THEWORLD BANK 44995 Doing Business: An Independent Evaluation Taking the Measure of the World Bank- IFC Doing Business Indicators IEG PUBLICATIONS 2006 Annual Report on Operations Evaluation Annual Review of Development Effectiveness 2006: Getting Results Addressing the Challenges of Globalization: An Independent Evaluation of the World Bank's Approach to Global Programs Assessing World Bank Support for Trade, 1987­2004: An IEG Evaluation Books, Buildings, and Learning Outcomes: An Impact Evaluation of World Bank Support to Basic Education in Ghana Brazil: Forging a Strategic Partnership for Results--An OED Evaluation of World Bank Assistance Bridging Troubled Waters: Assessing the World Bank Water Resources Strategy Capacity Building in Africa: An OED Evaluation of World Bank Support China: An Evaluation of World Bank Assistance The CIGAR at 31: An Independent Meta-Evaluation of the Consultative Group on International Agricultural Research Committing to Results: Improving the Effectiveness of HIV/AIDS Assistance--An OED Evaluation of the World Bank's Assis- tance for HIV/AIDS Control Country Assistance Evaluation Retrospective: OED Self-Evaluation Debt Relief for the Poorest: An Evaluation Update of the HIPC Initiative A Decade of Action in Transport: An Evaluation of World Bank Assistance to the Transport Sector, 1995­2005 The Development Potential of Regional Programs: An Evaluation of World Bank Support of Multicountry Operations Development Results in Middle-Income Countries: An Evaluation of the World Bank's Support Economies in Transition: An OED Evaluation of World Bank Assistance Engaging with Fragile States: An IEG Review of World Bank Support to Low-Income Countries Under Stress The Effectiveness of World Bank Support for Community-Based and ­Driven Development: An OED Evaluation Evaluating a Decade of World Bank Gender Policy: 1990­99 Evaluation of World Bank Assistance to Pacific Member Countries, 1992­2002 Extractive Industries and Sustainable Development: An Evaluation of World Bank Group Experience Financial Sector Assessment Program: IEG Review of the Joint World Bank and IMF Initiative From Schooling Access to Learning Outcomes: An Unfinished Agenda--An Evaluation of World Bank Support to Primary Education Hazards of Nature, Risks to Development: An IEG Evaluation of World Bank Assistance for Natural Disasters How to Build M&E Systems to Support Better Government IEG Review of World Bank Assistance for Financial Sector Reform Improving Investment Climates: An Evaluation of World Bank Group Assistance Improving the Lives of the Poor Through Investment in Cities Improving the World Bank's Development Assistance: What Does Evaluation Show? Maintaining Momentum to 2015? An Impact Evaluation of Interventions to Improve Maternal and Child Health and Nutri- tion Outcomes in Bangladesh New Renewable Energy: A Review of the World Bank's Assistance Pakistan: An Evaluation of the World Bank's Assistance Pension Reform and the Development of Pension Systems: An Evaluation of World Bank Assistance Poland Country Assistance Review: Partnership in a Transition Economy The Poverty Reduction Strategy Initiative: An Independent Evaluation of the World Bank's Support Through 2003 The Poverty Reduction Strategy Initiative: Findings from 10 Country Case Studies of World Bank and IMF Support Power for Development: A Review of the World Bank Group's Experience with Private Participation in the Electricity Sector Putting Social Development to Work for the Poor: An OED Review of World Bank Activities Small States: Making the Most of Development Assistance--A Synthesis of World Bank Findings Social Funds: Assessing Effectiveness Sourcebook for Evaluating Global and Regional Partnership Programs Water Management in Agriculture: Ten Years of World Bank Assistance, 1994­2004 World Bank Assistance to the Financial Sector: A Synthesis of IEG Evaluations The World Bank in Turkey: 1993­2004--An IEG Country Assistance Evaluation World Bank Lending for Lines of Credit: An IEG Evaluation All IEG evaluations are available, in whole or in part, in languages other than English. For our multilingual selection, please visit http://www.worldbank.org/ieg W O R L D B A N K I N D E P E N D E N T E V A L U A T I O N G R O U P Doing Business: An Independent Evaluation Taking the Measure of the World Bank-IFC Doing Business Indicators 2008 The World Bank http://www.worldbank.org/ieg Washington, D.C. ©2008 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 5 11 10 09 08 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. Cover photo: Women sort roses for export to Europe at fair-trade company Kiliflora (Tanzania). Photo courtesy of Jorgen Schytte/Still Pictures. ISBN-13: 978-0-8213-7552-5 e-ISBN-13: 978-0-8213-7553-2 DOI: 10.1596/978-0-8213-7552-5 World Bank InfoShop Independent Evaluation Group E-mail: pic@worldbank.org Knowledge Programs and Evaluation Capacity Telephone: 202-458-5454 Development (IEGKE) Facsimile: 202-522-1500 E-mail: eline@worldbank.org Telephone: 202-458-4497 Facsimile: 202-522-3125 Printed on Recycled Paper Contents v Abbreviations vii Glossary and Conventions Used in this Report ix Acknowledgments xi Foreword xiii Preface xv Executive Summary xix Management Response xxiii Chairperson's Summary: Committee on Development Effectiveness (CODE) 1 1 The Ideas Behind the Indicators 3 Role of the Investment Climate in Private Sector Growth 6 Three Principles Underlying What DB Measures 8 Two Principles Underlying DB's Methodology 11 2 Collecting Information and Constructing the Rankings 13 The Number of Informants 14 Qualifications and Motivations of Informants 16 Validating the Data 17 Publishing and Revising the Data 18 Constructing the Rankings 21 3 What Do the Indicators Measure? 23 General Characteristics of the Indicators 27 Key Features of Selected Indicators 39 4 Communicating and Using the Indicators 41 Presentation Style 42 Communications Strategy 43 A Tool for Regular Cross-Country Benchmarking 44 A Catalyst for Dialogue 44 A Guide to Policy Reform 46 A Research Tool 47 A Criterion for Operational Decisions 47 An Addition to the Bank's Toolkit 49 5 Findings and Recommendations 51 The Framework Underlying the DB Indicators 52 The Scope of the Indicators 52 Reliability of Information iii D O I N G B U S I N E S S : A N I N D E P E N D E N T E VA L U AT I O N 53 Motivating and Designing Reforms 53 Implications for the Bank Group 54 Recommendations 55 Appendixes 57 A: Methodology 61 B: How Equitably Do the Rankings Reward Reforms? 67 C: Differences between Data in 2007 DB Report and DB Web site (October 2007) for Same Data Collection Period 69 D: Common Law/Civil Law Analysis 77 E: Standard Interview Protocols 81 Endnotes 87 Bibliography Boxes 4 1.1 A Good Investment Climate Balances Private and Societal Interests 24 3.1 Civil and Common Law Approaches to Regulation 26 3.2 Can a Civil Law Country Succeed in a "Doing Business" World? 28 3.3 A Paper "Reform" in Afghanistan 29 3.4 Does Simplifying Business Registration Encourage Formalization? 30 3.5 Can a Tax Haven Be a Global Leader on Taxation? 32 3.6 Does Top-Ranked Imply "Well Regulated" . . . or "Unregulated"? 33 3.7 Measures on the Costs and Difficulty of Firing Workers and the ILO Conventions 42 4.1 Key Features of DB Communications 43 4.2 Keeping up with the Neighbors: DB Indicators Foster Benchmarking 53 5.1 If DB Were to Be Extended to Other Topics Figures 4 1.1 DB Measures Selected Aspects of Investment Climate 15 2.1 The Majority of DB Informants Are Lawyers and Accountants 15 2.2 Why Do You Participate? For Prestige and to Share Expertise 26 3.1 Countries with Similar GNI Can Have Different DB Scores Tables 5 1.1 DB Covers Only Some of the Top Constraints to Business 7 1.2 The 10 DB Indicators and Their Components 9 1.3 What DB Covers 10 1.4 The Informal Economy Casts a Long Shadow 14 2.1 Average Number of Completed Questionnaires per Indicator in Each Country Is Low 18 2.2 Large Changes in 2007 Rankings Resulting from Data Revisions 25 3.1 Do Civil Law Countries Score Lower Than Common Law Countries? 27 3.2 The Starting a Business Indicator 29 3.3 The Paying Taxes Indicator 30 3.4 The Employing Workers Indicator 31 3.5 Employing Workers: Highest- and Lowest-Ranked Countries 34 3.6 The Enforcing Contracts Indicator 36 3.7 The Getting Credit Indicator iv Abbreviations AAA Analytical and advisory activities BEE Business Enabling Environment CAE Country Assistance Evaluation CAS Country Assistance Strategy CPIA Country Policy and Institutional Assessment CPS Country Partnership Strategy DB Doing Business DFID Department for International Development (United Kingdom) EODB Ease of doing business ESW Economic and sector work FDI Foreign direct investment FIAS Foreign Investment Advisory Service FPD Financial and private sector development GDP Gross domestic product GNI Gross national income ICA Investment Climate Assessment ICR Implementation Completion Report IDA International Development Association IEG Independent Evaluation Group IFC International Finance Corporation ILO International Labour Organization MCC Millennium Challenge Corporation (U.S.) MIGA Multilateral Investment Guarantee Agency NGO Nongovernmental organization OECD Organisation for Economic Co-operation and Development OHADA L'Organisation pour l'Harmonisation en Afrique du Droit des Affaires/Organization for the Harmonization of Business Law in Africa PEP Private enterprise partnership PSD Private sector development PwC PricewaterhouseCoopers TA Technical assistance USAID United States Agency for International Development VAT Value added tax WBG World Bank Group WDI World Development Indicators WDR World Development Report v Glossary and Conventions Used in this Report DB 2007 data The data gathered during 2006 and published in late 2006 in the report entitled Doing Business 2007 DB 2007 revised The data set posted on the DB Web site replacing the DB 2007 data as originally published DB 2008 data The data gathered during 2007 and published in October 2007 in the report entitled Doing Business 2008 Index A type of subindicator composed of several binary criteria (such as the difficulty of firing index) Indicator One of the 10 main DB indicators (for example, starting a business) Ranking The ordinal ranking of a country Rating The cardinal number or index score of a country on a particular indicator (such as 16 procedures or 430 days) Subindicator One of the 32 elements comprising the 10 indicators (for example, number of days to start a business, number of procedures to start a business) that go into the calculation of the ease of doing business aggregate ranking. vii Acknowledgments This evaluation was conducted by a team led by The peer reviewers for this report were: Deepa Chakrapani and Victoria Elliott. Shonar Lily Chu (World Bank), Neil Gregory (IFC), Lala led the evaluation in its initial stages. Other Stephen Golub (Swarthmore College), Vijaya team members were Victoria Y. Chang, John Ramachandran (Center for Global Development Eriksson, Giancarlo Marchesi, and Stoyan Tenev. and Georgetown University), and Guven Sak A background paper was prepared by Simon (TEPAV, Economic Policy Research Foundation Commander (director, Center for New and of Turkey). Emerging Markets at the London Business School) with Katrin Tinn. Additional inputs were The evaluation team thanks its numerous provided by James Heyes and Salvatore Spada. interviewees among Bank Group staff, individu- Marinella Yadao and Yezena Yimer provided als in client countries, and Doing Business administrative support. William Hurlbut and informants. It also appreciates the cooperation it Caroline McEuen edited the report. received from the Doing Business team. Director-General, Evaluation: Vinod Thomas Director, Independent Evaluation Group­World Bank: Cheryl Gray Group Manager, IEG Corporate and Global Evaluation and Methods: Mark Sundberg Task Managers: Victoria Elliott and Deepa Chakrapani ix Foreword Institutions matter a great deal for development, DB has developed an impressive system for and a country's regulatory institutions are vital gathering standard information from lawyers and for the pace and quality of economic growth. The other informants in more than 170 countries. Doing Business (DB) indicators deal with the However, the number of informants on each part of the regulatory regime that governs the topic and country is small, making it difficult to start-up, operation, and growth of businesses. measure confidence levels around the country Improvements in the climate for businesses can rankings. The evaluation recommends that DB potentially generate jobs and incomes. DB is recruit more, and more diverse, informants; built on the premise that these firms are more disclose the number of informants; and be more likely to flourish if they have to abide by fewer, transparent about changes in published data. cheaper, and simpler regulations. The evaluation found that DB has often sparked By ranking countries on selected dimensions of constructive debate among country authorities business regulation, the DB report has attracted and business interests about ways of making considerable attention and has become one of the regulation simpler and lighter on firms. Some Bank Group's flagship knowledge products. And, fear that it can distort the policy priorities of like any rating exercise, it has also provoked authorities or the Bank Group by extending the important questions and concerns, both inside and encouragement for less regulatory burden, to outside of the Bank Group. This evaluation takes the discouragement of good and valuable regula- an independent look at the relevance, reliability, tions. Even though this seems not to have and usefulness of this innovative exercise. happened, the context and perspective on what DB really measures or addresses are crucial for DB assesses the burden of regulation on firms policy makers and practitioners to keep in mind. without aiming to capture the social or economy- wide benefits that regulations yield, such as In coming years, regulatory actions will become safety, environmental protection, worker protec- increasingly important as countries address tion, or transparency. DB offers a consistent challenges such as migration, health, and climate yardstick for comparing countries on regulation change. It will be crucial to emphasize both the as seen from the firm's private point of view. But need for efficiency in the implementation of a a complete appreciation of the quality of the regulation and the benefits that a good regula- business climate must also measure the quality tion can bring. That is why it is important to of infrastructure, labor skills, competition discuss the usefulness of DB and how it can be policies, and other determinants and outcomes improved, along with the context in which good of investment and profitability. regulations need to be implemented. Vinod Thomas Director-General, Evaluation xi Village shop at dusk, lit by solar panels, Sri Lanka. Photo courtesy of Dominic Sansoni/World Bank. Preface Doing Business (DB), the annual World Bank-IFC Directors of the World Bank Group and others. benchmarking exercise launched in 2004, is one The evaluation assesses the methods and of the Bank Group's flagship knowledge processes underlying the construction of the products. It aims to measure the costs to firms of indicators; the relevance of the indicators to business regulations in 178 countries and ranks desired intermediate outcomes; and their use by the countries along 10 dimensions. It also aims World Bank Group staff, policy makers, and other to advance the World Bank Group's private relevant stakeholders. To the extent that sector development agenda by motivating and countries may seek to implement efficient informing the design of regulatory reforms, regulatory frameworks to guide other areas of enriching international initiatives on develop- development, such as health, environment, ment effectiveness, and informing theory. By energy, and climate change, the Bank Group ranking countries and spotlighting both leaders could use lessons from the DB initiative to help and laggards, DB has attracted the interest of countries develop ways to benchmark the senior policy makers and is claimed to have soundness of their regulatory framework and inspired reforms on business climate issues. DB's track improvements over time. lively communications style has helped give the DB indicators an international profile. To carry out the evaluation, the IEG team commissioned a literature review, analyzed the DB has critics as well as fans. Some have ratings and underlying data published by DB, questioned the reliability and objectivity of its validated DB's methodology by interviewing a measurements. Others doubt the relevance of sample of their informants, and analyzed the DB the issues it addresses or fear it may unduly indicators' relevance and use in 13 randomly dominate countries' reform agendas at the selected countries by interviewing Bank and IFC expense of more crucial development objectives. staff, country officials, and experts. Appendix A And the attention given to the indicators may explains the methodology in detail. inadvertently signal that the World Bank Group values less burdensome business regulations The report has five sections: chapter 1 reviews more highly than its other strategies for poverty the intellectual underpinnings of the DB indica- reduction and sustainable development. tors. Chapter 2 reports on how DB collects and assembles data. Chapters 3 and 4 discuss the This IEG evaluation of the DB indicators relevance of the dimensions measured by provides an independent view of DB's strengths the exercise and their use inside and outside and weaknesses, in response to interest the Bank. Chapter 5 presents findings and expressed by members of the Board of Executive recommendations. xiii Women sort roses for export to Europe at fair-trade company Kiliflora (Tanzania). Photo courtesy of Jorgen Schytte/Still Pictures. Executive Summary D oing Business (DB), an annual World Bank-IFC publication launched in 2004, is one of the Bank Group's flagship knowledge products. It measures the burden of selected business regulations in 178 countries and ranks the countries on 10 dimensions. The program's stated objective is to advance the World Bank Group's private sector development agenda in four ways: motivate reforms through country benchmarking; inform the design of reforms; enrich international initiatives on development effectiveness; and in- form theory. This independent evaluation of the DB indica- difficult to isolate. Since regulations generate tors assesses the methods and processes social benefits as well as private costs, what is underlying the construction of the indicators; good for an individual firm is not necessarily the relevance of the indicators to desired good for the economy or society as a whole. intermediate outcomes; and their use by World Therefore, policy implications are not always Bank Group staff, policy makers, and other clear-cut, and the right level and type of regula- stakeholders. It finds that the indicators have tion is a matter of policy choice in each country. been highly effective in drawing attention to the burdens of business regulation, but cannot by The DB exercise reflects the limitations inherent themselves capture other key dimensions of a in the underlying research. As an exercise in country's business climate, the benefits of cross-country comparison, DB is not intended regulation, or key related aspects of devel- to, and cannot, capture country nuances. Firms' opment effectiveness. Thus, the Bank Group investment decisions also depend on variables and stakeholders need to consider the DB indica- not measured by the DB indicators, such as the tors in a country context and interpret them cost and access to finance and infrastructure, accordingly. labor skills, and corruption. Different aspects of regulation have varying degrees of economic The Underlying Framework of the importance depending on countries' income DB Indicators levels, legal regimes, and other characteristics. The DB exercise is anchored in research that Seven of DB's 10 indicators presume that lessen- links characteristics of a country's business ing regulation is always desirable, whether a environment to firm performance, and thence to country starts with a little or a lot of regulation. macroeconomic outcomes. The regulatory Reform as measured by the DB indicators framework--the part of the business environ- typically means reducing regulations and their ment that DB measures--has been shown to be burden, irrespective of their potential benefits. associated with firm performance, but its associ- ation with macroeconomic outcomes is less The evaluation confirmed that the DB indicators clear. Many other factors affect macroeconomic primarily measure laws and regulations as they outcomes, and the direction of causality between are written. But the relevance of each indicator regulation and economic outcomes is very in a given country depends on the extent to xv D O I N G B U S I N E S S : A N I N D E P E N D E N T E VA L U AT I O N which the law is actually applied, which DB does the questionnaire may also help to encourage not aim to measure. Likewise, the pay-off of a more informants to contribute. particular regulatory reform will depend on how significant a burden the regulation poses in Second, although DB makes available a great deal practice. These limitations underscore the need of information about its data and methods, it for DB to be interpreted cautiously and used in remains insufficiently transparent about the conjunction with complementary tools such as number and types of informants for each indica- Investment Climate Assessments. tor, the adjustments its staff make to the data received from informants, and the changes made Overall, the indicators objectively and reliably to previously published data and their effects on measure what they set out to measure, with a the rankings. DB needs to adequately explain to few qualifications. The controversial employing users the possibilities for errors and biases. workers indicator is consistent with the letter of relevant International Labor Organization (ILO) Third, DB makes much of its country rankings. conventions, but not always their spirit, insofar as The rankings entail three weaknesses. First, it gives lower scores to countries that have because most of the indicators presume that less chosen policies for greater job protection. regulation is better, it is difficult to tell whether Systematic differences in the country rankings for the top-ranked countries have good and efficient a few indicators are associated with countries' regulations or simply inadequate regulation. legal origins in civil or common law, but these Second, the small informant base makes it patterns have little impact on the overall rankings difficult to measure confidence in the accuracy of or the validity of the exercise. The paying taxes the individual indicator values, and thus in the indicator includes an anomalous subindicator-- aggregate rankings. Third, changes in a country's the total tax rate--which does not simply ranking depend importantly on where it sits on measure administrative burden to firms, but the distribution: small changes can produce large rather reflects a country's overall fiscal policy ratings jumps, and vice versa. These factors derived from social preferences. Finally, inaccu- contribute to anomalies in the rankings. rate nomenclature and overstated claims of the indicators' explanatory power have provoked These issues alone may not jeopardize the DB considerable criticism from stakeholders. indicators' reliability. But the lack of transparency about them undermines DB's credibility and Methodology and Data Reliability goodwill. DB's documents and presentations DB collects its information from expert inform- should include full explanations and cautions on ants in each country, mostly lawyers, who these points. provide information free of charge. This process can generate reliable data, but three areas of Motivating and Designing Reforms vulnerability need to be addressed. The DB indicators have motivated policy makers to discuss and consider business regulation First, the data are provided by few informants, issues. Its active dissemination in easy-to- with some data points for a country generated by understand language permits widespread press just one or two firms. Of particular concern is the coverage and generates interest from businesses, paying taxes indicator--DB relies exclusively on nongovernmental organizations (NGOs), and a single firm to provide both the underlying senior policy makers. methodology and the data for 142 countries. The number and diversity of informants for all indica- DB has had less influence on the choice, scope, tors need to be increased and their information and design of reforms. Most Bank Group staff validated more systematically. An increase in the and country stakeholders interviewed for this informant base will require a systematic vetting evaluation report that they draw on a range of process to reduce self-selection bias. Simplifying analytical material to determine the nature, xvi E X E C U T I V E S U M M A RY sequence, and direction of reforms; the DB Second, the DB exercise has demonstrated that indicators have limited use in this regard. As a cross-country ranking can be effective in cross-country benchmarking exercise, DB spurring dialogue and motivating interest and cannot be expected to capture the country- action. It could potentially be applied to other specific considerations involved in prioritizing, development issues--those for which actionable sequencing, and designing policy reforms. Each indicators can serve as proxies for the target year DB spotlights countries that have demon- outcomes and for which the direction of strated the largest gain in the overall ranking and improvement is uniform for all countries. an improvement on at least three indicators. Such an approach, while transparent, does not Recommendations capture the reforms' relevance and their poten- tial impact on the binding constraints to the 1. To improve the credibility and quality of the rank- investment climate in the country. ings, the DB team should: a. Take a strategic approach to selecting and in- IEG did not find evidence that the DB indicators creasing the number of informants: have distorted policy priorities in the countries ­ Establish and disclose selection criteria or in the Bank Group's programs, or that for informants. countries have made superficial changes for the ­ Focus on the indicators with fewest in- sole purpose of improving their rankings. formants and on countries with the least reliable information. In summary, DB measures the costs but not the ­ Formalize the contributions of the sup- benefits of regulation. Despite its methodologi- plemental informants by having them fill cal limitations, it has contributed to develop- out the questionnaire. ment by providing countries with a basis for ­ Involve Bank Group staff more actively to international comparisons of their regulatory help identify informants. regimes. It has helped to catalyze debates and b. Be more transparent about the following as- dialogue about investment climate issues in pects of the process: developing countries. For the Bank Group, it is a ­ Informant base: Disclose the number key global knowledge product. Most of the of informants for each indicator at the methodological limitations can and should be country level, differentiating between addressed promptly, lest they undermine its those who complete questionnaires and credibility. Inaccurate nomenclature should be those who provide supplemental rectified and the DB reports should not information. overstate claims of causality and the indicators' ­ Changes in data: Disclose all data cor- explanatory power. rections and changes as they are made. Explain their effect on the rankings, and, Implications for the Bank Group to facilitate research, make available all The evaluation notes two broader implications previously published data sets. for the Bank Group. ­ Use of the indicators: Be clear about the limitations in the use of the indicators for First, the Bank Group, by prominently recogniz- a broader policy dialogue on a country's ing DB's highly ranked countries, may inadver- development priorities. tently be signaling that it values reduced c. Revise the paying taxes indicator to include regulatory burdens more than other develop- only measures of administrative burden. Since ment goals. The Bank Group's approach entails the tax rate is an important part of the busi- helping countries achieve a wide range of ness climate, DB should continue to collect objectives, yet it has no comparable way of and present simple information on corpo- celebrating improvements in other important rate tax rates, but exclude it from the rank- development outcomes. ings (as it does for information on nonwage xvii D O I N G B U S I N E S S : A N I N D E P E N D E N T E VA L U AT I O N labor costs in the employing workers indi- 3. To plan future additions or modification to the cator). A wider range of informants should indicators, the DB team should: also be engaged for the paying taxes a. Use Bank analyses to drive the choice of DB indicator. indicators. Business Enterprise Surveys, In- 2. To make its reform analysis more meaningful, the vestment Climate Assessments, and other DB team should: work can help determine stakeholders' pri- a. Make clear that DB measures improvements to orities for domestic private sector growth. regulatory costs and burdens, which is only one The DB team should use such analyses to dimension of any overall reform of the in- determine the choice of new indicators and vestment climate. periodically reassess its current set of b. Trace the impact of DB reforms at the country indicators. level. The DB team should work with coun- b. Pilot and stabilize the methodology before in- try units to analyze the effects of imple- cluding new indicators in rankings. Frequent menting the reforms measured by the DB changes in methodology make comparison indicators (such as revised legislation or across time less meaningful. New indicators streamlined processes) on: (i) firm per- should be piloted (that is, data collected formance, (ii) perceptions of business man- and published for comment, but not fac- agers on related regulatory burdens, and tored into the rankings) until the method- (iii) the efficiency of the regulatory envi- ology is validated and stabilized. ronment in the country. xviii Management Response M anagement welcomes this Independent Evaluation Group (IEG) review of the World Bank/International Finance Corporation (WB/IFC) Doing Business (DB) indicators. It notes the finding that the DB exercise has been effective in motivating interest, spurring dialogue on reforms, and stimu- lating action. Suggestions and recommendations in the review will be used to strengthen the DB process going forward. That said, management has a set of observations it would like to make on the analysis. Specific responses to IEG's recommendations are given in the attached Management Action Record table. Concurrence with the Broad Thrust ers index complies with the core labor stan- of the Analysis and Recommendations dards and all other relevant conventions of The evaluation contains a number of important the International Labor Organization. conclusions that management finds most · The conclusion that the legal origin, whether helpful. Specifically, these include: civil or common law, does not determine a country's score in the DB indicators. A hypo- · An acknowledgement that the DB exercise has thetical civil law economy based on best prac- been highly effective in spurring dialogue on tices would rank third in the global ease of reforms and motivating interest and action. It doing business. has also informed a large academic literature · Concrete suggestions on improving the trans- on regulatory reform and the impact of regu- parency of the data collection and analysis and lation on economic and social outcomes (nearly the respondent selection process. 800 academic articles as of June 2007). Such re- · A recommendation to use other World Bank search can aid policy makers, particularly in de- analyses, most importantly the Enterprise Sur- veloping countries, in the search for the optimal veys and Investment Climate Assessments, to kind and level of regulation to ensure that the inform the choice of topics in DB and enrich majority of the population can participate in the analysis in future reports. economic activity and benefit from legal cer- · A concrete proposal on piloting methodologies tainty and social protection. on new indicators before including them in the · The recommendation to apply similar bench- aggregate ranking on the ease of doing marking to other development issues and to business. encourage the development of actionable cross-country indicators that can track im- Management Observations provements over time. Management has four issues that it would like to · The finding that while effective in catalyzing re- raise with regard to the analysis and recommen- forms debates and dialogue, the DB indica- dations in this review. tors have not distorted policy priorities or encouraged policy makers to make superficial Paying Taxes Indicator. The IEG review recom- changes to improve rankings. mends revising the paying taxes indicator to · The conclusion that the DB employing work- include only measures of administrative burden. In xix D O I N G B U S I N E S S : A N I N D E P E N D E N T E VA L U AT I O N management's view, this recommendation is not the text of laws and regulations. DB respondents consistent with another important recommenda- provide references to the relevant texts of the tion, on the use of the other World Bank analyses to laws and regulations. This is unlike the method- determine the priorities for regulatory reform. In ology of perceptions-based surveys, which the World Bank Enterprise Surveys, for example, depend on having large samples of representa- tax rates are considered a top obstacle in twice as tive respondents. To ensure accurate interpreta- many countries as tax administration. In the tion of regulations and time estimates, DB works Enterprise Surveys done in fiscal 2007, 17 of 40 find with local experts who routinely administer or the tax rate to be among the top 3 obstacles, and 33 advise on legal and regulatory requirements. of 40 find it to be a bigger obstacle than tax adminis- Since 2004, 10,270 local experts have contributed. tration. More generally, taxation is a regulatory tool Management agrees to further increase the and there is a trade-off between regulation and respondent pool, and has taken action, including taxation. It is important to note that DB measures through visits to 151 countries. In addition, business taxes only, and therefore does not reflect management has hired a respondents' manager the country's overall fiscal policy and revenue as a member of the DB team to select and collection. increase the number of respondents, focusing in particular on the poorest countries and other Making Available Previously Published Data sets. economies with the fewest number of respon- The IEG review recommends making available dents. Further, management commits to increase all previously published data sets, not corrected the involvement of Bank Group staff in recruiting for errors and methodology changes. This respondents; to conduct annual data collection practice is unorthodox and is not followed by visits to the 50 economies with the fewest other major primary data providers. Instead, DB number of respondents; and to expand the follows the practice of other data providers and piloted practice of giving out awards to the makes available back-calculated data series, respondents who have contributed high-quality corrected for errors and methodology changes. data over a sustained period of time. These data are made available on the "Get Full Data" page of the DB Web site. In addition, the Level of Regulation. The IEG review states that DB data used in the background research for DB are presumes less regulation is always better. This published on the "Research" page of the DB Web is incorrect. Six of the 10 indicators reward site. These two data sources have been widely countries for having more regulation or a simpli- used by researchers, with more than 800 fied way of implementing existing regulation. Top academic papers utilizing the DB data. Manage- reformers in DB 2007 implemented stricter ment agrees to make more information available regulations (for example, China, Mexico, and on reasons for data changes to facilitate the Tanzania) or simplified their implementation (for distinction between methodological changes, example, Croatia, Guatemala, and Romania). The systematic changes in coding rules, and errors. top 10 countries in the ease of doing business are All methodology changes are described in detail Singapore, New Zealand, the United States, Hong on the Web site at http://www.doingbusiness Kong, Denmark, the United Kingdom, Canada, .org/MethodologySurveys/. All revisions that Ireland, Australia, and Iceland. Countries with no affect the data published in the DB 2007 report regulation receive a "no practice" score in the integrated as of the time of publication of the DB relevant area and the lowest ranking. 2008 report are being made available at the "Get Full Data" page of the Web site. Conclusion Overall, management welcomes this evaluation Increasing the Number of Respondents. The IEG from IEG. Management generally accepts IEG's review recommends increasing the number of recommendations, with some caveats. Detailed DB respondents. Management notes that the DB responses to the recommendations are outlined methodology fundamentally depends on reading in the attached Management Action Record. xx M A N A G E M E N T R E S P O N S E Management Action Record Recommendation Management Response To improve the credibility and quality of the rankings, the Mostly Agreed. Bank Group management mostly agrees with DB team should: this recommendation, noting that it primarily points to the im- (a) Take a strategic approach to selecting and increasing portance of intensifying the rigor of recruiting and maintaining the number of informants: a large pool of expert respondents. ­ Establish and disclose selection criteria for informants. ­ To implement the first part (point a) of this recommendation, ­ Focus on the indicators with fewest informants and coun- management has hired a respondents' manager on the DB tries with the least reliable information. team. The task of the respondents' manager is to select and ­ Formalize the contributions of the supplemental inform- increase the number of respondents, focusing in particular ants by having them fill out the questionnaire. on the poorest countries and other economies with the fewest ­ Involve Bank Group staff more actively to help identify number of respondents. In addition, management commits to informants. increase the involvement of Bank Group staff in recruiting re- (b) Be more transparent on the following issues of process: spondents and to conduct annual data collection visits to the ­ Informant base: Disclose the number of informants for 50 economies with the fewest number of respondents. Thirdly, each indicator at the country level, differentiating between management will expand the piloted practice of giving out those who complete questionnaires and those who pro- awards to the respondents who have contributed high-quality vide "supplemental" information. data over a sustained period of time. Such awards serve to ­ Changes in data: Disclose all data corrections and changes express gratitude for the respondents' efforts and to main- as they are made. Explain their effect on the rankings, and, tain the pool of respondents. to facilitate research, make available all previously pub- ­ To implement the second part (point b) of this recommenda- lished data sets. tion, management commits to disclosing the number of re- ­ Use of the indicators: Be clear about the limitations in the spondents for each indicator at the country level, starting with use of the indicators for a broader policy dialogue on a the launch of Doing Business 2009. Management is also country's development priorities. making available details on data corrections/changes and (c) Revise the paying taxes indicator to include only meas- methodology changes that have been made in the year fol- ures of administrative burden. Since the tax rate is an im- lowing the launch of the previous report. Lastly, management portant part of the business climate, DB should continue to commits to expanding the discussion on the limitations in the collect and present simple information on corporate tax rates, use of the DB indicators in the "Methodology" section of the but exclude it from the rankings (as it does for information report and on the Web site. However, management disagrees on nonwage labor costs in the employing workers indicator). with the recommendation to make available all previously pub- A wider range of informants should also be engaged for the lished data sets, not corrected for errors and methodology paying taxes indicator. changes. This practice is unorthodox and is not followed by other major primary data providers. The data used in the background research for DB are already published on the "Re- search" page of the DB Web site. The full time series of DB data, corrected for errors and methodology changes, is also available at the "Get Full Data" page of the DB Web site. These two data sources have been widely used by researchers, with more than 800 academic papers utilizing the DB data. ­ Management mostly disagrees with the last point (point c) of the recommendation. The tax rate is often identified as a major constraint to business activity in the World Bank En- terprise Surveys. Including a measure of overall tax burden xxi D O I N G B U S I N E S S : A N I N D E P E N D E N T E VA L U AT I O N Management Action Record Recommendation Management Response in the DB indicators provides a complete treatment for the topic of paying taxes. Focusing only on the administrative bur- den of paying taxes will take the DB methodology away from covering a broader spectrum of areas relevant to small do- mestic businesses. However, management commits to expand the range of respondents on the paying taxes' survey by re- cruiting a larger set of accounting and tax experts. To make its reform analysis more meaningful, the DB team Agreed. Bank Group management agrees with this recommen- should: dation and will strive to make it even clearer in future DB reports (a) Make clear that DB measures improvements to regu- and presentations that DB covers only some dimensions of the latory costs and burdens, which is only one dimension of overall reform of the investment climate. Management also any overall reform of the investment climate. commits to a measurement and evaluation agenda, in partner- (b) Trace the impact of DB reforms at the country level. The ship with WB country units and IFC regional facilities, to docu- DB team should work with country units to analyze the ef- ment the effect of DB reforms on a set of economic and social fects of implementing the reforms measured by the DB indi- indicators. The World Bank Enterprise Surveys in particular will cators (such as revised legislation or streamlined process) on: be used for this work. (i) firm performance, (ii) perceptions of businessmen on re- lated regulatory burdens, and (iii) the efficiency of the regu- latory environment in the country. To plan future additions to or modifications of the indica- Agreed. Bank Group management agrees with this recommen- tors, the DB team should: dation and will direct the DB team toward using other Bank (a) Use Bank analyses to drive the choice of DB indica- Group analyses, and in particular the Enterprise Surveys and In- tors. Business Enterprise Surveys, Investment Climate As- vestment Climate Assessments, for both determining the choice sessments, and other work can help determine stakeholders' of new indicators and periodically assessing the existing set of priorities for domestic private sector growth. The DB team DB indicators. Management also commits to publishing new should use such analyses to determine the choice of new in- sets of indicators in future DB reports for comment, while not fac- dicators, and periodically assess its current set of indicators. toring those in the rankings until their methodology is validated (b) Pilot and stabilize the methodology before including by academic research. new indicators in rankings. Frequent changes in method- ology make comparison across time less meaningful. New in- dicators should be piloted (that is, data collected and published for comment, but not factored into the rankings) until the methodology is validated and stabilized. xxii Chairperson's Summary: Committee on Development Effectiveness (CODE) B ackground. The Doing Business (DB) report measures the burden of busi- ness regulation and ranks countries on 10 dimensions. The objective is to advance the private sector development agenda by motivating reforms via benchmarking; inform the design of reforms; enrich international initiatives on development effectiveness; and inform theory. This evaluation of the DB report takes an independent look at how indicators are constructed and what they measure. IEG Main Findings. The evaluation finds DB indica- posted on the Web site; it should acknowledge tors have been effective in drawing attention to that its published data are subject to change and the burdens of business regulation, but cannot make available to researchers all versions of the capture other important dimensions of a data set. The DB makes much of annual changes county's business climate, the benefits of regula- in country rankings, but these need to be tion, or related aspects of development. This understood in context. The DB team should underscores the need for DB to be interpreted make clear that DB measures reductions in cautiously and used in conjunction with comple- regulatory costs and tracks reforms at the mentary tools such as Investment Climate country level, but is not a general indicator of Assessments. The number and diversity of DB investment climate quality. Lastly, the DB team informants need to be increased and their should use Bank analyses to inform the information better validated. The DB should take development of further DB indicators, and a strategic approach to selecting and increasing should pilot and stabilize methodology before informants; define and publish informant including new indicators. selection criteria; and be more transparent about its informant base and changes in data. DB Draft Management Response. Management wel- assesses regulations as they are written, not the comed the evaluation of the DB report, noting its extent or way in which they are applied. The DB acknowledgement that the DB exercise has been reports should not overstate the indicators' highly effective in spurring dialogue and action explanatory power. The total tax rate subindica- on reforms, and the recommendation that similar tor goes beyond administrative burden to also benchmarking be applied to other development reflect a country's fiscal policy choices. Thus IEG issues. Management highlighted three issues in recommends that the DB exclude it from the the IEG recommendations. IEG recommends calculation of the aggregate ranking but continue that DB revise the paying taxes indicator to to collect and publish this important informa- include only administrative burden measures and tion. The DB team routinely changes a large continue to collect and present information on share of the data after it has been published and the tax rate but exclude it from the rankings. xxiii D O I N G B U S I N E S S : A N I N D E P E N D E N T E VA L U AT I O N However, management finds that this is not should clarify what indicators are not intended to consistent with IEG's recommendation on use of measure; DB can help governments improve their Enterprise Surveys and Investment Climate investment environment; DB needs a clear Assessments to determine regulatory reform communication strategy to the public and use of priorities, as Enterprise Surveys regularly identify disclaimers; the indicators may promote regula- the tax burden as a major concern to entrepre- tions but do not capture effective enforcement of neurs. IEG recommends making available all the rules; there was recognition that DB indica- previously published data sets to facilitate tors help to highlight the importance of regula- research, which in management's view would be tions; IEG found some weaknesses in DB unorthodox; in this context management also methodology that management should address; notes that back-calculated data series, adjusted avoid using DB as ranking of countries that may for methodology changes and correction, are have an impact on resource allocations; and take made available on the DB Web site. Manage- note that benchmarks and regulations are not ment agrees with IEG's recommendation to unique to DB--as well as questions on how increase DB informants, and is actively engaged functional equivalence can be taken into account in this area. across common-law countries versus civil-law countries. DGE Statement. DB is a widely recognized product of the World Bank Group (WBG) and a prominent Most members agreed with IEG's recommenda- part of its work on private sector development. tions for DB to take a strategic approach to Being a rating exercise, DB has also generated selecting and increasing the number of inform- important questions and concerns. Just as it is ants. There were other comments on the important to disseminate what the DB indicators methodology of DB: the need to look not only at do, it is important to note what the DB indicators what the DB indicators do not do, but rather do not do. While measuring the regulatory focus on what they do; there should be emphasis burden that some firms in the formal sector face, on reliability of indicators; and the indicators it does not capture some of the most crucial should capture country advances or effective variables affecting the investment climate of a reforms. Some members cautioned against the country, such as macroeconomic stability, labor use of DB indicators to top-rank countries. They skills, access to credit, infrastructure, or corrup- questioned how to look at rankings. One tion. Going further, they do not touch on the member suggested including cost of regulations social or economy-wide benefits that regulations of FDI and DB indicators and to consider yield, such as safety, environmental protection, or subnational governments' regulations. Another worker protection. While a useful measure of the member felt the DB indicators do not lead to burden of legal regulations, they are not and necessary reform and do not consider the politi- should not be used as an index of the quality of a cal economy, and questioned the work of the DB country's business climate. indicators. On the paying taxes indicator, there were diverse views expressed by speakers, as Overall Conclusions and Next Steps. CODE mem- some do not agree with IEG's recommendations bers welcomed the IEG evaluation of the DB to exclude tax rate from the indicator. Others report as well as the Draft Management Response. questioned the rationale for including the tax Overall, members welcomed the IEG review of the rate as part of the paying taxes indicator. On the DB indicators and commended the quality of the employing workers indicator, some speakers report. While noting that the DB report cannot noted that it may overstate what it measures. capture all dimensions of a country's business climate, some members acknowledged its contri- One member suggested disclosing the DB report butions in promoting reforms in some countries. together with the IEG review. Management Members raised a wide set of comments--among stated that it has taken note of comments and them: DB is work in progress; the DB report suggestions raised during the meeting. xxiv C H A I R P E R S O N ' S S U M M A RY: C O M M I T T E E O N D E V E L O P M E N T E F E C T I V E N E S S ( C O D E ) The following main issues were raised at the tors measure. The DGE commented that the DB meeting: indicators had sparked constructive debate among country authorities and business IEG Evaluation and DB Report. Several members interests, while also provoking fears that it may welcomed the IEG evaluation of the DB indica- distort policy priorities among country author- tors. A member suggested it be published as part ities and in the WBG by emphasizing the private of, or alongside, the DB report. Management costs of regulation at the expense of social commented that publishing the evaluation with benefits. the report, and doing so annually, would not be the most practical approach. A speaker A member noted that perhaps the IEG report was proposed that the IEG evaluation be featured on underestimating DB users (i.e., policy makers), as the DB Web site. IEG confirmed that there they do not necessarily read DB as a document would be active dissemination of the DB that promotes "no regulations." In that sense, he evaluation. Several members noted DB had added that policy makers use the DB report when spurred debate and has been helpful in improv- developing policy and can compare with ing regulatory environments. At the same time, countries that have implemented similar regula- some members remarked that the DB report tions. Management agreed that policy makers had shortcomings and that it was a work in are faced with a wide range of stakeholders they progress. A member wondered how DB had have to respond to and should not be underesti- become a flagship document if so many mated. IEG noted that the evaluation finds that shortcomings existed, while a speaker queried policy makers used DB as one tool among others as to the quality checks needed to launch a WBG in developing policy reforms. The DB reports product. One member noted that for small should avoid claiming that specific reforms were states, this was a very good tool, allowing directly stimulated by the DB indicators. IEG countries to assess their business environment. further recommends that the DB team work with A speaker added that DB was a product that country teams to trace the impact of country- people wanted to read, understand, and apply. level reforms measured by the DB indicators. One member asked that the DB indicators not be DB Indicators. A member stated that the method- used in Bank operations, particularly resource ology behind the DB indicators needed clarifica- allocation, while another member noted that tion, while the indicators had to be simple and since these are partial indicators they should be easy to understand. Management noted that one used with prudence. A member noted that of the fundamental objectives of DB was to Operations Policy and Country Services should continuously improve indicators to make them look into how IEG's evaluations can affect the use relevant and accurate. Another member made of the DB indicators in the CPIA exercise. the case that there was no need to continue to Management noted that for background produce DB and suggested that comparative information, the CPIA draws on a number of studies on regulation issues be emphasized. A data sets, each of which covers only some aspects speaker commented that DB needed to be of economic performance. clearer as to how an indicator reflects specific outcomes (i.e., registering a business vs. number Regulation vs. Deregulation. A member com- of licenses), as some indicators have issues mented that improved indicators may not related to health and safety. Management noted necessarily lead to correct reforms since they are that the indicators do not claim to measure all designed with the assumption that the lighter aspects of the business environment, and that the regulation, the better. Management the decision to keep DB relatively focused was clarified that the DB report does not reject in keeping with an earlier Board discussion. regulation; instead it is the issues of quality and Management agreed that it is important to keep efficiency of rules that are the focus of the DB improving the description of what the indica- report. IEG noted that their evaluation supports xxv D O I N G B U S I N E S S : A N I N D E P E N D E N T E VA L U AT I O N that policy makers use the DB indicators ranking if you have no regulation (i.e., sensibly and in tandem with other data. One property registry). IEG pointed out that the member observed that the tension that exists introduction of the "Reformer's Club" in the DB with regulations that can either be promoted by marketing clearly signals a normative rent-seekers or to protect the public interest is interpretation to the rankings. A member always present. The DGE observed that the WBG added that countries have varied constraints has the responsibility to emphasize both the (e.g. scarce resources, limited capacity), and importance of efficiency in implementing these may be exacerbated by competition to regulations as well as their potential value improve rankings. He further added that the added. Bank is not in a position to rank its members. Management noted that countries have used Disclaimer and Transparency. Several members the indicators constructively, while also taking noted the importance of communicating what into account quality issues and country- the DB report measures. A member remarked specific limitations. that DB was not a business-climate ranking or indicator and that it was important to communi- One member asked that the request to withdraw cate (i.e., disclaimer) the meaning of these the rankings should be seriously considered. regulatory indicators. Management added that Another member suggested that a best practice it was very important for the Bank to communi- component could be added, while another cate what it is that we are measuring and what member noted that benchmarks are not unique we are not. Some speakers noted that they to DB and are part of a wave of international would like clear disclaimers and comments on standards. A speaker observed that measure- the social value of good regulation. Management ments for these rankings change after the remarked that the DB report has a "health release of the DB report. Management added warning." It added that it would look to further that when comparing the exercise to a similar explain that deregulation is not the main evaluation at the OECD (i.e., Product Market purpose of the DB report. Regulation for OECD Countries), in the case of regulatory complexity and costs, the OECD and Regulation Enforcement and Impact. A few mem- DB rankings are highly correlated. IEG noted bers commented that indicators measuring that improvement in DB country rankings regulation effectiveness must also measure should not be characterized as improvements impact on the ground. Management noted that in the business climate; rather, they should the DB report not only looks at the level of be interpreted as a partial indicator of a regulation but also at the compliance cost and reduction in the regulatory burden. how this affects local entrepreneurs. It also added that there is a dimension of enforcement Paying Taxes Indicator. A few members noted that and implementation that is being observed by the tax rate issue was an important one to flag and comparing DB data with data from the should be considered. A member noted that it Enterprise Surveys, which capture the experi- would be useful to retain the tax rate as part of ence of actual business owners. the paying taxes indicator, while another member commented that the term "total tax" was Country Rankings. Some members noted that misleading. IEG noted that the reliance on a sole country rankings needed further work, as a source for information underlying the paying change in rankings does not necessarily improve taxes indicator is risky. IEG recommends that the regulatory environment, thus making the DB continue to gather and publish important exercise arbitrary. Management commented information about taxes that firms pay, but that in some cases the DB indicators look for discontinue factoring this into the overall more regulation (i.e., protecting investors); rankings. IEG commented that depending on a while other indicators will give the lowest country's resources and fiscal requirements, xxvi C H A I R P E R S O N ' S S U M M A RY: C O M M I T T E E O N D E V E L O P M E N T E F E C T I V E N E S S ( C O D E ) lowering taxes could prove to be negative for the ongoing DB work is focusing on issues related to investment climate, and indirectly for individ- potential gender discrimination. ual firms. A member stressed that the tax system should be taken in its totality to evaluate the Increasing Informants. Members noted that manage- associated burden or ease for undertaking ment should use a strategic approach to selecting business in a country. A speaker disagreed with and increasing the number of informants (e.g., the proposed changes to the paying taxes indica- accounting, tax experts) for DB. A few speakers tor, as tax remains one of the major business added that these should include relevant constraints for business development (i.e., in the stakeholders (e.g., employers, consumers). IEG Africa Region) and argued that focusing only on commented that DB should disclose how many the administrative burden of paying taxes will informants are the sources for each indicator, weaken the DB methodology. Another speaker and that the reliability of the ratings would noted that the issue was not the tax rate per se, improve if this number were increased. but at which threshold it became a burden to business activity. Management noted that it did Nomenclature. Speakers noted that titles for DB not agree with excluding the tax rate. It noted indicators should more clearly reflect what they that in the Enterprise Surveys, one of the issues are measuring. IEG recommended that the DB most raised by business owners is the tax burden. report be precise in the language used to describe what is being measured, as the names Employing Workers Indicator. A few speakers of the many indicators may overstate the scope thought that the WBG should encourage of their coverage. In the case of employing countries to guarantee internationally recognized workers, it covers specific rules about the hiring worker rights and that DB should look to reflect and firing of workers and hours of work, but certain labor standards (e.g., collective bargain- does not cover other critical areas (e.g., union ing, forced labor). Management noted that the rights, child labor). employing workers indicator is consistent with ILO standards. IEG noted that there are New Indicators and Subnational Regulation. A anomalies in the employing workers ranking, member suggested including the cost of regula- as the top country is Singapore, with very good tions on foreign direct investment within DB. regulations, while the Marshall Islands, with no Management noted that it was in the early regulations, is second. A speaker commented stages of piloting a foreign direct investment that no one would confuse Singapore with the indicator. IEG and management agreed that it Marshall Islands when it comes to labor market is important to pilot-test and validate new regulation. Management noted that by focusing methodologies before introducing any new on the regulatory costs associated with employ- indicators. A member proposed that subnational ing workers, this indicator provides a basis for government regulation be looked into. Manage- analysis of how regulation relates to important ment agreed with the idea of looking into outcomes such as informality or higher levels of subnational-level regulation (i.e., across cities) women or youth employed. In that sense, a DB and noted that it already has such a program indicator triggers a conversation about a range and now covers over 200 cities globally with of issues related to it. Management added that subnational reports. Giovanni Majnoni Acting Chairman xxvii Chapter 1 Evaluation Highlights · The Doing Business (DB) indicators provide consistent cross-country data annually on 10 specific aspects of a country's regulatory framework. · DB indicators are based on research that associates better regulations with an improved investment cli- mate, and thence with economic growth--but this research is still nascent. · Seven DB indicators presume that less regulation is better. · Five emphasize aspects of debt enforceability and availability of collateral. · DB argues that regulatory reform will encourage informal businesses to formalize. · DB indicators do not aim to capture the potential benefits of regulation. Users must be mindful of what the DB indicators measure and what they do not. Woman standing outside restaurant, Keur Moussa, Senegal. Photo reproduced by permission of Philippe Lissac/Godong/Corbis. The Ideas Behind the Indicators T his chapter situates the DB indicators in the context of investment cli- mate and private sector development. It introduces the main principles that shape what the indicators measure and how they are constructed. tion of child labor. Others--such as requiring Role of the Investment Climate multiple official stamps on a document--deliver in Private Sector Growth little or no public benefit. They simply provide Private sector growth is essential for developing officials with opportunities for rent-seeking. The countries to create jobs and raise incomes. The policy maker's challenge is to find the level of rate and nature of private sector growth in a regulation where the desired level of public country is affected by many factors, including good--say, tax revenues or worker safety--can be macroeconomic and political stability, traditions obtained with the minimum loss of efficiency to and culture, physical infrastructure, availability of affected firms. Some countries may be over- capital, and human resources. Institutional, regulated; others may be under-regulated. The policy, and regulatory factors also play an level of regulation in any country should reflect a important role. They are often grouped together country's preferred trade-off between public under the rubric of "investment climate," as goods and private (firm) benefits (see box 1.1). depicted in figure 1.1. Substantial research literature has The DB indicators are DB measures selected aspects of the investment established an association between the premised on research that climate--namely, the laws and regulations characteristics of the business regula- associates characteristics governing how firms do business (see shaded tory environment and the performance of the investment climate areas of figure 1.1). Research suggests, broadly of firms, and thence to macroeconomic with growth, and the speaking, that the regulatory framework does outcomes (see Acemoglu and Johnson effects of laws and matter for economic outcomes, but it is 2005; Botero and others 2004; Djankov regulation on the inconclusive about which regulations matter and others 2002; Hall and Jones 1999; investment climate. most, and how much they matter compared with Kaufmann 2002; Kaufmann and Ziodo- other determinants (Dollar, Hallward-Dreimeier, Lobaton 1999; Kaufmann, Kray, and Mastruzzi and Mengistae 2005). 2006; Knack and Keefer 1995; Rodrik 2004). DB's Web site contains a comprehensive bibliography of Business laws and regulations are intended to this research, to which the DB team itself has generate benefits to society at large, but they also contributed.1 Generally, this work uses cross- inevitably impose costs on the individual firm. country comparisons to show that various proxy Some regulations on firms may deliver an indicators for governance or business regulation important public good--for example, the prohibi- are associated with the size or performance of 3 DOING BUSINESS: AN INDEPENDENT EVALUATION Figure 1.1: DB Measures Selected Aspects of Investment Climate Stability and Regulation and Finance and Workers and security taxation infrastructure labor markets Stability Regulation Finance · Foster a skilled and healthy · Reduce political instability from · Balance market and government · Foster competition in the banking workforce by expanding access civil wars, political conflict, etc. failures for a good institutional fit sector to education, improving education · Maintain macroeconomic · Address regulatory cost and · Control risk-taking by banks and quality, supporting life-long stability with low inflation, informality other financial institutions learning, and the like sustainable budget deficits, · Reduce uncertainty and risk · Secure rights of borrowers, · Help workers affected by large- and realistic exchange rates in interpretation and creditors, and shareholders scale restructurings by reinforcing implementation of existing · Improve credit information by social insurance mechanisms and Security of property rights regulations using credit bureaus and stronger reaching out to the large share · Reduce robbery, fraud, and other · Reduce regulatory barriers to data protection and credit- of workers in rural and crimes against property competition reporting laws informal economies · End uncompensated · Craft labor market interventions expropriation of property Taxation Infrastructure to benefit all (formal and informal) · Verify rights to land and other · Broaden tax base · Improve climate for investment in the process of setting wages, property · Increase autonomy of tax in infrastructure by securing regulation of working conditions, · Facilitate contract enforcement agencies investors' property rights, and hiring and firing of workers · Reduce corruption in tax fostering competition, and administration encouraging private participation · Confront informality · Improve public management of · Simplify tax structure infrastructure · Improve customs administration · Improve compliance through computerization Source: World Bank 2004. Note: Shaded areas are those the DB indicators attempt to measure. the private sector or overall macroeconomic and Business Enterprise Surveys and the World outcomes.2 Economic Forum's Global Competitiveness Reports ask business leaders to rank the most Business is affected not only by laws and regula- important constraints they face. In this evalua- tions, but also by a host of other variables tion's 13 case study countries, these assess- outside the scope of the DB indicators. The ments note 12 important constraints (see table Bank's Investment Climate Assessments (ICAs) 1.1).3 Business leaders most often mentioned Box 1.1: A Good Investment Climate Balances Private and Societal Interests The World Development Report 2005: A Better Investment Cli- and can therefore improve the investment climate and pro- mate for Everyone, addressed the trade-offs between private and tect other social interests [p. 6]. social interests: A good investment climate is not just about generating At the heart of the problem lies a basic tension. . . . Most profits for firms--if that were the goal, the focus could be firms complain about taxes, but taxes finance public ser- limited to minimizing costs and risks. A good investment vices that benefit the investment climate and other social climate improves outcomes for society as a whole. That goals. Many firms would also prefer to comply with fewer means that some costs and risks are properly borne by regulations, but sound regulation addresses market failures firms [p. 2]. 4 THE IDEAS BEHIND THE INDICATORS access to and/or cost of financing, corruption, country's idiosyncratic issues. Nor Investment decisions also lack of infrastructure, inefficient government can the DB indicators be expected to depend on other bureaucracy, and tax rates. Four of these 12 help determine policy actions, variables not measured constraints are reflected in the DB indicators: because they cannot situate the by the DB indicators-- inefficient government bureaucracy (that can regulatory constraints within a notably, the cost and include regulatory constraints); tax rates and tax country's policy context and macro- access to finance and administration; and restrictive labor regula- economic framework. ICAs and other infrastructure and labor tions.4 Interviews with Bank Group staff and surveys are better suited to playing skills. stakeholders broadly confirm this analysis: they both these roles. rated lack of infrastructure, access to and cost of credit, and a shortage of human capital as the But research overall is inconclusive about the three most important constraints to private direction of causation. While it is typically sector development. 5 hypothesized that better regulations spur better economic results, causality may also run the The DB indicators--confined as they are to a opposite way, insofar as citizens in more subset of these factors--do not and cannot be advanced economies demand more efficient expected to identify priority action areas across regulations. There may also be unidentified the business climate as a whole. For example, in causal factors. For example, if cross-country low-income and post-conflict countries, such as analysis finds that higher labor productivity is Burundi, political insecurity was an overarching associated with less onerous business start-up constraint. As a standardized cross-country data procedures, it may be that a third factor (such set, the DB indicators also cannot elicit any one as the quality of human capital) is driving indica- Table 1.1: DB Covers Only Some of the Top Constraints to Business (Constraints mentioned by business leaders; those in bold are covered by DB) World Bank Group Investment Climate Assessments Top constraints and Enterprise Surveys;a Global Competitiveness Reportb Access and/or cost of financingc Algeria, Burundi, China, Moldova, Nigeria, Rwanda, Tanzania, Vietnam Corruption Albania, Algeria, Burundi, China, Moldova, Mongolia, Nigeria, Vietnam Inefficient government bureaucracyd Albania, Algeria, China, Moldova, Mongolia, Netherlands, Peru, Spain Infrastructure (such as electricity, transportation) Albania, Netherlands, Nigeria, Rwanda, Tanzania, Vietnam Tax rates Albania, Moldova, Mongolia, Rwanda, Spain, Tanzania Tax administration Burundi, Mongolia, Peru Anticompetitive or informal practicesd Albania, Peru, Spain Restrictive labor regulations Netherlands, Spain Skills and education of available workers Spain, Vietnam Political instability Burundi, Peru Macroeconomic instability Moldova Economic and regulatory policy uncertainty Peru Sources: World Bank Investment Climate Assessments and Business Enterprise Surveys (2004­07); World Economic Forum 2007/08. Note: Management notes that tax rates are the fifth most widely cited constraint by businesses. This highlights the importance of having a measure of tax burden in the paying taxes in- dicator. In all World Bank Enterprise Surveys done since 2006, 17 of 40 find the tax rate to be among the top 3 obstacles and 33 of 40 find them to be a bigger obstacle than tax administration. a. Respondents were given a list of 18 factors and asked to rate them on a scale of 1 to 5: 0 = no obstacle, 1 = minor obstacle, 2 = moderate obstacle, 3 = major obstacle, and 4 = very severe obstacle. The top three factors--with the most respondents commenting that the factor was either a major obstacle or very severe obstacle--are displayed in the table. b. Respondents were given a list of 14 factors and asked to select the 5 most problematic for doing business in their country and to rank them between 1 (most problematic) and 5. Re- sponses were weighted according to their ranking. The top 3 constraints by country are displayed in the table. c. Includes several factors, such as access to banking and credit services, interest rates, as well as availability of collateral as measured by the DB indicators. d. Constraint is not offered as a potential response in one of the surveys. 5 DOING BUSINESS: AN INDEPENDENT EVALUATION tors for economic performance and quality of Three Principles Underlying What DB public administration in the same direction Measures (based on Altenburg and von Drachenfels The DB model is anchored in the research 2006). described above that associates firm perform- ance and economic growth with characteristics Research affirms Recent research has begun to test the of the regulatory environment.9 DB puts these associations between links between the DB indicators and associations into a normative framework by outcomes and the economic outcomes, although this selecting 10 categories of laws and regulations to regulatory environment... research has been constrained by be measured, devising a procedure (discussed in DB's short time series. Djankov, chapter 2) for rating and ranking countries, and McLiesh, Ramalho, and Shleifer (2008) looked at deriving diagnoses and recommendations for reforms in getting credit, and found that credit policy makers to use (see chapter 4). rises after improvements in creditor rights and information. Commander and Svejnar (2007)6 The content of the DB indicators as a group found little evidence that the DB indicators have embodies three important ideas: a robust relationship with business environment constraints and firm performance, as measured 1. Less regulation is preferable across all parts of by revenue efficiency.7 the distribution and in all countries.10 The rat- ings do not allow for a minimum desirable ...but suggests a need for A background paper commissioned level of regulation needed to ensure public caution in attributing for this evaluation (Commander and benefits. This principle is embedded in 7 of the economic outcomes to Tinn 2007) found no statistically signif- 10 indicators11 and is especially prominent in changes in DB indicators. icant relationships between the 2004 the following 3 indicators:12 DB indicators and growth rates. It Employing workers: The fewer the restric- found few significant relationships with interme- tions on hours of work and the more easily a diate outcomes.8 For example, better legal rights firm can lay off redundant workers, the better for creditors and debtors were positively associ- the ranking. The 10 top-ranked countries in- ated with private credit, capital inflows, and clude 5 developed countries with high-quality foreign direct investment, but not with private labor laws, but also 5 small island states, some bank credit. There was a weak association with inadequate labor protections (see box between investment and dealing with licenses 3.6). Thus the indicator cannot capture the and enforcing contracts. No significant associa- possible offsetting benefits of job protection. tion emerged between registering property and Dealing with licenses: The fewer the steps construction, trading across borders and needed to get a permit to construct a build- exports and imports, starting a business and the ing,13 the higher the score. Possible benefits size of the informal economy, or employing from safety and environmental checks are not workers and employment. considered. Paying taxes: The lower the overall tax rate Finally, a recent analysis found no significant as a share of a firm's profit, the higher the relationship between reforms as measured by score. Among the 10 top-rated countries on this changes in the DB indicators and aggregate invest- indicator are Maldives, Oman, Singapore, and ment and unemployment rates (Eiffert the United Arab Emirates. Each of these has Seven of the 10 DB 2007). Because of the relative newness special characteristics that make it an unsuit- indicators exhibit a of DB data, as well as other limitations able role model for other countries seeking an preference for less described in chapter 2, this research optimal level of corporate taxation (see box 3.5 regulation--3 most cannot be considered definitive, but it in chapter 3). For instance, Maldives derives al- notably: employing suggests a need to be cautious in most all of its revenue from resort leases. The workers, paying taxes, and attributing economic outcomes to indicator overlooks each country's fiscal re- dealing with licenses. changes in the DB indicators. quirement to raise revenue, as well as the 6 THE IDEAS BEHIND THE INDICATORS Table 1.2: The 10 DB Indicators and Their Components Indicator Subindicators Indicator Subindicators Starting a business Procedures (number) Getting credit Strength of legal rights index (0­10) Time (days) Depth of credit information index (0­6) Cost (% income per capita) Public registry coverage (% of adults) Minimum capital (% income per capita) Private bureau coverage (% of adults) Dealing with Procedures (number) Enforcing Procedures (number) licenses Time (days) contracts Time (days) Cost (% income per capita) Cost (% of claim) Employing workers Difficulty of hiring index (0­100) Trading across Documents to export (number) Rigidity of hours index (0­100) borders Time to export (days) Difficulty of firing index (0­100) Cost to export (US$ per container) Firing cost (weeks of salary) Documents to import (number) Nonwage labor cost (% salary) Cost to import (US$ per container) Time to import (days) Registering property Procedures (number) Paying taxes Payments (number per year) Time (days) Time (hours per year) Cost (% of property value) Total tax rate (% of profit) Protecting investors Extent of disclosure index (0­10) Closing a Recovery rate (cents per dollar) Extent of director liability index (0­10) business Time (years) Ease of shareholder suits index (0­10) Cost (% of estate) Note: Italicized items are measured but not included in calculating the ease of doing business (EODB) ranking. equity implications of alternative sources forceability of debt contracts and availability of of revenue (Maldives government 2007). collateral. Getting credit: The fewer the restrictions on Defining the point at which the costs of regula- what can be counted as collateral (and the more tion exceed the benefits is difficult. It is likely that information lenders can obtain about borrowers' some developing countries impose more regula- credit histories), the more likely lenders are to tions than would be optimal for economic make loans and to be able to collect on them. development. Some level of regulation is helpful Enforcing contracts: The more efficiently the for ensuring a supply of social goods such as court system operates, the more easily a firm health and safety, environmental protection, and will be able to collect on a debt. transparency of business dealings. Regulations Registering property: The easier it is to regis- may even have benefits for individual firms-- ter a property, the more likely the owner can insofar as they ensure a level playing field with use it as collateral for a loan, and the competitors, for example. Yet seven of the DB more likely the lender will collect on Five of the 10 indicators indicators, taken to their logical conclusion, a bad loan. measure the would give the highest ranking to countries with Closing a business: The easier it is enforceability of debt the least regulation.14 to close a business through formal contracts and bankruptcy (instead of simply ceas- availability of collateral, 2. Property rights and debt enforceability are impor- ing operations), the greater the like- but leave out some tant determinants of lending and investment. Five lihood that creditors can collect on factors that affect firms' of the 10 indicators include measures of the en- their loans. use of credit. 7 DOING BUSINESS: AN INDEPENDENT EVALUATION Dealing with licenses: The easier it is to allow people and firms in the informal economy face a constructed warehouse to be used as collat- severe entry barriers caused by their low skills, lack eral, the higher the score. of access to capital, isolated location, and other structural factors. A recent review noted that DB's emphasis on collateral and debt enforceabil- "Empirical studies show that only a very small ity derives in part from the work of Hernando de number of micro-enterprises ever manage to Soto, which posits that poor property owners are upgrade and grow into larger units. The reasons locked out of the formal economy because they are manifold. Micro-entrepreneurs may, for lack legal rights to their land, so they cannot use it instance, lack information, technical skills, as collateral for loans to expand their businesses managerial competence, entrepreneurial spirit, or improve their properties. De Soto's work, and capital. . . . To graduate out of informality is however, omits the many other factors that affect thus a slow and difficult process of cultural change" firms' actual use of credit, such as interest rates, (Altenburg and von Drachenfels 2006, p. 406). value of the assets, degree of intermediation, culture, and the existence of viable entrepreneur- DB focuses on the idea that excessive regulation ial opportunities (see Galiani and Schargrodsky of private sector activity inhibits the transition 2005; Commander and Tinn 2007). Peru's large- from the informal to the formal economy. scale titling program, COFOPRI, did not induce Although some research shows that "countries the beneficiaries to solicit credit any with heavier entry regulation have lower firm The DB indicators argue more frequently than nonbeneficiaries, entry and lower growth . . . there is very little that lighter regulation and credit applications from beneficiar- evidence on the actual effects of business and less taxation ies were turned down more frequently registration reform" (Bruhn 2007, p. 1). Recent encourage informal firms than those from nonbeneficiaries research in Mexico finds that simplifying to move to the formal (Webb, Beuermann, and Revilla 2006). business registration procedures was associated economy, but the In Peru, lenders proved interested in with an increase in newly registered businesses, literature is inconclusive the applicant's repayment capacity, not although the new business owners were about whether these whether they had collateral, which formerly wage earners rather than unregistered factors can cause such often has low resale value (see Morris business owners (Bruhn 2007). change. 2004). Two Principles Underlying DB's 3. Lighter regulation and taxation can encourage non- Methodology formal firms to shift into the formal economy. DB's methodology has two distinctive character- Starting a business: Simpler procedures to istics. It uses a set of discrete indicators to create start a business will encourage informal en- an aggregate ranking, and it is applied exclusively terprises to formalize. to domestically owned firms in the formal sector. Paying taxes: The easier the tax-paying pro- cedures, the more likely a firm is to actually pay, Discrete and aggregable indicators rather than evade, taxes. DB separately assesses discrete dimensions of Employing workers: The fewer the restric- the regulatory environment. This is aimed at tions on hours of work (within limits) and the providing policy makers with specific, actionable more easily a firm can lay off redundant work- information. The 10 dimensions are not equally ers, the more likely it is that firms will employ important in all countries. To illustrate: workers on formal rather than on informal terms. · Protection of minority shareholders, as meas- ured by protecting investors, was deemed less The research literature is inconclusive important in several client countries with more DB aims to provide policy about why the informal sector exists pressing constraints, such as lack of infra- makers with specific, and persists. One explanation is simple structure and access to finance (see below). actionable information. tax avoidance. Another view is that · The rules measured by employing workers are 8 THE IDEAS BEHIND THE INDICATORS more important in countries with substantial single composite ranking called the The relevance of each formal employment and in countries with or- ease of doing business (EODB). Like indicator will necessarily ganized labor groups than in those where most any composite index, the EODB vary by country. people are small farmers. obscures its component information. The weights of the indicators are not important This evaluation asked Bank Group staff and since the ranking does not change much with country stakeholders to rank the importance of alternative weights.15 Rather, the change in the 10 DB indicators to private sector growth in ranking for any country is driven largely by their countries. While there were significant where a country is located on the distribution variations between countries, in aggregate over of countries on a specific indicator (discussed half the respondents rated eight dimensions as in chapter 2). Even within DB's own frame of very important or important to the growth of reference, the composite indicator would more the private sector in their countries. Only two accurately (though less attractively) be named indicators (protecting investors and closing a "index of regulatory burdens," since it does not business) were found to be slightly important or capture all dimensions of doing business. not important by more than half the respon- (Other issues of nomenclature are discussed in dents interviewed. chapter 3.) The indicators themselves cannot capture Covers formal, domestically owned firms country context, precisely because they are The DB exercise gathers information about a designed to allow cross-country comparisons on particular subset of a country's private sector the basis of uniform criteria. By the same token, activity--the regula-tory environment facing not all reforms will have an equal impact, and the domestically owned firms operating in the formal DB indicators are not designed to identify within- sector. The scope of this subset is defined by the country priorities. Users require supplemental specific information that informants are asked to information to determine the importance of provide (for example, what the law requires as each indicator or reform in a particular country distinct from what may actually happen) and the setting. Stakeholders interviewed for the evalua- characteristics of the hypothetical firms in the tion stressed the DB indicators' limitations in stylized cases (for example, ownership, annual helping countries select priority reforms (see turnover, or minimum number of chapter 4). employees). The DB reports appropri- The DB reports explain ately explain the scope and limitations the limitations of its DB aggregates country rankings for the 10 of this coverage (World Bank-IFC approach. discrete elements, weighted equally, into a 2007b, p. 67). Table 1.3: What DB Covers DB includes DB excludes Small and medium-size firmsa Microenterprises and state-owned enterprises Enterprises in the formal sector Enterprises in the informal sector Domestically owned firms and investors Foreign-owned firms and foreign investors Official and legal transactions and processes Illegal, corrupt, informal, and out-of-court transactions and processes Firms in the capital city Firms outside the capital city Limited liability companies Sole proprietorships a. The size of the firm varies, depending on the indicator. It ranges from 20 employees for dealing with licenses, 50 employees for starting a business and reg- istering property, 60 employees for paying taxes, to more than 201 employees for employing workers, trading across borders, and closing a business. Three in- dicators (getting credit, protecting investors, and enforcing contracts) do not specify firm size. 9 DOING BUSINESS: AN INDEPENDENT EVALUATION Table 1.4: The Informal Economy Casts a Long Shadow Region Number of countries Shadow economya (% GDP) Africa 24 41 Asiab 26 26 Latin America 17 41 Eastern Europe and Central Asia 23 38 OECD 21 17 Source: Schneider and Klinglmair 2004. a. The shadow economy includes unreported income from production of legal and illegal goods and services, either for monetary or barter transactions. b. This number is affected by the relatively low levels of informal activity in China (13.1 percent), Hong Kong (16.6 percent), Japan (11.3 percent), and Singa- pore (13.1 percent). Some regulatory DB's coverage explicitly excludes In summary, the thrust of the DB is broadly constraints are likely to some types of enterprises and transac- consistent with credible research that more be relatively unimportant tions (see table 1.3). Some of the laws efficient business regulation is associated with for informal and and regulations that apply to small and better private sector performance, and thence microenterprises . . . medium-size domestically owned macroeconomic outcomes. But the literature is firms may well apply to other kinds of necessarily partial, as it has not yet demonstrated businesses, such as large-scale enterprises or the direction of causality. Furthermore, regula- those with foreign ownership. But the regulatory tions deliver benefits as well as costs, and the constraints that DB measures are likely to be policy choices countries make are necessarily relatively unimportant for informal and microen- based on the trade-offs between the two. What is terprises, simply because they are more likely to good for a firm may not be good for firms as a conduct business without recourse to courts, group, or for the economy as a whole. formal credit providers, and taxes. The DB exercise reflects these inherent trade- . . . which in some In many countries, the informal sector offs. As a cross-country comparison, DB is not countries account for a accounts for a significant or even intended to, and cannot, capture country significant share of dominant share of private sector nuances and nonlinear relationships. It notes the private sector activity. activity, especially in low- and middle- costs of regulation, but not the benefits.16 Seven income countries (see table 1.4). of DB's 10 indicators presume that reducing Observers have criticized DB for not capturing regulation is equally desirable whether a country the important constraints on nonformal and starts with a little or a lot of regulation. While microenterprises. This observation, while true, is these limitations do not invalidate the exercise, somewhat off the mark. As discussed above, DB is they underscore the need to use the DB indica- based on the view that informal firms and transac- tors cautiously and in conjunction with comple- tions should eventually enter the formal mentary tools, such as Investment Climate economy, and that this is more likely to occur if Assessments, when measuring a country's invest- the burdens on firms in the formal sector are ment climate or related measures of develop- reduced. ment effectiveness. 10 Chapter 2 Evaluation Highlights · DB collects its data largely from lawyers and accountants deemed knowledgeable about a country's laws and regulations. · The number of informants who fill in questionnaires on any topic in a country is small. · DB does not keep track of those who are invited but do not participate; thus participant bias cannot be estimated. · The DB team validates and adjusts informants' data, which makes the data difficult to verify. · DB regularly updates its data to re- flect changes in methodology or cor- rect errors, but previously published data sets are not made available to users, and the impact of such changes on the overall and indica- tor rankings is not stated. · For a given amount of change in an indicator, a country's ranking may change a little or a lot, depending on its initial position in the ranking. Workers in a furniture factory, Cotonou, Benin. Photo reproduced by permission of Jorgen Schytte/Still Pictures. Collecting Information and Constructing the Rankings D oing Business creates its annual country rankings using information supplied by persons deemed knowledgeable about selected laws and regulations in each country covered. The DB team identifies individ- ual lawyers, notaries, officials, and firms and requests that they provide in- formation on one or more specified DB topics. Since the process is not based on a survey sent to a large group, but rather on information solicited from se- lected individuals, the term "informants" is used in this evaluation instead of "respondents." The validity of the DB indicators depends on naire; they are called "questionnaire informants"; how representative, reliable, and objective its others were consulted by DB to confirm or clarify process is for obtaining, recording, and analyzing selected points; they are called "supplemental information (see Dorbec 2006). This chapter informants." assesses DB's processes for interviewing inform- ants, reviewing and validating the information The number of informants on each Number of informants for they supplied, and constructing the rankings.1 topic in a country is small. For its 2007each topic in a country is report, DB received, on average, small. The Number of Informants between one and four completed The 2007 and 2008 DB reports note that about questionnaires per topic and consulted with up 5,000 individuals provided information for the to three supplemental informants per topic for indicators. It is frequently the case that several the 5 focus indicators in the 13 randomly individuals from the same firm or office help selected countries of this evaluation, as sum- prepare the firm's response to the questionnaire. marized in table 2.1.2 For example, junior staff may obtain data for a partner or principal to compile into the firm's The starting a business indicator has the most written submission. The DB reports list each of informants in each country--3.5 on average-- these individuals as informants, but in this perhaps because it has been used since 2004 and evaluation, each completed questionnaire is the questionnaire is relatively simple. But for the counted as one informant, irrespective of how paying taxes indicator, there is a single survey many individuals helped to prepare it. Some of informant in 142 countries--the local or the listed individuals completed a DB question- regional office of the global accounting firm 13 DOING BUSINESS: AN INDEPENDENT EVALUATION Table 2.1: Average Number of Completed Questionnaires per Indicator in Each Country Is Low DB dimensions Getting credit Employing Enforcing Legal Private/public Paying Starting a workers contracts rights credit bureau taxes business Average number of questionnaire informants 1.7 1.8 1.5 1.5 1.0 3.5 Average number of supplemental informants per country 0.2 1.3 0.6 0.0 3.0 1.5 Note: Averages are calculated by dividing total informants by 13 countries, except for getting credit - public/private credit bureau, which includes 10 countries because Albania, Moldova, and Tanzania do not have credit bureaus. PricewaterhouseCoopers LLC (PwC), with impair informants' objectivity. Informants in several people from each office contributing to Madagascar, for instance, are routinely in direct one questionnaire for each country. This is contact with the government department respon- because DB has established a partnership with sible for enhancing the investment climate. A PwC's global tax practice (described in chapter larger number of informants would dilute the 3), in which PwC is the sole informant on the influence of any self-interested responses. paying taxes indicator (except in the 33 coun- tries where PwC does not participate at all). The DB actively seeks new informants through information for the enforcing contracts and referrals from existing informants, IFC and Bank employing workers indicators is also based on contacts in borrowing countries, and local not more than 2 completed questionnaires in business directories. The DB Web site now each of the 13 countries. invites prospective informants to register their interest in becoming data contributors. But there Where information is To the extent that DB collects factual are no clearly stated criteria or processes for factual, the small information, as distinct from opinion seeking and selecting informants.3 To ensure that number of informants or perception, it arguably does not these efforts pay off in progressively more may not matter, but need a large number of informants to reliable data, DB should consider: (a) targeting subindicators on time lessen the source of error, as do them to the countries and indicators most and costs include perception or opinion surveys. But, as needing increased reliability and (b) establishing informed estimates. will be noted in chapter 3, not all of selection criteria and numerical goals for new DB's information is purely factual. The informants. It should report systematically on time and cost subindicators, for instance, require these activities. informants to make estimates based on their experience. Increasing the number of informants Qualifications and Motivations would reduce both the risk of erroneous factual of Informants information from a single informant and the More than two-thirds of DB informants in the 13 errors inherent in questions requiring informants' countries reviewed by this evaluation are lawyers judgments. An additional potential risk noted by from private firms, as shown in figure 2.1. Eight some Bank Group staff is that interested parties percent are accountants, most from PwC could seek to influence informants' responses to accountancy offices.4 improve their country's ranking. There is no evaluative evidence suggesting deliber- Informants interviewed for the evaluation were There are no clear ate manipulation of data, but close rela- professionally engaged on the topic on which selection criteria for tionships among lawyers and officials, they provided information.5 The DB team has informants. especially in smaller countries, could noted that in small and/or low-income countries, 14 COLLECTING INFORMATION AND CONSTRUCTING THE RANKINGS Figure 2.1: The Majority of DB tries with few specialized informants Generally, informants in Informants Are Lawyers and or a weak informant base. DB should the case study countries Accountants also consider developing a systematic were professionally procedure or set of criteria for assess- engaged on the topic they ing informants' qualifications. addressed. Government DB does not pay its informants. It simply Private official acknowledges the participating individuals in its sector 18% publications (except for the approximately 10 lawyer percent who do not wish to be publicly named).6 70% Assisting DB can require considerable effort; Accountant­ informants interviewed for the evaluation said PwC 7% they spent between one hour and one month on Accountant­ the exercise. The paying taxes and getting credit Other non-PwC 4% indicators are the most demanding of inform- 1% Note: Based on 141 questionnaire and supplemental informants. "Other" ants' time. includes members of the private sector and one administrative staff mem- ber and graduate student in China. Why, then, do the informants participate? About half of those interviewed for the evaluation, as it sometimes must rely on lawyers in general shown in figure 2.2, said participation would practice because there are no lawyers with enhance their firm's credibility or prestige (46 specialized practices in bankruptcy, civil claims, percent). Another third said they wanted to or other relevant areas. In Burundi, for example, share their experience (33 percent), and the rest one informant was primarily a criminal and family said they were interested in the intellectual lawyer with minimal experience in corporate law, exercise, had free time, or were asked and some informants for closing a business were to participate by somebody else. Informants generally unaware of the country's bankruptcy law. DB Several lawyers mentioned they had participate for prestige or could reduce the risk of error through consult- gained clients who heard of them to share experience. ing in advance with Bank and IFC country through the questionnaire, although counterparts in identifying informants and by attracting clients was not their objective in partic- undertaking greater quality assurance in coun- ipating. In some Organisation for Economic Co- operation and Development (OECD) countries, lawyers noted that the time spent on DB counts Figure 2.2: Why Do You Participate? toward the firm's commitment to provide pro For Prestige and to Share Expertise bono services. Prestige Not all those invited to participate actually do so; Requested by 46% colleague some decline, others agree but fail to follow 5% through. DB does not keep track of nonpartici- pants and why they do not participate. If nonpar- ticipants are systematically different from the actual informants, there is a possible selection Share bias, whose direction depends on the reasons for experience participation. For example, if informants are more 33% Intellectual likely than nonparticipants to be concerned exercise 10% Free about excessive business regulation in their time country, the data may overstate the regulatory available burden. If nonparticipants from small firms tend 5% to decline because they lack junior staff to help 15 DOING BUSINESS: AN INDEPENDENT EVALUATION complete a complex questionnaire, the data may with the informants and/or review of legislation. understate the regulatory burden, because the In the case of time estimates for enforcing informants will tend to be those with more contracts, the median value did not appear to capacity to help clients navigate the bureaucracy. have been selected.8 Informants in Nigeria gave Finally, if nonparticipants tend to be those too broad estimates of time, so DB staff calculated it busy to do pro bono work, while informants are based on a review of the current changes in those with a lighter workload, the responses may legislation. In Mongolia, one informant's re- be less reliable to the extent that the informants sponse was discarded in favor of another that are those less well established in their field. reported no change since last year. Cost indica- tors may also require DB to select from different DB does not keep track of Like most surveys and polls, DB estimates. Doing Business 2007 stated that in those who are invited to involves participant bias, and it needs Spain, enforcing a contract cost 15.7 percent of participate but do not, to systematically learn and disclose the disputed debt, while the two informants had which makes it difficult more about the magnitude and estimated 15.1 percent and 18.9 percent. DB to identify and disclose directions of nonresponse. In this explained that the higher estimate was the nature and direction regard, it may be useful to systemati- disregarded because it came from a large interna- of participant biases. cally collect and track information on tional law firm. The estimate has subsequently the number of informants who were been revised to 17.2 percent, close to the median contacted, but who did not qualify as informants, of the two original responses. were not interested in participating, or refused to participate for other reasons. DB should also DB's close attention to individual data points and consider diversifying its informant base to its resolution of differences and anomalies include business consulting firms, associations, undoubtedly help improve the quality of the and think tanks that meet the selection criteria database. The process also helps identify and to be developed. weed out any unreliable informants. The DB team's organization along topic lines permits Validating the Data staff to develop a feel for plausible levels and Once the informants' questionnaires ranges of the indicator values. At the same time, The DB team validates are received, the DB team validates a risk is created by the considerable reliance and adjusts the data the information based on documents placed on the decisions of DB staff to accept, based on documents and and consultations with supplemental overrule, or select among informants' replies, supplemental informants, informants. Typically there are four because this makes it difficult to verify or but this makes it difficult rounds of interaction between the DB replicate the data. This risk is partly mitigated by to verify the data. team and the informants, involving a validation process in which DB sends proposed conference calls, written correspon- country data to relevant Bank and IFC staff and dence, and in some cases a country visit. When to country authorities (through their executive informants' estimates of time differ, DB states directors) for comment. that it selects the median value (World Bank-IFC 2006b, p. 61). Bank Group staff and in-country stakeholders in 6 of the 13 countries reported dissatisfaction The evaluation reviewed the differences be- with the DB's process for validating the data.9 tween the information provided in the com- Albanian officials, for instance, considered that pleted questionnaires and the data the rankings for starting a business, enforcing Country stakeholders and points as reported.7 For employing contracts, and protecting investors do not Bank Group staff in 6 out workers and getting credit, in all seven square with the facts on the ground and found of 13 countries reported case study countries, Doing Business the response of the DB team to their rebuttals dissatisfaction with DB's 2007 published values that differed "not satisfactory." DB received 115 specific process for validating from at least one questionnaire, based challenges and clarifications to its 2007 report data. on further consultation by the DB staff from 50 country teams; 21 percent of these 16 COLLECTING INFORMATION AND CONSTRUCTING THE RANKINGS challenges were accepted.10 Some Bank Group plained that for three of these indicators (pay- staff noted that despite recent improvements, ing taxes, enforcing contracts, and dealing DB still gives them insufficient time (just a few with licenses), some of the changes may also re- days in some cases) to review draft DB data. flect data corrections, but that "it is difficult to Some interviewees considered that their team's separate corrected errors from methodology challenges were rejected by DB without due revisions."12 consideration. DB should consider how to devise · Corrections account for 1,000 changes in the a more open and in-depth validation process to DB 2007 data (44 percent of total changes). The help increase both the quality and the credibility DB team has not indicated the rea- of the data. sons why information supplied by DB regularly revises informants in prior years is subject previously published Publishing and Revising the Data to retroactive correction. The DB data . . . DB publishes its data and country rankings in its team has explained that "minor" annual report each autumn and on its Web site. It changes to data for the prior year--that is, makes ongoing changes to previously published changes amounting to 10 percent or less of the data, and the Web site indicates that it contains original value--are made without further in- the most current version. The DB 2007 data vestigation. Of the 1,000 corrections, 222 (22 presented on the Web site in October 2007 had percent) were "minor."13 2,28411 differences (on the total 5,600 data points · The addition of three new countries affected used to calculate the EODB ranking) from the the rankings but not the underlying data being data originally published in the DB 2007 report discussed here. (see appendix C). The Web site does not provide nor link to the original data set. In the DB 2008 The practice of changing previously published report, the revised 2007 data were used as the data can be helpful in improving the comparator for the previous year. reliability and consistency of a data set. . . . but does not make At the same time, to fulfill its objective available previously The DB 2008 report (World Bank-IFC 2007b, pp. of facilitating research and informing published data sets. 67­69) notes that data changes have been made theory, DB should disclose all such and gives three reasons: corrections and changes and explain their effects on the rankings as explained below, and make · Changes in methodology for three indicators: available previously published data sets.14 enforcing contracts, dealing with licenses, and employing workers. (Separately, the report Effects of data changes on country rankings indicates that the methodology for paying and top reformers taxes was also changed [World Bank-IFC 2007b, The DB Web site presents 2007 EODB rankings pp. 78­79].) and a top reformers list derived from the revised · Corrections in 47 data points. data. It does not state how the changes in data · Addition of three new countries. affected these rankings. The evaluation finds that the 2,284 changes resulted in changes The evaluation team's review of the changes to to the rankings for 106 countries (even Some data changes have the DB 2007 data found that: after accounting for the addition of had significant impact on three new countries in 2008).15 Twenty- ratings. · Changes in methodology for the four indicators four countries improved 10 or more account for 1,284 changes in the DB 2007 data positions and another 24 dropped 10 or more (56 percent of total changes). DB has not indi- positions on the EODB ranking. The most signifi- cated how the 2008 methodology was retroac- cant changes are listed in table 2.2. The roster of tively applied to information that informants top reformers also changed. Latvia entered by supplied in prior years on the basis of different moving up from eleventh to tenth, and Ghana assumptions or definitions. The DB team has ex- exited by falling from ninth to nineteenth. 17 DOING BUSINESS: AN INDEPENDENT EVALUATION Table 2.2: Large Changes in 2007 Rankings Resulting from Data Revisions Five biggest winners Five biggest losers EODB EODB EODB EODB October August October August Country 2007 2007 Change Country 2007 2007 Change Guyana 96 136 40 Nicaragua 85 67 18 Italy 49 82 33 Samoa 59 41 18 Turkey 64 91 27 Tajikistan 151 133 18 St. Kitts & Nevis 61 85 24 Papua New Guinea 79 57 22 Bhutan 120 138 18 Uruguay 87 64 23 DB should fully explain Some data changes nullify reforms ranked fifty-ninth and sixtieth (Israel and Mozam- the nature and extent of cited in the text of the DB 2007 report. bique) are separated by just 0.1 of a percentage periodic data changes For example, 23 countries earlier point. and their implications for identified as having reduced their the rankings. corporate taxes show no changes in A given change in a cardinal value (for example, a tax rates using the revised data. DB reduction in the number of days needed for a highlights reforms to getting credit in Italy and procedure) is more likely to advance a country's trading across borders in China, but the revised rank, holding other countries' actions constant, if data revealed that the relevant subindicators the country starts from a more concentrated actually deteriorated (World Bank-IFC 2006b, pp. segment of the distribution than if it starts from a 3, 30). DB should make clear that its rankings are more dispersed section. This arithmetic means subject to change and fully explain the extent, that countries at the more dispersed parts of the nature, and implications of these changes on distribution have to work harder to change their country rankings. overall ranking. Countries can make significant changes, yet fail to improve their rankings, if they Constructing the Rankings are at the dispersed sections of the distribution The DB process first establishes cardinal values for that indicator. The following three examples for each subindicator: time, costs, number of illustrate this asymmetry by simulating the change procedures, and the like. Countries in rankings for a subindicator, holding the actions The EODB is an average are ranked on each subindicator. The of the other countries constant (see appendix B). of 10 rankings. subindicator percentiles are averaged to come up with an indicator-level · How does reducing the minimum capital re- ranking. The 10 indicator percentiles are then quirement affect ranking on starting a busi- averaged to generate the overall EODB ranking. ness? The DB 2008 report notes that Egypt drastically reduced its minimum capital re- DB's reliance on successive stages of ordinal quirement from 695 percent of income per rankings obscures the underlying cardinal values. capita to just 13 percent. Holding other coun- The magnitude of the difference between the tries' actions constant, this reduction would countries is not the same on all points of the distri- have boosted its ranking by 33 positions.16 Al- bution. For example, on total tax rate, there is a 5.1 though Gambia, Macedonia, and Saint Kitts percentage point difference between and Nevis all reduced the minimum capital re- A country's location in the top two performers, Maldives and quirement much less than Egypt in absolute the distribution affects Vanuatu, and a 4.7 percentage point terms, they would have boosted their simulated how a given reform will difference between the bottom two, rankings much more than Egypt. By eliminat- change its ranking . . . Gambia and Burundi. But the countries ing the minimum capital requirement, these 18 COLLECTING INFORMATION AND CONSTRUCTING THE RANKINGS 3 countries would tie with the 66 others for first the areas where it can most improve its . . . but IEG did not find place on this subindicator. ranking for the least reform effort. If evidence that countries · How much does the tax rate have to fall to im- this were the case, one would expect use this characteristic of prove ranking on paying taxes? With a 43 per- the highly concentrated subindicators the ranking system to centage point reduction in its total tax rate, to be associated with more reforms.17 manipulate their DB Sierra Leone would improve only one posi- But the correlation between tightness indicator ranking. tion in the simulated ranking for paying taxes. of distribution and frequency of But Latvia, by reducing the total tax rate by just reforms is almost nonexistent (0.01).18 The total 10 percentage points, could improve 17 posi- tax rate is the third most frequent area of reform, tions because it is situated in the most con- and it has the tightest distribution of all the centrated segment of the distribution. subindicators, with 94 percent of the countries' · How does reducing the time to open a business rankings within one standard deviation from the improve starting a business? Doing Business mean. But the two most popular areas for 2008 notes that the Republic of Lao reduced the reform--number of procedures to start a business time to start a business by 60 days (36 percent and legal rights of creditors and debtors--are not of its starting value). Yet this change would not among the most tightly distributed. This quantita- affect its simulated ranking for starting a busi- tive analysis is fully consistent with the finding ness. Mauritius, by contrast, reduced the time reported in chapter 4--that IEG did not find by 41 days (89 percent of its starting value), and evidence of countries making superficial changes would thereby advance 20 positions on the for the sole purpose of improving their rankings. simulated ranking for starting a business. Nevertheless, the DB team may wish to consider ways of making the rankings more informative, It has been suggested that DB's use of rankings perhaps by establishing country groupings that might create an incentive for a country to reform reflect the cardinal values of each indicator. 19 Chapter 3 Evaluation Highlights · For the most part, the indicators measure legal regulations and in- formed estimates of practice, as dis- tinct from opinion. · There are some systematic differ- ences in rankings associated with countries' legal origins and policy choices; these do not undermine DB's validity. · Inaccurate nomenclature and over- stated claims of the indicators' ex- planatory power have provoked criticism. · The subindicator on total tax rate is anomalous because it goes beyond regulatory burden. · The employing workers indicator is consistent with relevant ILO con- ventions, even though it gives higher scores to countries with lower job protections. Merchant buying oranges for his fruit stall, Epping Market, Cape Town, South Africa. Photo reproduced by permission of Gideon Mendel/Corbis. What Do the Indicators Measure? T his chapter reviews key characteristics of the DB indicators as a group, and then focuses on issues that emerge from five indicators--starting a business, paying taxes, employing workers, enforcing contracts, and getting credit. General Characteristics of the Indicators naires require a response that combines the law as written with an estimate of what happens in Is DB rules-based? practice. 4 For example, on enforcing contracts, DB states that it differs from other surveys informants are asked the cost of resolving a because it collects information about a country's commercial dispute, including both court fees laws and regulations, as distinct from people's and estimated average attorney fees.5 views, estimates, or perceptions.1 The evaluation team analyzed the questionnaires for the 2007 The 2008 questionnaires contain an additional and 2008 DB reports to determine the share of 188 questions (109 additional questions in 2007) questions that required responses based: that are not used in the calculation of ratings or rankings. The responses are used to provide · Solely on written laws or regulations ideas for future work and additional insights into · On a combination of written law and the in- country issues. Of the 188 supplemental formant's experience questions in 2008, 77 ask solely about the law, 67 · Solely on the informant's judgment or ask about a combination of law and practice, and experience. 44 ask for the informant's judgment or opinion. For example, the legal rights index questionnaire The 2008 questionnaires contain 87 questions for the getting credit indicator asks informants' that generate data used to calculate the rankings opinions about the main areas that require for the 10 indicators and the aggregate EODB. Of reform.6 For starting a business, it seeks inform- these, 70 questions (80 percent) ask solely about ants' opinions about how the registra- the law as written.2 For example, for getting tion process has changed since the Ratings and rankings are credit, a question about the strength of legal previous year, as well as their sugges- grounded in written laws rights index asks whether management is tions for reforms.7 and estimates of practice; allowed to remain in control of a company opinions are collected but during reorganization.3 The remaining 17 Thus, while DB's ratings and rank- do not feature in the questions on time and cost in the 2008 question- rankings. 23 DOING BUSINESS: AN INDEPENDENT EVALUATION ings are grounded predominantly in formal rules power than it delivers. For instance, a donor and regulations, DB also gathers informants' official noted, "dealing with licenses has been opinions or perceptions through these supple- problematic because it has a broad name for what mentary questions. The practice of asking extra is a very narrow focus on construction permits questions adds to the complexity of the question- and thus may be misleading." Bank staff and naire, which may be a deterrent for potential others who interpret the rankings to country contributors.8 It may also give informants and authorities have urged DB to give a more accurate readers the impression that DB rankings are based signal of what it measures. More precise more on opinions than they actually are. nomenclature could help. Nomenclature Legal origin The names of several indicators overstate what DB's 2004 inaugural report asserted that coun- they actually measure. tries' regulatory regimes are strongly determined by their legal origin, as noted in box 3.1. This · Gettingcredit does not measure firms' access to assertion, especially DB's contrast of the common credit, which depends largely on macro- and civil law systems, has spawned a debate on The names of some economic and structural factors such as DB's treatment of countries with a civil law indicators--getting depth of financial intermediation and in- tradition, specifically of French legal origin. credit, dealing with terest rates. It simply measures the avail- licenses, and employing ability of credit history information to Among the 175 economies covered by DB 2007, workers--overstate what lenders and the legal rights of lenders 76 trace their laws governing commerce and they measure. and borrowers should there be a default property to the Napoleonic Code, while 59 have on the repayment of a loan. a system based in common law.9 Consistent with · Dealing with licenses measures the ease of ob- the assertion in Doing Business 2004, countries taining a construction permit, and not the wide with a common law tradition occupy 8 of the top range of licenses, permits, and authorizations re- 10 spots for EODB. Of the 44 countries in the top quired in all sectors for a wide variety of reasons. quartile, 19 are common law and 14 are civil law · Employing workers measures the rules govern- countries. In the bottom quartile, 30 countries ing hiring, firing, and paying workers, but not have a civil law origin, and these include all 17 other aspects of the labor market such as members of Organization for the Harmonization wages, mobility, and qualifications. · Registering property measures the procedures Box 3.1: Civil and Common Law to transfer the property title of land and a Approaches to Regulation building between two businesses, and not the procedures to obtain a title for the property the first time. The inaugural Doing Business report stated, "When the English, French, Spaniards, Dutch, Germans and Such broadly framed names might be justified if Portuguese colonized much of the world, they brought the indicators were proxies for broad but difficult- with them their laws and institutions. After inde- to-measure phenomena; DB has not systemati- pendence, many countries revised legislation, but cally demonstrated that this is the case. Interviews in only a few cases have they strayed far from the with operational staff and stakeholders confirm original. These channels of transplantation bring that the simplicity of the names is helpful in about systematic variations in regulation that are getting the attention of ministers and other not a consequence of either domestic policy choice nonspecialist audiences. But even though DB's or the pressures toward regulatory efficiency. Com- documentation makes amply clear what DB mon law countries regulate the least. Countries in the actually measures, country authorities, Bank French civil law tradition the most." Group staff, and other stakeholders have also Source: Doing Business 2004, p. xiv. criticized the DB for promising more explanatory 24 WHAT DO THE INDICATORS MEASURE? of Business Law in Africa (OHADA).10 civil law countries can still score well On 13 of 32 subindicators, on the DB indicators, as outlined in civil law countries score The evaluation analyzed the specific issues on box 3.2. significantly lower than which the civil law countries as a group score common law countries . . . lower than common law countries.11 On 13 Are the DB indicators adding new subindicators listed on the left in table 3.1, civil information? law countries scored significantly lower. Six of A cross-country indicator whose rankings were these significant differences relate to the number perfectly correlated with per capita income (or of procedural steps, commonly regarded as some other underlying characteristic) would not excessive in the French system. Four differences add new information; one could predict a country's relate to the greater protection of debtors and ranking by knowing its per capita income. For DB the lesser protection of minority investors that indicators, per capita income levels only partly characterize the civil law;12 the DB indexes award explain the rankings. The overall EODB ranking and points for attributes found primarily in common income per capita have a relatively high law.13 Three differences relate to job protection, rank correlation coefficient of 0.77.14 . . . nonetheless, civil law which may derive not from legal origin, but There is lower correlation (0.65) with countries can still score rather from policy choices made by this set of per capita income in the low- and well on the DB indicators. countries (see appendix D). middle-income countries.15 As shown in figure 3.1, countries with similar levels of gross Even though some aspects of the civil law system national income (GNI) per capita, such as Peru and are ranked lower on the DB indicator criteria, Brazil or Kenya and Mauritania, can have very differ- Table 3.1: Do Civil Law Countries Score Lower Than Common Law Countries? Differences are significant Differences are not significant Indicator Subindicator Indicator Subindicator Highly significanta Dealing with licenses Procedures (number) Employing workers Difficulty of hiring index Dealing with licenses Cost (% of income per capita) Employing workers Rigidity of hours index Employing workers Firing costs (weeks of wages) Employing workers Difficulty of firing index Registering property Procedures (number) Getting credit Credit information indexb Registering property Time (days) Getting credit Legal rights index Registering property Cost (% of property value) Protecting investors Director liability index Protecting investors Disclosure Index Protecting investors Shareholder suits index Paying taxes Total tax rate (% profit) Starting a business Procedures (number) Trading across borders Documents for export (number) Starting a business Cost (% of income per capita) Trading across borders Time for export (days) Starting a business Min. capital (% of income per capita) Trading across borders Cost to export (US$ per container) Paying taxes Time (hours) Trading across borders Documents for import (number) Trading across borders Time for import (days) Significanta Trading across borders Cost to import (US$ per container) Paying taxes Payments (number) Enforcing contracts Procedures (number) Starting a business Time (days) Enforcing contracts Time (days) Dealing with licenses Time (days) Enforcing contracts Cost (% of debt) Closing a business Recovery rate (cents on the dollar) a. Significance level set at 95 percent. Highly significant differences set at 99 percent. b. All statistically significant differences favor English common law countries except for the credit information index. 25 DOING BUSINESS: AN INDEPENDENT EVALUATION Box 3.2: Can a Civil Law Country Succeed in a "Doing Business" World? Having a civil law regime does not prevent a country from scor- Even countries at the bottom of the rankings can improve. Of ing well on the DB rankings. France, for example, ranks twelfth the 26 Sub-Saharan countries with a civil law tradition, 24 are in in starting a business; it has fast and inexpensive processes and the bottom quartile of the overall ranking. Mali, ranked one-hundred no minimum capital requirement. It also ranks fourteenth in en- and fifty-eighth in EODB, could improve substantially by following forcing contracts. Countries can improve their scores and rank- the model of Morocco or Tunisia. For example, if Mali reduced the ings within a civil law framework. Take, for example, Tunisia, a number of procedures, days, and minimum capital requirement to middle-income civil law country ranked eighty-eighth on EODB: start a business to the level of Morocco, it would improve 13 po- sitions in the overall ranking. Likewise, improving the procedures · If it improved its score on difficulty of firing to the same level for enforcing contracts and trading across borders to the level of as Belgium, it could improve its EODB ranking by 16 positions. Tunisia would improve Mali's overall ranking by 26 positions, al- · Opening a business in Tunisia is fairly efficient in time and cost. lowing it to move out of the bottom quartile. If Tunisia eliminated the minimum capital requirement, like If a hypothetical civil law economy were constructed combining France, it would further improve its EODB ranking by 11 the scores of the highest-scoring civil law country on each indi- positions. cator, it would place third in the overall ranking. Note: Calculations based on the Doing Business 2008 data. Per capita incomes only ent regulatory environments as countries in the middle, it is 0.52. Many Bank Group partly explain the DB measured by the DB indicators. Indeed, clients thus have scope to reduce the burden of indicator rankings. the rank correlation between DB indica- business regulations even at their current income tors and GNI per capita is highest, at levels. 0.89, for the richest and poorest countries16; for Figure 3.1: Countries with Similar GNI Can Have Different DB Indicator Scores 180 160 140 Kenya Mauritania Nicaragua 120 Bolivia ranking 100 capita Peru 80 Ecuador per Brazil GNI 60 40 20 0 0 50 100 150 Ease of doing business ranking Sources: Doing Business 2007 and DB Web site. 26 WHAT DO THE INDICATORS MEASURE? DB includes 32 subindicators, as detailed in costly entry regulations make opening a business chapter 1. The rank correlation among them is more difficult, and thus fewer entrepreneurs will generally low, suggesting that they are capturing do so (at least in the formal sector). The DB different dimensions of the regulatory environ- reports note that "cumbersome entry procedures ment. 17 Out of a possible 509 pair-wise correla- push entrepreneurs into the informal economy, tions, only 6 have correlation coefficients above even after controlling for income per capita" 0.60, and 16 are between 0.5 and 0.6. Only the (World Bank-IFC 2004, p. 22). DB states that subindicators of the trading across borders informal businesses tend to lack worker protec- indicator (DB's most recently added indicator) tion and benefits, to have substandard product are so highly correlated with each other as to quality, and to face difficulty in securing bank suggest that some may be redundant. For credit and using courts to resolve disputes. example, time to import and time to export have Formalization is beneficial because "the establish- a correlation of 0.91. The DB team could consider ment of a legal entity makes every dropping any such overlapping subindicators. business venture less risky and The logic behind the increases its longevity and its likeli- starting a business Key Features of Selected Indicators hood of success" (World Bank-IFC indicator is that more The evaluation took an in-depth look at five DB 2004, p. 17). onerous and costly entry indicators to identify what they really measure regulations result in and to assess how the rankings track with reality The time subindicator captures only fewer entrepreneurs in the case study countries. the duration necessary to complete opening businesses. procedures; it excludes the time an Starting a business entrepreneur may spend gathering information, Starting a business is one of the five original DB which varies widely and cannot be reliably indicators and the first to have been developed estimated with few data points. Opaque systems by Djankov and others (2002).18 It aims to may require more time than transparent systems. measure how efficiently an entrepreneur can Analysis using firm-level data suggests that this complete all officially required procedures to "time tax" may be large and important (Hellman formally operate an industrial or commercial and Schankerman 2000). business. A country's ranking is the average of its percentile rankings on the four subindicators Doing Business 2004 asserts that just two shown in table 3.2. procedures ought to be sufficient for regulating business start-up: the notification of existence The underlying logic is that more onerous and and the tax and social security registration. While Table 3.2: The Starting a Business Indicator Standard Subindicator Mean Median Min Max deviation Number of required pre- and post-incorporation procedures officially required to formally operate a business 9 9 2 20 3 Time to complete the procedures (calendar days) 44 31 2 694 61 Cost to comply with procedures (percentage of the country's income per capita) 61 21 0 1,075 122 Paid-in minimum capital that the entrepreneur must deposit in a bank before registration begins (percentage of the country's income per capita) 116 8 0 3,673 353 Source: Doing Business 2008. 27 DOING BUSINESS: AN INDEPENDENT EVALUATION the report acknowledges that other procedures, Box 3.3: A Paper "Reform" such as registering with the statistical office, in Afghanistan obtaining environmental permits, or registering workers for health benefits "seem to be socially Afghanistan was the top reformer for starting a busi- desirable" (World Bank-IFC 2004, p. 21), DB ness in Doing Business 2006 because it reduced the awards higher ratings to countries without these number of procedures from 28 to 1, and time from 90 procedures than to those with them (holding all to 7 days. Literature and key informants note that other subindicators constant). At the extreme, the authorities simply pushed all important proce- the best performer for this subindicator would dures to a stage after the legal registration of a be a country that required just the above- business. mentioned two procedures, even though firms themselves may value the worker satisfaction or Sources: Arruñada 2007 and interviews. social benefits arising from other procedures. "one-more-stop shop." The paper also notes that Most popular indicator for reform where one-stop shops have worked (such as in Since Doing Business 2005, the starting a Ireland, Malaysia, and Singapore), the senior level business indicator has produced the most of government had committed to investment reforms annually as measured by DB. Reducing climate reforms and made increased foreign direct business entry regulations may be easier politi- investment (FDI) a central pillar of their develop- cally and less expensive than progress on other ment strategies. Thus, new one-stop agencies in indicators, such as employing workers, that these countries benefited from an environment require political trade-offs. Further, where fewer licenses, approvals, and permits were The starting a business since 2005, two subindicators--the deemed a necessary and important component of indicator has produced days and cost of starting a business-- investment climate reforms. the most reforms as have been used by the United States' measured by DB. Millennium Challenge Corporation Is it relevant and important for economic outcomes? (MCC) in its formula for determining countries' Doing Business 2007 claims two economic eligibility status for grants, and also features as a benefits of formally registered businesses: they guidepost under the Bank's Country Perform- grow larger and they pay taxes (World Bank-IFC ance and Institutional Assessment (CPIA) com- 2006b). Literature and cross-country studies ponent on the "business regulatory have established the importance of higher entry environment." The time and cost to start a rates of new businesses for increased competi- business are also used as 2 of the 14 "outcome" tion and economic growth (see, for example, indicators in the "IDA results framework"19 (see Klapper, Laeven, and Rajan 2006). To what extent chapter 4). do registration procedures affect the entry of new businesses? After all, no matter how com- One-stop shop: DB reports promote the creation of plicated and time-consuming they are, each one-stop shops as a single access point for business must endure them only once. entrepreneurs to comply with entry regulations.20 But a review of one-stop shops by the Foreign Cross-country studies have established a correla- Investment Advisory Service (FIAS) finds that tion between the number of small and medium- "such a mechanism works in barely any size enterprises or new firm registrations and less Literature is not country of the world" (Sader 2002, p. business entry regulation, but have not yet conclusive about whether 3). They tend to generate turf battles if demonstrated causality (De Sa 2005, p. 4). For business registration asingleagencygainscontroloverallthe example, while Klapper and others (2007) find reform encourages various licenses, permits, and that barriers to starting a business are signifi- greater formalization clearances formerly granted by differ- cantly and negatively correlated with the number and the creation of new ent agencies. In this instance, DB's one- of registered businesses and new registrations of formal businesses. stop shop often becomes a companies, they acknowledge that they cannot 28 WHAT DO THE INDICATORS MEASURE? Box 3.4: Does Simplifying Business average of its rankings on the three The total tax rate Registration Encourage Formalization? equally weighted subindicators, subindicator goes beyond shown in table 3.3. regulatory efficiency to Mexico implemented a program to simplify munici- include judgment about pal licensing, one of several registration procedures The lower the total taxes on firms, the fiscal policy. for selected types of businesses. Firm registration in- higher a country scores on the total tax creased between 4 and 5.6 percent in the eligible in- rate subindicator (see box 3.5). This subindicator, dustries. Kaplan and others (2007) found that the unlike most other dimensions of DB, does not increase was temporary and concentrated in the measure regulatory burden alone; it also involves first 10 months after program implementation, lead- implicit judgments on complex issues of fiscal ing the authors to conjecture that "the program efficiency and equity. The lower the taxes paid by mostly affects the existing stock of informal firms the corporate sector, the more revenues need to and has a smaller effect on the creation of `truly' be raised from other sources to reach a given new firms" (pp. 4­5) as the program cleared a back- revenue target. A country's preferred combina- log of applications. Using household data, Bruhn tion of corporate, sales, personal income, VAT, (2007) found that while employment in eligible in- and trade taxes should represent a blend of dustries rose, the increase in firm registration comes revenue-generating capacity, efficiency, equity, "exclusively from former wage earners opening busi- transparency, and reasonable overall tax burden nesses" (p. 3). Rather than encouraging movement (IMF 2007). A lower tax rate on the corporate from the informal to the formal sector, registration sector is not necessarily beneficial for the benefited those already in the formal sector. a economy as a whole; assessments of tax regimes "need to be undertaken in an intertemporal, Sources: Kaplan and others 2007; Bruhn 2007. a. The authors conjectured that greater effects may have been seen general equilibrium framework to capture the full if more comprehensive reforms had taken place. impact of each choice" (OECD 2007a). postulate on the direction of causality.21 Recent DB's partnership with PricewaterhouseCoopers country-level studies have begun to address (PwC): The total tax paid, as measured by DB, causality but depict a mixed picture. While in includes not only corporate profit taxes, but also Russia and Brazil,22 reform of specific regulatory the social security and labor, property, capital procedures was linked to enhanced formaliza- gains, and dividend taxes that are paid by the firm. tion and the creation of new businesses, respec- This comprehensive definition of taxes paid by the tively, two other studies in Mexico note the firm is based on a methodology initially devised by limited impact of business registration reforms, PwC, the global accounting partnership. PwC's as illustrated in box 3.4. objective in developing its "total tax contribution" framework was to persuade tax Paying taxes authorities--notably in the United The methodology is based Paying taxes measures both the total taxes paid Kingdom--that firms actually on a complex framework by a firm and the administrative efficiency of contribute more to the public coffers developed by making the payments. A country's ranking is the than is indicated by corporate tax rates PricewaterhouseCoopers. Table 3.3: The Paying Taxes Indicator Standard Subindicator Mean Median Min Max deviation Number of yearly payments 34 32 1 124 21 Time required (in calendar days) 323 242 0 2,600 322 Total tax rate as a share of firm profits 51 44 8 287 38 Source: Doing Business 2008. 29 DOING BUSINESS: AN INDEPENDENT EVALUATION Box 3.5: Can a Tax Haven Be a Global Leader alone.23 DB adopted a version of PwC's methodol- on Taxation? ogy in 2005. (PwC 2007; World Bank-IFC 2007b, p. 77). The methodology is the most complex of all Kuwait, Maldives, United Arab Emirates (UAE), and Vanuatu are DB's DB indicators, because it requires detailed calcula- four top-ranked countries on total tax rate in the 2008 report. Each has tions of a variety of taxes for a standard firm. a special characteristic that enables it to avoid imposing taxes on firms. Vanuatu, a small Pacific island that has set itself up as a tax haven, The data for paying taxes are furnished to DB by does not impose personal or corporate income tax, capital gains tax, PwC's partner offices in 142 countries.24 This or withholding taxes. It raises revenues mostly from indirect sources, work is done under the terms of an agreement principally import duties and value added tax (VAT). Maldives also has that PwC is DB's sole informant about taxes in all no corporate income tax. Instead, the government raises revenues countries where it does business. PwC donates from tourism by charging lease rents on the land occupied by resorts. this work in return for the opportunity to co- The government of Maldives is reviewing its dependence on these publish the rankings jointly with DB, as part of its lease rents. Kuwait and UAE, for their part, derive most of their public "thought leadership" activities.25 This relation- revenues from oil and do not impose corporate, personal, or VAT taxes ship offers DB the advantage of a ready-made on domestic firms. These top-ranked countries cannot feasibly serve stable of qualified informants. But DB's reliance as role models for other countries seeking an optimal level of corpo- on a single partner, combined with the complex- rate taxation. ity of the questionnaire, makes validation of information more difficult. Informants in Algeria, Sources: Heritage Foundation 2007; Maldives government 2007; Vanuatu government 2007. China, Mongolia, Netherlands, and Spain all Table 3.4: The Employing Workers Indicator Subindicator Standard (lower score is always better) Mean Median Min Max deviation Difficulty of hiring index (0­100) 32 33 0 100 27 Restriction on when term contracts can be used Maximum duration of term contracts Minimum wage (% of value added per worker) Rigidity of hours index (0­100) 39 40 0 80 23 Legal maximum of hours and days worked per week Restrictions on night and weekend work Paid vacation more than 3 weeks (score 1 for 22 days or more, 0 otherwise) Difficulty of firing index (0­100) 31 30 0 100 23 Steps required to fire 1 or a group of redundant workers ­ index contains 8 criteria dealing with required notifications and priority rules for redundancy Firing cost (weeks of salary) 48 35 0 446 50 Cost of advance notice required, severance payments, and penalties (score 0 if cost is 8 weeks' salary cost, or less; score number of weeks if more than 8 weeks salary) Nonwage labor costa 15 14 0 55 11 Social security and similar payments (percent of salary) Source: Doing Business 2008. a. Measured but not included in the calculation of rankings. 30 WHAT DO THE INDICATORS MEASURE? noted errors in the tax data; thus reliability Employing workers The reliance on a single cannot be guaranteed. Employing workers rates the laws and global partner for the tax regulations that govern how firms hire indicator makes The procedure for collecting complex data for the and fire workers: the length of the validation difficult and total tax rate subindicator entails two risks for DB: workday, week, and year and the cannot guarantee (a) the reputational risk arising from partnering minimum wage firms have to pay. The reliability. with an advocate of a particular policy stance, and underlying assumption is that less regulation will (b) the operational risk of depending on a single result in higher employment rates and, in some source of data.26 Since the subindicator itself is contexts, lower shares of informal to formal anomalous within the DB framework, it would be employment. advisable for DB to reformulate the paying taxes indicator to include only measurements of A country's ranking on employing workers is the regulatory burden such as the total cost of average of the percentile rankings on compliance. Since corporate tax rates are undeni- four equally weighted subindicators, The assumption ably important for business, DB should continue listed in table 3.4. Firing workers is the underlying the employing to collect tax rate information separately and focus of two of the subindicators, workers indicator is that more simply, but exclude it from the rankings. giving this dimension a 50 percent less regulation will result This would also help to simplify the question- weight in the indicator. Three subindi- in higher rates of naire, permit more informants to contribute, and cators are indexes, each containing employment and more make the data more understandable to users. several criteria. With 15 criteria in all, formal employment. Table 3.5: Employing Workers: Highest- and Lowest-Ranked Countries Top 20 DB 2008 rank DB 2007 rank Lowest 20 DB 2008 rank DB 2007 rank Singapore 1 1 Peru 159 160 United States 1 1 Senegal 160 163 Marshall Islands 1 1 Niger 161 161 Tonga 4 4 Mozambique 161 162 Brunei 4 n.a. Gabon 163 166 Georgia 4 4 Luxembourg 164 n.a. Maldives 7 7 Morocco 165 165 Australia 8 8 Slovenia 166 159 Palau 9 9 Congo, Rep. 167 167 Denmark 10 10 Ecuador 168 168 Uganda 11 11 Sierra Leone 169 169 Micronesia 12 12 Panama 170 170 New Zealand 13 13 Congo, Dem. Rep. 171 174 Bhutan 14 n.a. Angola 172 171 Samoa 15 14 Paraguay 173 172 Fiji 16 15 Guinea-Bissau 174 173 Japan 17 20 Equatorial Guinea 175 175 St. Kitts and Nevis 18 16 São Tomé and Principe 176 176 Canada 19 17 Venezuela 177 177 Switzerland 20 39 Bolivia 177 178 Source: Doing Business 2008. 31 DOING BUSINESS: AN INDEPENDENT EVALUATION The indicator measures employing workers is one of the more some countries that have made policy choices costs but not benefits of complex DB indicators. favoring extensive job protection (see table 3.5). regulation or other dimensions of labor The indicator and its components The 20 lowest-ranked countries, displayed in the market flexibility. measure the costs of selected regula- table, include Luxembourg, and France and tions but not their benefits. Since the Germany also rank low at 144 and 137, respec- stylized case involves a firm with 201 unionized tively. As for the 20 highest-ranked countries, they employees, it does not capture the laws and rules include a surprising number of small island that affect smaller and nonunionized firms. Nor states--indeed, the Marshall Islands shares the does it capture other dimensions of top spot with Singapore and the United States. Numerous small island labor market flexibility, such as This may reflect the small islands' poorly states among the top 20 information, enforcement, and tied developed labor legislation, as illustrated in box may reflect poorly versus monetized benefits such as 3.6. Yet countries with well-developed labor developed labor housing, pensions, worker health and legislation also appear in the top 20, including legislation, yet they appear safety regulations, and so on. Employ- Australia, Canada, Denmark, Japan, and New alongside countries with ing workers has been DB's most Zealand. This reflects their mature public and well-developed labor controversial indicator, perhaps private labor market institutions, which facilitate legislation. because it delivers low rankings for mobility with minimal regulation. Box 3.6: Does Top-Ranked Imply "Well Regulated" . . . or "Unregulated"? Because DB counts the number of regulations, it is difficult to the employing workers indicator, Singapore and the Marshall Is- tell whether top-ranked countries have efficient and responsi- lands are both ranked number 1 (along with the United States), ble regulations or simply inadequate regulation. For example, on but their labor regulations are very different. Singapore has strong labor regulations, effectively Marshall Islands, a recent member of ILO, has impor- enforced tant gaps in labor legislation The Right of Association Singapore's constitution provides all citizens the right to form as- The law provides for the right of free association in general, and sociations, including trade unions. But the Parliament may im- the government interpreted this right as allowing the existence of pose restrictions based on security, public order, or morality labor unions, although none has been formed. With few major em- grounds. In 2004, approximately 20 percent of the national labor ployers, there were few opportunities for workers to unionize, and force was represented by 68 unions. the country has no history or culture of organized labor. Prohibition of Forced or Compulsory Labor The law prohibits forced or compulsory labor, including by children, The law does not specifically prohibit forced and compulsory labor and there are no reports of such practices. by children; however, there were no reports that such practices occurred. Prohibition of Child Labor and Minimum Age for Employment The law prohibits the employment of children under the age of 12, There is no law or regulation setting a minimum age for employ- and restrictions on employing children between ages 12 and 16 are ment of children. Children typically were not employed in the wage rigorous and strictly enforced. economy, but some assisted in family enterprises. Acceptable Conditions of Work The law sets the standard legal workweek at 44 hours and one rest There is no legislation on maximum hours of work or occupational day for each week. safety and health. Source: Bureau of Democracy, Human Rights, and Labor, U.S. State Department 2007 . 32 WHAT DO THE INDICATORS MEASURE? Eight countries are reported to have made The evaluation examined this assertion This indicator is positive reforms on employing workers overall and found it generally valid because: associated with the fewest in Doing Business 2007, making this the DB reforms. indicator with the fewest reforms (World Bank- · The ease of hiring index measures IFC 2006b, p. 4; 2007b, p. 4). Georgia, though, the ease and flexibility of using term contracts improved dramatically from a rank of 71 to 6 as a to employ workers. This issue is not covered result of a new labor code and flexible labor by any ILO conventions. rules.27 Reforms in this area are difficult because · The rigidity in hours of work index consists of where organized labor is strong, they typically 5 components that are all consistent with the engage competing interests. Even where it is provisions of the ILO conventions. weak, as in many low-income countries, labor · The firing cost and ease of firing index are market reforms simply take low priority. based on responses to 10 questions, of which 6 are fully consistent with the respective ILO Is employing workers consistent with accepted conventions, and 4 are consistent with the let- labor standards? ter of the relevant ILO provisions, but not with Critics of DB have argued that the employing their spirit, as shown in box 3.7. workers indicator rewards employment practices that are inimical to workers' interests (see, for The ILO provisions, with which DB is consis- example, Berg and Cazes 2007). They consider tent, represent a baseline degree of labor that regulations on job conditions, hiring, and protection agreed to by the international firing are needed for the protection of workers. In community. Many countries' response to these criticisms, DB has stated that laws offer more extensive or The employing workers the components of this indicator have recently generous job protections as a indicator is consistent been made consistent with the core labor matter of national policy, and with the letter of ILO standards of the International Labor Organization DB penalizes more generous provisions, but four (ILO) (World Bank-IFC 2007a). provisions. For example, measures do not reflect France and Germany, whose their spirit. Box 3.7: Measures on the Costs and Difficulty of Firing Workers and the ILO Conventions There are four measures of DB's employing workers indicator How much severance pay must a redundant worker get? The that are consistent with the letter, but do not reflect the spirit, ILO stipulates that a redundant worker should be provided sepa- of the relevant ILO provisions: ration or severance pay based on seniority, wage level, and other unspecified criteria. DB sets its own cut-off, and in DB 2008, a coun- Must the employer consider reassignment or retraining ac- try requiring up to eight weeks of severance pay gets the best score. tivities before redundancy termination? The ILO convention does All countries that require severance pay greater than eight weeks not require, but asks the employer to provide, in accordance with get the worst score. national laws and practice, an opportunity for consultation with Must the employer notify a third party before terminating a worker representatives about measures to mitigate the adverse group of workers? ILO Convention 158 requires that the employer effect of termination. DB gives a higher rating to countries that do notify a competent authority about a termination of a group of not require the employer to make such consultations. workers. The convention does not specify the cut-off number or Are there clearly established criteria applying to redundan- percentage of workers and leaves this to be determined in ac- cies? Although the ILO does not require application of clearly es- cordance with national laws and practice. Again, DB sets its own tablished criteria, it does recommend that employers select workers cut-off. In DB 2008, a country gets the worst score if it requires no- to be made redundant on that basis. The DB gives a higher rating tification for terminating a group of fewer than 25 workers. In 2007, to countries that do not require such criteria. DB's cut-off was a group of 20 workers. 33 DOING BUSINESS: AN INDEPENDENT EVALUATION laws require 22 or more days of vacation, score example, World Bank-IFC 2007b, p. 62), and the worse on rigidity of hours than Australia or Italy, methodology for enforcing contracts has which require 21 days or fewer. This is what changed more than most. In the 2007 report, the gives France, Greece, Spain, and other scenario was changed from a bounced check to a European countries their low overall rankings commercial dispute,28 which prevents the con- on employing workers. struction of a consistent time series for this indicator. Further changes were made for the Enforcing contracts 2008 report, and the 2007 ratings were retroac- The enforcing contracts indicator tively revised.29 These changes had a significant The enforcing contracts aims to measure how efficiently a impact on the ratings and rankings. For example, indicator considers that commercial dispute can be resolved. Tunisia, praised in Doing Business 2006 as one of it is always better to have The underlying logic is that a higher the easiest places to enforce contracts (World fewer procedures. degree of contract enforceability Bank-IFC 2005, p. 61) and ranked fortieth in encourages firms to develop relation- Doing Business 2007, now ranks a mediocre ships with a larger number of suppliers and eightieth in Doing Business 2008. These changes customers, fostering profitability and incentives were explained by changes to the methodology. to engage with more advanced technologies Such large changes in the data from one year to (Commander and Tinn 2007). This indicator (like the next make data on individual countries and most DB indicators) is linear; that is, fewer the rankings less than fully reliable, as discussed procedures are always considered better. It does in chapter 2. not consider that some judicial procedures may help ensure transparency, accountability, and Most disputes do not wind up in court fairness to the parties. All the DB indicators attempt to measure the law as distinct from actual practice. In the case of The informants--lawyers and notaries--base enforcing contracts, the gap between law and their answers on a scenario involving a seller of practice is particularly wide. The indicator goods suing a buyer who does not pay for the measures only contract enforcement through goods, citing poor quality. The amount in dispute the court system, and not other formal and is 200 percent of the country's per capita GNI. informal resolution methods commonly used in The ranking on enforcing contracts is many countries. In practice, a lawyer will select The methodology for the the average of the country rankings on the most cost-effective legal strategy to help a enforcing contracts the three subindicators, as shown in client recoup funds. The 2005 Business indicator changed table 3.6. Enterprise Surveys (covering 38 countries) substantially, resulting in found that two-thirds of business owners said large changes in DB periodically changes the method- they had resolved their most recent dispute over rankings. ology for some indicators (for the payment of an overdue bill without resorting Table 3.6: The Enforcing Contracts Indicator Standard Subindicator Mean Median Min Max deviation Number of required procedures between filing a suit and enforcement of judgment 38 38 20 55 7 Time taken to resolve the dispute (in calendar days) 605 543 120 1,800 308 Cost to defendant and plaintiff including attorney and court fees, expressed as a percentage of the claim value 34 26 0 163 28 Source: Doing Business 2008. 34 WHAT DO THE INDICATORS MEASURE? to the courts. In the United States, only about 10 capital fixed formation, domestic bank Stakeholders note that the percent of the civil cases in state courts go to credit, gross capital inflows, or foreign scenario for this court (Davis and Kruse 2007), while in Romania, direct investment (see Commander indicator diverges about one-third of such cases go to court. 30 The and Tinn 2007). The analysis also significantly from indicator, therefore, tends to overstate the found little association between the practice. burden of court procedures across the board, enforcing contracts indicator and and disproportionately more in those countries measures obtained from firm-level where noncourt mechanisms are used the most. surveys on aspects of the legal systems IEG analysis found no To partially correct this, the 2008 DB survey for enforcing contracts and loans significant association introduced a one-procedure "credit" to the score given with collateral. between the enforcing of countries with specialized commercial courts. contracts indicator and a This change reduced the counted number of While there may be some time lag range of intermediate procedures in 11 countries. between improvements in the DB outcomes. indicator and outcome measures, the Stakeholders confirmed that DB data on contract DB report needs to be cautious when making enforcement diverge from practice. Moldova's associations or implying causality between the high rank of 17 does not reflect the reality that enforcing contracts indicator and outcomes few disputes wind up in court. Indeed, a Bank- such as enhanced foreign direct investment (see supported analysis finds that the main reasons World Bank-IFC 2006b, p. 48, for example). Moldovan businesses do not resort to courts are the long duration of the process for settling Getting credit disputes and the high cost of legal services The getting credit indicator measures two things: (World Bank 2007d, p. 68). In Peru, although the legal rights of borrowers and specialized commercial courts have reduced the lenders (strength of legal rights index) The getting credit number of procedures and time needed to and the availability of credit informa- indicator measures resolve a dispute, interviewees noted that the tion about firms and individuals borrower and lender type of case specified under the DB methodol- (depth of credit information index). legal rights and the ogy would not necessarily go to court. The strength of legal rights index availability of credit measures how well collateral and information. Is it important for economic outcomes? bankruptcy laws facilitate lending. It The DB 2006 and 2007 reports suggest that the assigns a country one point for each of seven ease of enforcing commercial disputes in courts attributes of collateral law and three of is important because it is associated with higher bankruptcy law, based on information provided lending from commercial banks, increases in the by financial lawyers. Hong Kong and the United number of new firms and new hires in Kingdom have the highest score of 10, and established firms, and reduced demands on Afghanistan and Cambodia have the lowest score court budgets. The underlying research, Djankov of 0. and others (2002), finds excessive formalism in judicial procedures in countries. Djankov and The depth of credit information index measures others (2006) construct a debt enforceability the quality, scope, and accessibility of credit index that is found to be correlated with income information through public and private credit per capita, credit market development, and legal registries. The data are derived from banking origins. supervision authorities and credit registries. The depth of credit information index assigns one A background paper commissioned for this point for each of six features of a credit informa- evaluation found no significant association tion system. Twenty-one countries have the between the enforcing contracts indicator and a highest score of 6, and 56 countries have the range of intermediate outcomes--research and lowest score of 0. The top 10 countries for getting development, investment, domestic credit, gross credit are all high-income countries, except 35 DOING BUSINESS: AN INDEPENDENT EVALUATION Table 3.7: The Getting Credit Indicator Standard Subindicator Mean Median Min Max deviation Strength of legal rights index (0­10) 5 4 0 10 2 Index contains 10 criteria dealing with collateral and bankruptcy laws to protect borrowers and lenders Depth of credit information index (0­6) 3 3 0 6 2 Index contains 6 criteria dealing with the scope, accessibility, and quality of information available through public or private credit registries Source: Doing Business 2008. Malaysia and the Slovak Republic. that credit information sharing between lenders increases the supply of financing, decreases Additionally, DB gathers data on the share of the defaults, and enhances monitoring of the risks population covered by public credit registries taken by the financial system. A background and private bureaus, but these data are not paper commissioned for this evaluation found included in the calculation of the overall EODB better legal rights of borrowers and lenders to be ranking. positively associated with private credit, gross private capital flows, and net foreign direct According to Doing Business 2004, two factors investment (Commander and Tinn 2007). expand access to credit and improve its alloca- tion: credit information registries and creditor A 2007 World Bank review of financing constraints rights in the country's secured-transactions and suggests that information-sharing mechanisms bankruptcy laws. Good credit institutions matter most in low-income countries, while protect both creditors and debtors and make enforcement of creditor rights is more important everyone better off. Because credit histories are in high-income countries (World Bank 2007a). available, borrowers benefit from lower interest Other nonprice factors include: more important rates, as banks compete for good clients. limitations on access to credit, including geogra- phy (or lack of physical access); lack of proper There is research support for this customer documentation for identification, The literature supports argument. Love and Mylenko (2003) especially in low-income countries; and high the argument that credit found that private credit registries are minimum account balance requirements (World information registries positively related to availability of Bank 2007a). and creditors' legal bank financing for small and medium- rights help expand access size firms and that stronger rule of law In sum, the DB indicators are designed to to credit. is associated with more effective measure dimensions of the regulatory environ- private credit registries. Nevertheless, ment that are indeed important, although not they note that there is no evidence of causality equally important in all countries. The total tax between the creation of private registries and rate is anomalous, because although it is their effects on financing constraints. Jappelli important to business owners, it does not and Pagano (2002) found that bank lending is measure regulatory burden like the rest of the higher and credit risk lower in countries where DB indicators. For the most part, the indicators lenders share information, regardless of the measure actual legal rules and regulations and private or public nature of the information- informed estimates of practice, as distinct from sharing mechanism. Dorbec (2006) concluded opinion. Their relevance in a particular country 36 WHAT DO THE INDICATORS MEASURE? setting depends on the extent to which the laws There are a few systematic differences in country are applied, which DB does not measure. rankings associated with legal origins. These are Although in many circumstances it is the law on consistent with the ideas behind the DB the books that causes inefficient outcomes, framework and they have little impact on the understanding what actually happens on the overall rankings or the validity of the exercise. The ground is essential (La Porta and others 2007). employing workers indicator is consistent with The impact of a given reform will likewise vary relevant ILO conventions, but it does give higher across countries. scores to countries with lower job protections. 37 Chapter 4 Evaluation Highlights · DB's simple and bold communication is integral to the product, but at times simplicity comes at the expense of rigor. · DB has successfully spurred debate and motivated dialogue and addi- tional analyses on regulatory bur- dens and investment climate issues in developing countries. · As a cross-country benchmarking tool, the DB indicators cannot fully capture country-specific nuances and policy idiosyncrasies. Thus they have had less influence on design- ing reforms than on spurring debate. · The DB indicators' utility for re- search could be enhanced by ex- plaining the extent of data changes and making available previously published data sets. · The DB indicators are an important addition to the Bank's knowledge toolkit: they introduced benchmark- ing based on actionable indicators. · The DB indicators appropriately do not drive the Bank's operational or resource-allocation decisions. Women pack table grapes for export, South Africa. Photo reproduced by permission of Kip Ross/National Geographic Image Collection. Communicating and Using the Indicators T he DB indicators are designed to encourage policy makers to use them as an aid to decision making. Accordingly, DB makes communication and dissemination part and parcel of the core product. This chapter re- views how the DB indicator team communicates with audiences and how the DB indicators have been used in a variety of settings. Presentation Style disproportionately hurts smaller The DB reports are DB reports are presented in commendably simple businesses, especially in the services admired for their and straightforward terms. For example, "Egypt's sector where most women work."2 simplicity, but this reforms went deep" and "Thirty-nine countries Intercountry differences in female sometimes undermines made start-up simpler, faster or cheaper" (World unemployment rates actually reflect rigor. Bank-IFC 2007b, pp. 2, 3). A Foreign Investment many social, macroeconomic, and Advisory Service (FIAS) official said that countries business factors; the ease of paying corporate become interested in the DB issues because "at taxes plays only a small part. Another example: last they can understand a Bank report. Everybody "Each additional day that an export product is understands a ranking."1 A former prime minister delayed reduces exports by more than 1 per- interviewed for the evaluation said: "The World cent" (World Bank-IFC 2007b, p. 44). No source Bank is in the stone age. The public relations is cited for this statement, which in any case techniques are primitive. But IFC [International should be expressed as an association, since Finance Corporation] has done well with DB." But there are many other variables that affect dif- the drive for simplicity sometimes results in ferences in export volumes. inaccuracies or statements that are inadequately · The reports present information correlating supported by evidence. For example: performance on an indicator with broad eco- nomic outcomes such as increased foreign di- · Simple causal relationships are asserted where rect investment, although such links have not the evidence supports only association and been fully documented in literature. where the causal factors are complex. For ex- ample, Doing Business 2008 states, "Countries Lack of rigor in presenting information that make it easier to pay taxes have lower needlessly risks undermining DB's credibility. rates of unemployment among women. The The DB data and messages can and should be reason is simple: a burdensome tax system presented readably without sacrificing rigor. 41 DOING BUSINESS: AN INDEPENDENT EVALUATION Communications Strategy from the sales of its reports, estimated at $100,000 Doing Business stands out among per annum. A media strategy of Bank Group products for the variety maintaining high and innovativeness of the communica- The DB team monitors and reports on a range of visibility to promote tions tools it uses, as illustrated in box outputs from its communications program, debate and action is 4.1. The aim of the media strategy is to including its press citations, media events, Web integral to DB. promote action through increased site hits, downloads, and citations. The research public debate and competition among papers that are the basis for the DB indicators countries based on DB's annual benchmarking have been cited in 676 academic papers, accord- exercise. To achieve this, the primary effort of the ing to DB's count (World Bank-IFC 2007a). In communications strategy is simply to increase June 2007, the DB Web site was the World Bank media coverage and maintain a high and visible Group's most visited online database, with over international profile for the product. 120,000 hits. The results of this monitoring effort attest to DB indicators' high media coverage and The DB Web site is central Three stages of communication activi- public awareness. But the monitoring does not to its communications ties carried out following each year's ascertain systematically what results are achieved efforts. publication are: in changed public opinion and/or country policies. · Global and Regional pre-launch and post- launch virtual press conferences DB is effective in reaching audiences. Stakehold- · Road shows involving media events and pre- ers in 6 of 12 countries mentioned the DB indica- sentations to domestic policy makers and a tors unprompted when asked to recall economic diverse group of stakeholders (organized and and sector work that had been helpful or influen- sponsored in large part by Bank Group coun- tial.5 Bank Group staff noted that the DB indica- try offices)3 tors' extensive press coverage attracts the interest · Two-day workshops on DB findings and of senior policy makers, government officials, and methodology and local media events in 40 the business community in its messages. countries. Communications are primarily targeted to the 10 Box 4.1: Key Features of DB top reformers identified each year. In other Communications countries, road shows are held on the basis of demand from Bank and IFC country offices. An · Messages expressed in straightforward style additional key part of the communications · Report translated into up to 5 other languagesa strategy is the development and maintenance of · High-quality Web site with interactive capabilities an updated and interactive Web site. In fiscal 2007, · In-person and video presentations to country the DB team reported spending approximately stakeholders and decision makers $1,000,000 for dissemination events led by DB · Road shows and media presentations hosted by team management and members, support from a Bank Group country offices communications team, and maintenance of the · Customized country reports Web site.4 Not included in these estimates are the · Launches of spin-off publications, translations, time and costs incurred by country units in and topical reports (a) reviewing the DB reports and providing · Innovative approaches using social media comments, (b) explaining and address- · Active participation of product team in marketing The communications ing government comments on the DB and communication. outputs are monitored, data and methodology, and (c) techni- but not their influence on calassistance/trainingrelatedtotheDB a. Doing Business has been translated into French, Spanish, and Por- tuguese (2005, 2006, and 2007); Russian (2004, 2005, and 2006); Ara- public opinion or indicators paid for by the country bic (2004 and 2005); and Chinese and German (2005). policies. teams. The DB team receives revenue 42 COMMUNICATING AND USING THE INDICATORS How does this intensive dissemination translate of an overly simplistic connection The DB provides countries into practical use? Six ways in which DB has been between DB indicator ranking and with a cross-country used are reviewed below, along with the foreign investment may be greater in benchmarking tool. Its strengths and risks of each. countries with very limited capacity to comparative nature spurs undertake reform. policy debate. A Tool for Regular Cross-Country Benchmarking Even stakeholders who found DB indicators Stakeholders in all 13 countries reviewed6 useful for benchmarking questioned or criticized consider the chance to benchmark their country aspects of the methodology and process. Each against neighbors, peers, or competitors to be a year the DB team receives numerous queries and main motivator for dialogue about the business complaints from governments, both directly and environment. Country policy makers and channeled through Bank and IFC country staff, stakeholders use the DB indicators to compare about the rankings and how they are calculated. aspects of their regulatory framework with those "Often Country Management Units are called of neighboring or competitor countries and to upon by counterparts (often very irate counter- diagnose their weaknesses. This was the most parts) to explain the basis of scores or frequently and favorably cited use of DB indica- ranking."8 Apart from numerous Even stakeholders who tors noted by the evaluation. Seventeen of 29 challenges and debates about details of find DB benchmarking stakeholders (59 percent) interviewed in the 13 fact, country and Bank staff interview- useful question its countries as well as 24 of 42 Bank Group staff (57 ees raised methodological concerns methodology and process. percent) interviewed ranked the DB indicators about some of the areas discussed in "very useful" in enabling cross-country bench- chapter 2. Stakeholders noted that: marking,7 as box 4.2 illustrates. · Data informants are too few or represent Bank staff working in the Africa Region prominent law and accounting firms that are commented that the aggregate ranking may more likely to have primarily large and/or for- motivate governments to reform because they eign firms as clients. perceive it as a signaling device for potential · DB data is collected for the capital city and investors, especially foreign investors. The risks may not be valid for other parts of the country. Box 4.2: Keeping up with the Neighbors: DB Indicators Foster Benchmarking · Bank Group staff in Africa commented that DB indicators are · Algerian policy makers monitor how Algeria fares compared very useful because they provide cross-country bench- to Morocco and Tunisia. (Bank Group staff) marking data previously unavailable in many countries. Bank · Ranking with peers provides incentives for reforms, not the Group staff in Burundi noted that the DB indicators are the survey itself. I see the value of DB indicators in Albania when "only source out there" that allows for cross-country com- we have policy dialogue and tell the authorities Serbia did bet- parisons related to the business environment. ter last year and jumped X steps in the ranking because they did X, Y, and Z. (Bank Group staff) And elsewhere... · DB was used by donors, Bank, and others to point out the de- · By providing information about other countries, DB shows the ficiencies that the Investment Climate Assessment had potential for improving regulations and legislation. (Vietnam pointed to earlier by benchmarking Mongolia against other government official) countries. This really helped open the eyes of the government · The DB indicators help to raise greater awareness within the and Mongolians and galvanized them to take action. (Bank country on the need to improve our overall competitiveness. Group staff) (Tanzania government official) 43 DOING BUSINESS: AN INDEPENDENT EVALUATION DB rankings need to be · The firms depicted in the hypo- In both Moldova and the Netherlands, for interpreted with care. thetical cases are larger than the typ- instance, efforts to reduce the regulatory burden ical firm in middle- and low-income on business had begun before the advent of the countries, so the rankings may fail to DB indicators in 2003, but the DB indicators pick up improvements to the busi- increased awareness of regulatory issues and ness climate for micro and small increased the pressure for further reforms. enterprises. · Changes in methodology, data, and rankings Rwanda's Economic and Finance Commission make it difficult to explain and interpret DB asked DB to explain its methodology after the indicators to legislators and the public. country failed to make the top reformers list in · DB indicators lack a systematic validation the 2007 report. The presentation led to a process that draws on the experience of coun- workshop that involved over 70 participants try teams and country stakeholders. including legislators, officials, business persons, · The indicators omit measures of important con- and donors. The resulting task force remains straints on business, as discussed in chapter 3. under the aegis of the president's office. Clients' doubts about aspects of DB's methodol- Tanzania's multidonor Business Enterprise ogy could, if not allayed, jeopardize the use and Strengthening in Tanzania (BEST) program to impact of the report, as audiences question the streamline licensing and registration procedures relevance to their country's reality. As an Albanian got off to a slow start. The Bank's involvement, official expressed it, when there are as many along with the publication of the DB indicators, disagreements on indicators and overall rankings drew the attention of the president and other as there are now, the report is seen as a "dis- senior officials, helping Tanzania gain a top motivator" in international conferences on reformer spot in 2007. foreign direct investment. Bank Group staff observed that the Chinese authorities pay less DB has also inspired some countries to do attention to DB than to other cross- additional diagnostic work. In Peru, Doing DB has led to additional country benchmarks on rule of law, the Business 2006 drew the attention of Lima's diagnostic work in Peru investment environment, global com- mayor to the difficulties of starting a business in and Nigeria. petitiveness, trade and logistics, and the capital. Drawing on diagnostic work by FIAS corruption, because they consider and technical assistance from the IFC, the China's poor DB rankings inconsistent with its municipality reformed the process for obtaining strong private sector growth. More generally, a business license. The reform template is now stakeholders in all the case study countries noted being promulgated by the National Council for that they find the general findings of the DB the Simplification of Municipal Procedures for cross-country benchmarking useful, but do not Businesses. In Nigeria, the United Kingdom's always rely on the exact numbers. Department for International Development (DFID) is supporting the collection of DB indica- A Catalyst for Dialogue tors for every state. The data will be used for In many countries, discussion around DB, even diagnostic analysis and as benchmarking by the when contentious, has opened up a productive government and donors. dialogue between policy makers and other stakeholders about the business climate. DB's A Guide to Policy Reform active dissemination and simple communications While the vast majority (85 percent) of interview- style permits widespread press coverage, foster- ees affirmed DB's usefulness for motivating ing interest from business and NGO reforms, less than half (44 percent) considered it DB has successfully communities, and attracting the helpful as a guide to action because it offers little stimulated dialogue on attention of the most senior policy guidance about the priorities, sequencing, and business climate issues. makers. policy coherence needed to implement a 44 COMMUNICATING AND USING THE INDICATORS successful reform program. Moldova's home- DB's direct impact is thus difficult to Because the DB grown Cost of Doing Business assessment, which determine and appears limited, even indicators cannot and do surveys perceptions of 600 small and medium- in countries designated as top reform- not capture country- size enterprises, finds that businesses are most ers. For instance, in Tanzania, Doing specific policy nuances, concerned about arbitrary interference from Business 2007 noted improvements in they have had less police and uniformed services in their daily trading across borders because of influence in designing operations--a topic outside DB's ambit. Bank modernization of customs proce- reforms. Group staff working on China noted that a dures. Staff and stakeholders noted perception-based survey of businesses in 120 that while DB motivated authorities to look at cities was helpful in highlighting the constraints the issue, the process of modernizing pro- faced by businesses and encouraged reform in a cedures relied heavily on other Bank diagnostics number of cities. A Bank Group staff member and the country's own detailed studies. To guide working on Albania stated, "We cannot build its business climate reforms, the Netherlands projects or TA [technical assistance] programs uses its indigenously developed Standard Cost on the DB indicators. It is just indicative about Methodology, with an added emphasis on the business climate and is used to provide regulatory burdens that require the incentive to countries to improve the business firm to undertake activities outside DB identifies countries as climate. Our counterparts understand the limita- the scope of regular business reformers based on tions of the methodology." operations. The DB indicators are changes in country seen as a useful tool for monitoring rankings, without regard DB annually designates 10 countries as top progress, but not a principal source to the relevance and reformers. The "Reformers Club" provides a for prioritization of government quality of the reform . . . forum for recognizing countries that have made actions or policies. the largest changes in ranking in a given year. These are countries that have both improved As noted in chapter 3, since the DB indicators their rankings on at least three individual in- cannot capture country-specific policy nuances, dicators--indicating "breadth of reform"--and they cannot and do not help counties to situate improved the most on their overall EODB particular improvements within broader reform ranking from the previous year--indicating efforts nor ensure adequate sequenc- "depth of reform." This method rewards the ing and policy coherence, needed to . . . it is thus not suited to quantity of rankings changes and does not implement and sustain the changes in designing reform attempt to assess whether the changes consti- legislation. For instance, in Algeria programs targeting tute important or meaningful reforms. While the and Moldova, the governments critical bottlenecks. approach is practical and transparent, as experimented with the creation of Hausmann, Rodrik, and Velasco (2005, pp. 5­6) one-stop shops for licensing, which in Algeria note, "We cannot be assured that any given included construction permits and was counted reform taken on its own can be guaranteed to be by DB as a reform. But the pilot in Algeria is welfare promoting, in the presence of multitudes underutilized, and Moldovan ministries contin- of economic distortions. . . . and welfare may not ued to require other forms of revenue- be increasing in the number of areas that are generating activities such as "authorizations" reformed." An alternative approach would be to and "permits" before procuring a license. The design reforms to address the most "binding one-stop shop simply added another regulatory constraints" in order to produce the biggest bang layer. for the reform buck (Hausmann, Rodrik, and Velasco 2005, p. 7). Because the DB indicators An obstacle to using the DB indicators as a guide neither prioritize among the 10 dimensions nor to action is that some indicators measure special- provide detailed country-level analysis, they are ized aspects of a larger problem, as discussed in not suited to designing reform programs chapter 3. Many stakeholders mentioned that targeted at critical bottlenecks. DB's dealing with licenses indicator, which 45 DOING BUSINESS: AN INDEPENDENT EVALUATION Some DB indicators relates to construction permits, has measures that are "strictly DB-related."10 While measure specialized had limited usefulness in guiding the this unit does aim to introduce and involve aspects of a larger broad reforms of licensing. For experts in the relevant areas from other parts of problem. example, Niger reduced the number the Bank, it expects that "broader reform sugges- of procedures measured by dealing tions" will be managed by the other relevant with licenses from 27 to 19 by eliminating units. To avoid an implied endorsement of quick numerous security inspections at the construc- fixes, the Rapid Response Unit will need to ensure tion site, but this reform is unlikely to result in its advice to a country is integrally aligned, and real improvements to the licensing regime as perceived by clients to be aligned, with the Bank a whole. Similarly, in Tanzania, Nigeria, and Group's overall Private Sector Development Rwanda, access to credit and cost of financing (PSD) Assessment and other recommendations. are important constraints not measured by DB's getting credit indicator; while the credit bureau The DB indicators do not capture the extent to information it does measure is not considered which changes in legislation or streamlining of relevant by stakeholders in Rwanda. procedures are actually implemented. There is no clear articulation of the impact of the DB- Have countries tried to improve their ratings by measured reforms on firm performance, percep- changing the letter of the law without making tions of regulatory burden, or the overall serious reforms? Interviewees with Bank and regulatory environment in a country. In Vietnam, IFC country teams, FIAS, and the Millennium Bank Group staff reported using DB's getting Challenge Corporation (MCC) cited credit indicator to open a dialogue with the IEG did not find evidence instances of country officials asking Ministry of Justice on a plan for improving the that countries took an how to increase their DB rankings collateral lending environment and to advocate unduly narrow approach (India) or making the increase in DB for the creation of a private credit bureau in the solely to increase their rankings a goal in itself (Georgia and State Bank of Vietnam. But both staff and an rankings. Madagascar). The ratings simulator in informed stakeholder noted that only a small the DB Web site encourages users to number of individuals and businesses were using see the effects of possible changes. And the DB the collateral registry. It is not clear what the team has provided country officials with sugges- effects of the private credit bureau and collateral tions of specific actions they could take to system have been. improve their indicators upon request. But none of the 13 countries reviewed by this evaluation A Research Tool took an unduly or cynically narrow approach or DB's near-universal country coverage, combined "easy" steps purely to affect the ratings. Even with the accessibility of the data and methodol- where officials initially aimed to reform only ogy notes on the Web site, make it a useful tool those aspects measured by DB, they were for analyzing regulatory issues. But yearly persuaded by Bank and IFC staff to take a changes in methodology and retroactive changes comprehensive approach to business climate re- to prior year data without making available forms. FIAS and MCC staff noted that they use previously published data sets makes it difficult such inquiries to open a dialogue on genuine for research to be validated and replicated. This reform options.9 Often this involves detailed disadvantage could be attenuated if the Web site explanations by Bank staff and other donors to fully disclosed and explained all corrections and country counterparts about the changes and their effects on the rankings and The FIAS Rapid Response methodology of DB and what each provided previously published data sets.11 Unit will need to ensure indicator measures. its advice is integrally A second disadvantage is the small number of aligned with the Bank FIAS has recently created a Doing informants supplying the underlying data, as Group's advice and Business Rapid Response Unit with a discussed in chapter 2. Given the very small recommendations. mandate to help countries adopt number of completed questionnaires on each 46 COMMUNICATING AND USING THE INDICATORS indicator in a given country, it is not possible The "IDA results framework"--a tool Continuous revisions to to calculate meaningful standard errors or to help IDA donors track development the DB data and confidence intervals. DB needs to make transpar- results in IDA countries--reports DB unavailability of ent the number of completed questionnaires that numbers on the time and cost to start previously published data form the basis for each indicator in a country. a business as 2 of the 14 "outcome" sets limit its usefulness indicators (World Bank 2007b). This for research. A Criterion for Operational Decisions results framework is an ex-post report- The MCC uses two DB subindicators--days and ing mechanism and is used neither to allocate cost to start a business--in its formula for select- resources nor to guide IDA programs ex-ante. ing countries eligible for grants. Together, these subindicators account for 6 percent of a The DB indicators are used to monitor DB plays a role in country's score. In 2008, MCC will add DB's progress of lending operations, includ- determining eligibility for subindicators of time and cost to register ing development policy loans that deal grants made by the property, raising DB's weight to 9 percent. The with private sector development issues. United States' Millennium DB indicators have, according to MCC officials, In 6 of 11 countries reviewed by this Challenge Corporation. sparked more interest by ministers than most evaluation, such operations used DB other parts of their scorecard because they are indicators as one of the key monitoring indicators easy to understand and convey to the public and for specific components.14 For instance, the time point to specific areas that may need improve- and cost of starting a business is used to monitor ment. On the down side, MCC officials noted that progress of one component of Tanzania's Private eligibility decisions had been made on the basis Sector Competitiveness Project. of DB data that were subsequently changed. The high stakes of MCC eligibility make it all the more An Addition to the Bank's Toolkit important for DB to stabilize the methodology, DB has helped to define a new role for the Bank make clear that posted data are subject to in development assistance. A majority of the change, and make available both original and stakeholders interviewed noted that DB is one of modified data sets. the first initiatives to develop objective (that is, non-perception-based) cross-country data, and In the World Bank Group, DB plays an indirect thus fill a critical gap in knowledge.15 It is a role in assessing countries' policy frameworks. Six "knowledge" product, as distinct from analysis of the 10 DB indicators are used as "guideposts" done to support lending and related conditional- (along with Investment Climate Assessments and ity. DB draws on the Bank's unique position to other sources) to assist country teams in assemble information on a global scale. While determining country scores on "Business Regula- other indicators, such as the World Development tory Environment," one of the 16 criteria of the Indicators, cover many countries, DB DB does not affect the Country Policy and Institutional Assessment incorporates indicators that are defined Bank's resource (CPIA), the most important (but not the only) specifically enough to determine allocation decisions. determinant of allocations to International actionable steps. Development Association (IDA) countries.12 In addition, the DB's employing workers indicator is The DB model--use of a standard case methodol- one of several guideposts for the CPIA's social ogy, expert informants, and rankings--is being protection and labor criterion, even though this replicated in other indicators. For instance, the indicator captures only the administrative burden Logistical Performance Index devel- to firms of issues such as retraining and severance oped by the World Bank's Trade Group Among the Bank's pay, rather than a broad assessment of a country's ranks the quality of infrastructure, knowledge tools, DB is social protection policies. The extent to which customs procedures, and logistic costs one of the first to the guideposts (including the DB indicators) in 150 countries based on information introduce indicators influence the CPIA scores will be reviewed in a from freight forwarders, transporters, aimed at defining forthcoming IEG evaluation of the CPIA.13 and officials. Other efforts to replicate actionable steps. 47 DOING BUSINESS: AN INDEPENDENT EVALUATION this model in the financial sector are under way. linear trajectory for improvement and a clear definition of what constitutes a reform. The Bank To what extent can the Bank scale up or replicate should leverage the DB methodology in areas that the DB model in other areas of development? share these characteristics. Many development issues lack the DB indicators' critical characteristics of a widely accepted and 48 Chapter 5 Wig shop owner using laptop. Photo reproduced by permission of Cat Gwynn/Corbis. Findings and Recommendations D oing Business has contributed to the development landscape in three main ways. For country authorities, it sheds a bright, sometimes un- flattering, light on regulatory aspects of their business climate. For busi- ness interests, it has helped to catalyze debates and dialogue about reform. For the Bank Group, it demonstrates an ability to provide global knowledge, independent of resource transfer and conditionality. The annual exercise gen- erates information that is relevant and useful. But it has several weaknesses in process, content, and presentation that should be rectified soon if it is to maintain its credibility and usefulness. The Framework Underlying The DB exercise reflects these inherent limita- the DB Indicators tions. As an exercise in cross-country compari- The DB indicators are anchored in research that son, it is not intended to capture country nuances links characteristics of the regulatory environ- and nonlinear relationships. It measures selected ment to firm performance, and thence to dimensions of the regulatory environment, some macroeconomic outcomes. Although some of which are bound to be irrelevant in some research has convincingly demonstrated these countries. It notes the costs of regulation but not associations, any research relating the regulatory the benefits. Seven of DB's 10 indicators presume environment to economic outcomes is necessar- that lessening regulation is always desirable, ily partial. It does not capture the influence of all whether a country starts with a little or a lot of the other determinants, nor can it pin down the regulation. These limitations do not invalidate the direction of causality. Even where an association exercise, because the scope and thrust of DB are is demonstrated, the policy implications are not consistent with a credible view that less burden- self-evident, since regulations deliver benefits as some business regulation is associated with well as costs. What is good for a firm (or firms) better private sector performance. But they may not be good for firms at large, or the underscore the need to use caution in interpret- economy and society as a whole. The right ing the results and for the DB indicators to be balance for any country is a matter of political used in conjunction with complementary tools choice. such as Investment Climate Assessments. 51 DOING BUSINESS: AN INDEPENDENT EVALUATION The Scope of the Indicators process based on a global network of volunteer The indicators are consistent with their claim of informants. This process is capable of generating measuring the content of rules and laws, as reliable data, but three areas of vulnerability need distinct from perception. The dimensions of the to be addressed. regulatory environment that they measure are important in the aggregate--but not all indica- First, the database is too dependent on a small tors are important in all countries. The relevance number of informants, with some data points of an indicator in a particular country setting will generated by just one or two firms. For the depend partly on the extent to which the law is information about taxes, DB's exclusive reliance actually applied, which DB does not measure. on a single global firm for both the underlying The impact of a given reform will likewise vary methodology and the data from 142 countries across countries. While the addition of new poses extra risks. The number and diversity of indicators would expand the coverage of issues informants should be increased and their addressed, by itself this would not make the DB information validated more systematically. An indicators more reliable or useful. The more increase in the informant base will require a immediate challenge is to enhance reliability of systematic vetting process. Simplifying the the underlying information, as discussed below. questionnaire may also help to encourage more informants to contribute. The employing workers indicator is consistent with the letter of ILO provisions, but four Second, DB makes easily available a great deal of measures do not reflect their spirit. Beyond these data and explanatory material--arguably more minimum standards, the DB criteria give lower than most comparable exercises. Yet it remains scores to countries that have opted for policies of insufficiently transparent about the number and greater job protection. There are a few systematic types of informants for each indicator in a country, differences in country rankings associated with the adjustments staff make to the information legal origins in civil or common law. These are supplied by informants, and the changes made to consistent with the stated ideas behind the DB previously published data. It does not adequately framework and they have little impact on the point out the possibilities of errors and biases. overall rankings or the validity of the exercise. DB's measurement of the total tax rate is Third, DB makes much of its country rankings. anomalous because unlike DB's other subindica- The rankings entail three weaknesses. Because tors, it does not measure regulatory burden most DB indicators presume that less regulation alone. It derives from a country's fiscal require- is better, it is difficult to tell whether the top- ments and policy context. Moreover, the ranked countries have good and efficient regula- complexity of this subindicator necessitates DB's tions or simply inadequate regulation. The small reliance on PwC as virtually the sole informant-- informant base makes it difficult to measure a reliance that entails risk to the exercise. Because confidence in the accuracy of the individual tax rates are important for investors, information indicator values, and thus in the aggregate about them should be collected and presented, rankings. Finally, changes in a country's ranking but not included in the rankings. depend importantly on where it sits on the distri- bution; small changes can produce large ranking Inaccurate nomenclature and overstated claims jumps, and vice versa. These factors contribute of the indicators' explanatory power have to anomalies in rankings. provoked considerable criticism from stakehold- ers. DB reports should seek to retain their clarity These issues may not in and of themselves while using less sweeping language. jeopardize DB's reliability, but the lack of transparency about them undermines DB's Reliability of Information credibility and goodwill. In addition, the lack of DB has created a unique information-gathering stability in the data and the failure to make 52 FINDINGS AND RECOMMENDATIONS available prior versions of data that have been some risk to the MCC because of the continuous adjusted limit DB's usefulness for research. DB's revision of DB data after publication. documents and presentations should include full explanations and cautions on these points. Implications for the Bank Group In addition to the findings for the design and use Motivating and Designing Reforms of the DB exercise itself, the evaluation has The DB indicators have been influential in generated two implications for the Bank Group motivating policy makers to discuss and consider more broadly. business regulation issues. Its active dissemina- tion in easy-to-understand language permits The Bank Group, by so prominently recognizing widespread press coverage and generates interest DB's highly ranked countries, may be inadvertently from businesses, NGOs, and senior policy makers. signaling that it values reduced regulatory burdens more than its other development goals. Although The DB indicators have had less influence on the the Bank Group's approach entails helping choice and design of specific reform programs. countries achieve a wide range of objectives, it has Most Bank Group staff and country stakeholders no comparable way of celebrating improvements in report that they draw on a range of analytical other important development outcomes such as material to determine the nature, sequence, and poverty reduction, public sector effectiveness, or direction of reforms; the DB indicators have the Millennium Development Goals. limited use in this regard. There is little evidence that the DB indicators have distorted policy The DB exercise has demonstrated that a cross- priorities in the countries or in the Bank Group's country ranking exercise can be effective in programs, or that countries have implemented spurring dialogue and motivating interest and reforms with insincere motives. The DB indica- action. Can it be extended to other topics and tors do not play a role in IDA's resource alloca- issues? It can be used for issues that meet two tion process. Their use by the United States' conditions. There must be measurable indica- MCC as a basis for resource allocation poses tors to serve as agreed proxies for the target Box 5.1: If DB Were to Be Extended to Other Topics Bank Group management may consider building on DB's expe- · Identify target audience: Country benchmarking can be an rience by creating indicators on additional development topics. effective door-opener and motivate a wider dialogue. If so, the evaluation offers five lessons: Consider in advance who the indicators should aim to influ- ence and who could participate in the dialogue. · Choose what to measure and start small: Use existing or new · Create and maintain competitive pressure: Any indicator research to identify a few issues within a sector/theme that can be effective only to the extent it is widely communicated can serve as at least partial proxies for development. Then and understood by the target audience and can generate specify some quantitative variables that can be measured rel- competition among countries and pressure to reform. The atively easily, have an intuitive appeal, and are easily un- DB's assertive marketing and communication strategy com- derstood. This implies accepting that the indicators will be bined with its use of rankings helped to generate and main- limited in scope, not comprehensive. tain country interest. · Look for efficiency in data collection and processing: Data · Do not overstate the implications of the rankings: Cross- collection methods need to be simple. Use an appropriately country rankings inherently miss country-specific issue nu- diverse range of expert informants and provide informants ances. They have to be used in conjunction with other with a common reference point such as a hypothetical analyses to help countries determine the direction, nature, scenario. and sequence of reforms. 53 DOING BUSINESS: AN INDEPENDENT EVALUATION outcomes. And the direction of improvement present simple information on corporate must be the same across countries starting at tax rates, but exclude it from the rankings (as widely different levels. For many development it does for the information it collects on issues, the trajectories for change are not linear, nonwage labor costs in the employing work- but U-shaped. (For example, automated teller ers indicator). A wider range of informants machines are an indicator of efficient financial should also be engaged in supplying infor- services. At an earlier stage of development, more mation for the paying taxes indicator. automated teller machines are better, but in 2. To make its reform analysis more meaningful, the mature economies, too many can be a sign DB team should: of inadequate interbank networking). Further a. Make clear that DB measures improvements to lessons from the evaluation are distilled in box 5.1. regulatory burdens and costs, which is only one dimension of any overall reform effort of Recommendations the investment climate for private sector growth. The DB indicators measure reduc- 1. To improve the credibility and quality of the rank- tions in regulatory burdens and should be rec- ings, the DB team should: ognized and rewarded as such. These a. Take a strategic approach to selecting and in- improvements should not be characterized as creasing the number of informants: reforms of the overall business climate, which ­ Establish and disclose selection criteria reflects a number of non-DB-measured as- for informants. pects, as noted in figure 1.1. ­ Focus on the indicators with fewest in- b. Trace the impact of DB reforms at the country formants and countries with the least re- level. The DB team should work with coun- liable information. try units to analyze the effects of imple- ­ Formalize the contributions of the sup- menting the reforms measured by the DB plemental informants by having them fill indicators (such as revised legislation or out the questionnaire. streamlined processes) on: (i) firm per- ­ Involve Bank Group staff more actively to formance, (ii) perceptions of business man- help identify informants. agers on related regulatory burdens, and b. Be more transparent on the following aspects (iii) the efficiency of the regulatory envi- of the process: ronment in the country. ­ Informant base: Disclose the number 3. To plan additions to or modifications of the indica- of informants for each indicator at the tors, the DB team should: country level, differentiating between a. Use Bank analyses to drive the choice of DB in- those who complete questionnaires and dicators. These would include Business those who provide "supplemental" EnterpriseSurveys,InvestmentClimateAssess- information. ments,andotherrelevantBankanalysestoas- ­ Changes in data: Disclose a list of all data sess what stakeholders deem to be important corrections and changes as they are made. priorities for domestic private sector growth. Explain their effect on the rankings, and, The DB team should use such analyses to de- to facilitate research, make available all termine the choice of new indicators and pe- previously published data sets. riodically reassess its current set. ­ Use of the indicators: Be clear about the b. Pilot and stabilize the methodology before in- limitations in the use of the indicators for cluding new indicators in rankings. Frequent a broader policy dialogue on a country's changes in methodology make comparison development priorities. across time less meaningful. New indica- c. Revise the paying taxes indicator to include only tors should be piloted--that is, data col- measures of administrative burden. Since the tax lected and published for comment, but not rate is an important part of the business cli- factored into the rankings--until the mate, DB should continue to collect and methodology is validated and stabilized. 54 Appendixes Village shop at dusk, lit by solar panels, Sri Lanka. Photo courtesy of Dominic Sansoni/World Bank. APPENDIX A: METHODOLOGY This evaluation covers the period from the first (c) Patterns by legal system: The evaluation DB report published in 2004 to the report analyzed patterns in the values of the subindica- published in September 2007. Data analysis is tors for countries with particular legal systems based on a download of the full data set from the according to legal origin. The results of this DB Web site in August 2007 and, where noted, as analysis are presented in appendix D. subsequently revised in October 2007. Where appropriate, updating references are made to 2. Country Case Studies the 2008 report. In all, the evaluation interviewed Thirteen country case studies were used as the 167 individuals: 72 Bank Group staff, 40 DB basis for detailed quantitative analysis and to informants, 22 government officials, and 33 obtain qualitative information from interviews other stakeholders, including representatives with Bank and IFC staff, private sector represen- from the private sector, international donor tatives, government officials, and donors (see agencies, and academia. table A.1). Seven of the countries were randomly selected from the total 175 countries covered in The evaluation used the following methods to Doing Business 2007. An additional 6 were gather evidence: randomly selected from the subset of 19 countries that DB identified as "top reformers" 1. Analysis of DB Ratings and Underlying in the 2006 and 2007 reports.1 Raw Data (a) Range, means, and distribution of subindi- For all the case studies, evaluators interviewed key cators and indicators and simulation of Bank and IFC staff and stakeholders in person, by reforms: The evaluation calculated the range, telephone, and/or by e-mail, using uniform means, frequency distribution, and other charac- interview protocols developed by the evaluation teristics of DB data. The pair-wise correlations team (see appendix E for a sample of the interview among indicators and subindicators were protocols). Telephone calls were used as appropri- calculated. A simulation was conducted of how rankings would vary for a given change in the underlying indicator (see appendix B for details). Table A.1: Case Study Countries (b) Revisions in prior data: The DB team period- ically revises data for prior years. The evaluation Country case studies Top reformer case studies assessed the revisions made to the data Albania China published in the DB 2007 report as part of the Algeria Netherlands process of the DB 2008 report. It assessed the Burundi Peru volume and reasons for the changes and their Moldova Rwanda impact on the indicators and overall EODB Mongolia Tanzania ranking, as well as on the identification of reformer countries. This analysis in reflected in Nigeria Vietnam appendix C. Spain 57 DOING BUSINESS: AN INDEPENDENT EVALUATION ate to clarify and supplement information received tion, and the Economist Intelligence Unit. The by e-mail. In addition, the evaluation visited team reviewed internal correspondence from Moldova and conducted 12 face-to-face interviews operational staff commenting on the DB process with governmental and nongovernmental stake- and indicators for the 2007 and 2008 reports. holders. The mission observed the DB team's videoconference presentation of the 2008 report 3. Validation Exercise to an audience in Chisinau on November 2, 2007. The evaluation reviewed the data collection The mission also visited the Netherlands and process in the seven country case study countries interviewed four country stakeholders. For the through a review of the completed question- case studies, the evaluation conducted a total of naires and comparison with the final published 100 interviews: 55 Bank and IFC staff, 22 govern- data, and interviews with informants based on ment officials, and 23 other stakeholders, includ- standard guidelines. ing representatives from the private sector, international agencies, NGOs, and research think In the seven country case study countries, a total tanks. The evaluation team interviewed IFC staff of 68 informants are listed by DB for the 5 focus working on investment climate issues in the indicators (see table A.2). The evaluation team Private Enterprise Partnership (PEP) facilities and made at least three attempts to contact each of FIAS,aswellasBankstaffworkingonprivatesector them and succeeded in contacting and interview- development issues and relevant projects and ing 59 percent (40 informants) by phone or by e- analytical and advisory activities (AAA), as well as at mail. Of the 28 informants who could not be least one person from the country management contacted, 19 had unusable contact information team. These staff directed IEG to the two to three or did not respond after repeated attempts, 7 people in the government and donor community had left their position, and 2 had died. most knowledgeable about the DB exercise. The evaluation also analyzed the composition The case studies also included reviews of Bank and characteristics of the informants for all 175 documents, including Country Assistance Strate- countries in Doing Business 2007 (see chapter 2 gies, Investment Climate Assessments, economic for details on the findings from the validation and sector work, and project documents related exercise). to private sector development, as well as other assessments of the business environment from 4. In-Depth Analysis of Five Indicators the World Economic Forum, Heritage Founda- For assessing the relevance of the indicators to Table A.2: Reach of the Validation Exercise Percent of Percent of all informants Percent of Questionnaire Supplemental all questionnaire (questionnaire all informants Country informants informants Total informants and supplemental) (68 total) Albania 5 2 7 62 88 10 Algeria 3 1 4 38 36 6 Burundi 2 2 4 33 36 6 Moldova 3 1 4 43 57 6 Mongolia 2 1 3 25 38 4 Nigeria 7 3 10 70 63 15 Spain 7 1 8 35 38 12 Total 29 11 40 AVG 44 AVG 51 AVG 59 58 APPENDIX A: METHODOLOGY countries and relevant intermediate outcomes, themes that correspond most directly with the the evaluation focused its analysis on five broadly investment climate issues covered by DB. As representative DB dimensions: starting a these themes cover all sectors, the review identi- business, employing workers, enforcing con- fied 130 projects that were mapped to the tracts, getting credit, and paying taxes. The team Financial and Private Sector Development Sector reviewed relevant literature and interviewed 8 Board and approved between fiscal years 2004 (non-country-specific) Bank Group staff and 10 and 2007. other subject matter experts. As depicted in figure A.1, the Bank provided $9.8 5. Portfolio Review billion in loans and grants for the 130 projects The evaluation reviewed the portfolio of Bank mapped to the Financial and Private Sector investment operations and IFC technical Development (FPD) Sector Board. Not all of this assistance and advisory services to identify funding was related to strictly DB-measured patterns and trends in the Bank's support of indicators. Regulation and competition policy, private sector development, and specifically the small and medium-size enterprise support, and areas related to the 10 dimensions of the export development and competitiveness have business environment measured by DB between the most funding and account for nearly three- fiscal years 2004 and 2007. quarters (72 percent) of the total $4.8 billion allocated to the 11 DB-related themes. Project descriptions do not explicitly identify the costs related to the dimensions covered by DB. To estimate how much IFC allocated to technical To estimate the volume of Bank operations assistance and advisory services for DB-related related to the 10 dimensions covered by DB, the areas, the evaluation reviewed the six subareas of evaluation team selected 11 (of a total of 71) business lines that correspond most directly with Figure A.1: Financial and Private Sector Development (FPD) Sector Board Projects by Theme, Fiscal 2004­07 Total funding: 9.8 billion USD Improving labor markets Corporate governance Export development and competitiveness Macroeconomic management and international financial architecture 1% Personal and property rights Other themes 17% Regulation and competition policy DB-related themes 50% (4.8 billion USD) Small and medium-size enterprise support Other FPD themes 32% Tax policy and administration Trade facilitation & market access Judicial and other dispute resolution mechanisms, law reform, and legal institutions for a market economy 59 DOING BUSINESS: AN INDEPENDENT EVALUATION Figure A.2: IFC Technical Advisory Funding, Fiscal 2004­07 Total funding: 646.7 million Credit bureau Value addition (access to finance) to firms Dispute resolution 23% Subareas in BEE unrelated to DB 2% Diagnostic and M&E Infrastructure DB-related subareas 20% 16% (102 million USD) Policy, regulation, and institutions Environment and social sustainability 11% Access to finance Subnational 28% Cross-border BEE Note: BEE = Business Enabling Environment business line; M&E = monitoring and evaluation. the DB indicators. These were one from the nings for the approach adopted by DB. The same Access to Finance business line (credit bureau) review also undertook a cross-country econo- and five from the Business Enabling Environment metric analysis to: (1) assess the consistency of (BEE) business line (dispute resolution, diagnos- the indicators with other Bank and externally tic and monitoring and evaluation [M&E], policy, generated indicators of investment climate and regulation and institutions, subnational, and business regulation, and (2) determine correla- cross-border). As shown in figure A.2, of the 906 tions between the DB indicators and the technical assistance projects undertaken by IFC economic variables that one may expect to be between 2004 and 2007, $102 million (16 percent affected using both aggregate and firm-level data. of a total of $647 million) were spent on these six This background paper is available upon request subareas. Diagnostic and M&E and policy, regula- and will be made available on the IEG Web site. tion, and institutions account for more than two- thirds of this amount. 7. Use and Communications The evaluation interviewed staff at the Bank, IFC- 6. Literature Review FIAS, and the MCC and reviewed pertinent The evaluation commissioned a review of litera- documents in connection with how the DB indica- ture on the theoretical and empirical underpin- tors are used in various operational contexts. 60 APPENDIX B: HOW EQUITABLY DO THE RANKINGS REWARD REFORMS? Each of DB's 10 indicators uses cardinal values necessarily the same as that between those ranked for its subindicators: time, cost, number of first and second. Figure B.1 illustrates this point by procedures, and so on to create a ranking. These showing the frequency distribution for the total cardinal values are ranked according to their tax rate as a share of profits, a subindicator of respective percentiles in each of the subindicator paying taxes. There is a 5.1 percentage point distributions. The subindicator percentiles are difference between the top performer, Maldives, then averaged to come up with an indicator-level and the runner-up, Vanuatu. There is a 4.7 percentile; the 10 indicator percentiles are then percentage point difference between the last and averaged to generate the overall ease of doing next-to-last countries in the distribution, Gambia business (EODB) ranking.1 and Burundi. However, the countries ranked fifty- ninth and sixtieth, Israel and Mozambique, are The use of several levels of ordinal rankings separated by just 0.1 of a percentage point (39.1 obscures the underlying cardinal values. That is, percent and 39.2 percent respectively), while the magnitude of the difference between the there are 13 other countries accompanying them countries ranked, say, fifty-ninth and sixtieth is not in the range between 37 percent and 40.3 percent. Figure B.1: Difference between Ranks Can Vary Total tax rate ­ frequency distribution Israel and Mozambique 12 Latvia 10 Maldives 8 countries of 6 Number 4 Gambia 2 Vanuatu Burundi 0 9 21 32 44 55 66 78 89 101 112 123 135 146 158 169 180 192 203 215 226 238 249 260 272 283 Total tax rate as a share of profits (%) 61 DOING BUSINESS: AN INDEPENDENT EVALUATION Table B.1: Countries in the Bottom Quartile on the Paying Taxes Indicator Need to Reduce Taxes More to Increase Rankings Relative to Countries in the 2nd and 3rd Quartiles Simulated Difference Total tax Total tax Rank paying rank with in paying Country rate 2007 (%) rate 2008 (%) taxes 2007 2008 value taxes rank Latvia 43 33 52 35 17 Botswana 53 17 67 18 49 Kuwait 56 14 41 8 33 Belarus 186 144 175 175 0 Sierra Leone 277 234 138 137 1 A given change in a cardinal value, such as a Almost all the countries (165, or 94 percent) fall reduction in the time needed for a procedure, is within one standard deviation from the mean. more likely to advance a country's rank (holding Table B.1 presents the results of simulations2 after other countries' actions constant) if the country improvements in the total tax rate. Sierra Leone is starts from a more concentrated segment of the in the dispersed segment at the bottom of the distribution than if it starts from a more total tax rate distribution, right before Burundi dispersed section. This arithmetic means that and Gambia. Despite a 43 percentage point countries at the more dispersed parts of the reduction in total tax rate, the country improved distribution have to work harder to see changes only one position in the simulated ranking for in their overall rankings. Put differently, countries paying taxes. Belarus's substantial tax reduction can make significant changes that do not likewise did not affect the simulated ranking. improve their rankings if they are at the Latvia, by contrast, despite only reducing the total dispersed sections of the distribution for that tax rate by 10 percentage points, improved 17 indicator. The following three examples illustrate positions because it is situated in the most this asymmetry by simulating the change in populated segment of the distribution. Kuwait rankings for a subindicator, holding the actions and Botswana received an even stronger boost of the other countries constant. from their tax reduction because of the same effect. Example 1: How much does the tax rate have to fall to improve ranking on paying taxes? As seen in figure Example 2: How does reducing the minimum capital B.1, the frequency distribution for total tax rate as requirement affect ranking on starting a business? In a share of profits for all countries ranges from 9.3 2008, Egypt drastically reduced its minimum percent in Maldives to 291.4 percent in Gambia. capital requirement--from 695 percent of Table B.2: Despite Egypt's Efforts In Reducing the Minimum Capital Requirement, St. Kitts and Nevis, Gambia, and Macedonia Will Gain More on DB Rankings for Lower Reductions Minimum Minimum Simulated rank, capital capital Rank, starting starting a Difference in requirement requirement a business, business, with starting a Country 2007 (%) 2008 (%) 2007 2008 value business rank Finland 27 8 19 13 6 St. Kitts and Nevis 45 0 105 61 44 Gambia 120 0 124 70 54 Macedonia, FYR 112 0 76 27 49 Egypt 695 13 125 92 33 62 APPENDIX B: HOW EQUITABLY DO THE RANKINGS REWARD REFORMS? Figure B.2: Distribution of the Minimum Capital Requirement Subindicator for Starting a Business Minimum capital requirement ­ frequency distribution 70 60 50 40 Finland Frequency 30 20 St. Kitts & Nevis Macedonia and Gambia 10 Egypt, Arab Rep. of 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 85 95 150 200 250 300 350 400 450 500 550 600 650 700 800 900 1050 1100 1900 2600 4250 Minimum capital requirement as a share of GNI per capita income per capita to just 13 percent. Holding boosted their rankings more than Egypt would other countries' actions constant, it would have have. By eliminating the minimum capital generated a 33-position boost in the starting a requirement, these three countries tied with the business ranking (see table B.2). The distribution other 66 countries for first place in this subindica- of this subindicator, as shown in figure B.2, is tor. In turn, this substantially reduced their total concentrated around zero. More than a third of average percentile for starting a business, the countries (66 of them) do not have a improving their ranking for this indicator. Finally, minimum capital requirement. Although Gambia, a country such as Finland was also able to advance Macedonia, and Saint Kitts and Nevis all reduced in the rankings, although less than the other the minimum capital requirement much less than countries, because of the relative lack of concen- Egypt in absolute terms in 2008, they would have tration around it in the distribution. Table B.3: Countries in the Bottom Quartile on the Minimum Capital Requirement Subindicator Need to Do Much More to Increase Rankings Relative to Countries in the 2nd and 3rd Quartiles Simulated rank, Rank, starting starting a Difference in Time (days) Time (days) a business, business, with starting a Country 2007 2008 2007 2008 value business rank Estonia 35 7 51 27 24 Honduras 44 21 138 121 17 Mauritius 46 7 30 10 20 Mauritania 82 65 164 164 0 Lao PDR 163 103 73 73 0 63 DOING BUSINESS: AN INDEPENDENT EVALUATION Figure B.3: Distribution of the Time to Start a Business Time for starting a business ­ frequency distribution Honduras 11 10 Mauritius 9 Estonia 8 7 6 5 Frequency Mauritania 4 3 Lao PDR 2 1 0 2 10 19 26 34 43 50 59 68 76 92 109 145 233 Number of days Example 3: How does reducing the time to open a EODB is calculated from the average of the business improve starting a business? The Republic percentile scores for the 10 indicators. This final of Lao reduced the time to start a business by 60 percentile average, the EODB percentile, is a days in 2008, yet such a change did not affect the distribution of cardinal values ranging from 0.08 simulated ranking for starting a business (table for Singapore to 0.82 for the Democratic B.3). Mauritania experienced a similar result. Republic of Congo (DRC). These values are then Mauritius, by contrast, reduced the time by 41 ranked in order, with the first position belonging days, thereby advancing 20 positions on starting a to Singapore and the last to DRC. Figure B.4 business. Honduras and Estonia, both in the shows the distribution of the EODB percentiles middle segment of the distribution and close to for 2007. the majority of countries, also made significant progress in the ranking for starting a business. The simulations presented in tables B.1­B.3, Lao and Mauritania are at the bottom end of the aside from causing changes in the indicator distribution and fairly isolated (see figure B.3). A ranking, also produced changes in the EODB change in the sparsely populated bottom end will ranking. Table B.4 summarizes some of these be less likely to improve the percentile ranking of changes for selected countries. Mauritania and a country in that subindicator. In turn, it will have Sierra Leone are at the most dispersed part of little effect on the average of the percentiles of the the distribution and did not improve in the subindicators, which gives the indicator ranking. overall ranking, despite the improvements in time to start a business and total tax rate, How Do Reforms Affect the EODB respectively. Finland and Estonia also show no Distribution? improvement. Botswana, in contrast, improved As mentioned above, the overall ranking of 6 positions thanks to its tax reform, and 64 APPENDIX B: HOW EQUITABLY DO THE RANKINGS REWARD REFORMS? Figure B.4: Average Percentile of 10 DB Indicators Ease of doing business percentile ­ frequency distribution First quartile Second quartile Third quartile Fourth quartile 8 Botswana Macedonia 7 6 Mauritania 5 Sierra Leone 4 Frequency 3 Estonia 2 1 Finland 0 0.08 0.16 0.23 0.30 0.33 0.37 0.41 0.44 0.47 0.49 0.52 0.57 0.59 0.62 0.65 0.67 0.70 0.73 Percentile Macedonia improved 9 positions because of reform effort. If this were the case, one would the elimination of the minimal capital require- expect the highly concentrated subindicators to ment, because both countries are located in be associated with more reforms in a given the more tightly distributed portions of the year.3 Table B.5 ranks DB's subindicators from indicator. most to least concentrated and shows the number of reforms associated with each of Thus, a considerable improvement in the them in 2007.4 absolute value of a subindicator might not be enough to cause an improvement at the indica- Table B.4: Despite Positive Changes, Countries at tor level if that country is starting from a very low the Bottom and Top Quartiles Did Not Improve in base. Countries in the most dispersed part of the Overall Rankings distributions will need sizeable relative improve- ments in their subindicator values to catch up Simulated Difference with the rest. This is the case for most of the EODB EODB rank in EODB countries in Africa. Country 2007 with 2008 reform rank Finland 13 13 0 Does the Ranking System Distort Reform Estonia 17 16 1 Priorities? Botswana 48 42 6 It has been suggested that DB's use of rankings Macedonia, FYR 92 83 9 might create an incentive for countries to reform the areas where they are most likely to Mauritania 148 148 0 move up in the EODB ranking for the least Sierra Leone 168 168 0 65 DOING BUSINESS: AN INDEPENDENT EVALUATION Table B.5: No Apparent Relationship between Tightness of Distribution and Reforms Number of Doing Business countries in 1 standard Percentage of Frequency of indicator Subindicator deviation range countries in range reform in 2007 Paying taxes Total tax rate 165 94 23 Starting a business Minimum capital requirement 164 94 7 Time 164 94 7 Employing workers Firing costs 162 93 5 Dealing with licenses Cost 158 90 9 Starting a business Cost 158 90 5 Protecting investors Time 158 90 2 Registering property Time 150 86 6 Dealing with licenses Procedures 149 85 12 Employing workers Rigidity of hours 146 83 4 Starting a business Procedures 138 79 28 Dealing with licenses Time 135 77 3 Trading across borders Time to export 134 77 17 Getting credit Legal rights index 131 75 24 Enforcing contracts Time 131 75 7 Trading across borders Documents for export 129 74 10 Paying taxes Payments 128 73 12 Registering property Cost 127 73 14 Procedures 126 72 7 Protecting investors Disclosure index 124 71 9 Getting credit Credit information index 123 70 13 Closing a business Recovery rate 117 67 12 Protecting investors Director liability index 117 67 4 Shareholders suits index 116 66 3 Enforcing contracts Procedures 115 66 20 Employing workers Difficulty of firing index 113 65 1 Difficulty of hiring index 109 62 2 Average 137 78 10 Median 131 75 7 Correlation between % of countries in range and number of reforms 0.01 The total tax rate is the third-most-frequent area tightly distributed. The correlation between of reform, and it has the tightest distribution of tightness of distribution and frequency of all the subindicators, with 94 percent of the reforms is almost nonexistent (0.01), offering no countries' rankings within one standard support to the hypothesis that the ranking deviation from the mean. But the two most arithmetic is distorting reforms. Alternative popular areas for reform--number of proce- hypotheses are that governments implement dures to start a business and legal rights of reforms that are politically or administratively creditors and debtors--are not among the most easier, or the ones they think most relevant. 66 APPENDIX C: DIFFERENCES BETWEEN DATA IN 2007 DB REPORT AND DB WEB SITE (OCTOBER 2007) FOR SAME DATA COLLECTION PERIOD Number of Differences of differences Reason Written explanation 10% or less Starting a business Procedures (number) 24 Data corrections 10 Time (days) 32 Data corrections 13 Not found Cost (% of income per capita) 19 Data corrections 5 Minimum capital (% of income per capita) 11 Data corrections 0 Dealing with licensesa Procedures (number) 130 Methodology change/ -- data corrections Time (days) 148 Methodology change/ -- Page 68 of Doing Business 2008 data corrections Cost (% of income per capita) 106 Methodology change/ -- data corrections Employing workers Difficulty of hiring index (0­100) 44 Methodology change -- Rigidity of hours index (0­100) 40 Methodology change -- Page 68 of Doing Business 2008 Difficulty of firing index (0­100) 46 Methodology change -- Firing cost (weeks of salary) 28 Methodology change -- Registering property Procedures (number) 10 Data corrections 1 Time (days) 17 Data corrections Not found 4 Cost (% of property value) 29 Data corrections 9 Getting credit Depth of credit information index (0­10) 12 Data corrections -- Not found Strength of legal rights index (0­10) 46 Data corrections -- Protecting investors Extent of disclosure index (0­10) 34 Data corrections -- Extent of director liability index (0­10) 23 Data corrections Not found -- Ease of shareholder suits index (0­10) 33 Data corrections -- (continues on the following page) 67 DOING BUSINESS: AN INDEPENDENT EVALUATION Number of Differences of differences Reason Written explanation 10% or less Paying taxesa Payments (number per year) 127 Methodology change/ -- data corrections Time (hours per year) 54 Methodology change/ -- Page 78 of Doing Business 2008 data corrections Total tax rate (% of profit) 160 Methodology change/ -- data corrections Trading across borders Documents to export (number) 104 Data corrections 9 Time to export (days) 109 Data corrections 38 Cost to export (US$ per container) 121 Data corrections 40 Not found Documents to import (number) 124 Data corrections 19 Time to import (days) 114 Data corrections 37 Cost to import (US$ per container) 116 Data corrections 34 Enforcing contractsa Procedures (number) 166 Methodology change/ -- data corrections Time (days) 78 Methodology change/ -- Page 68 of Doing Business 2008 data corrections Cost (% of claim) 157 Methodology change/ -- data corrections Closing a business Recovery rate (cents on the dollar) 22 Data corrections Not found 3 a. According to the DB team, for these three indicators, the methodology changes affect so many countries that it is difficult to separate corrected errors from metholodogy revisions. 68 APPENDIX D: COMMON LAW/CIVIL LAW ANALYSIS There is a great body of literature hypothesizing · Three of the indicators for employing workers: that differences in economic prosperity can be rigidity of hiring index, rigidity of hours index, traced to the legal systems of countries. Some and rigidity of firing index research has posited that countries with a legal · The legal rights subindicator under the get- system originating in the English common law ting credit indicator. tradition have enjoyed greater per capita growth than countries whose legal systems originated in In addition, the number of procedures and time the French civil law tradition, deriving from the under the paying taxes and the time under European civil codes, especially the Napoleonic dealing with licenses indicators are significantly Code. This appendix explores whether legal different, favoring common law countries. The origins affect the performance of countries on only indicator that favors countries with a civil the DB indicators. The results show that law origin is the credit information index in common law countries perform better in four getting credit. This, according to Djankov and indicators, yet differences wane in two of them others (2006), can be attributed to the presence as additional control variables are included.1 of a public credit registry in countries with a French civil law tradition. Differences in all other Regression analysis was performed using the 32 subindicators are not statistically significant. subindicators that feed the 10 indicators. The subindicators served as the dependent variable. What Explains the Differences? The controls variables included were income per The four subindicators in starting a business-- capita and a dummy variable for civil law legal number of procedures, time, cost, and minimum origin. The results are displayed in table D.2. The capital requirement--are significantly higher in 175 countries in Doing Business 2007 were French-origin countries. It is plausible that in the coded into five categories according to legal case of the first three, the differences are a result origin2: common law (59), civil law (76), German of the participation of notary publics in the (20), Nordic (5), and Socialist (11). Four of the business registration process. 175 countries were excluded because their legal origin was not clear. When testing for differences The differences in protecting investors and between common and civil law origin, the getting credit could also be attributed to legal sample was limited to those 135 countries. origin, since the Napoleonic Code deals with commercial procedures, among other issues. There are 4 indicators and 13 subindicators However, there are no statistically significant where civil law countries perform significantly differences between the two groups of countries worse than common law countries. These are: in any of the subindicators for enforcing contracts, which could have been plausibly · The four subindicators that comprise the start- attributed directly to differences in legal origin ing a business indicator as well. · The director liability index and shareholder suits index that comprise protecting investors The differences in employing workers are not as 69 DOING BUSINESS: AN INDEPENDENT EVALUATION easy to understand, since the Napoleonic Code Controlling for Additional Factors does not delve deeply into this issue. A general As a second stage of the analysis, additional hypothesis could be that, on average, countries control variables were introduced to test the with a civil law tradition favor direct supervision robustness of the differences in DB indicator of markets. In this case, civil law countries would rankings, specifically for the employing workers prefer more government regulation to protect and paying taxes indicators. For example, the the rights of workers. difference in ratings for the employing workers indicator may reflect the preference for greater The differences in paying taxes are also not social welfare, specifically in continental European easy to understand, since the number of countries. Similarly, on paying taxes, the differ- payments and the time it takes to file taxes ences may reflect the level of efficiency of the would depend more on the efficiency of tax state. In sum, the differences based on legal origin collection than legal origin. For instance, DB for employing workers are somewhat less robust, rewards countries with full online filing by and disappear for paying taxes once other factors counting the tax as paid once a year, even if the are accounted for. (The analysis is summarized in payment is more frequent. table D.1 and detailed in Attachment D.1.) Table D.1: Differences between Countries Based on Legal Origin on Employing Workers and Paying Taxes Wane after Adding Other Control Variables Employing workers Controls Difficulty of hiring Difficulty of firing Rigidity of hours 1) None Significant (99%) Significant (99%) Significant (99%) 2) Welfare variables (individually and together) Not significant Significant (99%) Significant (99%) 3) Welfare variables (excluding small countries) Significant (99%) Not significant Significant (99%) 4) Continental Europe Significant (95%) Significant (95%) Significant (95%) 5) Income group (with welfare controls) · High income Significant (95%) Significant (95%) Significant (95%) · Upper-middle income Not significant Not significant Significant (95%) · Lower-middle income Significant (95%) Not significant Significant (95%) Paying taxes Controls No. of procedures Time Total tax rate 1) None Significant (95%) Significant (95%) Not significant 2) Revenue collection proxy Not significant Not significant Not significant 70 APPENDIX D: COMMON LAW/CIVIL LAW ANALYSIS ATTACHMENT D.1: RESULTS OF REGRESSION ANALYSIS FOR TEST DIFFERENCES BASED ON LEGAL ORIGIN ON EMPLOYING WORKERS AND PAYING TAXES Employing Workers Controlling for welfare preferences. A possible ually in the regression, the difference between explanation for the differences in employing common and civil law legal origin countries is workers could be the preferences for more social statistically significant (99 percent level) on two welfare in countries with a civil law tradition. To subindicators--difficulty of hiring and rigidity proxy for this, aside from income per capita, of hours. The significance of the differences in three additional control variables were intro- the difficulty of firing index depends on the duced into the regression: (a) revenue as a share control variable. The results do not change if of GDP,3 (b) tax revenue as a share of GDP, and revenue or tax revenue is used simultaneously (c) public health and education expenditures as with health and education spending in the a share of GDP.4 When these are included individ- regression. Significance of difference between countries of common and civil law Common vs. civil law legal origin legal origin, controlling for subindicators for employing workers income per capita and­ Public health and Common Civil Total revenue Tax revenue education Scale law law as share as share spending as Subindicator (0 is best) average average Difference of GDP of GDP share of GDP Difficulty of hiring index 0 ­ 100 17.0 46.2 29.2 0.99 0.99 0.99 Rigidity of hours index 0 ­ 100 20.7 48.7 28.0 0.99 0.99 0.99 Difficulty of firing index 0 ­ 100 20.4 40.0 19.6 Not significant Not significant 0.99 Firing costs (weeks of wages) 0 ­ infinity 58.3 51.3 7.0 Not significant Not significant Not significant Number of observations 59 76 135 135 135 Creating a continental Europe origin group. The adds civil law legal origin countries with the analysis has so far excluded Nordic (Denmark, German and Nordic countries of Europe. When Finland, Iceland, Norway, and Sweden) and this group is paired against common law German legal origin countries (Austria, Ger- countries, all differences in the values of the many, Switzerland, and Eastern European subindicators remained statistically significant, countries). It can be argued that these countries at least at a 95 percent level. That is, on average, might have similar preferences for the level of countries with common law legal origin taxation and the provision of public goods as do continue to perform better in all three French origin countries. Therefore, a new subindicators--difficulty of hiring, of firing, and group was created, continental Europe, which rigidity of hours. 71 DOING BUSINESS: AN INDEPENDENT EVALUATION Significance of difference between common law and continental European Continental European legal origin countries, controlling for vs. English legal origin income per capita and­ Continental Public health and Common European Total revenue Tax revenue education Scale law legal origin as share as share spending as Subindicator (0 is best) average average Difference of GDP of GDP share of GDP Difficulty of hiring index 0 ­ 100 17.0 43.3 26.3 0.99 0.99 0.99 Rigidity of hours index 0 ­ 100 20.7 49.2 28.5 0.99 0.99 0.99 Difficulty of firing index 0 ­ 100 20.4 38.3 17.9 0.95 0.95 0.99 Firing costs (weeks of wages) 0 ­ infinity 58.3 45.7 12.6 Not significant Not significant Not significant Number of observations 59 98 81 79 93 Stratifying by income group. The analysis was also between legal origins for the three subindicators performed by income group because associations are still statistically significant. The hypothesis that with the welfare variables could be influenced by a there are differences between the rigidity of labor country's revenue-collecting capacity. When laws in common law and continental European controlling for the three welfare state proxies for tradition high-income countries, as measured by the high-income-country group, differences DB, cannot be disproved. Significance of difference between common law and continental European Limiting to high-income legal origin countries, controlling for countries, N=30 income per capita and­ Continental Public health and Common European Total revenue Tax revenue education Scale law legal origin as share as share spending as Subindicator (0 is best) average average Difference of GDP of GDP share of GDP Difficulty of hiring index 0 ­ 100 5.5 33.3 27.8 0.95 0.95 0.95 Rigidity of hours index 0 ­ 100 13.3 48.9 35.6 0.99 0.99 0.99 Difficulty of firing index 0 ­ 100 5.9 32.2 26.3 0.99 0.99 0.99 Firing costs (weeks of wages) 0 ­ infinity 37.6 29 8.6 Not significant Not significant Not significant Number of observations 12 18 23 24 24 When the analysis is performed on the upper- index continues to be statistically higher in middle-income group, only the rigidity of hours countries with a continental European origin. Significance of difference between common law and continental European Limiting to upper-middle-income legal origin countries, controlling for countries, N=34 income per capita and­ Continental Public health and Common European Total revenue Tax revenue education Scale law legal origin as share as share spending as Subindicator (0 is best) average average Difference of GDP of GDP share of GDP Difficulty of hiring index 0 ­ 100 0 ­ 100 15.7 40.1 24.4 Not significant Not significant Rigidity of hours index 0 ­ 100 0 ­ 100 16.7 51.8 35.1 99 95 Difficulty of firing index 0 ­ 100 0 ­ 100 19.2 39.1 19.9 Not significant Not significant Firing costs (weeks of wages) 0 ­ infinity 0 ­ infinity 43.7 42.4 1.3 Not significant Not significant Number of observations 12 22 20 18 72 APPENDIX D: COMMON LAW/CIVIL LAW ANALYSIS However, in the case of lower-middle-income of the welfare state. The significance of differ- countries, it is the difficulty of hiring index that ences for the rigidity of hours index and the continues to be statistically greater in continental difficulty of firing index wane when welfare law origin countries after controlling for proxies proxies are added. Significance of difference between common law and continental European Limiting to lower-middle-income legal origin countries, controlling for countries, N=47 income per capita and­ Continental Public health and Common European Total revenue Tax revenue education Scale law legal origin as share as share spending as Subindicator (0 is best) average average Difference of GDP of GDP share of GDP Difficulty of hiring index 0 ­ 100 12.6 44.9 32.3 0.99 0.99 0.95 Rigidity of hours index 0 ­ 100 20 43 23 Not significant Not significant 0.95 Difficulty of firing index 0 ­ 100 12.9 39.1 26.2 Not significant Not significant Not significant Firing costs (weeks of wages) 0 ­ infinity 40.9 58.6 17.7 Not significant Not significant Not significant Number of observations 14 33 25 24 24 In the low-income group, only the rigidity of Nevertheless, in this particular group, the hours index continues to be statistically higher in information for the control variables is scarce, countries with a continental European origin which led to only using 13 or 17 observations in after controlling for proxies of the welfare state. the regressions. Significance of difference between common law and continental European Limiting to low-income legal origin countries, controlling for countries, N=46 income per capita and­ Continental Public health and Common European Total revenue Tax revenue education Scale law legal origin as share as share spending as Subindicator (0 is best) average average Difference of GDP of GDP share of GDP Difficulty of hiring index 0 ­ 100 27.2 51.4 24.2 Not significant Not significant Not significant Rigidity of hours index 0 ­ 100 27.6 55.2 27.6 0.95 0.99 0.99 Difficulty of firing index 0 ­ 100 34.3 40.8 6.5 Not significant Not significant Not significant Firing costs (weeks of wages) 0 ­ infinity 90.1 44.2 45.9 Not significant Not significant Not significant Number of observations 21 25 13 13 17 The results of these regressions do not Controlling for small-country outliers. Some small change substantially when comparing countries in the sample have unusually high common law versus civil law origin instead of values for the welfare control variables. continental European. Although some of the Therefore, countries with a population of less differences remain despite the inclusion of than 2 million (the Bank's suggested definition the control variables, the disappearance of of a small country) were excluded from the some could be evidence that other factors analysis. Once the proxies for the welfare state aside from legal origin are important for were added and small countries were excluded, explaining performance on the employing the differences in the difficulty of firing index workers indicator. were not statistically significant. 73 DOING BUSINESS: AN INDEPENDENT EVALUATION Significance of difference between common law and continental European Excluding countries with population legal origin countries, controlling for of less than 2 million income per capita and­ Continental Public health and Common European Total revenue Tax revenue education Scale law legal origin as share as share spending as Subindicator (0 is best) average average Difference of GDP of GDP share of GDP Difficulty of hiring index 0 ­ 100 18.9 44.2 25.3 0.99 0.99 0.99 Rigidity of hours index 0 ­ 100 23.3 49.4 26.1 0.99 0.99 0.99 Difficulty of firing index 0 ­ 100 25.3 36.1 10.8 Not significant Not significant Not significant Firing costs (weeks of wages) 0 ­ infinity 76 43.8 32.2 Not significant Not significant 0.95 Number of observations 36 84 71 68 74 Paying Taxes: Controlling for Additional countries. However, these differences could be Factors attributed to the government's efficiency in tax Two of the subindicators for paying taxes, number collection. When an additional control variable, of payments and time, are statistically significantly tax revenue as a share of GDP, is introduced into higher in civil law countries than in common law the regression, the differences cease to exist. Controlling for income per capita and tax Common law Civil law revenue as Subindicator Scale average average Difference share of GDP Payments (number) 0 ­ infinity 28.9 37.2 8.3 Not significant Time (hours) 0 ­ infinity 207.1 314.5 107.4 Not significant Total tax rate (% profit) 0 ­ infinity 46.9 57.3 10.4 Not significant 74 APPENDIX D: COMMON LAW/CIVIL LAW ANALYSIS Table D.2: Regression Results for Common and Civil Law Countries at the Subindicator Level Significance of difference after Common Civil controlling for law law income per Indicator Subindicator Scale average average Difference capita - Oct. 2007 Starting a business Procedures (number) 0 ­ infinity 8.2 10.9 2.6 0.99 Time (days) 0 ­ infinity 37.8 64.2 26.4 0.95 Cost (% of income per capita) 0 ­ infinity 44.4 96.3 51.9 0.99 Min. capital (% of income per capita) 0 ­ infinity 16.0 154.1 138.1 0.99 Dealing with licenses Procedures (number) 0 ­ infinity 16.5 18.6 2.1 Not significant Time (days) 0 ­ infinity 190.8 231.4 40.6 0.95 Cost (% of income per capita) 0 ­ infinity 539.6 693.7 154.1 Not significant Employing workers Difficulty of hiring index 0 (best) ­100 (worst) 17.0 46.2 29.2 0.99 Rigidity of hours index 0 (best) ­100 (worst) 20.7 48.7 28.0 0.99 Difficulty of firing index 0 (best) ­100 (worst) 20.4 40.0 19.6 0.99 Firing costs (weeks of wages) 0 ­ infinity 58.3 51.3 7.0 Not significant Registering property Procedures (number) 0 ­ infinity 6.2 6.5 0.3 Not significant Time (days) 0 ­ infinity 78.3 88.7 10.4 Not significant Cost (% of property value) 0 ­ infinity 6.9 8.4 1.5 Not significant Getting credit Credit information index 0 (worst) ­ 6 (best) 1.9 2.8 0.9 0.99 Legal rights index 0 (worst) ­ 10 (best) 5.3 3.4 1.9 0.99 Protecting investors Disclosure index 0 (worst) ­ 10 (best) 4.9 4.8 0.1 Not significant Director liability index 0 (worst) ­ 10 (best) 5.5 3.3 2.1 0.99 Shareholder suits index 0 (worst) ­ 10 (best) 6.5 4.7 1.8 0.99 Paying taxes Payments (number) 0 ­ infinity 28.9 37.2 8.3 0.95 Time (hours) 0 ­ infinity 207.1 314.5 107.4 0.99 Total tax rate (% profit) 0 ­ infinity 46.9 57.3 10.4 Not significant Trading across borders Documents for export (number) 0 ­ infinity 7.1 7.7 0.6 Not significant Time for export (days) 0 ­ infinity 25 29.6 4.6 Not significant Cost to export (US$ per container) 0 ­ infinity 1,128.1 1,298.6 170.5 Not significant Documents for import (number) 0 ­ infinity 8.3 9 0.7 Not significant Time for import (days) 0 ­ infinity 30.2 35.7 5.5 Not significant Cost to import (US$ per container) 0 ­ infinity 1,340.4 1,529.7 189.3 Not significant Enforcing contracts Procedures (number) 0 ­ infinity 38.1 39.1 1.0 Not significant Time (days) 0 ­ infinity 609.2 672.7 63.5 Not significant Cost (% of debt) 0 ­ infinity 33.2 40.9 7.7 Not significant Closing a business Recovery rate (cents on the dollar) 0 to $1.00 32.2 24.1 8.0 Not significant Note: N = civil law, 76: common law, 59; significant levels set at 95 percent or higher. 75 APPENDIX E: STANDARD INTERVIEW PROTOCOLS Appendix E.1: Interview Protocol for Validity of Assumptions Doing Business Informants e) The DB survey presents a business case or a standard firm as the basis for your responses. [Greeting] I am calling on behalf of the World In your opinion, are the assumptions described Bank's Independent Evaluation Group (IEG), in the survey representative of a typical firm in which reports directly to the Board of Directors of your country? Why or why not? the World Bank. The IEG is undertaking an evalua- f) In your judgment, how many firms fitting this tion of the World Bank Group's DB indicators. assumption have used your services? g) If you had to change the assumptions to make I'm calling/contacting you because you are listed them more consistent with your country's re- as an informant to the DB survey in Country X. alities, which assumptions would you change As part of the evaluation, we are reviewing the and why? And how would these changes affect process for collecting the data used in the DB your answers? report. We would very much value your views about the process and information collected. Survey Content and Structure Your contribution is important for enhancing the h) In your view, do the questions asked in the sur- future work of the World Bank Group. vey capture the essence of the business cli- mate challenges on the topic? Are the questions This interview will about 20 minutes. Please be focusing on the right aspects? assured that your views will remain anonymous, i) Do you have any other comments about the and responses to this survey will not be attrib- structure of the survey? uted to you personally, or to your organization. Validity of Information in DB Report Background Information j) Have you seen the data published in the last a) What are the topics/questions that DB asks DB report for your topic(s) or your country? you to provide information on? What is your Do you agree with the information? professional experience with these topic(s)? k) Inyourview,doyouthinktheDBreportcaptures b) How you were approached to participate? the changes in laws and regulations from one When did you first participate and how many year to the next appropriately? Why or Why not? times have you taken part? c) Why do you participate? Closing: d) How long did it take you to answer the survey, l) How useful has the Doing Business exercise including time spent by colleagues or been in your country? Please explain. subordinates? m)Is there anything else you would like to add about the DB survey process or report? 77 DOING BUSINESS: AN INDEPENDENT EVALUATION Appendix E.2: Interview Protocol for aspect of our work is to determine the relevance Policy Makers and Senior Government and the use of the DB indicators to the govern- Officials ment and policy makers in developing countries. Your contribution is important for enhancing the Introduction future work of the World Bank Group. We are writing on behalf of the World Bank's Independent Evaluation Group (IEG), which Our survey will take about 45 minutes to reports directly to the Board of Directors of the complete. Please be assured that your views will World Bank. The IEG is undertaking an evaluation remain anonymous, and responses will not be of the World Bank's DB indicators. An important attributed to you personally. I. Background: a) In order of importance, please tell us, what, in your view, are the three factors affecting or impeding the growth of domestic private sector enterprises? b) What issues have you or your government raised with donors, including the World Bank Group, regarding the development of the domestic private sector? II. Relevance of the DBI: c) Are you aware of the Doing Business indicators published by the World Bank Group? (Yes/No) d) The Doing Business indicators, the subject of this evaluation, present information on 10 aspects of the business climate. For each, please tell us how important each of these are to enhancing the environment for domestic enterprises. Please use a scale of 1­4 where 1 = Very Important, 2 = Important, 3 = Slightly Important, and 4 = Not important. 1. Very 3. Slightly 4. Not Aspects important 2. Important important important Comments Starting a business Getting credit Enforcing contracts Employing workers Paying taxes Dealing with licenses Registering property Protecting investors Trading across borders Closing a business Any other (please list) e) Do you have any comments about the methodology underlying the DB indicators? f) Overall, your country is ranked A out of B by the DB 2007 report. Do you agree with this ranking? Why or Why not? Indicator Ranking Indicator Ranking Comments 1. Starting a business 6. Registering property 2. Employing workers 7. Dealing with licenses 3. Getting credit 8. Trading across borders 4. Enforcing contracts 9. Investor protection 5. Paying taxes 10. Closing a business 78 APPENDIX E: STANDARD INTERVIEW PROTOCOLS III. Use of the DB indicators g) Have you ever used the DB indicators in the course of your work? How have you used them? Please specify. (If not, skip to Q10). h) Please rank the use of the DB indicators specifically in: 1. Very 3. Slightly 4. Not useful 2. Useful useful useful Comments Motivating reform Starting dialogue with country policy makers Creating consensus among stakeholders Other (please specify).... Designing reforms Suggestions on changes in legislation Prioritization of reform areas Other (please specify).... i) Please rank the usefulness of the following characteristics of the DBI? 1. Very 3. Slightly 4. Not useful 2. Useful useful useful Comments Specific indicators? (Please list) Use of country benchmarking In-depth analysis of laws Media coverage of the DB indicators Other? j) What other indicators did you find to be useful when designing policy or activities for developing domestic private enterprises? In you view, what is the relative value of the DB indicators to these other indicators? k) Please tell us about your involvement, if any, with the Bank group's Doing Business team. · During preparation of the report? · Commenting on the indicators? IV. Impact of DB indicators: l) In your view, in order of importance, what have been the major reforms that have aided or hin- dered the development of the domestic private enterprise in your country over the last 5 years? m)The DB reports over the last 3 years list the following reforms in your country (see table). In your view, how significant are these reforms to the development of domestic enterprises and why? 1. Very 3. Slightly 4. Not Reforms noted by DB significant 2. Significant significant significant Comments n) In your view, to what extent did the DB initiative, including DB reports, contribute to these reforms? Thank you. 79 ENDNOTES Chapter 1 tensive state ownership of land, and excessive over- 1. Key articles include Djankov and others (2002), sight and regulation of private sector activities. Botero and others (2004), Dollar and others (2005). See 6. This study used the Bank's firm-level data provided also Djankov (2008) for a more complete list. by the Business Environment and Enterprise Perform- 2. See Commander and Tinn (2007), pp. 3-4. Djankov ance Surveys (BEEPS) for 26 countries in Europe and (2007) also notes "nearly all the work [on the effects of the former Soviet Union. reform using ease of enterprise indicators] is cross-sec- 7. Revenue efficiency measures how much revenue tional, or uses panel analysis with an aggregate measure a company needs to take in to produce its net earnings. of economic freedom that may exaggerate the effects of It is the ratio of net earnings and revenue. reform. And researchers generally lack good microeco- 8. Made available upon request and will be posted nomic outcome indicators--like new business start-ups, on the IEG Web site. number of newly registered properties, job created, in- 9. The DB team has produced background papers on creases in productivity--so much of the work makes im- 8 of the 10 topics; 3 are published in the Quarterly Jour- plausible attempts to link specific regulatory reforms to nal of Economics, 2 in the Journal of Financial Eco- overall investment, employment rate and growth" (p. 10). nomics, and 3 as National Bureau of Economic Research 3. Enterprise surveys are not available for Algeria, WorkingPapers.Acompletelistofthesepapersisavailable China, Netherlands, and Nigeria for the years 2004­07. at http://www.doingbusiness.org/MethodologySurveys/ Rwanda is not covered in the 2007/2008 Global Com- 10. Except for protection of minority shareholders. petitiveness Report. 11. The seven indicators are: starting a business, 4. This evaluation finds that the overall EODB rank- dealing with licenses, registering property, paying ing is highly correlated with the World Development taxes, trading across borders, enforcing contracts, and Forum's Global Competitiveness Index (0.81) and the closing a business. Economist Intelligence Unit's Business Environment 12. Management notes that the starting a business in- Rankings (0.88), both perceptions-based indicators. dicatorrewardscountriesforsimplifyingthewaythatreg- Mas (2006) found similar results using the Heritage ulationsareimplemented,notforcuttingregulation.What Foundation's Economic Freedom Index and IMD's counts as simplification is unifying procedures or putting World Competitiveness Scoreboard, among others. them on the internet so there is less hassle and fewer op- DB's 10 indicators are highly correlated with compara- portunities to extract bribes. Djankov (2008) gives the ex- ble subcomponents of the Global Competitiveness ampleofstartingabusinesswheresimplifyingregulations Index or the Business Environment Rankings only in increases legal certainty. high-income countries (around 0.70). In middle- and IEGnotesthatDBreportsanddataonstartingabusi- low-income countries, the DB indicators are weakly ness do not consistently distinguish between eliminating correlated with perception-based surveys. However, procedures and simplifying them through unification be- given the differences in the methodology underlying cause they refer to some steps and procedures as being both data sets, a correlation analysis alone may not be "cut," "eliminated," and "lifted." sufficient to provide generalizable conclusions on 13. The indicator measures the steps needed to get whether or not the DB adds new information. a construction permit to build a warehouse; it does 5. Other constraints raised by stakeholders include not deal with licenses, permits, and authorizations in gen- political and macroeconomic instability, corruption, ex- eral. This point is discussed further in chapter 3. 81 DOING BUSINESS: AN INDEPENDENT EVALUATION 14. Management notes that the starting a business team ensured that the data collection would be guided indicator rewards countries for simplifying the way that by the latest theory and empirical research in econom- regulations are implemented, not for cutting regulation. ics and finance." What counts as simplification is unifying procedures or putting them on the internet so there is less hassle and Chapter 2 fewer opportunities to extract bribes. Djankov (2008) 1. The evaluation attempted to contact all 68 in- gives the example of starting a business where simpli- formants who provided information on the evaluation's fying regulations increases legal certainty. five focus indicators in the seven countries (Albania, Al- IEG refers to its comment in endnote 12. geria, Burundi, Moldova, Mongolia, Nigeria, and Spain). 15. Estimates by the evaluation team found correla- Of these, 57 had provided a completed questionnaire, tions greater than 0.90 between the original rankings and and the other 11 were supplemental informants con- others produced with alternative weighting schemes. sulted by the DB team in person, by telephone, or by 16. Management notes that the benefits of regulation e-mail to validate or clarify particular issues. The eval- can only be assessed in empirical analyses that link the uation counts each of these as a separate informant. The costs that regulations incur on businesses (what DB evaluation team made at least 3 attempts to contact each measures) to economic and social outcomes. The back- of the 68 informants and succeeded in contacting and ground research provided by the DB team, and available interviewing 59 percent (40 informants) by phone or by at http://www.doingbusiness.org/MethodologySurveys/, e-mail. Of the 28 informants who could not be contacted, does precisely that; as do numerous academic papers 19 had unusable contact information or did not re- listed on the same Web site http://www.doingbusi- spond after repeated attempts, 7 had left their position, ness.org/documents/Citations_of_Doing_Busines_re- and 2 had died. search_papers.pdf. The development of the DB indicators 2. The evaluation chose 7 countries at random from has made such research possible. This is illustrated in the the 175 countries covered by DB. In addition, it chose 2006 evaluation of World Bank research and flagship 6 countries at random from the list of countries classi- publications, commissioned by then-Chief Economist and fied by DB as "top reformers" in 2006 and 2007. Senior Vice-President of DEC Francois Bourgignon. The 3. Management notes that the methodology chap- report states "In fact, I believe that Doing Business is one ter of each DB report states the selection criteria: the of the most influential research initiatives that the IFC contributors need to live in the country surveyed by DB and the World Bank have ever undertaken. It has put the and need to practice in the topical area under review. focus on improving the efficiency of government policy 4. Non-PwC accountants (in Spain and Burundi) and ignited a vigorous discussion in emerging markets. represented 1 percent. This cannot only be seen by the fact that the first three 5. In Burundi and Nigeria, Bank Group staff reported entries under `doing business' (which is even an ex- that at least one informant lacked professional expert- tremely generic word combination) on Google link to ise on the topic. the Doing Business Web site at the World Bank. More- 6. In Doing Business 2007, 17 out of 201 informants over, literally at any policy forum in developing countries in the 13 countries reviewed did not wish to be publicly I have heard reference to the reports." Furthermore, the named. Bourgignon report comments "Overall the implemen- 7. For the five focus indicators in the seven country tation and execution of the data collection was very case study countries. carefully conducted and has undergone several refine- 8. Four countries had only one informant. Where ments and improvements. The Doing Business reports there were multiple informants (Mongolia, Nigeria, and have created a very robust and reliable set of benchmark Spain), the published data were not directly the median measures on regulation which are being used world value of the responses of the questionnaires. wide by practitioners and academics alike. They have be- 9. Albania, Algeria, China, Moldova, Netherlands, come a major source of country indicators on the reg- and Tanzania. ulatory environment of businesses world wide. Moreover, 10. Internal correspondence on Doing Business 2007. by engaging a cadre of first-rate academics (such as 11. This figure excludes about 295 changes caused Oliver Hart, Andrei Shleifer, and others) the World Bank by revisions in GNI data, which affect data points ex- 82 ENDNOTES pressed as a ratio with per capita GNI in the denomi- laws and regulations are collected and answers checked nator. for accuracy." 12. Internal correspondence with DB team dated No- 2. Seventy-six percent of the 79 questions that make vember 14, 2007. up the ranking in Doing Business 2007 ask about laws 13. This calculation excludes the changes for the and formal regulations. protecting investors indicator. The nature of the in- 3. "By law, does management remain in control of dexes in this indicator makes the 10 percent rule inap- the company's assets upon the initiation of a reorgan- plicable. ization procedure?" Question 1, section 8, legal rights 14. The World Development Indicators (WDI), for in- index for the getting credit indicator. A score of 1 is as- stance, make available previously published data sets signed if management does not stay during reorgani- through annual CD-ROMs. Management notes that this zation and an administrator is responsible for managing example does not support the evaluation team's claim the business during reorganization (equivalent to a re- that DB should make available all previously published sponse of "no" to this question). data sets, uncorrected for errors and without updating 4. Doing Business 2007, p. 61, acknowledges that them with the latest methodology. The WDI annually "The measures of time involve an element of judgment publishes the time-series of its data, but each publica- by the expert respondents. When sources indicate dif- tion corrects errors found in previous years. This is ex- ferent estimates, the time indicators reported present actly what DB does when reporting the time-series data the median values of several responses." on its Web site: http://www.doingbusiness.org/Custom 5. Seven of the 10 DB indicators include a subindi- Query/. cator on cost that is used in the calculation of the EODB IEG notes that it recommends that DB disclose and ranking. All these cost subindicators include official legal make available the data it has previously published but fees and, in all but one case, also include informants' es- subsequently supplanted with revised or corrected timates of costs of professional fees charged by lawyers, data. This is the type of data provided on WDI's CD- notaries, accountants, and the like (the exception is the ROMs. firing cost for employing workers). For example, the cost 15. These changes were calculated after excluding of enforcing contracts subindicator includes court fees, the three new countries incorporated in Doing Business as well as attorney fees and enforcement fees necessary 2008. for the plaintiff to enforce judgment through a public sale 16. Simulations use the Doing Business 2007 report of the defendant's movable goods. The costs for regis- as the baseline. tering property include the cost of registration materi- 17. The standard deviation gives an estimate of the als, registration fees, property taxes, as well as professional dispersion within a distribution. A normal distribution fees for lawyers and notaries. contains 65 percent of observations within one standard 6. The getting credit questionnaire asks in question deviation from the mean in both directions. Greater val- 2, section 11, "What in your opinion are the main areas ues than 65 percent would suggest a tighter distribution of secured transactions law that require reform? Why?" skewed to one side, with few outliers on the opposite 7. The starting a business questionnaire asks in ques- tail. The greater the number of observations within one tion 6: "In your opinion, is the company registration standard deviation from the mean, the more concen- process more or less efficient now in comparison to trated the distribution is on one side, as seen in figure the previous year?" and in question 7: "If you were to ad- B.1 in appendix B. vise the government on how to reform business start- 18. Calculation derived using the reforms and data up, what would be your main suggestion and why?" from Doing Business 2007. 8. Yammarino, Skinner, and Childers (1991), a meta- analysis of 115 studies on techniques to induce mail sur- Chapter 3 vey response rates, found survey length to have a 1. Doing Business 2007, p. 61: "The DB methodol- significant effect on response rates, regardless of the tar- ogy....us[es] factual information about what laws and reg- get population. ulations say....Having representative samples of 9. The other legal origins are Nordic (5), German respondents is not an issue, as the texts of the relevant (20), and Socialist (11). The legal origin of the remain- 83 DOING BUSINESS: AN INDEPENDENT EVALUATION ing 4 countries is not available. This classification is Djankov and others (2008), NBER Working Paper 13756, based on Doing Business 2004 and completed for the which also details the differences between this method- missing countries using the original source of the clas- ology and a previous total tax rate methodology devel- sification, the CIA Factbook. oped by PwC. Management also notes that in accounting 10. Organization for the Harmonization of Business and auditing services it is the standard to have a sole Law in Africa, or L'Organisation pour l'Harmonisation provider: in 2007, PwC firms provided exclusive auditing en Afrique du Droit des Affaires in French. services for 368 of the companies in the Fortune 500 and 11. Based on a regression analysis of ratings on all 422 of the companies in the Financial Times Global 32 subindicators and the DB 2007 revised data, and 500.Havingasingleproviderdoesnotjeopardizethequal- controlling for per capita income. See appendix D. ity of data since the DB coding is based on the text of tax 12. On protecting investors, for instance, legal experts laws, which PwC provides to the DB team for verification. suggest that the DB's focus on allowing proxies by e-mail, IEG notes that page 29 states that "DB adopted a ver- use of cumulative voting, and the right of a shareholder sion of PwC's methodology in 2005" and that "the to sue management all reflect the common law per- methodology is the most complex of all DB indicators spective. because it requires detailed calculations of a variety of 13. Interview with external and Bank Group subject taxes for a standard firm." matter experts. 27. Reflected in improvements in the rigidity of 14. Calculation is based on the published Doing hours index from 60 to 20 and in the difficulty of firing Business 2007 data. GNI per capita data was obtained index from 70 to zero. The DB staff makes adjustments from the DB Web site for the corresponding period. in reported data from one year to the next, which are 15. Excluding the poorest 35 countries in the world. shown on its Web site. In this case, there was no such 16. Countries in the top and bottom quintiles of the adjustment. income per capita rank. 28. Until 2006, this indicator involved a dispute 17. Calculations are based on Doing Business 2007. around a bounced check or a simple debt default. But 18. See Djankov and others (2002). The article was since most countries have specific legislation sur- largely inspired by Hernando de Soto's study of entry rounding defaults on negotiable instruments such as regulation in Peru in which the high costs of establish- checks, the case study was revised for DB 2007 to en- ing a business denied economic opportunities to the tail a contractual dispute over the quality of goods. poor. See de Soto (1990). 29. The Doing Business 2008 report made four 19. This framework is used neither to allocate re- changes: 1) The list of procedures was revised to ac- sources nor to guide IDA programs ex-ante. See World commodate the fact that in civil law countries the judge Bank (2007b). appoints an independent expert, while in common law 20. Doing Business 2004, p. 21, discusses some countries parties send the court a list of their expert wit- caveats on one-stop shops. These caveats are less promi- nesses. 2) Two elements were added to the standard sce- nent in reports of later years. nario: one on attaching the defendant's goods prior to 21. This paper uses the World Bank Group Entre- judgment and another on providing expert opinions. preneurship Survey database, which does not contain 3) To reflect the overall efficiency of court procedures, information about the sustainability of firms, an im- one procedure is subtracted for countries that have portant consideration when looking at longer-term im- specialized commercial courts and one procedure is sub- pacts. tracted for countries that allow electronic filing of court 22. Yakovlev and Zhuravskaya (2007) on Russia and cases. 4) The cost indicator includes all fees for en- Monteiro and Assuncao (2006) on Brazil. forcing judgments. 23. Interview with PwC on October 16, 2007. 30. Based on the 2005 World Bank Business Enter- 24. For the 33 countries where PwC does not oper- prise Survey for Romania. ate, DB informants are local accounting firms. 25. Interview with PwC on October 16, 2007. Chapter 4 26. Management notes that the underlying method- 1. Interview with FIAS management. ology for the paying taxes indicator is provided by 2. Doing Business 2008, p. 39. The full quotation is: 84 ENDNOTES "Countries that make it easier to pay taxes and contri- publication of her research. The IEG recommendation butions also have higher rates of workforce participation, to make available uncorrected data and data unadjusted and lower rates of unemployment, among women. The for the latest methodology is unorthodox: this is not reason is simple: a burdensome tax system dispropor- practiced by major data providers. tionately hurts smaller businesses, especially in the serv- IEG notes that the evaluation's recommendation ices sector and this is where most women work." refers to data published periodically by DB (on its Web 3. The road shows are sponsored by the country of- site and in publications) that are subsequently removed fice with minimal contributions from headquarters and supplanted by revised or corrected data. DB does ($2,000 per road show). not currently make these data available. 4. In addition, since 2004, USAID has provided 12. The exact formula is: IDA country allocation per $211,000 to expand DB's coverage to 7 post-conflict annum = base allocation + f (Country performance rat- countries. The DB also received $75,000 from the sale ing 2.0, Population 1.0, GNI/capita-0.125) where CPIA of reports to USAID field offices. USAID provided accounts for 80 percent of the country performance rat- $1 million to finance state- and municipal-level DB ex- ing. See World Bank (2007c). ercises managed by FIAS and research on best practice 13. See Approach Paper for IEG's Special Study on reforms. Budget detail provided by Knowledge Man- the Bank's CPIA, February 5, 2008. agement and Outreach Team in the Office of the Vice 14. Only operations approved between fiscal years President, FPD, World Bank-IFC in an e-mail dated No- 2004 and 2007 are included. The six countries are: Al- vember 27, 2007. bania, Moldova, Nigeria, Peru, Rwanda, and Tanzania. 5. From the case studies for a forthcoming IEG eval- 15. Mentioned in Burundi, Moldova, Nigeria, Nether- uation of the Bank's economic and sector work (ESW). lands, Peru, Rwanda, and Tanzania. 6. The section is based on 100 interviews with Bank Group staff and stakeholders in all the case study coun- Appendix A tries, plus reports from interviews in other countries held 1. Serbia & Montenegro was removed from the ran- by a forthcoming evaluation reviewing Bank ESW, and dom selection process because the country split in interviews with international donor agencies including 2006. Georgia and Romania were top reformers in both MCC and USAID. years, but were included only once in the random se- 7. Respondents rated usefulness on a four-point lection process. scale: very useful, useful, slightly useful, or not useful. Thirteen percent of Bank Group staff and stakeholders Appendix B considered the use of country benchmarking to be ei- 1. The data used in this appendix corresponds to ther "not useful" or "slightly useful," either because Doing Business 2007. they felt the data and/or ranking for their country were 2. The simulation uses the indicator value in Doing inaccurate, or because they believed cross-country com- Business 2008 (after a reform) and measures the impact parisons were inappropriate because it obscured the im- of the reform on indicator rankings and the EODB rank- portance of country context. ing for 2007, holding other countries' actions constant. 8. Bank management comments on Doing Business 3. The standard deviation can give a good estimate 2008 report. of the dispersion within a distribution. A normal distri- 9. Interviews with FIAS Rapid Response Unit and bution contains 65 percent of observations within one MCC staff. standard deviation from the mean in both directions. 10. Interview with management of the FIAS Rapid Re- Greater values than 65 percent could suggest a tighter sponse Unit. distribution skewed to one side, with a few outliers on 11. Management notes that all data used in back- the opposite tail: the greater the number of observations ground research by the DB team are available at within one standard deviation from the mean, the more http://www.doingbusiness.org/MethodologySurveys/; if concentrated the distribution. a researcher wishes to replicate the work by another ac- 4. Five of the 32 subindicators, 4 of which are in trad- ademic who used previous versions of DB data, the lat- ing across borders, have been excluded because it is dif- ter is obliged to provide these data on request upon ficult to assign specific reforms to them. 85 DOING BUSINESS: AN INDEPENDENT EVALUATION Appendix D category "English" in Doing Business 2004, and the cat- 1. The data used in this appendix correspond to the egory "civil law" corresponds to the category "French." updated version for DB 2007, downloaded from the DB 3. This variable includes cash receipts from taxes, Web site in October 2007. social contributions, and other revenues such as fines, 2. The basis for this classification can be found on page fees, rent, and income from property or sales. 115 of Doing Business 2004. This database has informa- 4. A good variable to add to this analysis would have tion for 130 countries. Information for the additional 45 been public contributions to social welfare programs as countries was retrieved from the source used by Doing a share of GDP. However, such a variable is available for Business 2004--the CIA Factbook. The category "com- a limited amount of countries, which made it imprac- mon law" used in this report corresponds directly to the tical to use. 86 BIBLIOGRAPHY Acemoglu, Daron, and Simon Johnson. 2005. "Un- Working Paper No. 3204. World Bank, Washington, bundling Institutions." Journal of Political Econ- DC. omy 113(5): 949­95. Berg, Janine, and Sandrine Cazes. 2007. "The Doing Altenburg, Tilman, and C. von Drachenfels. 2006. "The Business Indicators: Measurement Issues and Polit- `New Minimalist Approach' to Private Sector Devel- ical Implications." Economic and Labour Market opment: A Critical Assessment." Development Pol- Paper No. 2007/6. Geneva: International Labor Or- icy Review 24 (4): 387­411. ganization. http://www.ilo.org/public/english/em Arruñada, Benito. 2007. "Pitfalls to Avoid When Meas- ployment/download/elm/elm07-6.pdf uring Institutions: Is `Doing Business' Damaging Biggs, Tyler. 2004. "The Investment Climate and Private Business?" Universitat Pompeu Fabra Economics Sector Development." Background Paper prepared and Business Working Paper No. 1040, August 24. for Improving Investment Climates: An Evaluation http://www.econ.upf.edu/docs/papers/downloads/104 of World Bank Group Assistance. IEG­World Bank, 0.pdf Washington, DC. Arvis, Jean-Francois, Monica Alina Mustra, John Panzer, Blanchet, Didier. 2006. "Exploratory Analysis of the In- Lauri Ojala, and Tapio Naula. 2007. Connecting to dicators Proposed by Doing Business Reports 2005 Compete: Trade Logistics in the Global Economy. and 2006 of the World Bank." Attractivité Economique Washington, DC: World Bank. du Droit Working Paper No. AED-EAL-2006-3. Paris. Ayyagari, Meghana, Asli Demirguc-Kunt, and Vojislav Botero, Juan, Simeon Djankov, Rafael Porta, and Flo- Aksimovic. 2005. "How Important Are Financing rencio Lopez-De-Silanes. 2004. "The Regulation of Constraints? The Role of Finance in the Business Labor." The Quarterly Journal of Economics 119(4): Environment." World Bank Policy Research Working 1339­82. Paper No. 3820. World Bank, Washington, DC. Bruhn, Miriam. 2007. "License to Sell: The Effect of Bakvis, Peter. 2006. "How the World Bank and IMF Use Business Registration Reform on Entrepreneurial the Doing Business Report to Promote Labour Mar- Activity in Mexico." MIT Job Market Paper (2007); ket Deregulation in Developing Countries." Pre- World Bank Policy Research Working Paper (2008). pared for the International Confederation of Free http://siteresources.worldbank.org/DEC/Resources/B Trade Unions, Washington, DC. http://www.gurn.info/ ruhnOct2007.pdf topic/prsp/doingbusiness.pdf Carlin, Wendy, and Paul Seabright. 2007. "Bring Me Sun- Beck, Thorsten, and Asli Demirguc-Kunt. 2006. "Small shine: Which Part of the Business Climate Should and Medium-Size Enterprises: Access to Finance as Public Policy Try to Fix?" Paper presented at the An- a Growth Constraint." Journal of Banking & Fi- nual Bank Conference on Development Economics, nance 30: 2931­43. May, Bled, Slovenia. http://siteresources.worldbank Beck, Thorsten, Asli Demirguc-Kunt, and Ross Levine. .org/INTABCDESLO2007/Resources/SeabrightCar- 2003. "Law and Firms' Access to Finance." NBER linPaper.pdf Working Paper 10687. National Bureau of Economic Commander, Simon, and Jan Svejnar. 2007. "Do Insti- Research, Cambridge, MA. tutions, Ownership, Exporting and Competition Ex- Beck, Thorsten, Asli Demirgu-Kunt, Luc Laeven, and Vo- plain Firm Performance? Evidence from 26 Transition jislav Maksimovic. 2003. "The Determinants of Fi- Countries." IZA Working Paper 2637. Institute for the nancing Obstacles." World Bank Policy Research Study of Labor, Bonn, Germany. 87 DOING BUSINESS: AN INDEPENDENT EVALUATION Commander, Simon, and Katrin Tinn. 2007. "Evaluating -----------. 2002. "The Regulation of Entry." The Quar- Doing Business." Background paper prepared for this terly Journal of Economics 117(1): 1­37. evaluation. Available at: www.worldbank.org/ieg Dollar, David, Mary Hallward-Dreimeier, and Taye Cunat, Alejandro, and Marc Melitz. 2007. "Volatility, Labor Mengistae. 2005. "Investment Climate and Firm Per- Market Flexibility, and the Pattern of Comparative formance in Developing Economies." Economic De- Advantage." NBER Working Paper 13062. National velopment and Cultural Change 54 (1): 1­31. Bureau of Economic Research, Cambridge, MA. Dorbec, Anna. 2006. "Credit Information Systems: The- Davis, Kevin E., and Michael B. Kruse. 2007. "Taking the oretical and Comparative Analysis." Working Paper Measure of Law: The Case of the Doing Business Proj- AED-EAL-2006-2. Attractivité Economique du Droit ect." Law & Social Inquiry 32(4): 1095­119. (AED), University of Paris X-Nanterre. De Sa, Lilliana. 2005. "Business Registration Start-up: A Du Marais, Bertrand. 2006. "Methodological Limits of the Concept Note." IFC Working Paper. International `Doing Business' Reports." Working Paper AED-2006- Finance Corporation, Washington, DC. http://www.ifc 1. Attractivité Economique du Droit, University of .org/ifcext/sme.nsf/AttachmentsByTitle/BEEBus Paris X-Nanterre. RegStartup.pdf/$FILE/BEEBusRegStartup.pdf Eiffert, Benn. 2007. "The Economic Response to Regu- de Soto, Hernando. 1990. The Other Path. New York, NY: latory Reform, 2003­06." Paper commissioned for the Harper and Row. Center for Global Development. http://beifert Djankov, Simeon. 2008. "A Response to Is Doing Busi- .googlepages.com/eifertinvestmentresponseto ness Damaging Business." http://www.doingbus reform10-2.doc iness.org/documents/Response_to_Arrunada_JCE.pdf Freeman, Richard B. 2007. "Labor Market Institutions (forthcoming, Journal of Comparative Economics) around the World." NBER Working Paper 13242. Na- ------. 2007. "Measuring the Ease of Enterprise." tional Bureau of Economic Research, Cambridge, http://www.doingbusiness.org/documents/ Measur- MA. ing_the_Ease_of_Enterprise.pdf Galiani, S., and E. Schargrodsky. 2005. "Property Rights Djankov, Simeon, Caroline Freund, and Cong Pham. for the Poor: Effects of Land Titling." Universidad Tor- 2007. "Trading on Time." http://www.doingbus cuato Di Tella, Centro de Investigación en Finanzas, iness.org/documents/trading_on_time_full_report.pdf Business School Working Paper Series, Buenes Aires. Djankov, Simeon, Caralee McLiesh, and Andrei Shleifer. Hall, Robert E., and Charles I. Jones. 1999. "Why Do 2007. "Private Credit in 129 Countries." Journal of Some Countries Produce So Much More Output per Financial Economics 84(2): 299­329. Worker Than Others?" The Quarterly Journal of Djankov, Simeon, Oliver Hart, Caralee McLiesh, and Economics 114(1):83­116. Andrei Shleifer. 2006. "Debt Enforcement around the Hammergren, Linn. 2002. "Performance Indicators for World." NBER Working Paper 12807. National Bureau Judicial Reform Projects." World Bank. http://site re- of Economic Research, Cambridge, MA. sources.worldbank.org/INTLAWJUSTINST/ Djankov, Simeon, Rafael La Porta, Florencio Lopez-De- Resources/Hammergrenperformance.pdf Silanes, and Andrei Shleifer. 2007. "The Law and Hausmann, Ricardo, Dani Rodrik, and Andrés Velasco. Economics of Self-Dealing." http://www.doingbusi 2005. "Growth Diagnostics." John F. Kennedy School ness.org/documents/Protecting-Investors-Self- of Government, Harvard University, Cambridge, MA. Dealing.pdf (forthcoming, Journal of Financial http://ksghome.harvard.edu/~drodrik/ barcelona Economics) finalmarch2005.pdf Djankov, Simeon, Caralee McLiesh, Rita Ramalho, and Hellmann, Joel, and Mark Schankerman. 2000. "Inter- Andrei Shleifer. 2008. "The Effect of Corporate Taxes vention, Corruption and Capture: The Nexus be- on Investment and Entrepreneurship." NBER Work- tween Enterprises and the State." Economics of ing Paper 13756. National Bureau of Economic Re- Transition 8(3): 545­76. search, Cambridge, MA. http://www.nber.org/papers/ Heritage Foundation. 2007. 2007 Index of Economic w13756 Freedom. Washington, DC: Heritage Foundation. Djankov, Simeon, Rafael La Porta, Florencio Lopez-De- IMF (International Monetary Fund). 2007. Revised Guide Silanes, and Andrei Shleifer. 2003. "Courts." The on Resource Revenue Transparency. Fiscal Affairs De- Quarterly Journal of Economics 118(2): 453­517. partment. Washington, DC: IMF. 88 BIBLIOGRAPHY Jappelli, Tullio, and Marco Pagano. 2002. "Information The World Bank's Approach to the Core Labor Stan- Sharing, Lending and Defaults: Cross-country Evi- dards and Employment Creation." October 3. dence." Journal of Banking & Finance 26: 2017­45. http://www.house.gov/apps/list/hearing/finan Kaplan, David S., Enrique Seira, and Eduardo Piedra. cialsvsdem/lazarus.pdf 2007. "Entry Regulation and Business Start-Ups: Ev- Love, Inessa, and Nataliya Mylenko. 2003. "Credit Re- idence from Mexico." World Bank Policy Research porting and Financing Constraints." World Bank Pol- Working Paper 4322. World Bank, Washington, DC. icy Research Working Paper 3142. World Bank, Kauffmann, Daniel. 2002. "Governance Crossroads." In Washington, DC. Global Competitiveness Report, 2002­2003, ed. Peter Loayza, Norman V., Ana Maria Oviedo, and Luis Serven. K. Cornelius. New York: Oxford University Press for 2004. "Regulation and Macroeconomic Perform- the World Economic Forum. ance." World Bank Policy Research Working Paper Kaufmann, D., A. Kraay, and M. Mastruzzi. 2006. "Gov- 3469. World Bank, Washington, DC. ernance Matters V: Governance Indicators for Maldives government, Ministry of Planning and Na- 1996­2005." World Bank Policy Research Working tional Development. 2007. "Seventh National De- Paper 4012. World Bank, Washington, DC. velopment Plan 2006­2010--Creating New Kaufmann, D., A. Kraay, and P. Ziodo-Lobaton. 1999. Opportunities." http://www.planning.gov.mv/en/ "Governance Matters." World Bank Policy Research images/stories/wdp/seventh-ndp.pdf Working Paper 2196. World Bank, Washington, Mas, Ignacio. 2006. A Comparison of the World DC. Bank/IFC's Doing Business Regulatory Indicators Kenyon, Thomas. 2007. "How to Encourage Enterprise with Other Business Environment Indicators. Final Formalization: Some Practical Hints for Policymak- Report prepared for the Financial and Private Sec- ers in Africa." FIAS Policy Note. http://www.fias.net/ tor Vice Presidency of the World Bank Group. May ifcext/fias.nsf/AttachmentsByTitle/June2007Policy 31. NoteEnterpriseFormalization/$FILE/FIAS+Enter Messick, Richard E. 2005. "What Governments Can Do prise+Formalization+in+Africa.pdf to Facilitate the Enforcement of Contracts." Pre- Klapper, Leora, Luc Laeven, and Raghuram Rajan. 2006. pared for Reforming the Business Environment: "Entry Regulation as a Barrier to Entrepreneurship." From Assessing Problems to Measuring Results, Journal of Financial Economics 82(3): 591­629. Cairo, November 28­December 1. http://site Klapper, Leora, Raphael Amit, Mauro F. Guillen, and resources.worldbank.org/EXTGDLNREGION Juan Manuel Quesada. 2007. "Entrepreneurship and ECA/Resources/872782-1163537217246/april_24b.pdf Firm Formation across Countries." World Bank Pol- ------. 1999. "Judicial Reform and Economic Devel- icy Research Working Paper 4313. The World Bank opment: A Survey of the Issues." The World Bank Ob- Group, Development Research Group, Finance and server 14 (1): 117­36. Private Sector Team, Washington, DC. http://www- Monteiro, Joana, and Juliano Assunção. 2006. "Outgo- wds.worldbank.org/external/default/WDSContent ing the Shadows: Estimating the Impact of Bureau- Server/IW3P/IB/2007/10/05/000158349_200710051553 cratic Simplification and Tax Cuts on Informality and 44/Rendered/PDF/wps4313.pdf Investment." Pontifica Universidade Catolica, De- Knack, Stephen, and Phillip Keefer. 1995. "Institutions partment of Economics, Rio de Janeiro.http:// pro and Economic Performance: Cross-Country Tests fessores.ibmecrj.br/erg/wkshops/papers/2006 Using Alternative Institutional Measures." Econom- 0331.pdf ics and Politics 7(3): 207­27. Morris, Felipe, Victor Endo, and Rafael Ugaz. 2004. De- La Porta, Rafael, Florencio Lopez-de-Silanes, and Andrei velando el Ministerio: La Formalization de la Shleifer. 2007. "The Economic Consequences of propiedad en el Peru. Lima: COFOPRI-Banco Legal Origins." http://www.economics.harvard.edu/ Mundial. faculty/shleifer/files/jel.nov_combine.pdf.pdf (forth- OECD (Organisation for Economic Co-operation and De- coming, Journal of Economic Literature) velopment). 2007a. "Two Broad Approaches for Tax Lazarus, Suellen. 2007. Statement prepared for the Reform." Chapter 4 in Economic Survey of New House Committee on Finance Services hearing on Zealand 2007. Paris.http://www.oecd.org/dataoecd/ "The Fight against Global Poverty and Inequality: 39/62/38440934.pdf 89 DOING BUSINESS: AN INDEPENDENT EVALUATION ---------. 2007b. "Toward a More Efficient Taxation ----------. 2007d. "The Cost of State Regulation of the System in New Zealand." OECD Economics De- Enterprise Activity." Chisinau. World Bank, Wash- partment Working Paper 557. Paris. ington, DC. http://siteresources.worldbank.org/INT- Polaski, Sandra. 2007. "The World Bank's Approach to MOLDOVA/Resources/CoDB-2007_En.pdf Core Labor Standards and Employment Creation: Re- ----------. 2006. "Country Policy and Institutional As- cent Developments." Testimony submitted to The sessments: 2006 Assessment Questionnaire." Inter- United States House of Representatives Committee national Development Association, Operations Policy on Financial Services. Washington, DC. October 3. and Country Services, Washington, DC. http://site http://www.carnegieendowment.org/files/polaski- resources.worldbank.org/IDA/Resources/73153- testimony.pdf 1181752621336/CPIA06CriteriaA2.pdf PwC (PricewaterhouseCoopers). 2007. Paying Taxes ----------. 2004. World Development Report 2005: A 2008: The Global Picture. www.pwc.com Better Investment Climate for Everyone. New York: Rodrik, Dani. 2004. "Getting Institutions Right." CESifo World Bank and Oxford University Press. DICE Report, Harvard University, Cambridge, MA. World Bank-IFC. 2007a. "List of Academic Citations of http://ksghome.harvard.edu/~drodrik/ifo-institutions the Doing Business Background Papers." Prepared %20article%20_April%202004_.pdf by the Doing Business Unit. http://www.doingbusi- Sader, Frank. 2002. "Do One Stop Shops Work?" FIAS ness .org/documents/Citations_of_Doing_Busines_re Research Working Paper, Foreign Investment Advi- search_papers.pdf sory Service, Washington, DC. http://www.ifc.org/ ----------. 2007b. Doing Business 2008. Prepared by ifcext/fias.nsf/AttachmentsByTitle/doonestopshops the Doing Business Unit. Washington, DC: World work/$FILE/Do+One+Stop+Shops+Work.pdf Bank. Schneider, Friedrich, and Robert Klinglmair. 2004. ------. 2006a. Reforming Business Registration Regu- "Shadow Economies around the World: What Do We latory Procedures at the National Level: A Reform Know?" Center for Economic Studies & Institute for Toolkit for Project Teams. Small and Medium En- Economic Research (CESifo) Working Paper 1167. terprise Department. Washington, DC: World Bank. Munich. http://ssrn.com/abstract=518526 http://www.ifc.org/ifcext/sme.nsf/AttachmentsBy Vanuatu government. 2007. "Information about the Title/BEEReformBusRegNational/$FILE/Bus+Reg+ Government of Vanuatu." Web site: http://www.van- Book.pdf uatugovernment.gov.vu/vanuatugov.html and ac- ------. 2006b. Doing Business 2007. Prepared by the cessed on December 19, 2007. Doing Business Unit. Washington, DC: World Bank. Webb, Richard, Diether Beuermann, and Carla Revilla. ------. 2005a. Doing Business 2006. Prepared by the 2006. "La Construccion del Derecho de Propiedad Doing Business Unit. Washington, DC: World Bank. ­ El caso de los asentamientos humanos en el Perú." ------. 2005b. Doing Business 2005. Prepared by the Tarea Associatión Gráfica Educativa, Lima, Peru. Doing Business Unit. Washington, DC: World Bank. World Bank. 2007a. Finance for All Policies and Pitfalls ------. 2004. Doing Business 2004. Prepared by the in Expanding Access. Washington, DC: World Bank. Doing Business Unit. Washington, DC: World Bank. ----------. 2007b. Focus on Results: The IDA14 Results World Economic Forum. 2007. Global Competitiveness Measurement System and Directions for IDA15. In- Report. New York: Palgrave Macmillan. ternational Development Association­Operations Yammarino, Francis, Steven Skinner, and Terry Childers. Policy and Country Services. http://siteresources 1991. "Understanding mail Survey Response Be- .worldbank.org/IDA/Resources/Seminar%20PDFs/734 havior." The Public Opinion Quarterly 55(4): 613­39. 49-1172525976405/3492866-1175095887430/ Yakovlev, Evgeny, and Ekaterina V. Zhuravskaya. 2007. IDA15Results.pdf "Deregulation of Business." CEPR Discussion Paper ----------. 2007c. "IDA's Performance-Based Allocation 6610. CEPR, London. http://ssrn.com/abstract=965838 System: Simplification of the Formula and Other Outstanding Issues." International Development As- This bibliography excludes Bank Group and external sociation­Operations Policy and Country Services. documents--such as Country Assistance Strategies http://siteresources.worldbank.org/IDA/Resources/ and donor reports--reviewed as background for each Seminar%20PDFs/73449-1172525976405/3492866- of the thirteen case study countries. 1175095887430/PBA_Sept.2007.pdf 90 THE WORLD BANK GROUP WORKING FOR A WORLD FREE OF POVERTY The World Bank Group consists of five institutions--the International Bank for Reconstruction and Development (IBRD), the International Finance Corporation (IFC), the International Development Association (IDA), the Multilateral Investment Guarantee Agency (MIGA), and the International Centre for the Settlement of Investment Disputes (ICSID). Its mission is to fight poverty for lasting results and to help people help themselves and their environment by providing resources, sharing knowledge, building capacity, and forging partnerships in the public and private sec- tors. THE INDEPENDENT EVALUATION GROUP ENHANCING DEVELOPMENT EFFECTIVENESS THROUGH EXCELLENCE AND INDEPENDENCE IN EVALUATION The Independent Evaluation Group (IEG) is an independent, three-part unit within the World Bank Group. IEG-World Bank is charged with evaluating the activities of the IBRD (The World Bank) and IDA, IEG-IFC focuses on assessment of IFC's work toward private sector development, and IEG-MIGA evaluates the contributions of MIGA guarantee projects and services. IEG reports directly to the Bank's Board of Directors through the Director-General, Evaluation. The goals of evaluation are to learn from experience, to provide an objective basis for assessing the results of the Bank Group's work, and to provide accountability in the achievement of its objectives. It also improves Bank Group work by identifying and disseminating the lessons learned from experience and by framing recommendations drawn from evaluation findings. ISBN 978-0-8213-7552-5 THEWORLD BANK SKU 17552