. Thus, in the case of fixed lines and within the Central American region, El Salvador's density trails only that o f Costa Rica (Figure 3.14). Moreover, in the case of mobile phones, El Salvador i s behind Panama, but has a much higher density than Honduras, Guatemala, Nicaragua and Costa Rica (Figure 3.15). Nevertheless, there are still important challenges ahead, as the number o f phones per capita encountered in El Salvador i s still well below that o f countries such as Chile, Brazil, and Mexico. Figure 3.14. Fixed PhoneLines per 100 pop. in Latin American countries (2003) 99World Economic Forum, The Global Information Technology Report 2003 -2004,2004, New York, 149 Figure3.15. Mobile Phone Linesper 100pop.in LatinAmerican countries(2003). 51.1 205. El Salvador has also achieved a density of fixed lines well above that of countries with similar levels of income. When making international comparisons o f telecommunication densities it i s useful to take into account the large differences in per capita GDP that exist even across countries o f the same region, as well as the strong relationship existing between phone density and income per capita. Indeed, as reported in Figure 3.16, the latter variable explains between 87% and 97% o f the variation inthe number o f phone lines per capita in a large sample of countries. As seen in this figure, while in 1998 El Salvador had about the same number o f fixed lines per capita than its level of income would lead to expect, by 2003 the country exhibited a phone density that i s substantially above that o f other countries with the same level o f income. This figure also shows that, with the exception o f Guatemala, most o f the differences in phone density between El Salvador and its neighbors can be explained by the countries' levels o f per capita income, and the same i s true for comparisons with South American countries such as Argentina, Brazil and Chile. Figure3.16. FixedPhoneLinesper 100 pop.And GDPper capita, 1998 and 2003 MainLinesand GDP per capitain 1998 Main Linesand GDP per capita in 2003 6.007 5.00 -, 5.00 1 4.00 4 .9Argentina ."0 3.00 ; I . if:, 2.00 2 l o o ] .- f3 M L 000 7 800 goo 10'00 1100 4 bo II.oo -I00 -1.00 J ~ y = 0 8794x 4 6236 - y = 0 9 6 7 4 ~- 5 4175 Ri= 0 762 R'=08043 -2 00 1 -2.00 J Log GDP per capita Log GDP per capita Source: ITUdatabase 150 206. Competition has increased substantially, leading to drastic reductions in long distance tariffs. The structure o f El Salvador's telecom market has changed substantially inthe last five years. While in 1998 there was only one company providing fixed lines and one company providing cellular lines, in 2003 the sector had expanded to include nine companies in the fixed line business, 4 mobile companies, 11carriers, and 10Internet providers. Thanks to the increased competition, prices for telecom services have been drastically reduced. Thus, in the case o f international calls to the United States, the rates went from $2.00 per minute in 1997 to $0.22 in 2004. As shown in figure 4, during the last few years, El Salvador has become one o f the countries with the lowest long distance tariffs, both for both national and international calls, above only Costa Rica and Panama, respectively. Figure3.17. InternationalLongDistanceTariffs in CentralAmerica, 2003 (US$per min. to the U.S.) "Pd 0.87 El Salvador G y ~ l m l a Costa Rica Honduras Nicaragua Source: Regulatory Agencies 207. New lines are also less expensive and take less time to be installed than in neighboringcountries. The cost of fixed line installations has fell substantially, from US$335 in 1999 to US$61 in 2003.'00 As seen intable 2, in2003 the cost o f a commercial installation was substantially lower than in Guatemala, Honduras and Nicaragua, and also slightly lower than in Costa Rica. Moreover, data from the investment climate surveys shows that it takes on average one week to get a new phone in El Salvador, compared to 50 days inGuatemala, four months in Nicaragua and almost six months inHonduras. Figure3.18. NationalLong DistanceTariffs inCentralAmerica, 2003 (US$ per min.) 0.10 Costa Rics El Salvador Guateinala Nicaragua HOlldursJ Pa" Source: Investment Climate Survey 208. Local calls are priced competitively and are somewhat higher than in neighboring countries. Monthly rental rates for fixed telephone services have shown some slight increases: loo This is an average for residential installation fees, across all telecom firms inEl Salvador. 151 from US$6.84 in 1999 to US$8.27 in 2003 for residential lines (100 minutes per month), and from US$13.69 in 1999 to U S $14.06 in 2003 for commercial ones. As seen in Table 3.6, both residential and commercial monthly rates are higher than in other Central American countries - the only exception being Nicaragua in the case o f commercial lines - and so i s the cost o f local telephone calls. Thus, for instance, a typical company that i s billed for 1000 minutes o f local calls per month pays $35.28 per month in El Salvador, compared to $23.72 in Honduras and $12.50 in Nicaragua. I t must be noted, however, than in the latter countries the local phone calls are "cross-subsidized by long distance calls which, as seen above, are relatively less expensive in El Salvador. Residential Commercial Fixed line Monthly Charge Total Fixed line Monthly Charge Total Cost installation rate ldmin.), Cost 1600 /. installation rate 1u/min.) , ,fl000 min) (200 min.) min)(a) (200 min.) (a) Guatemala 78.91 5.65 ,026 16.71 II 189.4 5.65 0.026 28.03 El Salvador 61.54 8.27* ,023 22.58 61.54 14.07* 0.023 35.28 Honduras 25.82 2.29 .020 10.51 237.68 5.74 0.02 23.72 Nicaragua 156.24 5.94 ,024 16.84 237.68 15.85*** 0.024 41.83 Costa Rica 63.33 4.09** ,008 8.22 88.17 4.97** 0.008 12.50 209. Service reliability i s also higher in El Salvador than in the rest of Central America. According to data from the Investment Climate Survey, only six percent o f the manufacturing firms surveyed in El Salvador considered the telecommunications infrastructure to be a major or severe constraint to their operations and growth, compared to 7% in Guatemala, 12% in Nicaragua and 18% in Honduras. These smaller concerns regarding the telecom sector are justified not only by the easier access to new lines and the lower cost o f long distance calls, but also by the fact that service interruptions are much less frequent in El Salvador. Indeed, the average firm experienced less than one interruption per year, and was without service less than six hours per year. As seen in Table 3.7, firms in other countries o f the region report a much higher number o f service interruptions and are left without,phone services for much longer periods o f time. Table 3.7. Incidence, length and impact of telephone cuts (2002) %who had avg times total hours %sales lost phone cut phone cut /a phone cut /a phone cut /a El Salvador 16.0 0.8 5.8 0.03 Guatemala 19.6 2.6 18.0 0.2 Honduras 24.8 6.5 56.0 0.3 Nicaragua 26.2 7.0 79.5 0.7 Peru 16.0 0.6 50.0 n.a. Brazil 27.1 2.1 27.0 0.3 Ecuador 36.6 12.4 203.0 1.1 al all firms, regardless of whether or not subjectedto cuts Source: Investment Climate Survey 210. Despite the impressive expansion of El Salvador's telecommunication infrastructure, there is evidence that the country is not exploiting all the available 152 opportunitiesfor increasing the use of information and communicationtechnologies (ICT). The number o f internet accounts has not experiencedthe highrates o f growth that are found for fixed and mobile telephones. There are only 2.5 personal computers per 100 inhabitants, compared to almost ten in Costa Rica (Figure 3.19). Moreover, in 2003 there were only 6.3 internet hosts per 10,000 inhabitants, compared to more than twice that number in Nicaragua and Guatemala, and about four times that number in Panama and Costa Rica. It must be noted, however, that the density o f Internet hosts found in El Salvador and Costa Rica i s very close to what would be expected from comparisons with countries o f the same level o f income (Figure 3.20). Guatemala and Nicaragua, on the other hand, exhibit numbers o f internet hosts than what their incomes would lead to expect, and so do South American countries such as Argentina, Brazil and Chile. Figure 3.19. Internet Hostsper 10,000 pop. and personalcomputers per 100 pop. (2003) 25 94 i BIntemet Hosts mPersonalComputers i Source: ITU database 211. Despite the country's relatively low density of internet hosts, Salvadoran manufacturing firms use email and the internet with the same frequency of their Guatemalan counterparts and much more frequently than firms from Honduras and Nicaragua. About 63% o f the firms surveyed in El Salvador report regularly using email to communicate with suppliers or clients, and 30% report using the internet for the same purpose (Figure3.21). The use o f electronic mail and the internet is similar inGuatemala butmuch lower inHonduras and Nicaragua. These differences are found to persist when comparisons are made with firms o fthe same size and sector. 153 Figure3.20. Internet Hostsper 1000 pop. and GDPp/c in 2003 8 00 600 6 { .- 400 -Q 2.00 0.00 4 -2004 900 1000 11.00 -400 -$ - -600 y = 1.6444~ 12 70s R2= 0 6845 -8 00 GDP per capita Source : ITUdatabase 212. Within El Salvador,larger firms are more likely to use email and the internet, and so are exporters and the firms that use imported inputs. As seen in Figure 3.22, the percentage o f firms that report using email and the internet increases sharply with employment size. Moreover, econometric estimates reported in the appendix indicate that even after controlling for firm size, sector and province, exporters and firms that use imported inputs are much more likely to use those technologies. Maquilas, however, do not behave differently from other exporters or importers and firms from Santa Ana and other small Salvadoran provinces are less likely to use email or the internet. Finally, we find that the usage o f ICTs i s higher in the sectors o f metal products and textiles, even after controlling for sized, location and export and import status. Figure3.21: Internetand email use, by country (%) Figure3.22: Internet and email use, by size(%) 77.3 94.6 213. One important obstacle to a larger use of ICTs in El Salvador is the high cost of internetaccess. Eventhough Internet providers advertise "Free Access" to the Internet, all dial- up users pay regular local phone tariffs to the telephone companies - about 2 cents a minute-to access the Internet. Thus, for instance, a company that wishes to use the Internet during weekdays for one hour per day, would typically pay more than US$40 per month, and have only a low speed (56 Kbh) dial-up connection. New broadband access i s available through cable modem and DSL, but prices for those services are even higher (around US$50 per month). The 154 result is that only large companies, and companies that use the internet as a substitute for international calls - e.g. to communicate with foreign suppliers or clients - have the resources to pay for broadband access. As seen infigure 3.23, El Salvador shows high Internet dial-up prices compared to other Latin American countries. Figure3.23 InternetDial-up Prices"' per month, LatinAmerica (2004) 80.0 7 71.6 50.0- Source: Tariff data with calculations by the World Bank. 214. Another impedimentto a greater use of the Internet is the scarcity of local content. There are few Salvadoran web sites and there i s little useful local content in the web for Salvadoran users. The major Internet Service Providers (ISPs) are the telephone companies as no one can compete with their "Free Internet" plans. In many countries ISPs have significantly contributed to local content development, motivated by the desire to increase their revenues through the supply of higher value added services. However, in the case o f El Salvador neither the Government nor the telephone companies have promoted local content development. Some o f the few initiatives in that direction have been promoted by the private sector, Thus, for instance, the "@rroba de oro" initiative has promoted local content development, through a prize awarded annually to the best web site of the year. Another good initiative has been carried out through Infocentros. The latter have trained the staff o f 4,000 small companies, so that they can develop their own web pages, which can then be hosted in their system. These and other initiatives directed at developing local internet content are likely to lead to a greater exploitation of the opportunities provided by El Salvador's improved telecommunications infrastructure, Indeed, countries that have promoted local content development in the initial phase o f Internet development have successhlly increased the usefulness o f the net to businesses and residential users, as the examples of Singapore, Ireland, Israel and Chile attest. The increased perception o f usefulness has in turn led to further increases in the number o f internet users and a greater insertion o f those countries in the Information Society. The point here i s that El Salvador could also exploit the above described virtuous circle if it were to further promote the development o f local internet content. 215. The young and specially the adult population have a low level of ICT exposure, High Schools have recently acquired computer centers to train students. However, due to administrative reasons, these centers are not used to their full capacity and students do not Intemet Dial -Up prices = connection fee + month rental fee + (usage per minute* 30 hours permonth) 155 receive enough hours o f instruction in the use o f personal computers and the Internet. Thus, those technologies are not being used to their full potential, in terms o f expanding the possibilities o f study and research, and improving the students' analytical skills, particularly in the fields o f mathematics and science. As for the adult population, there are only a few adult training centers - mainly private - that teach computer skills. At the tertiary level, there are few programs to train technicians and programmers at the Associated Degree level (two year technical or vocational education), as opposed to a larger number o f University Degrees programs. As a result, El Salvador does not have enough technicians and programmers to meet demand, but has a surplus o f University graduates. BOX 3: E-governmentin Chile The Chilean Government has been a strong promoter o f ICTs through the development o f a number o f E- govemment initiatives. Among the latter, some o f the most important are the ones developed in the areas o f government procurement, tax retums, customs, and the on-line supply o f govemment services directed at SMEs. In the area o f procurement, Chile developed a system o f information on public purchases and contracting which operates with more than 100 public services (see www.comuraschile.cl). Companies that wish to do businesswith the public sector need only to register a single time and record the areas in which they offer goods or services (e.g., office fumiture, construction services, IT consulting, etc.). Whenever a public agency needs to purchase goods or contract a service, it fills out a request in the electronic system, specifying the kind o f operation and including all the documentation and information associated with the request. Automatically, the system sends an e-mail to all the private companies registered in that selected area, minimizing response time and providing an equal opportunity for all firms. The tax collection agency (Servicio de Impuestos Intemos) has also been leading the development o f public sector Intemet-based services solutions. In2000, it distributed free software through its www.sii.cl web site to facilitate preparation o f annual income tax retums for 1999. As a result, 26 percent o f all returns were delivered via the Internet and nearly 50 percent o f taxpayers made their income tax payments to the. national Treasury via the Internet. The customs services also operates wholly in electronic format and the Government has also been active in developing electronic services o f more direct interest to small businesses. It has launched a one-stop window service via its portal www.sitioemuresa.cl, which has information for businesses on more than 1,200 items and is available at 130 state offices. I t enables single- window access to 59 government procedures. Source: World Bank, Chile New Economy Study, 2004 216. Another factor that limits the use of ICTs in El Salvador is the fact that, despite some successful initiatives in some specific agencies, the government does not have a strategy for E-Government nor an Agency to implement it. The modernization o f the government's operations through the use o f ICTs has the potential to enhance service delivery, increase transparency, reduce costs and improve the relations between government and citizens. Moreover, international experiences show that the Government's leadership, by means o f a well coordinated and articulated E-government strategy, i s crucial for promoting the use o f ICTs by the private sector, thus promoting productivity growth both directly and indirectly. In El Salvador, some successful E-Government initiatives have been implemented and must be mentioned. The Customs Administration, for instance, implemented a modernization plan that now allows users to prepare their Customs Declarations and make the corresponding payments on line. Similarly, in 2003 the Ministry o f Finance started offering the option o f on-line Tax Preparation Forms, and for the first time many tax-payers filed their tax forms on line. These, 156 however, have been isolated efforts. There i s no E-Government strategy comparable, for instance, to the one implementedby Chile (see box 3). Moreover, The ICT Sector does not have a Head: there i s no Ministry, or high level Government Office in charge o f preparing, coordinating and executing a Strategy for E-Development, overseeing the efforts o f all the Ministriesand public entities, incoordination with the private sector, and the civil society. As a result, the initiatives directed at the modernization o f government operations have not been coordinated, and have thus had a smaller impact than what could have been expected in the presence o f a coordinated strategy. 217. El Salvador could also increase the use of ICTs by updating the regulatory framework that governs the telecommunication sector. The existing Telecom Law was drafted in 1997 when the World Wide Web was only 3 years old! Thus, El Salvador does not have an e-commerce legislation that regulates, for instance, the use o f digital signatures to enable online transactions. There i s also a need to update some o f the Telecom Regulations that were enacted in 1998 and that have now become obsolete. A draft E-commerce legislation was prepared in2003. However, the project was not successfully endorsed by all the sectors involved and as a result, the government did not introduce the bill to Congress. The obsolescence o f the telecom legislation is a major bottleneck to develop the sector. The lack o f an E-commerce legislation, for instance, i s one o f the main factors underlying the fact that few businesses are using the Internet to sell their products, receive payments, or place orders to their suppliers. Users in general do not perceive a large benefit from using the Internet to do business, buy products, make reservations, consult information, or interact with Government and other users. 218. Besides the needed changes in the Telecom legislation, S I G E T needs to update its own regulations and procedures, many of which are also obsolete. SIGET has not been able to tackle interconnection issues and will have limitedcapacity to deal with anti-monopoly issues inthe future. There is a need for addressing interconnection issues, and inparticular for finding ways o f reducing the corresponding charges, which limit the growth o f the sector. The current regulation does not allow SIGET to fix interconnection rates from fixed to mobile phones. There is, however, a market failure motivation for intervening in this area, illustrated by the current high interconnection rate o f 26.5 cents per minute. In addition, SIGET has to streamline the processing o f consumer complaints and increase public awareness o f consumer rights. It has a backlog o f license requests and there are many illegal transmitters in the country. SIGET does not have a radio frequency spectrum management and control system, and relies on an obsolete and manual procedure to track use o f frequencies. POLICY RECOMMENDATIONS ONTELECOMMUNICATIONS 219. S I G E T should establish a reduced rate for Internet Access. Apart from the current local charge for fixed phone calls, SIGET should establish a special rate for internet access that would allow consumers to use' internet services for much lower rates. This would stimulate purchases o f personal computers and the set up o f more Internet accounts, as well as the development o f more local content. Moreover, as a larger number o f people and businesses begin to use e-mail and visit web pages, the usefulness o f the service would increase, leading users to spend more time online, and stimulating further migration into broadband-based services. Even 157 though competition in the ISP market can be expected to increase, telephone companies would benefit from the increasednumber o f users and traffic. 220. The government should take complementary actions to stimulatethe development of local internetcontent.The above describedvirtuous circle that i s expected to follow a reduction inthe rates charged for accessing the internet would by itself stimulate the development of local internet content. However, the latter could be further promoted through government initiatives that reward local efforts in the area o f web site development, either through contests - in the lines o f the "arroba de oro" initiative - or by the promotion o f cooperation programs in which Universities and Technical Institutes would develop the ICT systems and content o f government agencies and private companies. In addition, the government could take advantage o f the opportunities provided by ICTs to promote more interactions with Salvadoran immigrants living abroad. For instance, the development o f local content could be stimulated by the possibility o f using the internet to make the goods and services offered by Salvadoran businesses available to the large population o f Salvadorans living inthe U.S.. 221. The government should create a single, high level entity in the Government with enough authority to develop and implement an ICT Strategy for the country. Specifically, this entity would supervise the implementation of an E-Government Strategy encompassing all ministries and government agencies. Eventually, E-Procurement should be made mandatory for all agencies, and most transactions involving businesses or the general public should be made available on-line. The objective would be that o f promoting greater transparency and increasing productivity and efficiency inboth the public and the private sector. 222. The Telecom legislation and regulations should be updated and SIGET should be strengthened. The government should prepare, discuss with all sectors and submit to congress legislation for E-Commerce. In addition, some aspects o f the 1998 telecom legislation and the corresponding regulatory framework should be updated. A modem, flexible but adaptable telecommunications regulatory framework i s needed to protect consumers but also allow further private investment in ICTs. A revision o f the interconnection regulatory framework i s needed to promote competition and fair prices to consumers. In this respect, the role o f SIGET should be redefined, so that it has the ability to take the necessary measures to ensure competition and avoid unfair monopolistic practices. Moreover, in order to better contribute to this reform process, SIGET should be strengthened by means o f a Technical Assistance program directed at reviewing the existing regulations, improving the procedures for processing consumer complaints, raising public awareness o f consumer rights, and purchasing o f a modem radio spectrum management and control system to improve licensing, reduce illegal transmitters and resolve interferences. 223. Finally, the Government should develop an adult ICT training initiative and improve the existing training programs for youth. More and better human resources are key to improve the productivity o f Salvadoran companies by means o f exploiting the opportunities provided by ICTs. In this respect, the existing educational and training programs should be reviewed to ensure that they respond to the needs o f the private sector. For instance, the Government should consider increasing the number o f technical training programs at the Associate Degree level (technician, programmer). 158 TRANSPORT 224. The lack of an Atlantic coastline increases the importance of developing and maintaining El Salvador's road infrastructure and increasing the efficiency of border- crossings in order to facilitate the country's internationaltrade. Even though in 2003 only 15% o f El Salvador's imports originated in Central America, and just 24% o f the total value o f exports was destined to the Central American region, a much higher fraction o f the country's international trade takes place through its land borders with Guatemala and Honduras.lo* Indeed, according to a SIECA study, in 1999 the ports that were primarily used for the movement o f Salvadoran goods were located in the Atlantic coast: Puerto Barrios and Puerto Santo Tomas de Castilla in Guatemala, and Puerto Cortes in H~nduras."~In the Pacific coast, the main Salvadoran port o f Acajutla also suffered the competition o f Puerto Quetzal, in Guatemala, which inthe late nineties appears to have attracted some o f Acajutla's business. As a result o f the intensive use o f foreign ports, El Salvador's border crossings are among the ones with more intense traffic inthe region (Figure 3.24). Figure3.24: Vehicle Traffic at CentralAmerican border crossings, 1999 Vehicle Trafflc at Border-Cro.rlng.~"erage Dally Trafflc (lSSS] Source ECAT(ZOO7, c)400-500ADT 300400 ADT -200-300 ADT The ECATstud rdenttfied 19 bordsrsmssi concentrated 14% (EST 1999)of the volume of!ntra-regronal bade 225. Only a minority of the surveyed Salvadoran manufacturingfirms reported using Acajutla as the mainport throughwhich they importtheir inputs or export their products. The IC survey included information on the main ports used by exporters and importers. As shown in Figure 3.25, only 37% o f the non-maquila firms with direct imports indicated that Acajutla was the main port used for their imports: 31% reported using Guatemalan ports, 2% mentioned Puerto Cortez, and the rest indicated that they used only land (21%) or air transport (9%). As for maquilas, only 27% reported using Acajutla as their main port for imports, while 50% mentioned Guatemalan ports, 17% responded that they use air transport only and 7% mentioned Puerto Cortez. In the case o f exporters the survey collected information on the main port chosen by the firms that use maritime transport: 48% o f non-maquilas and 20% o f maquilas reported that their main port for exports was Acajutla, with the rest using mainly Guatemalan ports (table 3.26). lo* According to the Central Bank o f El Salvador, the United States were responsible for 67.6% of total Salvadoran exports in2003, and 49.8% of imports.See http://www.bcr.gob,sv/estadisticas. I O 3EstudioCentroamericanode Transportes, SIECA 2001. 159 Figure3.25: Mainports usedby importers(Investment ClimateSurvey data) Non-Msquilaa Msuvilss SIo.T.ds Caolilla 18% 7 Amjutla S1o.T.deCastilla Acajulla 4 . - Land only/ 9% Quelzal 7% 3% 21% 7% Source: InvestmentClimate Survey Figure3.26: Main ports usedby exporters(Investment Climate Survey data) N u n - M s q u i l s i Sto TdeCartilla 31% metra S1o.T deCasUila 8% 63% pto Cortez/ pt=;rfos 5% Source: Investment Climate Survey 226. The relativelylow use of Acajutla by Salvadoranmanufacturing firms is consistent with the high maritime transport costs when shipping from that port. According to data from the U.S. Department o f Transportation, the share o f maritime transport costs in the total value o f garments exported in 2000 to the U.S. from the port o f Acajutla was 8.4% for the east coast, and 25% for the west coast (Figure 3.27). Shipping garments from Acajutla to the U.S. i s more expensive than doing it from all the other main ports o f Central America, and inparticular from Puerto Quetzal (Guatemala), which due to its location i s the most direct competitor o f Acajutla. Moreover, even Latin American and Asian ports located up to four times farther away from U.S. ports than Acajutla exhibit smaller transport costs, at least in the case o f garments. When the comparisons are made at a more disaggregated level (5 digit SITC codes), focusing only on the main garments products shipped from Acajutla to the U.S.,the results indicate that maritime transport costs are close to 5% o f the value o f the main products exported to the east coast - Jackets, Blazers and Trousers - and to 10% o f those exported to the west coast - Baby Garments, Pantyhoses and Tights.Io4As shown in Table 3.9, for the products shipped to the east coast Acajutla still exhibits significant disadvantages with respect to Central and South American ports, as well as with respect to ports from other continents. However, at least in the case o f the main product shipped to the west coast, Acajutla appears to have a cost advantage over the Guatemalan port o f Santo Tomas de Castilla. 227. The high maritimetransport costs when shipping from Acajutla are associatedwith the fact that this port is used mainly for imports and charges relatively highfees, which in turn is the result of low efficiency levels and a small scale of operations (at least in the The corresponding SITC-5 digits codes and complete descriptions for these products are: Jackets and Blazers (84230); Trousers, bib and brace overalls, breeches and shorts (84260); Babies' garments and clothing accessories, knitted or crocheted(84512); Pantyhosesand tights (84621). 160 movement of containers). Acajutla was built in 1952, and is owned and operated by the public sector Autonomous Executive Port Authority (Comision Ejecutiva Portuaria Autonoma - CEPA). According to CEPA, 4,546 metric tons were moved through Acajutla in2002, with 79% o f the movement corresponding to imports, and 5.5% to container^.'^^ As explained by FUSADES (2003), the fact that most of the port's movement consists o f imports increases maritime transport costs, as shipping companies charge one-way users for the total cost of the round-trip - they charge also for the return trip o f relatively empty ships. In addition, the low level o f containerization leads to diseconomies o f scale that are reflected in relatively high port fees for the movement o f containers. As seen in Table 3.8, the movement o f containers through Salvadoran ports (mainly Acajutla) grew 57% in 2003, but still represented a very small fraction o f total container movement through Central American ports (3.4% excluding Panama). Thus, compared to Acajutla, container movement was more than ten times larger in Costa Rican and Guatemalan ports (2.6 times larger in Quetzal only) and about seven times larger in Honduran ports. The low scale of container movement of Acajutla, coupled with inefficiencies at the port level, have translated into higher waiting times and higher port fees, which are usually internalized by shipping companies into the fares charged to importers and exporters.lo6 It i s worth noting, however, that the impact o f some of these problems has been minimized in recent years, as illustrated by the 26% growth in cargo movement through Acajutla reported by CEPA between 2002 and 2004 (compared to 2% growth between 2000 and 2002). Table3.8: Container Movement in CentralAmerican Ports, 2002-2003 Sourc:e: C:entral American Isthmus Port StatisticalSummary COCATRAM, 2004 Figure3.27: Maritime transport costs to the U.S.from selectedports as % of export value Maritime TmniporlCortr for Garments Exports lo the USkEsrt Cowl Maribme Transport Costl for Gsmentl Erpartr te the USA, West Coast 1 (Transpon Cost BS% ofExpon Value) i (Tmpon Cost s % ofExpotiValue) 2 1 1 2w. IS% IPA 5% 'r,. I Source: U.S. Department o f Transportation. IO5Acajutla has specialized historically in dry bulk, which was 57% o f total movement in 2001. In that same year, liquidbulk was 24% and general cargo 19% o ftotal cargo. FUSADES (2003), p. 207. See Clark, Dollar and Micco (2004) on factors that affect maritime transport costs. According to FUSADES, Acajutla is competitive in dry and liquid bulk but port fees for Evergreen containers are higher than in Puerto Bamos (116%), Corinto (95%), Santo Tomas de Castilla (44%), Cortes (31%) and Quetzal (17%). 161 228. Data collected by the ICs confirms that Salvadoran firms face higher costs for shipping containers to the east coast of the U.S. but seem to have a relative cost advantage for reaching the west coast. In the context of the investment climate surveys performed in Central America, firms with direct exports to the U.S.were asked about the cost of shipping 20 feet containers to different North American ports. Figures 3.28 and 3.29 report the results for the ports o f Miami and San Francisco. Salvadoran exporters report the highest costs for shipping containers to Miami, with average differences ranging from 4% with respect to the price paid by Guatemalan firms, to 16% with respect to the cost faced by their Honduran counterparts. Not surprisingly, given their easy access to ports in the Pacific coast, Salvadoran and Nicaraguan firms appear to pay less than their Honduran and Guatemalan counterparts to reach the west coast port of San Francisco. Table 3.9: MaritimeTransportCosts to the U.S.,for selectedgarments (YOof exportvalue) E x p o r r i t o 1 h c U S E i i t C o s s 1 E x p o r t s t o t h e U S W e s tC 0 . ~ 1 P o r t " a m c T r s n s p o r t c o s t P o r t n s m e T r a n s p o r t C O P ! . J a c k e t s a n d B l a z e r s ( S I T C = 8 1 2 1 0 1 B a b v ' s g a r m e n t s ( S I T C = 8 4 5 1 2 ) C o l o m b o 8 . 7 % S a n t o T o m a r d e C a s t i l l a 2 5 . 2 % B o m b a y 8 I % A c a j u t l i 1 0 . o 'h A e a j u t l l 5 , 6 % C o l o m b o 9 . 4 % C h i t t a g a n g 5 4 % B o m b a y 7 . 6 % S h a n g h a i 5 0 % C h i i l a g o n g 6 . 5 % I s t a n b u I 2 % B s n g k a k 5 6 % B a n g k o k 2 . 99% S h a n g h a i 1 . 5 % S a n t o T o m a r d e C n i t i l l a 2 . 8 % c a l l . . 4 1 % P u e r t a c O l t e S 2 4 % . C a s a b l a n c a 2 . o % I s t a n b u I 2 , 4 % T h o u a c r n a n d s h o i l s ( S i T C = 8 4 2 6 0 1 P a n t v h o s e a n d t i n h t s ( S I T C = 8 4 6 2 1 \ C h 1 1 1 s g o n g 5 .I% A c a j u t l a I0 . 0 % C o l o m b o 5 .I % S h a n g h s i 6 8 % B o m b a y 5 0 % A r i j u t l e 4 , 9 % S h a n g h a i 4 . 5 % C a r s b l a n s s 3 , 9 % S a n l o T o m a r d e C s r l l l l a 1 . 6 % P u e r t o c a r t e r 1 . 5 % B a n g k o k 2 . 7 % s a 0 P a u l 0 2 . 7 % 1 1 1 8 n b u I 2 . 6 % c a l l s o 2 . 5 % S o u r c e U S D e p o r l m e n 1 o f T r a n s p o r l o ~ i o n y, e a r2 0 0 0 . Figure3.28: Shipping cost, Miami(20 f. cont.) Figure3.29: Shipping cost, San Francisco (20 f. cont.) 3188 2071 2162 ;* 7 Honduras Nicaragua Guatemla El Salvador Honduras Nicaragua Guatemala El Salvador Source:Investment ClimateSurvey. Source: Investment ClimateSurvey 229. The Government of El Salvador has clearly recognized the need for improving the competitiveness of its ports in order to enhance the country's trading position. As a result, the government has embarked on a process o f reform and modernization o f the port sector in general. It passed a new General Maritime and Ports LawIo7, which reformed the governance lo' Legislative Decree No. 994, September 19, 2002 162 structure o f the sector, including the creation o f a new regulatory body - the Maritime Port Authority (AMP) - and established the framework for private sector participation in the sector. In the past, four years, CEPA has invested US$24m in upgrading Acajutla, and it expects the port to be showing profits in 2004 (US$9m) for the second consecutive year."* In early 2004, CEPA announced investment plans o f $8 million for the port o f Acajutla, which in 2003 achieved a 25% increase in cargo and traffic volumes, and a reduction in cargo unloading times. Moreover, since the Legislative Assembly approved the Basic Condition for a master concession o f the port'Og CEPA has beenworking to implementthe concession o f the port. Unfortunately in April 2004, the 25-year concession that had been awarded to a Chilean-Salvadoran-Peruvian consortium was annulled on the grounds that the winning bid failed to fulfill all the required legal and technical conditions. A new concession process i s currently being designed and a new bidding round has been announced for early 2005. During 2004, the government also implementeda biddinground for the concession of the Cutuco port, located inthe eastem part of the country and inactive since 1995. Even though CEPA announced its choice of a Japanese- Belgium consortium that made a $150 million bid, the government has revealed that it can only invest$125 million, and is currently considering alternatives for securing extra financing. TOTAL ROAD DENSITY PAVEDROADDENSITY 230. Compared to other Central American countries, El Salvador's road density is high with respect to territory but relatively low with respect to the country's population. El Salvador i s second only to Costa Rica interms o f its number o f km o f roads (paved or unpaved) per squared kilometer o f territory (Figure 3.30). However, due to its relatively small territory, the population density o f El Salvador's road network i s among the lowest o f the region, below that o f Honduras and Nicaragua. This i s a factor that may lead to extensive use o f the roads pavement and can thus be affect the quality o f the network, the vehicles' operating costs (gasoline and lubricant consumption, vehicle and tire wear etc.) and the travel times o f road users. These factors, in turn, can directly affect the tariffs and the quality o f the country's freight transport industry. 231. Thanks to accelerated reconstruction efforts, El Salvador's paved road network increased from 1984 km in 1999 to 3017 km in 2004."0 The Salvadoran road network lo*El Salvador-CountryReport, EIU, 2004-05 logLegislativeDecree No. 1014, October3, 2002. Ministry of Public Works 163 experienced particular destruction during the years o f the civil war and as a result o f the 2001 earthquake. During the 1990s reconstruction efforts were concentrated on the primary road network and in the cities. By regional standards, the pace o f reconstruction was slow during the mid 1990s but accelerated after 1999. Indeed, beginninginthat year the Government embarked on a substantial road investment program in which 40 percent o f the nation's principal road network was rehabilitated and repaired. This included a roughly 40 percent increase in the country's paved primary road network, as part o f the Central Government's Cuminos Rurules Sostenibles (Sustainable Rural Roads) program, the repair o f earthquake damaged roads, and partnerships between Fondo de Inversion Social para el Desarrollo Local (FISDL) and local governments to increase rural road access111.Furthermore, a Road Maintenance Fund(FOVIAL) was created in 2000 and i s providing steady funds for the maintenance o f the primary road network, financed partially through a gasoline tax. In2004, the primary road network comprised 6073 km - up from 5,555 in 2002 - o f which 3017 km were paved and 3056 km unpaved. In 2002, almost 60% o f the paved road network was considered to be ingood or regular condition, compared to just 30% for the unpaved network (see Table 3.10). It must be noted, however, that the quality o f the road network varies considerably across regions. Indeed, Ahuachapan, L a Union and Morazan have the highest proportion o f road network in bad condition - between 5 1% and 67% - while Santa Ana and Sonsonate have the lowest proportion o f road network in good condition-respectively 21% and 23%. l2 Table3.10: Conditionof the PriorityRoadNetwork, 2002 Type of Road Km Good Regular Bad Paved 2,418 43% 15% 42% Unpaved 3,137 2% 28% 70% Total 5,555 35% 24% 41% Source:Ministry of Public Works (MOP) 232. In the context of the Plan Puebla Panama (PPP), specific investments will be made to improve the roads that link the Honduran border at El Amatillo with the Guatemalan borders at PedroAlvarado and San Cristobal. El Salvador, as its Central Americanneighbors has embarked on the Plan Puebla Panama (PPP) project, which seeks to develop infrastructure investmentson a regional basis, inorder to improve the overall economic competitiveness o f the region and promote its physical integration. Inthis respect, a regional road network, called Red Internacional de Carreteras Mesoamericanas (RICAM), has been identified and investments are being channeled to develop, improve and maintain these highways. In El Salvador, the PPP highways comprise a section o f the Pacific Corridor that runs from El Amatillo (border with Honduras) through Sirama, L a Libertad and to Pedro Alvarado (border with Guatemala). In addition, the PPP covers a regional complementary corridor, running from Sirama to San Salvador, through Santa Ana to San Cristobal (border with Guatemala). All o f these highways are existing ones and are part o f the Salvadoran priority network whose maintenance i s assured by FOVIAL. Specific investments will be carried out with donor funding sponsored by the PPP initiative, mainly to reinforce bridges and pavements, and to construct deviations from cities. In this context, one of the most recent projects to be undertakenunder the sponsorship of the PPP is the San Salvador "beltway", that will serve both as a urban road and as a mean for deviating from San Salvador when traveling along the main PPP corridor. `I'ElSalvador PovertyAssessment - Strengthening SocialPolicy, World Bank, 2004 `I2Ministry o f Public Works. 164 233. As other countries in the region, El Salvador has a two-tiered freight transport system: while large international firms employ container-based multimodal transport services, the majority of the small and medium sized Salvadoran firms rely on relatively informal and unregulated service providers. According to the SIECA Central America Transport Study'I3, during the late 1990s the Salvadoran freight transport service sector was characterized by having an old vehicle fleet made up o f 34,000 units o f more than 2 tons, largely composed o f second-hand U.S. vehicles. O f the 10,000 freight transport companies existing at the time, only about 10% were affiliated to the Salvadoran Association o f Freight Transport Entrepreneurs(ASETCA), comprising a fleet of 6,000 units, or less than 20% o f the total fleet. While the industrytransported about 2.6 billion ton-km per year, the theoretical capacity of the system was estimated to be 8 billion ton-km, thus implying a massive over-supply of services. The majority o f the freight transport (70%) was destinedto the domestic market, and it operated largely unregulated, with no restrictions on the duration o f the drivers' workdays or weekly rest periods. Moreover, the majority o f the small transport freight firms - called "unitarios" -relied on third party agents, both to obtain freight - the so-called "comisionistas" - and to help them with bureaucratic procedures - the "tramitadores". Due to their scarce financial, organizational and technical capabilities, most o f their business was one-way, which together with the excess supply described above and the losses associated with the high levels of crime existing in the region contributed to relatively low profit margins. On the other extreme, large multinational companies tend to rely on multimodal transport services provided by shipping lines, which offer door-to-door container services to the U.S. and Europe, as well as air transport services for goods with high value-weight ratios. In both cases, frequencies are well programmed, modern equipment i s available and the different transport modes are well coordinated. Furthermore, the procedures at border crossings and customs are simplified so the overall quality and reliability o f the service is high. However, the cost o f these services is also consideredto be very high, which makes them inaccessible to most o f the small and medium firms. These high costs are in tum though to be due to the limited competition existing in this segment o f the freight transport market, to inefficiencies at the ports and to the large distances to be covered between El Salvador and the Guatemalan or HonduranAtlantic coast ports. Figure3.31: TransportConstrainedfirms (YO) Figure3.32: Transport Constrainedfirms ifCAFTA (YO) 40 8% 13.7% 12.8% Guatemala Nicaragua El Salvador Honduras Guatemala Nicaragua El Salvador Honduras Source: Investment Climate Survey. Source: Investment Climate Survey 234. Even though less than 10% of Salvadoranfirms are constrained by problemsin the transportation sector, the latter acquire much greater importance in the context of CAFTA. Maquilas are more likely to report transportation constraints to their operations `I3 EstudioCentroamericanode Transportes (ECAT), SIECA2001, Appendix 4/2 165 but they are less likely to indicatethat transportationproblems will hamper their abilityto benefit from CAFTA. Only 9.5% of the firms surveyed in El Salvador reported that they face major or very severe constraints in the area o f transportation (Figure 3.31). This proportion i s slightly higher than inHonduras (7.7%), but it is lower than inGuatemala andNicaragua (around 13%). However, a much larger proportion o f the surveyed firms - about 31% in the case o f El Salvador - indicated that transportation problems could create major or very severe constraints for taking advantage o f the potential benefits o f CAFTA (Figure 3.32). As before, that proportion was slightly larger than in Honduras (28.5%), but lower than that obtained for Guatemala and Nicaragua (respectively 41% and 38%). Econometric estimates reported inthe appendix indicate that maquilas are much more likely to report constraints associated with the transportation sector than are other exporters or importers o f the same size, and operating in the same sectors and locations - the only exception being firms that import inputs from other Central American countries, which are also more likely to report such constraints. With regard to the constraints for benefiting from CAFTA, however, muquilus are less likely to report constraints associated with transportation problems. Figure 3.33: Input delivery delays (days lost) Figure 3.34: Transport losses (% of sales) 5 9 _. 4 R 2 0% 1.9% 1.6% 15% Nicaragua Honduras Guatemala El Salvador Nicaragua Guatemala El Salvador Honduras Source: Investment Climate Survey. Source: Investment Climate Survey 235. Objective indicators of problems caused by failures in the transportation sector confirm that those problems are smaller in El Salvador than in neighboring Guatemala and Nicaragua. Moreover, those problems affect to a greater extent the firms that import from or export to other CentralAmerican countries, which suggest that failures in the land transportationsystem are those that generate higher losses to Central American firms. On average, Salvadoran firms loose 4.8 production days per year due to delays in the delivery o f inputs, which amounts to about one day less than what is reported inGuatemala and Nicaragua, and a half day less than in Honduras (Figure 3.33). However, within El Salvador our econometric estimates suggest that importers lose almost an additional day per year, and importers from Central American countries lose another 1.3 days, compared to other importers with similar characteristics. Another indicator provided by the ICs i s the percentage o f sales that firms lose due to either interruptions inthe transportation system or, more importantly, due to the breakage, spoilage or theft o f their products duringtheir transportation. As seen inFigure 3.34, Salvadoran firms lose about 1.6% o f their sales due to those reasons, compared to 1.5% for Honduran firms, and respectively 1.9% and 2% for Guatemalan and Nicaragua firms. However, within each o f those countries the firms that report the highest losses are, according to our econometric estimates, those that have a higher employment size and those that export to other Central American countries -exporters whose main clients are located outside the region are actually 166 less likely to experience such losses. The higher losses reported by larger firms are somewhat surprising in light o f the fact that, as argued above, they are more likely to have access to the modern segment o f the freight transport sector. However, that advantage appears to be compensatedby the fact that larger firms tend to reach a wider spectrum o f clients, often located farther away from their plants. 236. El Salvador has made great progress in streamlining and modernizing customs procedures, as well as in terms of harmonizing and coordinating those procedures with those of other Central American countries. Countries inthe region have increasingly adopted the Central American Unified Customs Code (CAUCA), which is expected to result in reduced customs clearance times. Inthe case o f El Salvador, Custom operations have been accredited and certified under the I S 0 9000 quality standard. The adoption o f the regional regulatory framework, along with reforms introduced to streamline operations within the General Directorate o f Customs (DGRA) and make them compatible with regional guidelines, has resulted in important benefits in terms o f the simplification o f procedures. The DGRA has introduced a self-declaration system in which merchandise recipients self-assess the payable custom duties and go to customs only to enter the necessary information about the transaction and withdraw the goods. The new system has simplified procedures and reducedbureaucratic red tape. Inaddition, a new form called "Declaration o f Merchandise" has been introduced, replacing the old "Import Document". This new form has improved customs operations, reducing the number o f control checks and easing administrative procedures, and has also allowed for the exchange o f informationwith other departments within the Ministry o f Finance. Concurrently, an internet based system has been also put into place, allowing users to fill out electronic custom declarations and process them online. In conjunction with the Directorate o f the Treasury, DGRA has also introduced an electronic payment window to allow users to pay for custom duties online. All these initiatives are part o f the modernization process being carried out by the DGRA and which aims at consolidating virtual customs operations and introducing completely paperless procedures. The DGRA has also awarded concession contracts for the handling and storage o f merchandise to qualified private firms (Almacenes Generales de Deposito). This initiative has resultedin better service, quicker dispatch times, and the availability o f insurance protection to stored cargo. Finally, as a result o f agreements signed with neighboring countries, Salvadoran Custom officials are operating on a pilot basis at peripheral control points in border crossings outside the country. 'I4 `I4Such control points include: Tecdn-Uman, at the border between Guatemala and Mexico; Peiias Blancas, at the border between Nicaragua and Costa Rica; and in the ports o f Quetzal, Barrios and Santo Tomas de Castilla in Guatemala and Puerto Cortes in Honduras. Inaddition, there are various integrated custom control points located at: El Amatillo (with Honduras and Guatemala); El Poy and Guasaule (with Honduras); and Valle Nuevo (with Guatemala). DGRA also operates at juxtaposed control points (control facilities adjacent to those o f the neighboring country) at the border crossings o fPedro de Alvarado-Hachadura and San Cristobal (both with Guatemala). 167 Figure3.35: Firms Constrained by customs (%) Figure 3.36: FirmsConstrained by customsif CAFTA (YO) 51.0% 36.8% 35.3% 23.7% 25 3% Guatemla Nicaragua El Salvador Honduras Guatemala El Salvador Nicaragua Honduras Source: Investment Climate Survey. Source: Investment Climate Survey 237. As in the case of the constraints associated with the transportation system, custom- related constraints tend to acquire much greater importance in the context of CAFTA. They are mentioned less frequently in El Salvador than in Guatemala, but slightly more often than in Honduras. Less than 10% o f the firms surveyed inEl Salvador reported that they face major or very severe constraints in the area o f customs (Figure 3.35). The percentage o f firms reporting such constraints i s practically the same than in Nicaragua, it i s larger than in Honduras (7.7%), but much smaller than in Guatemala (23.7%). A similar ordering o f countries i s obtained when firms are asked whether problems at customs could constrain their ability to benefit from CAFTA: 51% o f Guatemalan firms report major or very severe constraints, compared to about 37% for El Salvador, 35% for Nicaragua and 25% for Honduras (Figure 3.36). Moreover, our econometric estimates indicate that importers are more likely to report general customs constraints, while exporters are less likely to mention customs as an area that could constraint them after CAFTA is implemented. We also find that in the other three Central American countries exporters whose main clients are located within the Central American region are more likely to mention customs as a source o f constraints. This result, however, does not apply to the sample o f Salvadoran exporters. Figure3.37: Customdelays for Exports(investmentclimate data) A v e r a g e a n d longest periods to c l e a r customs, for exports i ( I n t e r n a t i o n a l eo m pa rison, days) 18 I 5 I B a v e r a g e period Olongest period Source: Investment Climate Survey. 238. Probably due to the recent modernization and simplification efforts, the delays faced by exporters at Salvadoran customs are very low not only by regional but also by world standards. However, importers face longer delays than inneighboring Honduras and 168 Nicaragua.According to data from I C surveys, Salvadoran exporters take an average of 1.6 days to clear customs, and report having taken at most 3.3 days to perform the corresponding procedures during the year preceding the survey (Figure 3.37). On the other hand the average and longest delays reported by importers were about 6 and 12 days, above Nicaragua and Honduras but well below Guatemala and Brazil (Figure 3.38). Inthis context, it i s worth noting that despite the progress achieved, private custom agents and freight transporters still see room for improvement. Indeed, the reduction in clearance times achieved with the internet based system is sometimes offset by the delays caused by stringent physical controls conducted by the national police and aimed at fighting smuggling and drug trafficking. Other problems arise as a result o f the excessive discretionality o f officials, the lack o f an adequate and enforceable code o f conduct, the importance that i s still given to the political affiliation o f candidates when filling positions, the lack o f modem risk analysis techniques and appropriate equipment for non- intrusive inspections and faster turnaround o f laboratory sample testing. Figure3.38: Customdelays for Imports(investmentclimate data) A v e r a g e and longest periods to clear customs, for imports (International comparison, days) 3 5 3 0 2 5 2 0 15 10 5 0 I =average period Olongert period Source: InvestmentClimate Survey. 239. Both simple tabulations and detailed econometric estimates suggest that the importers that are most affected by delays at customs are those that use the port of Acajutla and are not constituted as maquilas. Importers that report using only land transportation (that is, importing from neighboring countries) face average import clearance times o f 4.7 days, with peaks o f 7 days. Incomparison, those importing mainly through the port of Acajutla face clearance periods o f 9.9 days, with peaks o f 16 days. In addition, the econometric estimates reported in the appendix show that maquilas face substantially lower clearance times than other importers, and so do the firms whose inputs come mainly from other Central American countries. Inparticular, we also find that the firms that use mainly the port o f Acajutla to import their inputs tend to have longer delays at customs than do the firms whose imports enter the country through landborder crossings. 240. The uncertainty associated with the duration of customs procedures, particularlyin the case of importers,leads firms to holdlarger inventories.The difference inclearance times between imports and exports may be due to the number and type o f inspections carried out by customs on imports, which not only are intended to reduce contraband but also to contribute to a better recollection o f tariff revenues. In the IC sample, exporters declared that on average 37 percent o f their exports in 2003 were subject to inspection while importers declared that 41 o f 169 their imports were subject to inspections115.As for the finding that the longest period taken by customs i s twice the average period, it i s fair to say that it contributes to the uncertainty faced by importers, and may induce them to hold bigger amounts o f inventories o f imported products to provide for this uncertainty. This latter decision adds an extra cost for those firms (see Guasch and Kogan, 2001), which may affect SMEs to a greater extent, given that they usually face bigger liquidity constraints. According to the I C survey, firms that import directly more than 10 percent o f their inputs keep inventories for 33 days o f production while those that import less than 10 percent keep inventories for 14 days on average116.This positive relationship between the amount of inventory heldand the percentage of directly importedinputsremains statistically significant even after controlling by the amount o f inventory deliveries, sectors, sizes and locations o f firms. TRANSPORT:POLICYRECOMMENDATIONS 241. mainly Guatemala - as well as on the highway accesses to those ports suggests that the The great dependence that El Salvador has on the ports of neighboringcountries - country needs to continue to pursue aggressively its investment plans for improving and maintaining its road infrastructure.The importance o f improving the land transport system results both from geographical imperatives - namely the need to access the Atlantic coast ports of.other countries in order to reach the US. east coast - but also from the finding that the manufacturing firms that experience larger losses as a result o f failures in the transportation system are those that trade with other Central American countries. Thus, it i s crucial that El Salvador pursues the investment priorities that are set up in the Plan Puebla Panama (PPP), actively participating in this initiative, and ensuring that neighboring countries also comply as much as possible with their own road improvement agendas. It must be noted that El Salvador has been successfully implementinga road sector reform that has allowed it to establish a Road Fund(FOVIAL), inorder to extend and consolidate the maintenance of its existing roadnetwork. As a result, a priority networkhas beendefined and steady financing has beenmade available for its maintenance. All efforts and policies that support, promote, and expand these achievements should be encouraged. The priority network coincides with the PPP network and will thus ensure the sustainability o f future PPP investments. Notwithstanding these accomplishments, the historical neglect o f secondary and unpaved road network remains a major challenge for the country. In this respect, El Salvador should seek to improve the quality o f its unpaved priority network, o f which 70% i s currently inbad condition. 242. It is extremely urgent to develop adequate regulationsthat will add professionalism to the freight transport industry. Currently this sector is dominated by informal individual service providers - the so-called "unitarios" - that are mostly unregulated and relatively inefficient. In order to reduce transport costs and increase the competitiveness o f Salvadoran producers, the country needs to promote the development o f freight transport service companies that are capable o f providing more complex services, thus enhancing the overall quality and reliability o f the sector. These could be carried out with the development o f incentives to the `Is Ifwe consider the medianvalues insteadofthe means,the difference isbigger,with 10percentand20 percent of exports andimportsbeinginspectedrespectively. This relationis also valid when using a more restricted definition of importers: those importingmore than 40 percent of their inputs keep inventory levels equivalent to 35 days of production, while those importing less than 40% keep 16 days o finventones. 170 formalization and professionalization o f freight transport service providers, and the setting up of training programs that can help the sector on its way to diversifying its services and improving its level o f professionalization. Moreover, a reform should be envisaged regarding liability and accountability o f firms during transport, using a standard international transport contract that establishes clearly the responsibility o f all stakeholders involved, thus facilitating claims and reducing insurance costs. 243. It is clear the El Salvador must improve the efficiency of its ports and makefull use of the existing infrastructureifit is to cope with increasedtrade due to CAFTA. El Salvador i s located in a strategic position that could allow the country to reach the U.S. west coast at relatively low maritime transport costs. Despite these potential advantage, unidirectional trade flows, a low scale o f operations at least in terms o f container movement and low port efficiency levels have translated into high port fees and highmaritime transport costs for Salvadoran firms that import or export through the main port o f Acajutla. This situation, however, has begun to change thanks to a government strategy that has involved considerable government investments duringthe past four years and the introduction o fprivate participation inport activity. As a result the volume of traffic has increased and the port has reported profits for two years in a row. Knowing that hrther efficiency improvements would require more investments and additional improvements in administrative and operational procedures, the government i s currently in the process o f completing the implementation o f a 25-year management concession for Acajutla and i s also finalizing the details for choosing a contractor for building a new terminal inthe currently inactive port o f Cuhco. It must be noted, however, that it i s crucial to frame El Salvador's national port strategy within an overall regional transport strategy, taking into account the potential impact o f the improvement sought by the PPP in the land transport linkages between the region's ports and capital cities. Indeed, these PPP sponsored investments, coupled to improvements inborder crossings and customs, could greatly facilitate the access to Guatemalan and Honduran ports. 244. In the area of customs, El Salvador should continue its modernization and professionalization efforts, with a focus on reducing the delays faced by importers, particularly at the port of Acajutla. During recent years El Salvador has made considerable progress in improving the efficiency o f customs and facilitating border crossings. Nevertheless, there are still deficiencies in customs procedures and operations, some o f which are common to other countries in the region. In this respect, additional improvements could be achieved by a facilitating inspections through the incorporation o f modern equipment and risk analysis techniques, as well as by increasing the professionalization o f the customs agency. Moreover, particular efforts should be devoted to reducing customs delays at the port o f Acajutla which, as shown by the I C s data, are relatively longer than those encountered by importers that use land transportation. 171 ANNEX3.1: ECONOMETRIC ESTIMATES We estimated an econometric model aimed at investigating,in a multivariate framework, the characteristics of the firms that are more likely to stop production and experience monetary losses as a result of power outages. Our empirical estimated equation is of the following form: where Lossi is either a dummy variable for firms that report stopping production as a result o f power outages, or the monetary losses due to the outages as percentage o f sales (for firms that experienced outages). As explanatory variables we include employment size (Sizei), a dummy variable for maquilas, a dummy for firms that have a generator, the number o f power outages, as well as sector (Sectori) and province (ProvJ dummies. The results, with samples covering either just Salvadoran firms, or also firms from Nicaragua, Honduras and Guatemala are reported in Table A3.1. We also estimated a model aimed at characterizingthe firms that are more likely to use the internetand electronic mail. Our empirical estimated equation is o f the following form: where Interneti i s either a dummy variable for f i r m s that report using the internet to communicate with suppliers or clients, or a dummy for firms that report using email for the same purpose. As explanatory variables we include employment size (Sizei), a dummy variable for maquilas, dummies for exporters and importers, as well as sector (Sectori) and province (Provi) dummies. The results, with samples covering either just Salvadoran firms, or also firms from Nicaragua, Honduras and Guatemala are reported inTable A3.2. In the area of transport, we estimated a model aimed at investigatingthe characteristicsof the firms that are more likelyto report that they are constrained by transport- or customs- relatedproblems, and those of the firms that report delays or monetary losses derivedfrom those problems. Our empirical estimatedequation is ofthe following form: Transportj= p'Sizei + p2Sectorj +p3Provi +P4Maquilai +p'Tradei + ~i (A3.3) where Transporti can be either a dummy variable for firms that report major or very severe constraints associated with the transportation system or with customs, or a variables that 172 objectively measures the delays and monetary losses associated with those problems. l7 We control for employment size (Sizei) and include a dummy variable for maquilas, as well as sector (Sectori) and province (PrOVj) dummies in all specifications. Finally, Tradei i s a vector o f variables measuring the firms participation ininternational trade: whether they import and export directly, whether that trade i s mainly with other Central American countries, and in the case o f imports whether land or air transport are used exclusively and, if not, what are the main ports that are used. The results, with samples covering either just Salvadoran firms, or also firms from Nicaragua, Honduras and Guatemala are reported inTable A3.3 and A3.4. '" The above mentioned constraints can either be defined in general terms or in the event o f CAFTA implementation:those constraints are such that they respectively limit the firms' operations and growth, or their abilityto maximizethe benefitsfrom CAFTA. 173 Table A3.1: Correlates of production stoppages and monetary losses due to power outages (plant level data from ElSalvador, Guatemala, Honduras and Nicaragua) Production Production Losses (% o f Losses (% o f stoppage(a) stoppage(a)(c) sales)(b) sales) (b)(c) Employment (log) 0.041** 0.015" -0.004 -0.006*** (0 038) I (0.098) (0.130) (0.000) Maquila dummy -0.278""" -0.064 -0.009 0.004 (0.008) (0.107) (0.471) (0.524) Generator dummy -0.198*** -0.185""" -0,013" -0.010** (0.000) (0.000) (0.082) (0.026) Number o fPower Outages 0.005** 0.0005** 0.0004** 0.0002*** (0.016) (0.038) (0.017) (0.000) Chemicals -0.113 -0.093** -0.015 -0.004 (0 223) I (0.040) (0.228) (0.597) Rubber& plastics 0.127 0.055 0.000 -0.002 (0.152) (0.247) (0.981) (0.791) Mineral prod. -0.159* -0.000 -0.025** -0.012** (0.085) (0.995) (0.036) (0.040) Textiles 0.073 0.009 -0.010 -0.010 (0.505) (0.871) (0.577) (0.218) Metal prod. 0.013 0.009 -0.002 -0.002 (0.847) (0.813) (0.804) (0.783) Food & bev. -0.043 -0.056* -0.011 0.001 (0.490) (0.060) (0.224) (0.856) Ahuachapan -0.154 -0.136 -0.010 -0.011 (0.121) (0.157) (0.509) (0.469) La Libertad -0.085 -0.056 0.001 0.001 (0.162) (0.300) (0.886) (0.949) Other -0,001 -0.006 -0.009 -0.005 (0.988) (0.941) (0.489) (0.701) San Miguel -0.254** -0.328*** -0.021 -0.027* (0.0 16) (0.003) (0.156) (0.062) Santa Ana -0.236"" -0.164* -0.018 -0.017 (0.033) (0.084) (0.175) (0.221) Observations 329 1433 298 1051 Probability values in parentheses.* significant at 10%; ** significantat 5%; *** significant at 1%. (a) Probit estimates with robust standard errors (marginal effects reported) (b) Tobit estimates. Province dummies also for the other countries in columns (2) and (4). (c) Sample covers the four countries. 174 Table A3.2: Correlates of usingthe internet and email to reach clients or suppliers (plant leveldata from ElSalvador, Guatemala, HondurasandNicaragua) Internetuse(a) Internetuse(a) @) Emailuse(a) Emailuse(a)(b) Employment(log) 0.120*** 0.087*** 0.179*** 0.224*** (0.000) (0.000) (0.000) (0.000) Exporterdummy 0.141*** 0.131*** 0.172*** 0.133*** (0.007) (0 000) # (0.002) (0.000) Importerdummy 0.048 0.088*** 0.204* ** 0.209*** (0.352) (0.000) (0.000) (0.000) Maquiladummy -0.088 -0.034 0.028 0.087 (0.271) (0.355) (0.872) (0.325) Chemicals 0.072 0.126*** 0.012 0.152*** (0.411) (0,008) (0.908) (0,010) Rubber & plastics 0.054 0.050 -0.022 0.146* (0.554) (0.320) (0.857) (0.077) Mineralprod. 0.026 0.099"" -0.086 0.037 (0.781) (0.031) (0.365) (0.469) Textiles 0.320** 0.061 0.158 0.147* (0.020) (0.302) (0.251) (0.052) Metalprod. 0.162** 0.080* 0.098 0.128*** (0.045) (0.065) (0.157) (0.009) Food& bev. 0.021 0.031 0.065 0.052 (0 762) I (0.355) (0 324) I (0.231) Ahuachapan -0.120 -0.103 -0.088 -0.066 (0.194) (0.165) (0.565) (0 683) I L a Libertad 0.033 0.014 0.141* 0.160* (0.590) (0.788) (0.054) (0.064) Other -0.194* -0.138" -0.351*** -0.317*** (0,062) (0.055) (0,001) (0.001) SanMiguel -0.088 -0.093 -0.106 -0.160 (0.440) (0.251) (0.348) (0.172) Santa Ana -0.162** -0.123"" -0.189** -0.170" (0 040) I (0.043) (0.041) (0.078) Observations 465 1721 465 1771 Probabilityvalues inparentheses. significant at 10%; * ** significant at 5%; *** significant at 1%. (a) Probitestimateswith robust standarderrors (marginal effects reported) (b) Sample covers the four countries. 175 Table A3.3: Correlatesof constraintsand losses associatedwith TransportProblems (plant leveldata from ElSalvador, Guatemala,Hondurasand Nicaragua) Major Major Transport Transport Days lost dueDays lost due Transport Transport Transport Transport Constr.if Constr. if to delays in to delays in losses (% losses (% Constraint Constraint CAFTA(a)CAFTA(a3)delivery of delivery o f of sales) @) of sales) (a) (a)(c) (C) inputdb) inputs(b)(') @) (c) Employment (log) 0.006 0.008 0.006 0.028** -0.028 0.140 0.007** 0.009*** (0.646) (0.229) (0.764) (0.015) (0.870) (0.123) (0.038) (0.000) Maquila dummy 0.107 0.079** -0.143 -0.089* -0.303 -0.506 -0.007 -0.003 (0.139) (0.043) (0.144) (0.090) (0.729) (0.214) (0.695) (0.721) Direct Importerdummy -0.007 0.027 -0.060 -0.009 0.910** 0.347 -0.008 -0.005 (0.835) (0.186) (0.302) (0.777) (0.045) (0.161) (0.374) (0.376) Imp. from Cent. America0.139** 0.082** 0.049 0.073 1.346" 1.188*** 0.019 0.002 (0.041) (0.028) (0.621) (0.207) (0.052) (0.004) (0.161) (0.850) Direct Exporter dummy 0.023 -0.026 0.064 0.018 0.641 0.460 -0.022 -0.023*** (0.599) (0.362) (0.475) (0.717) (0.368) (0.205) (0.108) (0.008) Exp. to CentralAmerica -0.007 0.038 -0.022 0.018 -0.463 -0.203 0.035** 0.029*** (0.877) (0.260) (0.815) (0.731) (0.535) (0.595) (0.013) (0,001) Observations 465. 1758 465 1771 464 1765 465 1771 Robustp values inparentheses. significantat 10%; * **significant at 5%; ***significantat 1%. (a) Probit estimateswith robust standarderrors (marginaleffects reported) (b) Tobit estimates.Provinceand sector dummies includedinall regressions.(c) Sample coversthe four countries. 176 Table A3.4: Correlates of constraints and losses associated with Customs Problems (plant level data from El Salvador, Guatemala, Honduras and Nicaragua) Major Customs Major Customs Customs Constr. if Customs Customs Customs Customs Constr.(a) Constr. if CAFTA(a) Days Days Days Constr.(a) (') CmTA(a) (c) (imports)@) (imports)(b)(imports) (b)(c) Employment (log) 0.011 0.016** 0.043" 0.050*** -0.016 -0.070 -0.053 (0.412) (0.035) (0.062) (0.000) (0.839) (0.359) (0.149) Maquila dummy 0.011 0.024 0.005 -0.070 -0.815** -0.515 -0.501*** (0.867) (0.499) (0.967) (0.217) (0.015) (0.154) (0.000) Direct Importer dummy 0.093** 0.115*** 0.019 -0.035 (0.015) (0.000) (0.761) (0.290) Imp,from Cent. America -0.027 0.035 -0.033 0.076 -0.586** -0.375" -0.347*** (0.572) (0.313) (0.742) (0.201) ,(0.019) (0,090) (0,009) Direct Exporter dummy -0.012 -0.019 -0.199** -0.144*** 0.326 0.229 0.218 (0,819) (0.554) (0.032) (0.003) (0.247) (0.452) (0.102) Exp. To Central America 0.011 0.067" 0.096 0.122** -0.436 -0.382 -0.287** (0.845) (0,080) (0.371) (0.031) (0.127) (0.207) (0.028) Land imports only -0.947*** (0.000) Air imports only -0.787*** (0.005) Main port: St0.T. de Cast. -0.824*** (0.000) Mainport: Bamos -0.313 (0.316) Main port: Cortez -1.375*** (0.000) Main port: Quetzal -0.269 (0.509) Observations 43 1 1680 465 1771 171 171 580 Robustp values inparentheses. significant at 10%; * ** significant at 5%; *** significant at 1%. (a) Probit estimateswith robust standarderrors (marginal effects reported) (b) DependentVariable is inlogs, when positive. Tobit estimates.Province and sector dummiesincludedinall regressions.(b) Sample covers the four countries. 177 CHAPTER 4: ACCESS TO FINANCE 245. Financial constraints can limit the growth and investment potential of private firms. Inthe context of perfect capital markets, external funds provide a perfect substitute for internal capital and firms' investment decisions are independent o f their financial structure. However, to the extent that financial markets are characterized by the presence o f asymmetric information, the evaluation o f investment opportunities is costly, creating differences in the cost o f debt and equity as compared to internal finance. As a result, internal and external sources of finance are not perfect substitutes and credit constrained firms' investment decisions are dependent on the availability o f internal finance. The necessary reliance on the availability o f internal funds in order to take advantage o f investment opportunities i s an indicationo f financial constraints. The latter can lead firms to make sub-o timal investment decisions, by affecting the timing and amount o f investments. EmpiricallyE8, a large body o f research has shown that investment thus suggestingthat they are credit-constrained .For instance, a recent study by Beck, Demirgiiq- decisions o f smaller firms depend on the availability o f internal funds (e.g., retained earnings), Kunt and Maksimovic (2002) showed, using data on firms from 54 countries, that financial constraints in terms o f access and cost o f funds exert an influence on firm growth and that smaller firms are most adversely affected by those constraints.' I 9 246. This chapter summarizes the basic characteristics of the Salvadoran financial system and uses firm-level data collected through the ICs to provide evidence on two key issues: access to and costs of credit. The first section will provide a brief overview o f the current state o f the Salvadoran financial system. The second section will summarize the main findings o f the survey and interpret them inthe context of the current situation o f the financial system in El Salvador. The last section will provide preliminary policy recommendations, drawing from additional work conducted by the World Bank group. This chapter mainly presents the perspective of Salvadoran firms on the financial system (demand analysis). The study could benefit from complementary assessments o f the views o f and constraints faced by banks and other financial institutions (supply analysis). 247. Even though El Salvador's level of financial development is above that of neighboring countries, improvements could be made to strengthen the stability of the financial system and to overcome some microeconomic constraints that limit access to credit. The first half o f the 1990's witnessed high private credit growth rates reflecting the dynamism o f the economy in the post-war reconstruction period. However, since 1996 the economy has experienced lower GDP growth which led to a decrease in private credit growth. The slowdown o f the economy coupled with natural disasters has put some pressure on banks' portfolio performance and profitability. Duringthe last decade, El Salvador has introduced many positive features in the regulation o f its financial system including: consolidated supervision o f financial conglomerates, increased capital requirements for the establishment o f banks and to cover risky assets, tighter controls on lending to related parties, creation o f a deposit guarantee scheme and improvement in the bank failure resolution system. However, further measures are ~~ 'le Please see Fazzan, Hubbard and Petersen (1988). For applications to countries inLatin America see Galindo and Schiantarelli (2003). ' I 9For a comprehensive review o f the literature on finance and growth see Levine (1997) and World Bank Finance for Growth: Policy Choices in a Volatile World. 178 needed to enhance liquidity resources o f the banking system in times o f crisis and enhance supervision capacity o f lending activities (both domestically and abroad). Despite the problems described, El Salvador financial system sophistication i s considerable given its development level. As a result, survey data shows that firms inEl Salvador have better access to finance and at lower rates that in neighboring countries'*'. However, reforms are needed to expedite creditor rights enforcement and expand the assets available to be used as collateral. Both measures could help increase accessto finance. THESTATEOFTHE SALVADORAN FINANCIAL SYSTEM 248. The restructuringof the financial system and El Salvador's stable macroeconomic environment fostered rapid private credit growth during the first half of the 1990's. The period 1990-1995 was characterized by GDP growth rates averaging 6 percent. As mentioned in the El Salvador Country Economic Memorandum (2003), these high growth rates were a reflection o f a strong post-war reactivation o f investment and consumption, facilitated by a significant inflow o f remittances. In addition, a series o f structural transformations including trade and financial sector liberalization, re-privatization o f banks and other state enterprises, comprehensive tax reforms and improvements in the competitive environment for private investment provided further incentives for growth. In this macroeconomic context and as financial liberalization took place, private credit growth rates averaged 20.4 percent inthe 1990- 1995 period, The credit boom led to increasing inflationary pressures, with inflation rates averaging 14.8 percent inthe 1990-1995 period. 249. Since the second half of the 1990's economic growthgas slowed down, contributing to low rates of growth for both credit and deposits. Despite the series o f reforms, since 1995 the Salvadoran economy experienced lower GDP growth, which averaged 2.6 percent for the 1996 - 2003 period. Apart form a reduction in post war reconstruction activities, El Salvador experienced a series o f adverse external shocks, including a deterioration o f its terms o f trade, which seriously influenced the performance o f the agricultural sector, the 2001 earthquakes and more recently a slowdown o f the U.S.demand for Maquila products. Inthis context, during the second half o f the 1990's private sector credit expansion experienced a sharp reduction, with an average growth rate o f 7.9 percent inthe 1996-2000 period. Apart from the economic slowdown, slower credit growth resulted from increased caution inbanks' lendingpractices and the fact that the private sector reached a limit inits capacity to borrow after five years of credit boom during the early 1990s. Inaddition, starting in 1996 the Central Bank enacted a tighter monetary policy which, by sterilizing capital inflows to limit the supply o f credit, partially offset existing weaknesses inthe supervisory capacity o f the Financial Sector Superintendence. 250. During the second half of the 1990's' El Salvador enacted a series of reforms to strengthen its financial system. Despite the problems described in previous paragraphs, El Salvador has a relatively good level o f financial development. During the second half o f the 1990's, El Salvador enacted a series o f measures to strengthen the financial system, including the establishment o f a Superintendence for the financial system (1996) and a comprehensive banking law (enacted in 1999 and revised in 2002). Many positive features where introduced as a result Starting in 2001, El Salvador adopted the US Dollar as a legal tender which produced a reduction inUSD interest rates as foreign exchange risks were eliminated. 179 o f the reforms, including: consolidated supervision o f financial conglomerates, increased capital requirements for the establishment o f banks and to cover risky assets, tighter controls on lending to related parties, creation o f a deposit guarantee scheme and improvement in the bank failure resolution system. As a result o f this and other reforms inthe first half o f the 1 9 9 0 ' ~El~Salvador boasts one o f the most developed banking systems in Central America. As figure 4.1 shows, El Salvador has a level o f financial sophistication above neighboring countries and slightly above the world average as measured by the World Economic Forum (2004). As noted by FUSADES (2003), when one compares financial sector sophistication to GDP per capita ratios, El Salvador achievement inthe area o f the financial services is evenmore evident. Figure 4.1: Financial Market Sophistication, by country (0 to 7 Scale) 5.4 5.2 Source: World Economic Forum(2004). 251. Even though dollarization led to lower inflationary pressures and interest rates, the absence of a lender of last resort affects the ability of the economy to deal with liquidity crisis. Starting in 2001, El Salvador adopted the US Dollar as a legal tender which produced a reduction in inflationary pressures, as expectations o f monetary mismanagement disappeared. While during the 1990s inflation rates averaged 9.9 percent, they fell to 2.6 percent during the period 2001-2003. Dollarization also had effects on the downward trendo f interest rates inUSD. As devaluation expectations ended, the risk o f lending in U S dollars to companies that eam revenues in local currency was reduced significantly, thus contributing to a decline in interest ratesI2'. Despite these benefits, dollarization implies that the central bank cannot create liquidity intimes of crisis and serve as a lender of last resort. Thus, in a dollarized economy it is vital to set beforehand an adequate set o f mechanisms aimed at minimizing the risks o f liquidity crisis and alleviating the consequences o f such crisis if they were to occur. The government has already increased liquidity requirements in 2003, but further measures are needed to increase the availability o f funds duringtimes o f systemic crisis. It i s worth mentioning at this point, that the deposit insurance fund set up after the enactment o f the Banking Law o f 1999 is not intendedto provide liquidity to the system but to provide relief from specific episodes o f banking failure. Despite this, the deposit insurance fbnd'22only holds assets for about 2 percent o f the covered deposits and would probably be insufficient to deal with the consequences o f the failure o f a mediumor large bank. 12' Despite this, dollarization does not prevent movements inthe real exchange rate (relative prices o f traded versus non traded goods) which could affect the level o f eamings inUS dollars o f companies inEl Salvador. 122 The scheme cover deposits up to USD6,700 and i s funded through premiums paidby participants 180 252. Currently the banking system is characterized by slow credit and deposit growth. Average credit growth was 4.6 percent between 2001 and 2004. Duringthe same period, deposits grew at 2.3 percent per year and banks had to increasingly fund their operations through the use o f external credit lines in international markets, which they were able to do thanks to El Salvador's investment grade status. As shown inFigure 4.2, the low levels o f intermediation and credit demand led to a progressive decline o f active interest rates. Despite the fact that passive interest rates also decreased, the net interest margin has shrunk after 2000. As a consequence, financial institutions have begunto explore other areas o f business and have increasingly relied on other sources o f income (e.g. credit cards), which in 2003 represented around 20% o f total income. 1 2 0 - I::: --\ . ~6 0 - 4 0 - 0 3 2 0 4 1 2 0 - ~ 0 0 1 253. Despite recent improvements in loan quality indicators,the level of non performing loansinthe system remainshigh. Loanquality indicators improvedduring2003. However, due to the slow growth o f the economy and the above mentioned external shocks the banking system portfolio still shows a high level of non-performing loans. Overdue loans amount to 2.8 percent o f total loans as o f 2003. In addition to that, foreclosed assets represent 3 percent o f total assets inthe system or 4.7 percent oftotal gross loans. The highlevel of foreclosed assets inthe system i s partly explained by the Salvadoran procedural legislation that requires a valuation o f the assets to be sold during the enforcement o f secured claims. Anecdotal evidence shows that in many cases assets are appraised well above its market value and thus no bid i s placed. As a result, banks end up satisfying their claims by becoming owners o f the asset that could have been otherwise converted into cash. In addition to foreclosed assets, the extent o f loan quality indicators in the system is also understated due to the swap o f non performing loans to FICAFE'23. 254. Banks are shifting funds away from manufacturingand focusing on mortgage and consumer lending. Since the end o f 2001, mortgages and consumer lending have become increasingly important in banks' portfolios. As o f June 2004, mortgages and consumer lending FICAFE (FideicomisoAmbiental parala Conservacion del Bosque HumedoCafetalero) i s a specialpurposetrust comprised of weak bank credits to coffee producers that were exchanged for 20 year-FICAFE bonds. The total amount under FICAFE is USD 260 Millions (2.8 percent of total system assets). InMarch2003, Congress passed a resolutionthat gave a two year graceperiodto all interestandprincipal from the coffee sector due to FICAFE. 181 represented 38.2 percent o f the loan portfolio o f banks as compared to 27.1 percent in 2001124. At the same time, banks shifted away heavily from manufacturing which now accounts for 10 percent o f banks' portfolios as opposed to 21.9 percent at the beginning o f 2001'25. Several explanations have been given for this decreasing participation o f manufacturing in banks' portfolio including a less dynamic economy and the banks' search for new and more stable sources o f revenues. The latter factor, in particular, would explain the banks' focus on the consumer segment, which even during low growth periods benefits from the steady flow o f remittances from abroad'26. However, increased lending to the real estate and consumer segments also involves higher administrative costs and risks. For example, and as mentioned by FICTH(2003b)' the 2001 earthquakes and the slow growth o f the economy has forced banks to foreclose many real estate loans. In this respect, the Securitization law currently being debated would allow for the pooling o f loans and could thus potentially become an efficient mechanism by which risk associated to real estate lendingcould be allocated more efficiently among market participants. 255. Salvadoran banks are increasingly lending in other Central American Countries. Beginning in2001 and insearch o f additional sources o f income, Salvadoran Banks -mainly the larger ones - started to expand their lendingactivities inHonduras, Guatemala and Nicaragua.'27 Inthis respect, Salvadoranbankshavebeenmotivatedbythe relative comparative advantage that i s associated with lower local interest rates, as well as by the relatively depressed local credit demand. It is worth mentioning, however, that cross country lending activities raise some concerns in terms o f portfolio risk since the countries that are receiving the finds are relatively more unstable than El Salvador.'28 256. The banking system is highly concentrated: four banks hold 81 percent of total assets. As o f March 2004, fourteen banks operated inthe systemwith four of themconcentrating 81 percent o f total assets. This high concentration ratio is a result o f a series o f consolidations that began in the late 1 9 9 0 ~ ' ~and were accentuated by the diminishing intermediation margins ' that prevailed after the enactment o f dollarization in 2001, Foreign ownership in the banking system i s low with only two banks concentrating about 5.3 percent o f assets. Despite this minimum presence, local financial institutions face increasing pressure from foreign banks as they become service providers to large firms in the country. In addition, foreign banks with no local presence are also competing in the Salvadoran market by providing credit to large local firms. 130 There are two state banks (Banco de Foment0 Agropecuario and Banco Hipotecario El Salvador) which hold 4.1 percent o f total system assets. In addition, there are a series o f state owned financial intermediaries aimed at engaging in special purpose financing. They include: 124 Mortgages and consumer lending increased during the period at a CAGR of 14.3 percent and 30.2 percent respectively. The total lendingto Manufacturingdecreased during the period at a CAGR of 20.7 percent. In2003, remittancesamounted to 2,105.3 Million dollars (14.1percent of GDP). As reportedby Yang (2003), 91 percent of the remittancesare mainly usedfor consumptionexpenditures. As of December2003, 5.5 percent of the loans of Banco Agricola, 8 percent of the loans of BancoCuscatlan and 7.8 percent o f the loans of Banco Salvadorefio were domiciled abroad. Source: Fitch Ratings. 12* In2002, the superintendenceintroducedregulationsto limit the exposureof banks abroadto a maximumof 150 percentof abank's equity, althoughany exposure above 75 percent. 129 Apart from increased consolidation due to competitivepressures, irregular bank closings during the second half of the 1990contributedto a lesser extent to the trend. (for example FINSEPROandFINCOMER). I 3 OFUSADES(2003), p, 190. 182 Banco Multisectorial de Inversiones (BMl,), Fondo Social para la Vivienda (FSV), Fondo Nacional de Vivienda Popular (FNVP) and Fondo de Financiamiento y Garantia para la Pequefia empresa. The Banco Multisectorial de Inversiones (BM) i s a second-tier bank that provides funds to be on-lent to specific sectors o f the economy. For example, during 2003 BMI funds financed loans for USD 51 millions to the SME sector. The second most important special purpose institution i s the Fondo Social de la Vivienda, intended to provide home financing to low-income households. Given its target group and the recent evolution o f the Salvadoran economy, a high fraction o f the loans in its portfolio have become impaired (25 percent as o f June 2004) and considerable efforts are beingmade to alleviate this situation.131 257. Salvadoran banks exhibit better efficiency indicatorsthan their counterpartsfrom neighboringCentral American countries. Despite the highconcentration levels, the operating environment for banks in El Salvador remains competitive as banks search for clients in a low intermediation environment. These competitive pressures have also led banks to search for efficiency gains in their operations. As figure 4.3 shows, El Salvador boasts cost to assets ratios that are among the lowest in Central America and are close to those o f the more advanced countries in the region.132However, the recent slowdown in credit and the high level o f foreclosed assets and loans swapped into FICAFE imply that these ratios could be underestimated as the banks' revenue-producing assets may be lower than reported. Figure 4.3: Overhead Cost as a Share of Total Assets, average 1999-2001 Source: Beck and Levine (2003) 258. Micro Lending operations have a good environment for development in El Salvador. The high level o f concentration in the banking system might cause concern since international evidence indicates that firms, especially as size decreases, face higher financing obstacles in markets that exhibit larger banking concentration. 133 However, El Salvador has a well developed microcredit market, which is integrated to the financial sector through the direct and indirect financing o f banks and supervision o f the most important institutions. Furthermore, a securitization law currently under discussion could provide another vehicle to increase the participation o f financial institutions inthe micro credit segment. Currently, one bank specializes inmicrocredit -Calpia now Banco Procredit, holding 54.000 Loans valued at 66 MillionUSD - and two o f the four major banks have microcredit operations. Inaddition, there are six non bank 13'See FITCH (2004). 13*For further details see FUSADES (2003), p, 183. 133Beck et al. (2003). 183 financial intermediaries who are supervised by the SSF that serve as second floor banks to a series o f microfinance institutions. One o f these supervised entities i s Fedecrkdito134, which holds a portfolio o f 140,000 loans valued at $235 million and i s the largest micro finance operation in El Salvador. Despite the progress achieved, many microfinance institutions do not employ risk mitigation mechanisms to deal with common risk factors in their portfolio. Presently, only Fedecredito's system seems to have incorporated some degree o f insurance in its micro loans, and it is already buying reinsurance from insurance companies. Figure 4.4: Firms reporting major or severe obstacles related to finance, by size 3 8 3 9 A v a i l a b i l i t y o f C o s t o f F i n a n c i n g A c c e s s l o F i n a n c i n g F i n a n c i n g i B T o t a l - M i c r o O S m a l l O M e d - L a r g e I Source: InvestmentClimate Survey. ACCESSAND COSTOF CREDIT 259. Financial constraints are mentioned more frequently by micro, small and medium firms as obstacles to their day to day operations and growth. Access to finance is reported as a major or very severe constraint by 31percent o f the firms, and problems related to the cost and availability o f finance are mentioned by about 28% o f the surveyed companies. However, only about 20% o f the large firms covered in the I C s - with more than 75 employees -indicate that the availability o f external funds, or difficulties associated with accessing them (e.g. credit is available but firms have difficulties in getting access to it due to high collateral requirements) are major constraints to their operations, and only 15% mention the cost o f finance as an important constraint. In contrast, between 34% and 39% o f small firms - with between 10 and 25 employees - mention the issues related to the cost o f finance, its availability or access problems as major obstacles (see Figure 4.4). '34 It comprises 48 Cajas de Credit0 (credit unions), 7 Workers (cooperative) Banks, and the SSF-supervised Federation, which acts as a second-floor bank for the other institutions. 184 Figure4.5: Share of Firmswith Loans(%), InternationalComparison Source: Investment Climate Survey. 260. Except among small firms, exporters tend to report lower financial constraints than non-exporters, and so do foreign-owned firms. Only 10.5% o f the firms with at least 10% o f foreign capital ownership mention issues related to access to finance as important constraints, and 5% report that the cost o f finance i s a major problem, compared to respectively 33% and 30% for domestic firms. This could be due to the larger size o f most foreign firms, or to the fact that those firms can potentially use their parent companies to gain access to international financial markets - which would make them less sensible to the local availability o f external finance. Similarly, while among firms that produce exclusively for the domestic market, 34% report major or very severe constraints related to access to finance, 28% o f non-maquila exporters and 26% o f maquilas report such problems (see Table A4.1). Moreover, among large firms with more than 75 employees, only 15% o f non-maquila exporters report access to finance constraints, compared to 30% for non-exporters o f the same size. Among small firms, however, an inverse relationship i s observed, with more exporters than non-exporters - respectively 50% and 30% in the case o f availability o f finance - reporting financial constraints. One could speculate that this is result is driven by a larger demand for external finance among exporting firms in general - associated with wider investment and growth opportunities - and by a lower access to bank credit among smaller firms, regardless o f their exporting status. 261. Survey data shows that Salvadoran firms objectively have a higher level of access to finance than their counterparts from neighboring Central American countries. However, the differences are smaller for micro and small firms. Figure 4.5 shows that 63.2 percent o f the firms in El Salvador'35report having a loan from a financial institution, a level above that of other Central and Latin American countries. However, as can also be seen in Figure 4.5, the percentage o f firms with access to loans in El Salvador is much lower than that found in export oriented countries such as Thailand and Malaysia. Moreover, a more detailed comparison with neighboring Guatemala, Honduras and Nicaragua shows that while the better access statistics found for El Salvador apply to all size ranges, the overall difference i s driven mainly by the better access to bank credit o f medium and large Salvadoran firms. Indeed, in the latter group about 74% o f the firms report having loans from financial institutions, compared to 55% on average for medium and large firms from the three above mentioned countries. In contrast, a 135A similar figure is reportedby a March2003 survey carriedout by FUSADES.For more details see FUSADES (2003). 185 n with L 1 Xlt for bank loans are more likely to have them than are firms with some degree o f foreign ownership (at least 10%). 265. Our econometric estimates also show that firms that own real estate are lesslikelyto be credit constrained and so are firms with externally audited financial statements and high levels of capacity utilization. These results hold even after controlling for firm size, sector, location, and exporter and ownership status. We also find that firms operating in the textile and food and beverage sectors, and those located inthe provinces o f Ahuachapan and L a Libertadare less likely to be credit constrained. These results are apparent in simple tabulations - see tables A2.2 and A2.3 -but they emerge more strongly from the regression analysis (see annex 4.2). Figure4.8: Interest Rates for loansin USD grantedto largefirms after 1999,by country I 14.8 I El Salvador I ~ Honduras Nicaragua Guatemala Source:InvestmentClimate Survey. 266. Firms in El Salvador pay lower interest rates than their counterparts from neighboring Guatemala, Honduras and Nicaragua. In order to provide a basis for international comparison, Figure 4.8 shows U S dollar interest rates paid by large firms in other Central American countries. We take into consideration only large firms to avoid comparison problems due to different risk profiles among firm sizes and because in the rest o f Central America, especially in Honduras and Guatemala, access to U S Dollar loans is much less widespread than in El Salvador. As can be seen in figure 4.8, interest rates paid by large Salvadoran firms (7.6 percent per year) are lower than those charged in other countries in the region.137 267. As shown by simple tabulations and confirmed by econometric estimates, interest rates vary across sizes and exporting status, and they are lower for firms that are incorporated, have foreign ownership and own real estate. As expected, interest rates tend to be smaller for larger firms: as seen in Figure 4.9, the average interest rate for loans in USD granted after 1999 decreases from 11.7% for microenterprises to 7.6% for large firms. That picture also shows that exporters and particularly maquilas tend to pay lower interest rates than non-exporters. Regression results reported in annex 4.2 confirm that MSMEs tend to pay higher interest rates than larger firms (see annex 4.2). However, with regard to the differences ininterest rates related to exporter status we find that they are significant in the case o f El Salvador, after other firm characteristics are controlled for. Moreover, when the analysis i s performed 13'In Costa Rica, the average interest rate for manufacturing firms was 8.1% in 2003. Using Central Bank data, FUSADES (2003) also finds that the interest rates charged in El Salvador for loans in USD are among the lowest in Central America. 188 incorporating data from Guatemala, Honduras and Nicaragua, we find that exporters pay lower interest rates than non-exporters, but the differences are much smaller for "regional" exporters than for other exporters - i.e. the lower rates are those paid by exporters whose main clients are located outside Central America. Figure 4.9: Interest Rates for loans in USD granted after 1999, by size and exporter status 4 " 7 I Micro Small Med Large Maquila Non Maquila NOExports Exporter Exporter Source: InvestmentClimate Survey. ACCESSCONSTRAINTSREPORTED BY FIRMS 268. After the lack of demand for credit, the reason that is given most frequently for not applying to bank loans is the presence of high collateral requirements. Although high interestrates are the second most frequent reason mentioned by Salvadoranfirms, interest rates are a much smaller concern in ElSalvador than in other CentralAmerican countries. As much as 87.8% o fthe Salvadoran firms that report not having a loannever applied to one. For those that did apply to that type o f credit, the rejection rate suggested by the survey is 6.7%, which i s slightly lower than the figure obtained for the other three Central American countries for which we have data (7.9%). These low rejection rates probably reflect a self selection mechanism that prevents firm from applying to bank loans if they know that they have low chances o f obtaining them. Among those firms that never applied to bank loans, the lack o f demand for that type o f credit is mentioned by 60.3% o f Salvadorans firms - compared to 45% for the other three countries. The other reasons most frequently mentioned are high collateral requirements, high interest rates and cumbersome procedures. However, probably as a reflection o f the lower interest rates prevailing inEl Salvador, interest rates play a much less important role in deterring firms from applying to bank loans in El Salvador than in neighboring countries (Figure 4.10). Indeed, survey data shows that 52.5% of the firms that did not apply for bank loans - despite having a demand for them - report high collateral requirements as one o f the two main reasons for doing so, 42.3% cite high interest rates and 25.4% mention cumbersome procedures. While the first and the third reasons are mentioned by similar fractions o f firms in neighboring countries, problems related to high interest rates are much less frequent in El Salvador - e.g. they are mentioned by 82.4% o f Nicaraguan firms as a reason for not applying to bank credit. 189 Figure4.10: Constraintspreventingfirms from applyingto a loan I100 , , I 90 80 70 60 50 40 30 20 10 - -- Cumbersome Procedures Collateral Requirements High Interest Rates mHonduras .Nicaragua =Guatemala OEl Salvador Source: InvestmentClimate Survey. 269. Firms in El Salvador are required to provide collateral in an amount equivalent to 123% of the loans. Survey data shows that in El Salvador 86 percent o f the firms with loans were required to provide collateral. As can be seen in Figure 4.11, on average the value o f collateral represents 123 ercent o f the loan amount, a figure that i s in line with those obtained for other L A C countrieslY8, although higher than in Bangladesh, China, Malaysia and Pakistan. The data shows no variation inthe amount o f collateral requested across size ranges, regions or industries.It is worth noting that the enforcement o f secured claims inEl Salvador currently does not provide an efficient mechanism for the li uidation o f collateralized assets. As mentioned in the introduction, the Civil Procedures Code" requires a valuation o f the assets to be sold and anecdotal evidence suggests that many times appraisals are set well above market value. As a result, no bids are placed and banks are forced to absorb those assets. At the same time, delays in the enforcement of secured claims and inefficiency in the judicial system also prevent a fast liquidation o f assets. We will return to these topics in the policy recommendation section o f the chapter. 270. Collateral Sources: Real Estate i s what counts. As in many other countries in the region, the prima source o f collateral i s real estate, which represents 69 percent o f total collateral sources", Inthe case ofElSalvador, this is facilitated by a modemrealestate registry that has recently been unified. On the other hand, the use o f movable assets as collateral i s limiteddue to the non existence of a unifiedregistry, which creates the possibility o f duplicate pledges over the same asset. Moreover, the relatively antiquated legislation prevents the use o f certain assets as collateral and, as a result, machinery is not widely used as a form o f collateral and represents only about 14 percent o f the collateral used inbank loans (see Figure 4.12). Apart from anecdotal evidence collected in the field on the problems o f using movable assets as collateral, the results o f our regression analysis show that firms that report owning real estate are 13' It i slower than inNicaragua, Ecuador and Honduras but similar to Peru and Guatemala. ' 3 9Rules in the Civil Procedure Code establish that assets have to be valued prior to their auction and that offers for less than 2/3 o f that value are not admitted. If there are no offers above this level, the auction can be repeated as many times as needed but the 2/3 threshold can not be broken or ,alternatively, the debtor can seize the asset. Besides these complexities o f the Civil Procedure Code, current regulations allow banks to keep the seized assets for a period o f five years, although accounting practices require provisioning each year a percentage o f the value booked. This regulatory practice could also act as a des-incentive for the development o f an efficient market for the sale of assets. Inany case, it is clear that procedural complexities associatedwith liquidations o f assets are high inEl Salvador, not allowing for the development o f an efficient secondary market. 140 Real Estate represents 62 percent in the case o f Honduras, 38 percent inNicaragua and 62 percent in the case o f Guatemala. 190 9 percent more likely to obtain a loan than those who don't (see annex 4.2). This finding i s particularly relevant for the case o f micro-firms, which often do not own real estate, and which would particularly benefit from a more widespread use o f movable assets as collateral. Survey data also shows that in the case o f micro and especially small firms, there i s a greater use o f personal assets as a form o f collateral, which mainly reflects the lack o f distinction between the owner's and the company's assets and i s a common phenomenon in Latin America. Although the survey does not provide additional details on this issue, anecdotal evidence suggests that these personal assets are often in the form o f real estate, which would imply an even larger reliance on this type o f assets for collateral purposes. It i s also worth mentioning that the government has made efforts to increase access to collateral by allowing, since 2001, for the creation o f "Sociedades de Garantias Reciprocas". Nonetheless, the system has not experienced strong development yet. Figure 4.11: Average value of collateral (as YOof loan), by country 218 Source: InvestmentClimate Survey. Figure 4.12: Sources of collateral (as YOof loan), by size. El Salvador. 79 1 R e a W e Macbinerv lntanclible Pers nal and Others I ElMicro .Small U M e d OLarge =Total 271. Having audited statements increases the chances of obtaining a loan, but the costs of external audits prevent increases in their use among smaller firms. As previously mentioned, 25 percent o f the firms inthe sample report that a reason for not applying to a loan i s the cost and time o f the application process. Audited financial statements can improve the application process by expediting the review o f credit applications. Survey data shows that financial statements auditing is much more common in El Salvador than in neighboring countries. Around 79 percent o f the firms in the sample conduct auditing o f their financial statements, compared to less than 35% on average inthe other three countries. As expected, this number i s greater in the case of medium and large firm which audit their financial statement in 191 more than 95% o f the cases, compared to respectively 49 and 71 percent for micro and small firms. 14' As mentioned above, the econometric estimates reported inthe technical appendix show that after controlling for other firm characteristics (including size), firms that audit their financial statements are about 10% more likely to have a bank loan, provided that they have a demand for it. However, it is also worth noting external audits also have direct and indirect costs, with the latter being associated to the formalization o f the firms' activities (with the corresponding tax implications). Indirect costs, in particular, appear to be especially important, since only 17 percent o f the surveyed firms report that direct costs are a constraint for hiringexternal auditors. SOURCES OF FINANCE 272. Firms rely mainly on retained earnings for their working capital: they represent 45% of their total needs, compared to 25% for bank finance and 19% for supplier credit. Micro and small firms rely more heavilyon internal funds and use bank credit to a lesser extent. The composition of the sources o f finance in El Salvador is relatively similar to that found inother countries inthe region (see Figure4.13). On average, Salvadoran firms finance 45 percent o f their working capital needs with retained earnings while bank financing represents 25 percent o f the total sources o f funds for Salvadoran firms. This figure i s very similar to that found in Honduras but much higher than in Guatemala and Nicaragua - where banks cover less than 14% o f the firms' total working capital needs. However, as a reflection o f their decreased access possibilities, micro and small firms rely relatively more on retained earnings for their financing needs: banks represent only about 17% o f their working capital, compared to respectively 29% and 37% for medium and large firms (Figure 4.14). In addition, survey data shows that small firms - with between 10 and 25 employees - rely more on supplier credit, which represents 24.4 o f their working capital, compared to 12% for large firms, and respectively 17% and 19% for micro and medium firms. Moreover, given their more limited access to bank finance micro and small firms use to a greater extent funds form family and friends, as well as credit cards. Figure 4.13: Main Sources of finance for working capital, international comparison (YO) I 59 58 1 1 Retained earnings Banks Supplier Credit 1 Source: Investment Climate Survey. 14'The large differences with other Central American countries persist when comparisons are made by size range: the percentages o f micro and small firms that report external audits in the other three countries are respectively 14.3% and 27.9%, while the fraction o f medium and large firms with external audits is 58.1%. 192 Figure 4.14: Main Sources of finance for working capital, by size ( O h ) Micro Small Med Large Total '~ LEIRetained Earnings WCommercial Banks OSupplier Credit/ Source: Investment ClimateSurvey. 273. Firms rely more on commercial banks for their investment needs than for their working capital. However, retained earnings are still the main source of investment - - finance. Survey data shows that 68 percent of the firms inthe sample pursuednew investments. To fund these investments firms relied on similar sources o f finance as in the case o f working capital needs. Retained earnings were again the main source, representing 47.1 percent o f the funds usedfor new investments.As shown inFigure4.15, this percentage is lower than for most countries for which we have comparable data. In the case o f El Salvador, the lower share o f financing associate with retained earnings i s replaced by higher shares o f bank finance and supplier credit. Thus, 30 percent o f the funds used for new investments are from bank loans, compared to 14 percent in Brazil and 17 percent in Nicaragua and 20 percent in Guatemala. Moreover, supplier credit finances 12% o f new investments, compared to 6% on average in Guatemala, Honduras and Nicaragua. Within El Salvador, the share o f investmentfunds financed through commercial banks increases with firm size, while that funded through supplier credit i s smaller for large firms (Figure 4.16). Figure 4.15: Main Sources of finance for investment capital, by country (YO) Retained earnings Supplier Credit Source: Investment Climate Survey. 193 c P I 277. Besides procedural complexities, the Judiciary is highly inefficient, which brings more uncertainty to the resolutionof commercial disputes. Improving court facilities and judges' experience with and knowledge of commercial law would raise firms' trust in the judicial system and improve creditor rights. Users o f the Judicial system express concerns with its efficiency and the limited experience and training o f judges when dealing with commercial law cases. Apart from the cumbersome procedures already described, court facilities and equipment are insufficient or deficient. Moreover, the management o f cases i s also inefficient and there is not a well developed system for the selection and continuing training o f judges in commercial law. Recent cross-country studies suggest that judicial efficiency (both for creating security titles and enforcing contracts) and inflation are the main sources o f spread differentials across countries.143 Improving the efficiency o f commercial courts would provide a better framework for more efficient creditor rights enforcement, thus reducing litigation costs for banks and (possibly) interest rates. The government i s currently implementing a Judicial Modernization Project with the support o f the World Bank.'44 This program i s aimed at strengthening the institutional management capacity o f the Judiciary, modernizing the Court System, improving access to and transparency o f the Judiciary and developing the professional competence and quality o f officers and employees o f the Judiciary. 278. Insolvency procedures are rarely used in El Salvador as the system is plagued with inconsistencies and is slow and costly. The government should enact a unified legal framework to facilitate insolvency procedures. Although the Salvadoran law allows for the possibility o f bankruptcy as well for reorganization procedures involving suspension o f payments, the corresponding procedures are rarely used and are perceived as complicated, slow and costly.'45 The legislation for corporate insolvency is fragmented, as it i s governed by three different legal bodies which are applied simultaneously: the Code o f Civil Procedures, the Code of Commerce and the Law o f Mercantile procedures. This legal fragmentation results in inconsistencies that make the enforcement o f bankruptcy relatively unpredictable, and subject to a high degree o f discretionality on the part o f judges. The current state o f affairs does not facilitate the reorganization o f viable companies nor the liquidation o f insolvent ones and creates additional difficulties for effective creditor rights enforcement that ultimately translate in increased access constraints. The Government should thus enact a unified legal framework for the resolution of insolvency procedures. 279. The Government should provide support for the creation of a unified registry of movable assets. El Salvador has done extensive reforms inthe area o freal estate registration but the use o f movable assets as collateral is hampered by the lack o f a unified registry. Collateral availability i s reported by 52.5 o f firms as a reason for not applying to bank loans. As discussed, the main source o f collateral used in El Salvador is real estate. As a result, increasing the spectrum o f assets that can be used as sources o f collateral for firms that do not own real estate could increase their chances o f obtaining bank credit. The Real State registry i s now unifiedand '43Laeven and Majononi (2003). Cross section model on the determinants of bank and aggregate level spreads. Sample o f 106 countries. '44Judicial Modernization Project- El Salvador - July 5,2002 '45According to anecdotal evidence, there have been only one case o f bankruptcy and only a few (less than five) reorganization proceedings inthe past twenty years. 196 i s thought to be one o f the most advanced in the region.'46Besides, a large volume o f titles have been scanned and stored electronically. On the other hand, a unified system for pledges on movable assets does not exist in El Salvador, increasing the likelihood o f duplicating pledges over the same asset. A pledge may be subject to registration in the registry o f Property or the registry o f Commerce and pledges over vehicles should also be filed at the vehicle registry, which i s under authority o f the the Vice-Ministry o f economics. In addition, the pledge legislation i s antiquated and limits the use o f some assets to secure loans. These problems limit the possibilities for usingmovable assets as collateral thus reducing access to finance, especially for firms that do not own real state. The government should support the creation o f a single registry for pledges on movable assets in order to eliminate the current fears o f duplication o f pledges inthe system. 280. The government should make efforts to regulate adequately the market for the securitization of assets which could increase the funds available to Microfinance institutions. The newly created "Reciprocal Guaranty Societies" (Sociedades de Garantias Reciprocas) could also be a valuable tool to improve access to credit among companies that lack other sources of collateral. As mentioned earlier in the chapter the government has made important progress with regard to two new instruments for increasing access to finance: a Securitization law currently under discussion in Congress and a government decree creating the "Sociedades de Garantias Reciprocas" (SGR). The Securitization law would allow to pool together micro credit loans, thus providing new funding for micro finance institutions. To allow for an orderly development o f the market, the government should pay special attention to the regulation o f the standards for origination and administration o f the loan portfolios to be securitized under the new legal framework. As for the newly created "Sociedades de Garantias Reciproca~~''~~,they could contribute to increasing the sources o f collateral available to MSMEs that wish to apply to loans from banks or other financial institutions. Currently there i s only one SGR operating in the country (it was created inFebruary 2004). The SGR system could become a useful mechanism to enhance access to finance and a deeper understanding o f the state o f the system would help to provide advice on mechanisms to foster its development. 281. A constraint for further growth of micro lending is the treatment of individual credit risks as if they were statistically independent. Adding insurance to common risk factors could help minimize loan non-performance. As mentioned in the introduction, El Salvador has a well developed micro finance market. However, the credit risk practices o f the financial institutions involved in that market could be further improved by insuring debtors against common risks, such as natural disasters or commodity price shocks. Whenever the micro lender has a strong expertise in a specific segment, mitigating risk by lending to borrowers exposed to different types o f risk could be inefficient. Thus, in order to minimize loan non performance due to common factors affecting their portfolio, a better approach would be that including insurance inthe finance packages offered by micro-lenders, which should then reinsure `46 The real estate survey (catastro), the registry of Real Estate and Mortgages (Registro de la propiedad real e Hipotecas) and the registry o f Commerce (Registro de Comercio) are now functioning under a common national registrationcenter (Centro Nacional de Registros). 14' Decreto No 553. September 2001 197 themselves in the market instead - as op osed o f using self-insurance mechanisms by means o f inefficiently diversifying their portfolios. E 8 14'Presently, only FedecrCdito's system seems to have incorporatedsome degree of insurance inits micro loans, and that institution i s already buying reinsurance from insurance companies. 198 ANNEX4.1:TABLES Table A4.1: Major or severe obstaclesrelated to finance, by exporter status and ownership Source:InvestmentClimate Survey. Firmswith Loans Firms that don't Credit Constrained needloans Firms Ahuachapan 66.7 26.7 6.7 La Libertad 71.2 21.9 6.8 San Miguel 69.2 7.7 23.1 San Salvador 60.7 20.0 19.3 SantaAna 63.9 19.4 16.7 Other 58.8 17.7 23.5 Total 63.2 19.6 17.2 Firmswith Firms that don'tCredit Constrained Loans needloans Firms Source:InvestmentClimate Survey. 199 ANNEX4.2: ECONOMETRICESTIMATES T o investigate the determinants of access to bank credit we employ the framework proposedby Bigstenet al. (2003) to simultaneously modelthe demand for and the supply of bank loans. We estimate a two-step Heckman selection model that permits to study simultaneously the factors underlying the demand for and the supply o f bank credit. Both the demand and the supply equations control for a series o f firm characteristics, including size, sector, region, exporter status, foreign capital ownership status, the degree o f capacity utilization, real estate ownership, the presence o f externally audited financial statements, and whether the firm is incorporated. In addition, we also include in the demand equation the average level of education o f the workforce and variables that proxy for the general perspectives o f the firm regarding the country's economic and political stability, as well as the presence o f constraints related to access, availability, and cost o f finance. We also estimate a model o f the determinants o f the interest rates paid by Salvadoran firms in loans granted after 1999, using the same explanatory variables o f the credit supply model, as well as dummy variables for loans granted in US dollars and for loans granted ineach of the years after 2000. The results of both models are presented inTable A4.4. 200 Table A4.4 -Determinants of Supply and Demandfor Loans, and interestrates Firmhas Bank Demand for Firmhas Bank Demand for Interest Rate Interest Rate Loan (a) Loans(a) Loan (a)(c) Loans(a)(c) (log) (b) (log) (b)(c) Microenterprise -0.277*** -0.345 -0.316*** -0.179 0.159* 0.191*** (0.001) (0.246) (0.000) (0.222) (0.066) (0.005) Small firm -0.193*** -0.489" -0.229*** -0.252" 0.203*** 0.214*** (0.002) (0.051) (0.000) (0.055) (0.002) (0.000) Medium firm -0.100** -0.157 -0.135*** -0.050 0.172*** 0.166*** (0.044) (0.489) (0.000) (0.696) (0,001) (0.000) Exporter dummy -0.126** -0.181 -0.067" 0.021 -0.041 -0.127** (0.039) (0.408) (0.068) (0.854) (0.472) (0.034) Exporter to Central America 0.148** 0.129 0.141*** 0.124 -0.076 0.119** (0.017) (0.575) (0.000) (0.344) (0.174) (0.037) Foreign ownership dummy -0.143 -0.764*** -0.169*** -0.898*** -0.106** -0.003 (0.108) (0.002) (0.007) (0.000) (0.044) (0.969) Capacity Utilization 0.328*** 0.168 0.043 -0.083 -0.033 -0.065 (0.002) (0.65 5) (0.460) (0.651) (0.75 8) (0.459) Firmowns real estate 0.092" 0.151 0.121*** 0.183** -0.092* -0.090** (0.096) (0.362) (0.000) (0.030) (0.075) (0.035) Audited financial statements 0.096 -0.200 0.096*** 0.037 0.034 0.097** (0.149) (0.321) (0.002) (0.678) (0.690) (0.025) Incorporated firm 0.024 0.304 -0.014 -0.009 -0.189*** -0.107** (0.665) (0.103) (0.662) (0.925) (0.006) (0.019) Chemicals -0.030 0.124 -0.040 0.271 0.018 -0.035 (0.703) (0.661) (0.430) (0.112) (0.784) (0.635) Rubber & Plastics 0.018 0.450 0.044 0.199 -0.013 0.017 (0.837) (0.167) (0.488) (0.307) (0.876) (0.781) Mineral Prod. -0.056 -0.397 0.031 0.168 -0.248*** 0.029 (0.545) (0.128) (0.527) (0.233) (0.007) (0.746) Textiles 0.218*** -0.375 0.086 0.069 -0.006 0.185** (0,001) (0.399) (0.181) (0.718) (0.963) (0.035) Metal Prod. 0.020 0.094 -0.009 0.077 0.024 0.045 (0.788) (0.686) (0.858) (0.572) (0.750) (0.560) Food & Bev. 0.068 0.276 0.071* 0.281** -0.103* 0.046 (0.261) (0.203) (0.059) (0.015) (0.053) (0.460) Ahuachapan 0.122 -0.157 0.154* 0.082 0.088 0.093 (0.170) (0.676) (0.071) (0.838) (0.395) L a Libertad 0.100** -0.075 0.101** 0.018 -(0.400)8* 0.09 -0.083 (0.048) (0.717) (0.039) (0.930) (0.068) (0.131) Other 0.103 0.223 0.083 0.122 0.092 0.135 (0.309) (0.472) (0.3 80) (0.675) (0.395) (0.212) San Miguel 0.148 0.639 0.076 0.520 -0.019 -0.086 (0.179) (0.178) (0.436) (0.212) (0.829) (0.257) Santa Ana 0.069 -0,101 0.061 0.019 -0.140 -0.128 (0.381) (0.705) (0.451) (0.943) (0.112) (0.179) Loan is inUSD 0.039 -0.262*** (0.727) (0.000) Observations 459 459 1726 1726 255 804 Robust probability values inparentheses. significant at 10%; * ** significant at 5%; *** significant at 1%. (a) Heckman two-step estimates.(b) Ordinary least squares estimateswith robust standarderrors. (c) Sample includes firms from four countries (El Salvador, Guatemala, Honduras and Nicaragua). 201 CHAPTER 5: SKILLS,QUALITYAND TECHNOLOGY EDUCATION LABOR AND SKILLS 282. Between 1960 and 2000 El Salvador made significant progress in increasing the average schoolingof its population, which now surpasses that of Guatemala, Honduras and Nicaragua, but is still below that of most Latin American countries. During the past four decades, the average years o f schooling o f the Salvadoran population increased from 1.7 to 4.5. This increase was among the largest of Latin America. As a result, El Salvador is now above most o f its Central American neighbors in terms o f the level o f schooling o f its population - the exceptions being Costa Rica and Panama. However, despite this progress the average schooling found in El Salvador remains below that o f most LatinAmerican countries, suggesting that there are important challenges ahead, particularly if one considers that even countries such as Chile and Peru - with among the highest levels o f schooling o f the region - are still well below the levels o f educational attainment found inEast Asian countries (Figure 5.1). Figure 5.1: Average years of schoolingin 1960 and 2000- International comparison 10.5 Source: D e Ferranti et. al. (2003). 283. More than 80% of the Salvadoran population has at most primary education, while the fraction of the population that has tertiary education (10.6%) surpasses that with secondary (8.8%). In most developing countries with low levels o f average schooling the distribution o f educational attainment has the shape o f a pyramid, with a wide base o f individuals with at most primary schooling, and decreasing fractions of the population with respectively secondary and tertiary schooling. As average schooling increases the educational distribution usually maintains its pyramid shape, but with a shrinking base and a wider share o f individuals with secondary schooling. Eventually, as it is currently the case in several East Asian countries, the fraction o f the population with secondary schooling becomes greater than both the fraction with primary and with tertiary schooling. At a latter stage o f development, the share o f the population with tertiary education surpasses that with secondary education, but only after the latter has surpassed the fraction that has at most primary schooling (see Box 6 for the American experience). In El Salvador, however, as well as in a handful o f Latin American countries including Venezuela, Costa Rica and Ecuador, the educational distribution has evolved from the 202 traditional pyramid into an anvil shaped distribution, where the fraction o f the population with secondary schooling i s smaller than both the ones with primary and tertiary schooling (Figure 5.2). Moreover, in spite o f recent advances, in 2000 secondary school enrollments ere still only half o f what would have beenexpected, given El Salvador's income.149 Figure5.2: Educational distribution in2000 -International comparison . !BNoStudiesor Primarv DSecondarv DTertiarvI Source: De Ferranti et. al. (2003). 284. Regardless of its causes, the shortage of secondary educated workers has worrisome implications,not only for continuingto reduce poverty levels, but also to promote higher rates of economic growth. The absence o f a considerable mass of secondary educated workers has worrisome implications. Such a base i s crucial if El Salvador is to transit through a path o f sequential technological upgrading, which requires a high share o f the workforce with secondary schooling. Moreover, a considerable mass o f secondary educated workers is crucial for attracting trade and FDIwith a high technological content, and to benefit from those investments and their spillovers inthe rest o f the economy. BOX 4: The Great Transformationof American Secondary Education Between 1910 and 1940, secondary schooling expanded greatly in the United States. In 1910, 10 percent of youths were high school graduates, but by 1930the medianyouth had a high school diploma. This transformation gave the United States a large lead over other industrializedcountries in the level of education of its work force - a lead that persisted for a long time: As late as 1962, for example, only 15 percent of British 17-year-olds were inschool, while the high school graduation rate in the United States was almost 70 percent. What explains this "Great Transformation" of Americaneducation? Recent work by Goldin (1999,2001) provides anumber of insights. First, prior to mass secondary enrollment, the rate of retumto secondary school was very high inthe United States. This rate of retumwas high, in part, because high schools in the United States successfullytransformedthemselves from institutions that only prepared youths fro college, to programs that taught their students skills - such as accounting, typing, shorthand, and algebra - that were valued inthe workplace. Such high rates of retumprovided youths with the incentive to seek further schooling - even those who did not intend to go on to college. Second, school administratorsand otherswaged a successful campaign to convinceyouths andtheir parents that high schools offered them something of value. Third, the low levels of inequality and the egalitarian features of the schooling system in the United States, which did not encourage early, rigid tracking of students, facilitated mass secondary enrollments. Lastly, the Great Depression substantially reduced the employment prospects of youths, and thus reduced the opportunity cost of schooling. As Goldin explains, "the Great Depression may have had one positive effect: it enticedyouth to stay in school." Differences inthe opportunity cost of schooling also help explain why the "high school movement" laggedinthe agricultural areas of the South, as well as the industrial areas of the North and Midwest. Massive expansion of secondary enrollments in the United States substantially brought down the high school wage premium. As the premium fell, employersintum hiredmore high school graduates in occupations that hadpreviouslybeenfilled by unskilled workers. Source: De Ferranti et al. (2003),Closing the Gap inEducation and Technology. LV5 Figure5.3: Share of Skilledworkers by firm size(%) $so &a &9 8 ,&e rsnon Droductionworkers MmanaaersBDrofessionais I Source: Investment Climate Survey. 285. At least in the manufacturingsector, the firms that employ a larger share of skilled workers are relatively smaller, they report the use of imported inputs but they are either oriented towards the domestic market or export mainly to other Central American countries. As seen inFigure 5.3, the shares o f non-production workers and that of managers and professionals tend to be larger among smaller firms. This negative correlation between size and the share o f skilled workers employed by manufacturing firms is confirmed by econometric estimates reported in the appendix. Those estimates also show that for a given firm size, and in given sectors and locations, exporters tend to employ relatively fewer skilled workers except when their main foreign clients are located in other Central American countries - skilled labor shares and average years o f schooling are similar among the latter type o f exporters and for the firms oriented towards the domestic market. This result i s hardly surprising, since exporters are likely to specialize in the production o f products in which the country has a comparative advantage, and make an intensive use o f relatively abundant unskilled labor. Interestingly, we also find that there is a complementarity between the use o f imported inputs and the demand for skilled labor, which suggests that those inputs serve as a channel for the absorption o f foreign technologies that embody a bias toward skilled labor. Figure5.4: Firmsthat provideinternaland externaltrainingto their workers, by country (YO) 56 54 54 ElSalvador Guatemala Honduras Nicaragua I Otraining Ointernai training Oexternal training mtraining by lnsaforp (or equiv.) Source: Investment Climate Survey 286. With the exception of maquilas, exporters are more likely to provide training to their workers. Other characteristicsassociated with the provision of training are a larger employment size, the employment of workers with a higher level of education and the use of importedinputs. Salvadoran firms provide training to their workers more often than do their counterparts from Guatemala, Honduras and Nicaragua (see Figure 5.4). These differences are drivenby the larger fraction o f Salvadoran firms that provide external training, and inparticular by the larger percentage that use the services of the government's institution in the area o f 204 training - INSAFORP in the case o f El Salvador. Both in El Salvador and in the above mentioned neighboring countries, and with the only exception o f maquilas, a higher proportion o f exporters provides training to their workers (see econometric estimates in the appendix). Larger firms are also more likely to train their workers, and so are firms that already use more educated workers and report the use o f imported inputs. Once again, this suggests that an increase in the degree o f international integration o f Salvadoran firms, through higher imports and exports should -unless it takes place exclusively through the growth inthe maquila sector - lead to a larger demand for training services. 287. Only about one half of the firms that report paying the 1% contribution to INSAFORP do in fact use the services of that institution, either directly or through certified training providers. Compared to neighboring countries, the percentage o f firms that contribute to the official training institution and do infact use its services i s relatively high: that proportion i s about 50% in El Salvador and 36% on average in Guatemala, Honduras and Nicaragua. Among the firms that contribute to INSAFORP but do not provide training to their workers - neither internal nor external -89% indicate that on-the-job training i s sufficient for their particular needs, and 43% argue that they can hire workers with the necessary skills, acquired either through their formal education or by working in other companies. Moreover, 64% o f the surveyed firms consider that the external training services available in El Salvador are adequate to their needs, a proportion that increases to 82% for the companies that use the services o f INSAFORP. On the other hand, for the firms that consider the available services inadequate for their companies - and in fact 84% o f those companies do not use the services o f INSAFORP or its certified providers - the main problem o f the available training programs is that their content i s irrelevant for the companies' specific needs (reason mentionedby 76% o f the firms). Additional problems mentioned by those firms are the limited supply o f training programs (51%), the lack o f qualified instructors (48%) and costs that are considered too high (35%) Figure 5.5: Labor Market Rigidity Indexes from DoingBusinessDatabase ODifkulty of FiringIndex Source: Doing Business Database. 288. Indexes constructed by the World Bank to perform cross-country comparisons of labor market flexibility (Doing Business Database) suggest that El Salvador has relatively rigid regulations regardingthe hiring and firing of workers but more flexible regulations 205 concerning hours of work. The norms that govern the hiring and firing o f workers are more rigidthan inthe rest o f Central America and than inthe average Latin American country, which in turn displays less flexibility than high income OECD economies and developing countries from East Asia and the Pacific (Figure 5.5). Although hours o f work are relatively flexible, part time employment entails mandatory benefits and firing restrictions that are similar to those granted to full-time workers. Fixed-term contracts are also allowed, but only for up to 12 months, and they are restricted to fixed-term tasks. Minimumwages, however, are relatively lower than in Nicaragua and Honduras. Indeed, while they are higher than in those countries in absolute terms (Figure 5.6), after adding mandatory non-wage labor costs the Salvadoran monthly minimumlabor cost represents 11% o f annual per capita GDP, compared to 16% for Nicaragua and 23% for Honduras. InCosta Rica, Panama and Guatemala minimumwages - including non- wage mandatory costs - are similar to those found inEl Salvador: they also represent 11%o f p/c GDP (on average). In the area o f firing regulations, it i s worth noting that the Salvadoran legislation does not consider that redundancy i s a "fair" ground for dismissal - e.g. as in the case o f terminations motivated by economic reasons.lSo Figure 5.6: Minimumwages and minimumtotallabor cost (US%) mminimum wage "in. wage plus mandatory contributions Source: Anzueto et al. (2004) Figure5.7 :LaborRegulationsas constraints to business operations (YOof firms) 28.1 29.2 17 2 El Salvador Nicaragua Honduras Guatemala IDGeneral ConstraintW Constraint if CAFTAI Source: Investment Climate Survey I 5 OS e e the World Bank publication Doing Business in 2005 for data on firing c o s t s in El Salvador, c o m p a r e d to o t h e r d e v e l o p i n g countries. 206 Figure5.8: Share of firms with some presenceof workers' unions(%) 19.7 ElSalvador Guatemala Honduras Nicaragua IEaall firms Wmediumllarge (25 plus emp.) 1 Source: Investment Climate Survey 289. Notwithstanding the rigidities associated with the hiring and firing of workers, a very small fraction of firms (3.9%) rate the government regulations that affect labor contracts as sources of major or very severe problems for their operations. The fraction o f firms concerned with labor regulations was much lower in El Salvador than in neighboring countries (Figure 5.7), even when firms were asked about constraints that would limit their ability to benefit from a Free Trade Agreement with the U.S..One possible explanation i s that compliance with the legislation i s limited, due to low levels o f enforcement o f labor market regulations, an interpretation that would be consistent with the very low levels o f unionization found in Central America and especially in El Salvador. Indeed, as shown in Figure 5.8, only 1.9% o f the firms - 3.3% inthe case o f those with at least 25 employees - reported that some o f the workers were unionized. Moreover, inthe presence ofweak unions wage contracts canundue some o f the distortions incorporated in the labor legislation, for instance by lowering base earnings, so high mandated benefits do not necessarily translate into higher take-home pay. While it i s also possible for firms to use temporary work or outsourcing as mechanisms to minimize their legal obligations towards workers, the ICs data indicates that in El Salvador permanent workers represent on average 89% o f the workforce o f manufacturing firms, a fraction that is similar to that found in neighboring countries (Figure 5.9). Finally, the firms' limited concerns with labor regulations - and their probable ability to avoid full compliance - could be associated with the presence o f excess labor supply, as suggested by the relative ease with which firms seem to able to fill their vacancies, particularly in the case o f more skilled workers (Figure 5.10). Figure5.9: Share of permanentworkers in total employment(YO) 89.0 97 0 88 7 ElSalvador Guatemala Honduras Nicaragua Source: Investment Climate Survey 207 Figure 5.10: Time to fillvacancies of skilled and unskilled workers (weeks) 10unskilled I&1skilled 1 Source: Investment ClimateSurvey POLICY RECOMMENDATIONSEDUCATION LABORON AND SKILLS 290. One of the main challenges faced by Salvadoran policy makers in the area of education i s that of increasing the share of the workforce with secondary levels of schooling. Higher levels o f secondary schooling are crucial not only in terms o f social policy goals - to sustain the positive achievements in the area o f poverty reduction - but also to facilitate the technological upgrading o f local manufacturers, to attract FDI with high technological content, and to benefit from the potential spillovers o f those investmentsto the rest o f the e~onomy.'~' As indicated in the World Bank's poverty Assessment for El Salvador, in order to increase secondary enrollments and completion rates, it will be important to strengthen programs o f school construction and rehabilitation inareas where supply constraints appear to be binding and to develop public-private partnerships to exploit excess capacity inprivate schools. It is worth noting that currently the firms that produce for the domestic or the Central American market are the ones that exhibit a greater demand for more educated workers. However, our results also suggest that an increase in the use o f imported inputs following an increase in the degree o f international integration o f the Salvadoran economy should lead to a positive shift in the demand for skilledworkers, at least inthe manufacturing sector. 291. Even though almost two thirds of the surveyed firms consider that the services available in El Salvador in the area of training are adequate for their needs, only 50% of the firms that pay the mandatory contribution to INSAFORP actually use the services of that public institution. Efforts in this area should be directed at making the content o f the training programs offered by INSAFORP - or other certified providers - consistent with the needs o f the private sector, in order to overcome what the great majority o f the surveyed firms views as the main problem affecting existing training programs - their irrelevancy with respect to the firms necessities in this area. In this respect, our results show that firms involved in imports and exports tend to demand more training services, suggesting that new Free Trade Agreements should increase the demand for training services and create different needs interms o f the content o f the programs offered. Moreover, the supply o f training courses should also be integrated into a wider educational policy, so that vocational education and technical training can serve the goal o f complementing the formal schooling o f Salvadoran workers. Is' For a regional perspective of the challenges in the area of education, and international examples of educational policies followed by several Latin American countries, see D e Ferranti et al. (2003). 208 QUALITY 292. Many components of the national quality system are functioning in El Salvador, while others must be enhanced or developed. To support domestic industries and promote trade, many countries have established a national infrastructure for standards development, accreditation o f testing laboratories and certification organizations, and other standards-related activities. The complexity o f a national quality system i s dependent upon the country's level o f industrialization, the extent o f its dependence on international trade, and the technical requirements of its trading partners. Highly industrialized economies require more complex systems. The greater the dependence on foreign markets, the more the system must comply with internationally recognized criteria. Successful quality systems address all o f the components within a coordinated system'52. InEl Salvador the coordination of activities related to standards development, accreditation, testing, and certification i s the responsibility o f the Consejo Nacional de Ciencia y Tecnologia (CONACYT). As described inFigure 5.11, CONACYT i s an autonomous body under the responsibility o f the Ministerio de Economia (MINEC). It i s directed by a board of directors comprised o f representatives from government, the private sector, professional associations, and academia. So far, CONACYT has been successhl in developing the functions o f standards development and laboratory accreditation, but further efforts are needed in terms o f developing a local certification program, and instituting an independent accreditation authority. 293. CONACYT has a robust standards development program, especially when compared to some of the other countries in the region. CONACYT coordinates the development and adoption o f national technical standards with other government bodies and with the private sector and scientific institutions. By having all of the stakeholders representedinthe standards development process, the resulting standards should be more relevant for Salvadoran industry and government. Figure5.12 provides a regional comparison of the number of national standards developed in each country: with 920 standards developed by CONACYT, El Salvador i s ahead o f all other Central American countries, including Costa Rica and Guatemala, which follow more closely, respectively with 869 and 699 standards. Many o f the Salvadoran standards are based on international standards such as those developed by the International Organization for Standardization (ISO), the Codex Alimentarius Commission (CAC), and the International Electrotechnical Commission (IEC). Approximately 10% o f the Salvadoran standards are mandatory, while the rest are recommended standards. The subjects covered by the mandatory ones reflect the importance o f the food processing, beverage, petroleum, chemical, and textile industries in El Salvador Recommended standards are optional but become mandatory when referenced in any procurement o f goods or services by the government o f El Salvador, or inother commercial contracts. 15* See Annex 1 for a more detailed description of various components of a typical national quality system. 209 , Figure5.11: Currentquality system inElSalvador (plannedcertificationin dottedlines) 11 MINISTERIO DEECONOM~A 11 Consejo Nacional de Ciencia y Tecnologia I Laboratorio de Departamento de Metrologia Normalizacibn, Legal Metrologiay Certificacih de la Calidad r-----------I Standards Information Sprvire - I Source: Research by Thompson Consulting, Inc. 294. CONACYT is also responsible for accreditationof testing laboratories and it has made much more progress in this area than other countries in the region. Accreditation is the procedure by which an authoritative body gives formal recognition that a person or organization is competent to carry out specific tasks. Testing laboratories and certification organizations are most often accredited. An increasing number o f regulators responsible for trade policy use accreditation as a means for parties to demonstrate their compliance with trade or safety regulations. CONACYT i s a member o f the International Laboratory Accreditation Cooperation (ILAC) and the Interamerican Accreditation Cooperation (IAAC). Participation in these international and regional organizations provides opportunities for CONACYT to increase its knowledge about accreditation and establish contacts for hture negotiation o f mutual recognition agreements, which will benefit Salvadoran exporters by having the results o f tests done by accredited laboratories inEl Salvador accepted in other markets. Twelve laboratories (9 private, 2 public, and 1 academic) in El Salvador have been accredited by CONACYT. The accredited laboratories provide testing services for food products, potable water, cement, concrete, and drugs. Figure 5.13 indicates that El Salvador has a relatively large number o f accredited labs compared to the other countries in the region. This facilitates the testing o f products locally, reducing the time and cost for firms to have their products tested. 210 Figure5.12: Number of nationalstandards in CentralAmerica, by country 920 437 Belize Costa Rica El Salvador Guatemala Honduras Nicaragua Panama Source: Researchby Thompson Consulting, Inc. Figure5.13: Number of accreditedlaboratoriesin CentralAmerica, by country 12 Belize Costa Rica ElSalvador Guatemala Honduras Nicaragua Panama Source: Researchby Thompson Consulting, Inc. 295. Even though by law CONACYT is also responsible for the certification of products and services in El Salvador, including certification to I S 0 9001:2000, the certification program i s not yet functioning. Certification i s the practice o f providing assurance that a product, service, system, process, or material conforms to specific requirements. Conformity usually i s validated by a neutral third party, independent o f both supplier and purchaser, and i s used most often where safety, health, or environmental impact are an issue. Management system certification applies to what an organization does to manage its processes. Management system standards are models to follow in setting up and operating management systems. Quality management system standards include those developed by the International Organization for Standardization (ISO) - I S 0 9001:2000, and quality practices used inthe agriculture and seafood industries such as the Hazard Analysis Critical Control Point (HACCP) system, Good Agricultural Practice (GAP) guidelines, and Good Manufacturing Practice (GMP) requirements. Environmental management system standards, such as I S 0 14001,specify the requirements for a firm to manage processes influencing the impact of the firm's activities on the environment. Certificationo f a firm under I S 0 9001:2000, HACCP, GAP, GMP, and I S 0 14001 standards i s a mark o f approval recognized worldwide and may be required as a condition o f sale. Thus, the adoption o f internationally accepted quality standards facilitates entry into foreign markets, both for Salvadoran firms that seek to supply parts or components to foreign manufacturers, and for producers o f finished products searching foreign distribution networks. Because El Salvador's own certification program is not yet functioning CONACYT has partnered, in the interim, with 211 the national standards bodies of S ain (AENOR) and Colombia (INCONTEC) to provide certificationservices inEl Salvador.1p3 296. Accordingto the ICs, about 5.2% of Salvadoranfirms haveI S 0 9001:2000 certified quality standards and 1.5% are I S 0 14000 certified. I S 0 certification is much more frequent among larger firms. According to the ICs, the proportion of firms that have I S 0 9001:2000 certification i s slightly larger in El Salvador than in Guatemala, Honduras and Nicaragua. It is, however, much lower than what has been found using ICs surveys in countries such as Ecuador, Brazil, Pakistan, Malaysia (31% of I S 0 certified firms) and China (50%), which signals the importance o f the challenges ahead (Figure 5.14). Within El Salvador, as in other countries, the incidence o f quality certified firms increases greatly with size, from 1% among microenterprises in the case of I S 0 9001:2000, to about 3.5% for small and medium firms, and 12.5% for large f i r m s (Figure 5.15). Figure 5.14: Firms with I S 0 9001:ZOOO quality certification,by country (YO) 50.4 3.2 3.4 5.0 5.2 Source: Investment Climate Survey Figure 5.15: Firms with I S 0 quality certification,by size (YO) n n 9.4 4.7 El Salvador I S 0 9000 Gua.-Hon.-Nic. ISO- - - El Salvador - I S 0 Gu.-Hon.-Nic. ISO. - 9000 I4000 14000 'nmicro .small Omedium Olarae I Source: Investment Climate Survey 297. Exporters are more likely to have internationally recognized quality certifications and to employ a larger share of their workforce in quality control activities. These effects are stronger for exporters whose main clients are located outside Central America. As seen in Figure 5.16, both inEl Salvador and inthe other Central American countries for which ICs have It is interesting to note that a comprehensivereform of the El Salvador customs system, supported by projects such as the IDBProgramto StrengthenCustomSystems, has resultedinEl Salvador havingthe first customs system inLatinAmericato meetthe I S 0 qualitystandard. 212 been performed, the fraction o f firms with I S 0 9000 and I S 0 14000 quality certifications i s smaller for non-exporters. Among exporters, these simple tabulations show that the incidence o f I S 0 certification is larger for firms whose main foreign clients are located outside Central America, and it i s also larger for maquilas than for non-maquilas. However, econometric estimates reported inthe appendix show that the differences inthe frequency o f I S 0 certification between maquilas and other exporters tend to disappear when the comparisons are made controlling for other firm characteristics, such as sector, location, firm size and average worker schooling. When the estimates are obtained using data from the four countries pooled together, the results confirm that firms that export to countries located outside the Central American ' region have a greater probability o f being I S 0 certified, and employ a larger share o f their workforce in activities related to quality control. Moreover, maquilas also have a larger share o f their staffworking inthe area of quality control. Finally, the results also show a complementarity between the incidence o f I S 0 9000 quality certification and the use o f imported inputs, the provision o f training to workers and the employment o f workers with higher levels o f schooling. Figure 5.16: Firmswith I S 0 quality certification, by exportingstatus (YO) 12.5 10.6 El Salvador I S 0 9000 Gua.-Hon.-Nic. ISO- - El Salvador I S 0 - Gua.-Hon.-Nic. ISO- ~ ~ 9000 14000 14000 I Onon-exDorter main export to CA (non-maw.) main export out of CA (non-maw) maquila 1 Source: Investment ClimateSurvey POLICY RECOMMENDATIONS ON QUALITY 298. ElSalvador should develop its own certification program to promotethe adoption of internationally recognized quality standards, and respond to the demand of Salvadoran exporters. The results o f the survey show that compared to other firms o f similar sectors, locations and size, exporters are more likely to invest in quality certifications and quality standards. These effect i s stronger for firms that export mainly to clients located outside Central America, so that investments in the area o f quality certification and control should tend to increase in the context o f a Free Trade Agreement with the U.S. In order to respond to that increased demand, the country needs to strengthen its national quality system, by creating local certification bodies, and establishing a separate accreditation authority - the same institution cannot act as both accreditation authority and certification body. Inaddition, promotional efforts should be made to create awareness inthe manufacturing sector o f the benefits o f implementing appropriate quality and environmental management systems standards, the returns on investments in certification to these standards, and on how these measures can increase the firms' competitiveness in the world market. Finally, El Salvador should firther stimulate quality certification through public procurement, as well as by means o f specific programs directed at 213 promoting quality standards among small and medium firms. It i s worth noting that most o f these measures are mentioned in the National Action Plan for Trade Capacity Building for El Salvador.154 TECHNOLOGY155 299. While El Salvador i s evidently far from the technological frontier, the ICs results demonstrate that Salvadoran manufacturing firms are aware of the need to upgrade and renovate their product offerings in order to remain competitive. According to the ICs, inthe last three years 82percent o f Salvadoran firms claim to have upgraded their product line, 62% developed a new product line and 51% introduced new technology that substantially changed the production process. Such dynamic response i s similar to that o f firms located in other Latin and Central American countries - see Figure 5.17. Although these results suggest a considerable amount o f innovation among Salvadoran firms, the scope o f the product or process innovations introduced has not beenbroad enough to foster the firms' successful participation ininternational markets with a diversifiedportfolio o f highvalue addedproducts. Figure5.17: Innovatingfirms by type of innovation and by country I (YOof firms by innovationtype) Ways in which flrms Introduce innovations -~ _ _.. ... . . . . -.. .. . . . . .. Ei Sal,rdor Bollvii Honduras 72 Nicaragua 85 Eouador 84 Brazil 95 0 10 20 30 40 60 EO 70 80 90 100 Uppradrsproduotline INewproductilne aProcesschange Source: Investment Climate Survey 300. Self-reported innovations are much more frequent among larger firms with export activities. Econometric estimates also reveal that rates of innovation are lower among maquilas but larger among firms that have a more educated workforce, provide training to their workers and make use of imported inputs. The percentage o f firms that report process innovations during the two years preceding the survey almost doubles - from 34% to 65% - '54The National Action Planfor Trade CapacityBuilding: Meeting the Challengeof Globalization, datedJuly 2003, was preparedunder the direction of the Trade Policy Departmentwith the Ministry of Economyof El Salvador with technical support providedby the OAS-IDB-ECLAC Tripartite Committee. For the preparationofthe Action Plana broad consultationwas organized by the Government of El Salvador with public entities having responsibilitiesin 'the area of trade, as well as with representativesof the private sector, academia, andresearchinstitutions. '55For a detailed analysis of the structure and main components of El Salvador's National Innovation System see World Bank (2004), El Salvador Country EconomicMemorandum. 214 when one moves from microenterprises to large Salvadoran firms (Figure 5.18). The fraction of firms that report having introduced new product lines or improved their existing ones exhibits a similar pattern, with much lower rates o f innovation for smaller firms. However, for a given firm size, the rates o f self-reported innovation are significantly larger among exporters. These differences are also apparent in Guatemala, Honduras and Nicaragua, but they are much more pronounced in El Salvador (Figure 5.19). Moreover, as shown in the appendix, at least in the case o f El Salvador and after controlling for other firm characteristics, maquilas are an exception to the above pattern, as they are less likely to report innovations, not only in comparison with other exporters but also with respect to firms oriented exclusively towards the domestic market. On the other hand, the econometric evidence suggest a clear complementarity between innovation on one hand, and workers' skills - measured either by the average level o f education o f by the presence o f training activities - and the use o f importedinputson the other. Figure 5.18: M a i n source of Technological Innovations, by size (YO) 90 89 &G%O $*a *e&'+-+ ,& OUpgrade Line &lNew Line ONew Process Source: Investment Climate Survey 301. The main form of technology acquisition at the plant level is,thepurchase of new machineryand equipment. As in other countries of the region,foreigntechnology licensing, which is a common mean of innovationin East Asian economies, is mentioned by a small fraction of the surveyed firms. The purchase of machinery and equipment is indicated as the most important way o f introducing technological innovations by 64% o f the firms with self- reported innovations. It is one o f the three most important channels o f technology acquisition for 79% o f the firms. The two other channels quoted most frequently are the hiring o f new key personnel, by 48% o f the firms, and in-house technology development, which i s mentioned by 38% o f the firms (Figure 5.20). Although the primacy o f embodied technological change i s not unique to El Salvador, its importance appears to be even larger than in most other countries for which we have comparable data. The percentage o f firms that mention foreign technology licensing as an important source o f innovations i s only 6%, which is slightly higher than in other countries o f the region, but much lower than in China, where it i s mentioned by 24% o f the firms. China i s also the country where internal research and development activities are mentioned most frequently (by 79% o f the firms) as one o f the three most important sources o f technological innovations. Across firms o f different sizes, the acquisition o f technology embodied in new machinery and equipment and the licensing o f foreign technologies are more frequent among large firms than among MSMEs. Moreover, micro and small firms give less importance to the hiring o f new personnel as a channel for gaining access to new technologies (Figure 5.21). 215 Figure5.19: Self-reportedinnovationsby exporter status (% of firms) 88 El Salvador Guatemala Honduras Nicaragua 10Non-exporters Exporters Source: Investment Climate Survey Figure 5.20: Main sources of technologicalinnovations-InternationalComparison (YO) 79 OEmbodied in machinery BHiring key personnel HDeveloped in-house EILicensing Source: Investment Climate Survey 302. The incidence of technology licensing is lower than in neighboringGuatemala and Honduras,butit is muchhigheramongmaquilas and foreignownedcompanies. Inresponse to a direct question, 13% o f Salvadoran firms declare to use technology licensed from a foreign- owned company versus 9% in Nicaragua, 15% in Honduras and 19% in Guatemala. Higher licensing levels can be observed among large firms, maquilas, firms operating in the textile sector, and firms with foreign capital ownership (Figure 5.22). Investments in licensing foreign technologies have been historically low in El Salvador, although some advances were made in the late 1 9 9 0 ~ .Licensing activity as measured by royalty payments as a percentage of GDP, ' ~ ~ amounts to 0.19%, a performance slightly better than the LAC average (0.15%) but less than a third of that found among the so-called East Asian "tigers", which spend 0.47% of GDP in royalty payments.157 El Salvador: Country Innovation Brief, World Bank, 2003. 157DeFerranti et al. (2003). 216 Figure 5.21: Main source of Technological Innovations, by size (YO) 88 74 73 74 n @ e-$8 \*R 1OEmbodied in machinery MHirIngkey personnel UDevelopedin-house ElLicensing Source: InvestmentClimate Survey 303. In order to gain access to technologicalinnovations, Salvadoran firms rarely use any form of cooperation with clients, suppliers, industry associations or R&D institutions. Firms' weak integration inthe productive chain, which is a commonpattern inCentral and Latin American countries (see Figure 5.22), prevent firms (especially MSMEs), from benefiting from economies o f scale, knowledge spillovers and joint learning and innovation. In the case o f El Salvador it i s interesting to note that only 9% o f surveyed firms declare to have developed innovations by collaborating with suppliers, half the proportion encountered in Honduras, Guatemala or Nicaragua. Not surprisingly, technological linkages among firms and universities and other research institutions are also missing in El Salvador. In contrast, Salvadoran firms seem to use consultancy services more than their comparators, as a channel for incorporating new technologies (Figure 5.23). Figure 5.22: Technology licensing in El Salvador, by firm characteristics (YOof firms) I 35 - 30 - 28 25 - 20 - Source: InvestmentClimate Survey 304. Despite the importance given to the acquisition of new machinery and equipment as a source of technological innovations, imports of capital goods have been historically low in ElSalvador, The above finding confirms that inEl Salvador, as inmost low and middleincome countries, technology transfer and innovation mainly take place through the incorporation o f new capital equipment into the production process. In most developing countries, demand for new machinery i s generally met with imports. Thus, imported capital goods are a major source of technology transfer and productivity gains, which may occur either by introducing technologically improved inputs into the production process or by learning the know-how that 217 they embody. Infact, one o f the main potential advantages associated with trade liberalization is that can provide increased access to state-of-the-art technologies embodied in imported capital goods. In El Salvador total import penetration i s comparable to the Central American average and higher than the overall L A C average. Imports o f capital goods, however, represent only 16% o f total imports, which i s low even by Latin American standards: that share i s about 25% in Chile and Peru, and more than 40% inMexico and Malaysia (Figure 5.24). This suggests that El Salvador has not taken enough advantage o f its increased level o f integration in the global economy, at least in terms o f intensifying its imports o f goods that incorporate technologies developed abroad. Figure 5.23: Less frequent forms of technology acquisition. by country (% of firms) Honduras Nicaragua Guatemala El Salvador aNew equipment Equipment suppliers =Trade FairdStudy Tours 0 In cooperation with clients =Transferred from parent compani I Consultants Businesslindustry assoc clUniversities, public institutions Source: InvestmentClimate Survey 305. Even though Salvadoranfirms claim to innovate by acquiring new equipment, the technological content of such equipment appears low. While approximately 52% o f Salvadoran firms claim to innovate by introducing new equipment - 66.7% report innovations and 78% o f those mention the purchase o f new equipment as one o f the three main channels for such innovations - a substantially smaller fraction o f firms (24%) uses computer-controlled production equipment.This proportion i s close to that o f Guatemala and Honduras -respectively 24% and 21%, but much larger than that found in Nicaragua (11%). Similarly, on average 11% o f the workers o f the surveyed Salvadoran manufacturing firms use computers on their jobs: slightly less than in Guatemala - where 11-8% o f workers use computers - but above the 7.8% found in Nicaragua and Honduras. Most manufacturing sectors show relatively low levels o f automation, with the main exceptions given by the chemical, rubber and plastic and textile industries, where the proportion o f firms with computer-controlled production equipment reaches respectively 40%, 50% and 60%. Similarly, in the chemical industry 25% o f the workers reportedly use computers. 218 Figure 5.24: Imports of capital goods, by country (YOof totalimports) AA48 48 Source: D e Ferranti et. al. (2003). 306. The use of computers and computer-controlled production equipment is much higher among larger and exportingfirms that employ a larger fraction of skilled workers. Even after controlling for the firms' sector and location, our econometric estimates (see appendix) indicate that larger firms and firms that report exports are much better equipped than equipment. Moreover, we also find a strong positive correlation - evenafter controlling for other other firms, at least in terms o f the use of computers and computer-controlled production characteristics - between the use of those types o f modern equipment and the employment o f a more educated workforce, the provision o f worker training, and the use o f imported inputs. In addition, at least in the samples that incorporate firms from all four countries, we find that maquilas are less likely to use modern equipment than are other exporter firms operating in the same sectors and locations. Finally, firms located in L a Libertad are the most likely not only to report innovations but also to operate with more modem equipment. Figure5.25: R&D Expenditures (YOof GDP), by country (average 1980-1995) Source: D e Ferranti et. al. (2003). Figure5.26: :R&D Expendituresper worker (US$),by country (average 1980-1995) Source: Investment Climate Survey 219 307. El Salvador's expenditures in the development or adaptation of new technologies have also been small relative to its population and GDP. Even for countries that rely heavily on the transfer o f technology generated abroad, domestic expenditures in research and development (R&D) o f new products and processes are crucial for adapting imported technologies to the specific characteristics o f domestic industries and consumers. Considering both public and private expenditures, the resources devoted to R&D in El Salvador represent about 0.19% o f GDP, which i s much less than what i s observed in most Latin American countries - e.g. Costa Rica spends 0.25% o f GDP on R&D, and Chile and Brazil are close to spending 0.5% o f GDP (Figure 5.25). A potentially positive aspect o f the Salvadoran case i s that only 52% o f the R&D expenditures are performed by the public sector, a proportion that i s lower than in most countries o f the region - for which the public share i s 77% on average - and i s closer to the public share o f total R&D found in the OECD and in East Asia - about 46%. However, in absolute terms, considering both private and public resources, El Salvador spends about $11 dollars per worker inR&D, which i s about one third o f the regional average, and more than 20 times less than the expenditures o f the average East Asian country (Figure 5.26).15* Figure5.27: Patentsin the U.S. during 1996-2000, by country (patentsper millionpopulationper year) Source: Investment Climate Survey 308. About 17% of the firms surveyed in El Salvador claim to invest in research and development (R&D) activities and less than 3% of the firms' staff is allocated to that area. The incidence of R&D i s higher among exporters and smaller firms. The proportion o f Salvadoran firms that report R&D activities is smaller than in Guatemala (35%) but higher than in Honduras and Nicaragua (about 12%). However, within each of those countries the share of R&D staff in total employment is larger for smaller firms, as well as for firms that provide training to their workers, export some o f their output, use imported inputs and produce chemical products (see appendix). Among exporters, those that have their main clients outside o f Central America and are not organized as maquilas report the highest shares o f R&D staff in total employment. 309. Given the relatively small amount invested in technology development, it does not surprise to find a relatively small number of US. patents registered by Salvadoran inventors. Over the 1996-2000 period, there were only 0.1 patents registered per year by Salvadorans inventors for every million population. While most Latin American countries 15*These low rates of investment in R&D are not due to low returns. Lederman and Maloney (2003) estimate that the economic returns to R&D in countries of El Salvador's level o f income are high at around 70%. Their calculations suggest that El Salvador should be investing between 4 and 10 times more in R&D than it did in the 1990s. 220 exhibit rates o f innovative activity that are much lower than those o f the East Asian economies - Korea has 81 patents registered in the U.S. for every million population, compared to 2.1 for Costa Rica - El Salvador's performance i s among the lowest o f the region (Figure 5.27). This, however, i s not surprising given that El Salvador allocates relatively less resources to R&D activities than do other LatinAmerican countries with higher levels o f patenting activity (Figures 5.25 and 5.26). Moreover, statistical analysis show that El Salvador's patenting activity i s lower than that o f other countries with the same level o f income, the same labor force size, and the same value o f merchandise exports to the UnitedStates since the 1 9 6 0 ~ . ' ~ ~ POLICYRECOMMENDATIONSTECHNOLOGY ON 310. The government of El Salvador should work to improve the conditions for Salvadoran firms to benefit fully from the existing world stock of knowledge, thus increasingthe absorption and diffusion of foreigntechnologies in the Salvadoran economy. Measures to promote this goal include those aimed at facilitating imports o f capital goods, by reviewing bureaucratic requirements for imports and improving the access to financing mechanisms for the acquisition o f machinery and equipments. Inthis respect, specific measures could be designed to facilitate the purchase o f capital goods - including the acquisition o f computers and related goods - by micro, small and medium firms. Licensing could also be positively affected by additional investments and efforts in the enforcement o f intellectual property rights. Moreover, to fully exploit the potential benefits offered by foreign direct investment, the Government o f El Salvador should focus on attracting selected value added FDI which could improve the productivity o f the local industries that present the highest growth potential160. Additionally, specific efforts should be made to strengthen collaboration between existing foreign and local companies by creating specific linkages through a combination o f different activities (e.g., providing market or business information, matchmaking services, managerial and technical assistance to foreign and local firms operating in the same production chains). These initiatives should be coupled with measures aimed at strengthening the specialized investment promotion agency (PROESA). 311. To encourage technology diffusion - especially among MSMEs and firms operating in industries in which El Salvador has a comparative advantage161,the Government should support the development of a modern innovationinfrastructure,includingcenters offering information,technical assistance, trainingand quality services. Lack o f informationor access to certain technological services is typically an important obstacle to technology acquisition and 15'World Bank(2003), CountryInnovationBrieffor El Salvador. I6OThese actions should take into account those sectors identified by the Competitive Intelligence Unit of the Ministry of Economy, whose responsibility is to evaluate the competitiveness of various sectors, and to develop international benchmarking. In fact, alignment of overall economic policies and FDI policies is an essential requirementto ensure successful implementationof the Government's strategy. 16'Accordingto World Bank's Country InnovationBrieffor El Salvador, the country has an innovativecomparative advantage mainly in textile mill products, rubber and other plastic products, engines and turbines and agricultural chemicals, Thus, targeted public policies designed to stimulate R&D in these sectors could be a fruitful avenue for future policy. 221 diffusion, especially for micro, small and medium firms.'62 Inthis sense, technology extension services mightbe a useful tool. Technology services may offer different emphasis interms o f the type and purpose o f the service provided, as illustrated by Table 5.1. Bundling services -such as quality, training and technology extensions- may bring economies o f scale and scope in provision. Currently, some o f these services, especially training and consulting, are being offered by Salvadoran public institutions, but rivate firms perceive them as inadequate (too general and o f low quality) to meet their needs.194 312. The Government should promote an increase in the current level of public and private investments in research and development activities. To that end, El Salvador should gradually increase public expenditures on R&D for activities with commercial return, as well as guarantee their stability. Moreover, the Government could consider introducing tax rebate schemes and competitively allocated matching grants to both domestic and foreign firms investing in industrial R&D. The allocation o f R&D resources, both profit and not-for-profit, should favour experimental development if a more output-driven innovation system i s to be achieved. The Government should provide incentives to strengthen cooperation inR&D between the private and public sectors, and encourage more direct involvement of private firms in this area. To ensure that El Salvador better exploits its R&D potential, the Government should provide specific incentives to support information sharing, joint research and consulting ~ 162The firms surveyed by FUSADES in 2003 report that the main obstacles to develop innovation activities are the lack of technicalinformation (56%), the lack ofprofessionaland technical staff (49.5%) and the lack of skilled labor (47%). The mainbarriersrelate to the difficulties to access and absorb knowledgeandpoint to the needto adjust the educationaland training programs to the demands of the productive sector. At the same time, a large proportion of firms (80%) declare to spendnot much or nothing intraining. Inparticular, 56% of small firms do not invest at all in training. 163The creationof private-public business development and technology centers which have the specific purpose of supporting firms --especially MSMEs or firms operating in clusters-- is an option that has developed recently in countries such as Peru, Colombia, Mexico and are about to start inEcuador or Honduras. These centers, often called CITES(Centros de Znnovacidn Tecnoldgica),provide specific training and advisoryservices (e.g., on equipment use, quality management and improvement of production processes) as well as access to and information on new technologies (e.g., for raw material quality control, design, manufacturing, production and testing techniques, and packaging and labeling). See a detailed description of the objectives and structure of these Centers in the World Bank's Honduras Trade and Productivity Loan (Annex 2) and in Fuster J. (2004), background paper for the HondurasTrade Project. Accordingto results inthe surveyby FUSADES, 2003. 222 activities between firms and universities.Inaddition, research funds should be allocated through transparent, competitive bidding processes that encourage participation by private institutions and firms. Finally, the Government could support a program o f technological intermediaries that would facilitate development o f links betweenMSMEs and technology service providers. 313. The Governmentshould promoteoutput-based and market-oriented R&D in public research institutes and universities. Possible measures include those directed at gradually reducing the access o f those institutions to earmarked funds'65while at the same time expanding their autonomy to look for new sources o f funding, particularly through different partnerships with the private sector. In addition, the government should improve the existing economic incentives for technology transfer from public research institutes and universities to the private sector, by allowing researchers and their institutions to benefit from their discoveries through flexible assignment rules o fproperty rights. 314. The Government should take actions to improve the governance of the Salvadoran national innovation system (NIS) by building the capacity to formulate, implement, and evaluate coherent, incentive-compatible and cost-effective innovation and technology policies. Budget allocation for R&D, innovation program design and evaluation, as well as capacity and institution building for the promotion o f science and technology, have not been a main priority for the Government o f El Salvador over the past decades, CONACYT, the main authority in charge o f the formulation, design and implementation o f R&D strategies and programs and the promotion o f collaboration among various agents o f the National System o f Innovation, has had a very limited impact inthe 12 years since its creation. To date the Council i s not highly active, partly due to weak organizational infrastructure and very small budget, o f barely US$0.5 million. Measures to improve the governance o f El Salvador's' N I S include: the definition o f a coherent national technology and innovation policy, establishing a set o f programs designed to achieve measurable objectives, as well as its main funds; the strengthening o f the policy-making, coordination and evaluation capacities o f CONACYT; an increase in private sector participation in CONACYT (at the Commission level); and finally an improvement inthe national R&D statistics, in order to increase the transparency and accountability o f technology and innovation policies. 16' CONACYT estimatedthat public R&D expenditures in 1998 were about US$6.2million and representeda 0.3% of the budget. This amount is much lower than the total public resources dedicated to science and technology activities (which cover payroll and operating expenses of universities and public research and technology centers) amountingto US$51million and 2.7% of the budget. According to FUSADES, these expenditures have increased considerably inthe last few years reachingUS$72million in2002, which representsa2.9% of the budget. 223 ANNEX5.1: ECONOMETRICESTIMATES We estimated a series of econometric models aimed at characterizingthe firms that have a higher demand for skilled labor, and invest more in quality and technology. Our empirical estimated equation is o f the following form: where Techi represents either the employment share o f skilled labor, dummies for firms that provide training to their workers or have I S 0 quality certification, the employment share o f staff working in quality control or research and development activities, the employment share o f workers that use a computer on the job, or a dummy variable for firms that report having computer-controlled production equipment. As explanatory variables we include employment size (Sizei), a dummy variable for maquilus as well as sector (Sectori) and province (Provi) dummies inall specifications. Inaddition, Skilli is a vector o f variables that measure the level of education o f the workforce and whether training i s provided to workers, and Tradei i s a vector o f variables measuring the firms participation in international trade: whether they are exporters, whether their exports are mainly to other Central American countries, and whether the firms use any imported inputs. The results, with samples covering either just Salvadoran firms, or also firms from Nicaragua, Honduras and Guatemala are reported intables A5.1, A5.2 and A5.3. 224 Table A5.1: Correlates of the demand for skilled labor and provision of training to workers (plant level data from ElSalvador, Guatemala, Honduras and Nicaragua) Non-prod. Non-prod. Avg. Avg. Training Training Empl. Share Empl. Share Schooling Schooling (1og)dummy (b) dummy (b) (a) (a) (log) (a) (4 Employment (log) -0.023*** -0.014*** -0,009 -0.016"" 0.135*** 0.128*** (0.005) (0,001) (0.450) (0.017) (0.000) (0.000) Average Schooling 0.047*** 0.040*** Exporterdummy -0.019 -0.028* -0.034 -0.042* 0.221** 0.089* (0.535) (0.074) (0.466) (0.078) (0.013) (0.056) Export.to CentralAmerica 0.054* 0.030* 0.099** 0.041* -0.120 -0.004 (0.097) (0.065) (0.035) (0.097) (0.235) (0.943) Use of importedinputs 0.012 0.025*** 0.072** 0.091*** 0.084 0.097*** (0.531) (0.008) (0.021) (0.000) (0.149) (0.001) Maquila Dummy -0.016 -0.061*** 0.044 -0.009 -0.219* -0.129" * (0.662) (OIO0O) (0.444) (0.769) (0.089) (0.037) Chemicals 0.233*** 0.228*** 0.213*** 0.189*** 0.036 0.092" (0.000) (0,000) (0.000) (0.000) (0.701) (0.097) Rubber& Plastics 0.124*** 0.082*** 0.037 0.086*** 0.209* 0.138* (0.003) (0.000) (0.487) (0.005) (0.055) (0.051) Mineral Prod. 0.045" 0.050*** -0.036 -0.038 -0.104 -0.054 (0.093) (0,001) (0.508) (0.179) (0.257) (0.268) Textiles 0.054 0.021 0.067 0.064* 0.114 -0.004 (0.212) (0.263) (0.356) (0.068) (0.506) (0.954) Metal Prod. 0.074*** 0.057*** 0.108** 0.117*** 0.014 -0.039 (0.001) (0.000) (0.013) (0.000) (0.871) (0.431) Food& Bev. 0.180*** 0.123*** 0.062* 0.023 0.121* 0.137""" (0.000) (0.000) (0.087) (0.272) (0.093) (0,001) Ahuachapan 0.069* 0.067* -0.056 -0.041 0.044 0.047 (0.061) (0.065) (0.334) (0.475) (0.766) (0.748) La Libertad 0.059** 0.059"" -0.046 -0.038 0.024 0.030 (0.023) (0.018) (0.171) (0.250) (0.751) (0.691) Other 0.018 0.012 -0.220*** -0.235*** 0.177* 0.171* (0.542) (0.680) (0,005) (0.002) (0.089) (0.083) San Miguel 0.018 0.012 -0.319*** -0.330*** 0.218** 0.158 (0.660) (0.767) (0.000) (0.000) (0.037) (0.121) Santa Ana 0.058* 0.051 -0.160*** -0.157*** 0.123 0.107 (0.081) (0.126) (0.002) (0.002) (0.165) (0.205) Observations 453 1695 462 1741 462 1741 Probability values inparentheses. significant at 10%; * ** significant at 5%; *** significant at 1%. (a) Ordinary least squares estimateswith robust standarderrors. (b) Probit estimates with robuststandard errors (marginal effects reported). Provincedummies are also includedfor other countries when applicable. 225 Table A5.2: Correlates of quality certification and quality control (plant level data from El Salvador, Guatemala, Honduras and Nicaragua) I S 0 certif. (b) I S 0 certif. (b) Quality Staff Quality Staff Share (c) Share (c) Employment (log) 0.015** 0.009*** -0.026*** -0.030*** (0.024) (0.003) (0.000) (0.000) Average Schooling 0.000 0.003* 0.002 0.002** (0.900) (0,090) (0.257) (0.012) Training dummy 0.016 0.011 0.004 0.004 (0.289) (0,142) (0.594) (0.235) Exporterdummy 0.042* 0.023** 0.017" 0.014** (0.060) (0.048) (0,099) (0,019) Export.to CentralAmerica -0.007 -0.015* -0.013 -0.010 (0.726) (0.087) (0.256) (0.103) Use of importedinputs -0.008 0.017** -0.002 -0.003 (0.632) (0.021) (0.766) (0.405) Maquila Dummy -0.010 -0,001 0.017 0.029*** (0.612) (0.924) (0.239) (0.000) Chemicals 0.009 0,009 -0.013 -0.011 (0.759) (0,546) (0.282) (0.154) Rubber & Plastics 0.047 -0.003 -0.028** -0.023*** (0.193) (0.829) (0.040) (0,009) MineralProd. 0.017 0.019 -0.006 -0.023*** (0.625) (0.205) (0.555) (0,000) Textiles -0.016 -0.007 -0.002 (0.197) (0.717) (0.798) Metal Prod. 0.006 0.006 -0.002 -0.003 (0.816) (0.648) (0.814) (0.664) Food& Bev. 0.004 0.002 -0.008 -0.011** (0.847) (0.879) (0.334) (0.032) Ahuachapan -0.015 -0.013 (0.363) (0.463) La Libertad -0.003 0.002 0.010 0.011 (0.851) (0.895) (0.240) (0.209) Other -0.025 -0.016 -0.004 -0.004 (0.155) (0.186) (0.769) (0.744) San Miguel 0.001 0.007 (0.936) (0.603) Santa Ana -0.006 -0.005 (0.628) (0.707) Observations 442 1665 Probabilityvalues inparentheses. significantat 10%; *462 1741 ** significant at 5%; *** significantat 1%. (b) Probit estimateswith robust standard errors (marginal effects reported). (c) Tobit estimates. Province dummies are also includedfor other countries when applicable. 226 Table A5.3: Correlatesof innovation,R&Demployment and use of computers and computer-controlled productionequipment (plant leveldata from ElSalvador,Guatemala,Hondurasand Nicaragua) Self- Self- Computer Computer Comp. Comp. R&D Staff R&D Staff reported reported use use contr. equip. contr.equip. Share Share Innov.(b) Innov. (b) (%empl.)(c) (%empl.) (c)(b) (b) (%empl.)(c) (%empl.)(c) Employment (log) 0.125*** 0.058*** 0.020** 0.031*** 0.124*** 0.083*** -0.015*** -0.016*** (0.000) (0.000) (0.018) (0.000) (0.000) (0,000) (0.000) (0.000) Average Schooling 0.032*** 0.027*** 0.019*** 0.024*** 0.031*** 0.020*** 0.002 0.005*** (0.002) (0.000) (0.000) (0.000) (0.000) (0.000) (0.272) (0,000) Training dummy 0.086* 0.148*** 0.011 0.029*** 0.066 0.062*** 0.019** 0.030*** (0.083) (0.000) (0.539) (0.003) (0.103) (0.001) (0.027) (0.000) Exporter dummy 0.138" 0.066 0.052* 0.044*** 0.095* 0.054"" 0.032** 0.023*** (0.097) (0.106) (0.058) (0.003) (0.091) (0.043) (0.013) (0.005) Export. to Centr. America-0.133 -0.022 -0.031 -0.020 -0.025 -0.017 -0.022 -0.019** (0.194) (0.649) (0.289) (0.213) (0.644) (0.519) (0,100) (0.031) Use o f imported inputs 0.106** 0.078*** 0.050** 0.044*** 0.039 0.077*** 0.023** 0.019*** (0.046) (0.003) (0.011) (0,000) (0.344) (0,000) (0.012) (0.001) Maquila Dummy -0.372*** -0.077 -0.002 -0.045"" -0.072 -0.043 -0.020 -0.016 (0.007) (0.200) (0.950) (0.016) (0.238) (0.114) (0.267) (0.138) Chemicals 0.060 0.029 0.159""" 0.156*** 0.063 0.022 0.033** 0.034""" (0.521) (0.549) (0.000) (0,000) (0.385) (0.524) (0.023) (0.001) Rubber & Plastics -0.008 0.062 0.099*** 0.107*** 0.274*** 0.156*** -0.016 0.008 (0.938) (0.3 18) (0.006) (0.000) (0.008) (0.002) (0.375) (0.528) Mineral Prod. -0.028 0.032 -0.005 0.051*** 0.102 -0.001 -0.021 0.003 (0.728) (0.415) (0.890) (0.003) (0.216) (0.985) (0.173) (0.767) Textiles -0.063 -0.037 0.005 0.064*** 0.514*** 0.242*** 0.005 -0.009 (0.640) (0.558) (0.922) (0.005) (0.001) (0.000) (0.815) (0.523) Metal Prod. -0.032 0.005 0.066** 0.079*** 0.040 0.008 0,010 0,010 (0.677) (0.903) (0.016) (0.000) (0.546) (0.812) (0.434) (0.291) Food& Bev. 0.109 0.105*** 0.090*** 0.083*** 0.029 0.029 0.016 0.019** (0.105) (0.002) (0,000) (0.000) (0.617) (0.279) (0.156) (0.011) Ahuachapan 0.020 0.055 0.042 0.050 -0.034 -0.042 -0.009 -0.008 (0.881) (0.666) (0.339) (0.257) (0.621) (0.464) (0.673) (0.776) La Libertad 0.119* 0.115"" 0.056** 0.055** 0.105** 0.105** -0.004 -0.003 (0.068) (0.047) (0.015) (0.013) (0.050) (0.022) (0.724) (0.847) Other 0.068 0.028 -0.092** -0.091** -0.066 -0.037 -0,019 -0.018 (0.422) (0.726) (0.022) (0.021) (0.374) (0.540) (0.269) (0.371) San Miguel -0.228" -0.275** -0.022 -0.021 0.156 0.103 -0.020 -0,026 (0.061) (0.013) (0.617) (0.614) (0.240) (0.293) (0.303) (0.262) Santa Ana 0.011 0.001 0.031 0.035 0.007 -0.003 0.008 0.013 (0.889) (0.992) (0.345) (0.273) (0.916) (0.949) (0.574) (0.451) Observations 462 1727 459 1735 Probability values inparentheses. significant at 10%; * ** significant at 5%;***462 1741 458 1731 significant at 1%. (b) Probit estimateswith robust standard errors (marginaleffects reported). (c) Tobit estimates.Province dummies are also included for other countries when applicable. 227 ANNEX5.2: MAIN COMPONENTS OF A TYPICALNATIONAL QUALITYSYSTEM A Quality Coordination Councilis responsible for standards, metrology, testing, certification, and accreditation services. The Quality Coordination Council i s an autonomous body comprised o f representatives from both the public and private sectors. The Quality Coordination Council establishes policy, sets goals and priorities, and oversees the national quality system. It coordinates the metrology, standards, and quality information services. An accreditation authority, reporting to the Quality Coordination Council, evaluates and recognizes laboratory and certification accreditationbodies that meet established requirements. The Metrology System includes legal and scientific metrology systems. The legal metrology system is usually a consumer-protection agency that monitors measuring equipment used in commerce (e.g. checking that scales at supermarkets are correct). Scientific systems include centers responsible for maintaining national measurement reference standards (e.g. for length, mass and volume). These are utilized by private and public sector calibration services that ensure precision o f equipment usedby producers and testing laboratories. National Standards include standards for specific sectors (e.g. IEC standards for electronics, ASTM standards for steel and metal) and general standards related to management processes (e.g. I S 0 quality management system standards for manufacturing and service industries). The aim o f National Standards i s to protect human health, safety, and the environment; they also facilitate interoperability, increase efficiencies, and reduce the production o f defective parts and materials. National standards should be defined on the basis o f the World Trade Organization's (WTO) Code of Good Practice for the Preparation, Adoption and Application of Standards, includedinthe WTO's Agreement on Technical Barriers to Trade (TBT). The WTO requires its members to giver first preference to international standards rather than developing unique domestic standards. QualityInformationCenters help firms and consumers identifyrequirements for both domestic and export markets. Many Quality Information Centers also serve as the Enquiry Points required under the WTO TBT Agreement. Among its duties, the EnquiryPoint responds to requests for information from other WTO members concerning technical regulations, standards and certification systems in effect in its country. In return, information from the other 146 WTO members is available at the EnquiryPoint. The AccreditationAuthority-is the national authority responsible for verifying the competence o f the accreditation bodies operating in the country. Depending on the country, accreditation o f inspection bodies, laboratories, quality system certification bodies, and product certification bodies may be undertaken by a single organization or specialized accreditation bodies for each function. If a single authority i s responsible for accreditation, the accreditation authority acts as the accreditationbody and there is no need for both. AccreditationBodies give formal recognitionthat a personor organization is competent to carry out specific tasks. Testing laboratories and certification bodies are most often accredited. 228 Accreditation bodies can be public or private and can be recognized by a National Accreditation Authority. Laboratories that determine conformity of products to recognized standards and procedures are important elements o f a national quality system. Testing services are provided by private laboratories as well as university,government, and industryfacilities. Laboratories test regulated and unregulated agricultural and manufactured products for safety, health, and technical requirements. For these tests to be o f value, they must be performed inaccordance with standard procedures from reputable laboratories usinginstruments and equipment o f appropriate accuracy. Certification Bodies certify conformance o f firms' products and services with National Standards. Certification bodies can be public or private and can be accredited by a National Accreditation Authority. Figure A5.1: Institutional Setting of a Typical National Quality System I QUALITYCOORDINATIONCOUNCIL ACCREDITATION AUTHORITY I I ACCREDITATIONBODIES 1 I I National Quality Standards Information Laboratories Certification Centers Bodies Source: InvestmentClimate Survey 229 REFERENCES Acero, Hugo.2004. "Propuesta para el desarrollo de una politica de seguridad en El Salvador." PowerPoint Presentation basedon data from the UnitedNations Development Program. Aguilar, Jeannette. 2002. Metodologias para la cuantijkacidn del delito. FESPAD. San Salvador, El Salvador Alvarenga, Patricia. 1996. Culturay e`tica de la violencia. El Salvador 1880-1932. EDUCA San Jose. Anmeto, Maria Isabel de, Lisardo Bolaiios, Sigfrido LCe and Hugo Maul. 2004. Politicas laborales en Centroamkrica 2Oportunidades o barreras para el desarrollo? Centro de Investigaciones Economicas Nacionales (CIEN). Guatemala City, Guatemala. Ayres, Robert. 1998. Crime and Violence as Development Issues in Latin America and the Caribbean. The World Bank. Washington, DC. Bigsten, A., Collier, P., Dercon, S., Fafchamps, M., Gauthier, B., Gunning, J.W., Soderbom, M., Oduro, A., Oostendorp, R., Patillo, C., Teal, F., Zeufack, A. 2003. "Credit Constraints in ManufacturingEnterprises inAfrica," Journal of African Economies, 12 (l), 104-125 Bolaky, Bineswaree and Caroline Freund. 2004. "Trade, Regulations and Growth." World Bank Policy ResearchWorking Paper No. 3255. The World Bank, Washington D.C. Broadman, Harry G., James Anderson, Constantjn A. Claessens, Randy Ryterman, Stefka Slavova, Maria Vagliasindi and Gallina A. Vincelette. 2004. Building Market Institutions in South Eastern Europe: Comparative Prospectsfor Investment and Private Sector Development. The World Bank. Washington, DC. Buvinic, Mayra, Andrew Morrison and Michael Schifter. 1999. Violence in Latin America and the Caribbean: A Frameworkfor Action. Inter American Development Bank. Washington, DC Calderon, Cesar and Luis ServCn. 2003a. "Macroeconomic Dimensions of Infrastructure inLatin America." Mimeo, The World Bank, Washington D.C. Calderon, Cesar and Luis ServCn. 2003b. "The Output Cost o f Latin America's Infrastructure Gap." Forthcoming in: Easterly, W., Serven, L.,eds., The Limits of Stabilization: Infrastructure, Public DeJicits, and Growth in Latin America. Stanford University Press and The World Bank. Washington, DC. Carrion, Fernando. 2002. De la violencia urbana a la convivencia ciudadana. En: Fernando Carrion (ed.). Seguridad ciudadana, iespejismo o realidad? FLACSO. Quito, Ecuador. Carrion, Fernando. 2004. La inseguridad ciudadana en la comunidad andina. InJosC Luis Rhi- Sausi (ed.)El desarrollo local en Amkrica Latina. Logros y desafios para la cooperacidn europea. Recal/CeSPI/Nueva Sociedad Caracas. 230 Centro de Informacion, Documentacion y Apoyo a la Investigacion (CIDAI). 2000. "El fracas0 policial." Proceso, 904-2-3. San Salvador, El Salvador. Clark, Ximena, David Dollar and Alejandro Micco. 2004. Port Eficiency, Maritime Transport Costs and Bilateral Trade. NBER Working Paper No. 10353. National Bureau of Economic Research.Cambridge, Massachusetts. Clarke, George. 2001. "Does Internet Connectivity Affect Export Performance? Evidence from Transition Economies." Mimeo, The World Bank, Washington D.C. COCATRAM. 2002. "Resumen Estadistico Portuario Regional del Istmo Centroamericano, Aiio 2001. Redde EstadisticasPortuarias Centroamericanas". COCATRAM. 2004.. "Resumen Estadistico Portuario Regional del Istmo Centroamericano, Aiio 2003. Redde EstadisticasPortuarias Centroamericanas" COCATRAM. 2003. "Resumen Estadistico Portuario Regional del Istmo Centroamericano, Aiio 2002. Redde EstadisticasPortuarias Centroamericanas". Cbrdova, Ricardo and JosC Miguel Cruz. 2004. La cultura politica de la democracia en El Salvador, 2004. IUDPO-UCA. San Salvador, El Salvador. Cruz, JosC Miguel and Gonzblez, Luis Armando. 1997. "Magnitud de la violencia en El Salvador". Estudios Centroamericanos 588,953-966. Cruz, Jose Miguel; Alvaro Trigueros, and Francisco Gonzblez. 2000. El crimen violento en El Salvador. Factores sociales y econdmicos asociados.: Instituto Universitario de Opinicin Publica and The World Bank. San Salvador, El Salvador. De Ferranti, David, Guillermo Perry, Indermit Gill, Jose Luis Guasch, William Maloney, Carolina Sanchez-Paramo and Norbert Schady. 2003. Closing the Gap in Education and Technology. The World Bank, Washington D.C. Djankov, Simeon, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer. 2002. "The Regulation of Entry." Quarterly Journal of Economics, 117:1,pp. 1-37. Djankov, Simeon, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer, 2003, court^,'^ Quarterly Journal ofEconomics, 118, 2: 453-517. Dollar, David and Aart Kraay. 2002. "Institutions, Trade, and Growth." Carnegie Rochester Series on Public Policy. Dollar, David, Mary Hallward-Dreimeier, Anqing Shi, Scott Wallsten, Shuilin Wang and Lixin Colin Xu. 2003a. "Improving the Investment Climate in China." Mimeo, The World Bank, Washington D.C. Dollar, David, Mary Hallward-Dreimeier and Taye Mengistae. 2003b. "Investment Climate and FirmPerformanceinDeveloping Countries." Mimeo, The World Bank, Washington D.C. Dollar, David, Anqing Shi, Shuilin Wang and Lixin Colin Xu. 2003c. "Improving City Competitiveness through the Investment Climate: Ranking 23 Chinese Cities." Mimeo, The World Bank, Washington D.C. Easterly, W. and S. Rebelo. 1993. "Fiscal Policy and Economic Growth: an Empirical Investigation." Journal of Monetary Economics, 32:3, pp. 417-458. 231 Easterly, W. and R. Levine. 2001. "It's Not Factor Accumulation: Stylized Facts and Growth Models." The WorldBank Economic Review 15:2, pp, 177-219. Escribano A. and J. L. Guasch. 2004. "Econometric Methodology for the Evaluation o f the Impact o f Investment Climate (ICA) Variables on Productivity at the firm level: The Case o f Guatemala, Honduras and Nicaragua." Mimeo, The World Bank, Washington D.C. Esfahani, Hadi and Maria Teresa Ramirez. 2002. "Institutions, Infrastructure and Economic Growth: an Empirical Investigation." Journal of Development Economics, 70:2, pp. 443-477. Economist Intelligence Unit.2003. El Salvador Country Report. Economist Intelligence Unit (EIU). 2004a. Country Profile 2004: El Salvador. Economist Intelligence Unit (EIU).2004b. Country Report, El Salvador at a Glance: 2004-05. Fazzari, S., Hubbard, G., Petersen, B.1988. "Financing Constraints and Corporate Investment" Brookings Papers on Economic Activity, (I),I41-206. Fundacion de Estudio para la Aplicacion del Derecho (FESDAP). 2004. Estado de seguridad pfiblica y la justicia penal en El Salvador: julio 2002-diciembre 2003. FESDAP Ediciones. San Salvador, El Salvador. Fundacih GCnero y Sociedad. (GESO). 2004. El riesgo de ser victima de homicidio en Costa Rica, Guatemalay El Salvador. San JosC: GESO. Mimeo. Fischer, Stanley. 1993. "The Role o f Macroeconomic Factors in Growth." Journal of Monetary Economics, 32:3, pp. 485-512. FitchRatings.2004a. El Salvador: Can the SacaAdministration Deliver? Fitch Ratings.2004b. Salvadorian Banks: Year EndResults. Fitch Ratings.2004~.Banco Multisectoral de Inversiones. Fitch Ratings.2003a. The Salvadorian Prudential Regulations. Fitch Ratings.2003b. The Salvadorian Banking System. Frankel, Jeffrey and David Romer. 1999. "Does Trade Cause Growth?" American Economic Review, 89:3, pp. 279-96. Fusades.2003. Informe de Desarrollo Econ6mico y Socia1:"Competitividad para el desarrollo".El Salvador. Fuster, J. 2004 Background Paper for the Honduras Trade Project. Washington, DC, The World Bank. Gamez. Salguero, E .2001, "Iniciativa Mesoamericana de Facilitacion del Intercambio Comercial: Modernizacionde Aduanas y Pasos Fronterizos", InterAmerican Development Bank, Washington, DC. Glaeser, Edward L., and Andrei Shleifer, 2003, "The Rise o f the Regulatory State", Journal of Economic Literature, 41,2: 401-425. Godnick, William, Robert Muggah and Camilla Waszink. 2002. Stray Bullets: The Impact of Small Arms Misuses in Central America. Small Arms Survey and International Action Network on Small Arms, Geneva and London. 232 Gonzklez, Luis Armando. (1997). "El Salvador en la posguerra: de la violencia armada a la violencia social." Realidad, 59,441-458. Gramlich, E. 1994. "Infrastructure Investment: A Review Essay." Journal of Economic Literature, 32, pp. 1176-1196. Guasch, Luis and Joseph Kogan. 2001. "Inventories in Developing Countries: Levels and Determinants-a Red Flag for Competitiveness and Growth." Policy Research Working Paper No. WPS 2552. The World Bank. Washington, DC. Hall, Robert and Charles Jones. 1999. "Why do Some Countries Produce so MuchMore Output per Worker than Others.'' Quarterly Journal of Economics, 114:1, pp. 83-116. Henderson, Keith, Angana Shah Sandra Elena and Violaine Autherman. 2004. "Regional Best Practices: Enforcement of Court Judgements, Lessons Leamded from Latin America." IFES Rule of Law White Paper Series. IFESandUSAID.Washington, DC. Hendley, Kathryn, andPeterMurrell,2003, "Which Mechanisms Support the Fulfillmentof Sales Agreements? Asking Decision-Makers inFirms," Economics Letters, 78, 1:49-54. Instituto Universitario de Opini6n P6blica. 2003. Encuesta de evaluacion del aiio 2003. Serie de informes 102. IUDOP-UCA. San Salvador, El Salvador. InterAmerican Development Bank. 2004, "El Salvador Competitiveness Support Loan(ES- 0151)" Loan Proposal. Washington, DC. Jansson, Tor and Geoffrey Chalmers. 2001. "The Case for Business Registration in Latin America." Mimeo, The InteramericanDevelopment Bank, Washington D.C. Jaramillo, Felipe and Daniel Lederman. 2004. "CAFTA, Trade and Development in Central America: Overview of Analytical Issues and Findings." Unpublisheddraft, The World Bank, Washington D.C. Kaufmann, D., A. Kraay, and P. Zoido-Lobaton. 1999. "Govemance Matters." World Bank Policy ResearchWorking PaperNo. 2196. Kauffman, Kraay and Mastruzzi. 2003. "Govemance Matters 111: Govemance Indicators for 1996-2002." Mimeo, The World Bank, Washington D.C. Knack, S., and P. Keefer. 1995. "Institutions and Economic Performance: Cross-country Tests Using Alternative Institutional Measures." Economic and Politics 7(3): 207-27. Krug Etienne H.,Linda L.Dahlberg, James A. Mercy, Anthony B.Zwi and Rafael Lozano (eds). 2002. World Report on Violence and Health. World HealthOrganization. Geneva. Lawrence, Robert Z. 2004 "Capitalizing on CAFTA." Mimeo Lederman, Daniel. 1996. "The Sources of Economic Growth: A Survey from a Latin America Perspective." Office o f the Chief Economist o f Latin America and the Caribbean Region, World Bank, Washington, D.C. Mimeo. Lederman, Daniel and William Maloney. 2003. "R&D and Development." Office of the Chief Economist of Latin America and the Caribbean Region, World Bank, Washington, D.C. Mimeo. 233 Lederman, Daniel, William Maloney and Luis Semen. 2003. Lessons from NAFTA for Latin America and Caribbean Countries: A Summary of Research Findings. Office of the Chief Economist o f Latin America and the CaribbeanRegion, World Bank, Washington, D.C. Levine, R. 1997. "Financial Development and Economic Growth: Views and Agenda." Journal of Economic Literature 35,2, pp. 688-726. Levine, R.,N.Loayza, and T. Beck. 2000. "Financial Intermediation and Growth: Causality and Causes." Journal of Monetary Economics 46, No.1:31-77. Loayza, Norman, Pablo Fajnzylber and CCsar Calderon. forthcoming. Economic Growth in Latin America and the Caribbean: Stylized Facts, Explanations and Forecasts. The World Bank. Washington, DC. Londofio-Kent, M.and Paul Kent.2003. "A Tale of Two Ports, The Cost o f Inefficiency". ResearchReport, World Bank. Lothian, Tamara.. 2003. The FUSADES Strategy for El Salvador: Priority Areas for Judicial Reform andthe Problem o f Corruption. FUSADES. Lucas, Robert. 1988. "On the Mechanics of Economic Development." Journal of Monetaw Economics 22:3-42. Martinez Julia Evelyn and Beltran de Vieytez 2002. DesaJios y oportunidades de las PyME salvadoreiia: Construyendo una agenda de desarrollo. ANEP, FUNDES, FUNDAPYME. San Salvador, El Salvador. Mauro, P. 1995. "Corruption and Growth." Quarterly Journal of Economics 110: 681-712. Moser, Caroline and Ailsa Winton. 2002. "Violence in the Central American Region: Towards and Integrated Framework o f Violence Reduction." Working Paper No. 171. Overseas Development Institute. London. PanamericanHealthOrganization (PAHO) 2002. La salud en las Amkricas 2002, VolumenI. Washington, D.C. Rico, JosC Maria. 2004. Inseguridad ciudadana en CostaRica: balance de la situacidn. San Jos6: PNUD. Proyecto de Informe de Desarrollo Humano de Costa Rica. Mimeo. Romer, Paul. 1986. "Increasing Returns and Long-Run Growth." Journal of Political Economy, 94:1002-1037. Seminario Regional. 2003. Integracion de 10s Sub-sectores de Transporte en la Regiondel Plan Puebla Panama.Boletin OrdinarioNo. 198. SIECA. Schneider, Friedrich. 2002. "Size and Measurement of the Informal Economy in 110 Countries Around the World." Mimeo, The World Bank, Washington D.C. Shleifer, Andrei, 2004, "Understanding Regulation", Forthcoming inEuropean Financial Management. SIECA, CLACDSDNCAE. 1999. "El Corredor Logistic0 Centroamericano, UnSalto Cualitativo en la Dotacion de Servicios a1Comercio Exterior de la Region." SIECA. 2001. Estudio Centroamericano de Transportes.Mimeo. 234 Stern, Nicholas. 2001. "A Strategy for Development." Mimeo, The World Bank, Washington D.C. - Superintendencia de Pensiones.2004.Revista de Estadisticas Previsionales. El Salvador. Svenson, Jakob. 2003. "Who Must Pay Bribes and How Much? Evidence from a Cross Section of Firms." The Quarterly Journal of Economics, Harvard, 2003. Umafia, Mario 1999.Infraestructura Avanzada de Transportes para el Desarrollo Sostenibley la Integracion de CentroamCrica. United Nations Development Program (UNDP) (2002), Parente, Stephen and Edward Prescott. 2000. Barriers to Riches. Cambridge, Massachusetts: M I T Press. Velasco, A., Rodriguez, C. 2003. Finance inEl Salvador: Avoiding Crisis, Nurturing Growth. Wayne and Croes. 2004. "Impacts o f Investment Climate Constraints on Accommodation Sectors inEcuador, Guatemala, Honduras andNicaragua." The World Bank. Washington, DC. Wilson, John S., Catherine Mann, and Tsunehiro Otsuki (2004). "Assessing the Potential Benefit o f Trade Facilitation: A Global Perspective." World Bank Policy Research Working Paper No. 3224. The World Bank, Washington D.C. World Bank 2001a. "Argentina: Legal and Judicial Sector Assessment," Washington, DC. The World Bank Legal Vice Presidency. Washington, DC World Bank 2001b. "Administration o f Justice and the Legal Profession inSlovakia." The World Bank, Poverty Reduction and Economic Management Unit, Europe and Central Asia Region. Washington, DC. World Bank 2002a. Improving the Investment Climate in India. The World Bank. Washington, DC World Bank. 2002b. Project Appraisal Document :Judicial Modernization Project.Washington, D.C. World Bank 2003a. Improving the Investment Climate in Bangladesh. The World Bank. Washington, D C World Bank 2003b. "El Salvador: Social and Institutional Context, Profile and Challenges of the Judiciary and the Judicial ModernizationProject." The World Bank Washington, DC World Bank 2003c. Country Innovation Brief.The World Bank. Washington, DC. World Bank. 2003d. El Salvador: Country Economic Memorandum. Mimeo, Washington, D.C. World Bank. 2004a. El Salvador Development Policy Review. The World Bank. Washington, DC World Bank. 2004b. Republic o f El Salvador Country Economic Memorandum.. The World Bank. Washington, DC. World Bank, 2004c. World Development Indicators: The World Bank. Washington, DC World Bank. 2004d. "Good Practice inTrade Facilitation: Lessons from Tunisia." PREM Notes No. 89: The World Bank. Washington, DC. 235 World Bank. 2004e. "Korea's Move to E-Procurement." PREMNotes No. 90.: The World Bank. Washington, DC World Bank 2004f. El Salvador: Poverty Assessment - StrengtheningSocial Policy. The World Bank. Washington, DC. World Economic Forum.The Global Competitiveness Report 2004 -2005, New York Yang, Dean. 2003. "Salvadorans Overseas: The Foundation o f a Pro-Poor Growth Strategy." Mimeo,The University ofMichigan, Ann Arbor. 236