Report No. 12313-GU Guatemala An Assessment of Poverty April 17, 1995 Country Department If Human Resources Operations Division Latin America and the Caribbean Regional Office U~~~~~ V #W:~~~~~~~~2; - j -*a I~~~~~~~~~~~~~~~~~~~T~4 Currency Equivalents (as of December 14, 1994) Currency Unit = Quetzal (Q) US$ 1.00 = Q 5.78 Fiscal Year January-December GUATEMALA: AN ASSESSMENT OF POVERTY LIST OF ABBREVIATIONS AND ACRONYMS ARI - Acute Respiratory Infections AVANCSO - Association for the Advance of the Social Sciences (Asociaci6n para el Avance de las Ciencias Sociales en Guatemala) BANDESA - National Bank for Agricultural Development (Banco Nacional para el Desarrollo Agriculo) BOG - Bank of Guatemala CACM - Central American Common Market CEPAL - Economic Commission for Latin America and the Caribbean (UN) CDUR - Urban and Rural Development Councils (Consejos de Desarrollo Urbano y Rural) CG - Central Government CISMA - Center for Mayan Social Research (Centro de Investigaci6n Social Maya) DIGEBOS - National Extension Service--Forestry (Direcci6n General de Bosques) DIGEPA - Project Support Office (Ministry of Education) (Direcci6n General de Proyectos de Apoyo) DIGESA - National Extension Service--Agriculture DIGESEPE - National Extension Service--Livestock DTP - Department of Technical Planning ENSD - National Socio-Demographic Household Survey (Encuesta Nacional Sociodemografica) FAFIDESS - National Financial Consulting Foundation for Development and Social Service Institutions (Fundaci6n de Asesoria Financiera a Instituciones de Desarrollo y Servicio Social) FECOAR - Federation of Regional Farming Cooperatives (Federaci6n de Cooperativas Agricolas Regionales) FIS - Social Investment Fund (Fondo de Inversi6n Social) FOGUAVI - National Housing Fund (Fondo Guatemalteco para la Vivienda) FONAPAZ - National Fund for Peace (Fondo Nacional para la Paz) FONATIERRA - National Fund for Agricultural Lands (Fondo Nacional de la Tierra) FSDC - Community Development Fund (Fondo de Solidaridad para el Desarrollo Comunitario) FUNDACEDI - National Education and Integral Development Foundation (Fundaci6n para la Educaci6n y el Desarrollo Integral) FUNDESA - National Foundation for Guatemalan Development (Fundaci6n para el Desarrollo de Guatemala) GAiT - General Agreement on Trade and Tariffs HC - Community Daycare Centers (Hogares Comunitarios) ICTA - National Agricultural Research Institute IDB - Inter-American Development Bank IFAD - International Fund for Agricultural Development IGSS - Guatemalan Institute of Social Security (Instituto Guatemalteco de Seguridad Social) HICA - Cooperative Institute for Agriculture (Instituto de Cooperaci6n para la Agricultura) INE - National Statistics Institute (Instituto Nacional de Estadistica) INTA - National Institute for Agrarian Reform (Instituto Nacional de Transformaci6n Agraria) ISI - Industrialization Strategy Institute LCPE - Local Currency Price Equivalent MNIDES - Ministry of Development (Ministerio de Desarrollo) MOE - Ministry of Education MOF - Ministry of Finance MOH - Ministry of Health MSPAS - Ministry of Public Health and Welfare (Ministerio de Salud Publica y Asistencia Social) NFPS - Non-Financial Public Sector NGOs - Non-Government Organizations NTBs - Non-Tariff Barriers OECD - Organization for Economic Cooperation and Development OPS - Panamerican Health Office (Oficina Panamericana de Salud) PAHO - Pan-American Health Organization PANE - School Food and Nutrition Program (Programa de Alimentaci6n y Nutrici6n Escolar) PLADES - Action Plan for Social Development of Guatemala POAs - Annual Operation Plans PREN - National Reorganization Program PRES - Short Run Program of Economic and Social Adjustment PROGETTAPS - Technology Generation and Transfer and Seed Production Program PVOs - Private Voluntary Organizations SEGEPLAN - National Planning Secretariat (Secretarfa General del Consejo Nacional de Planificaci6n Econ6mica) SIMME - Commission to Support Microenterprise (Comisi6n Nacional para el Fomento de la Microempresa y Pequefia Empresa) UNICEF - United Nations Children's Fund USAID - United States Agency for International Development The Poverty Assessment is intended to stimulate dialogue with the Government and other sectors on a poverty alleviation strategy for Guatemala. It is also intended to complement Bank operations and sector work. This report was prepared by Claire Lucas under the guidance of Mr. John Newman with inputs from Claudio Sapelli and George Monserrat (Macroeconomic Policy), Dominique van de Walle, Harry Patrinos, and Haeduck Lee (Poverty Profile) Zafiris Tzannatos and Allesandro Cigno (Labor Market), Madalena dos Santos and Luis Crouch (Education), Ted Downing and Elizabeth Katz (Agriculture), Christina Wood and Diane Steele. Alex Panagides participated in many stages of data analysis and report writing. The peer reviewers were Helena Ribe, Dominique van de Walle, Theresa P. Jones and Karen Lashmnan. The report also benefited from the comments of Kye Woo Lee and Eduardo Velez. Bill Wood, Kevin Rost and Vicki Macintyre provided editorial assistance. Barbara Diallo, Carmen Alcantara and Maria Colchao provided computing and formating support. The Department Director was Mr. Edilberto L. Segura (LA2DR). The Division Chief was Mr. Kye Woo Lee (LA2HR). pp *suoisnpauo, (t) ... svaio I;mtn u! dQtoviadsa 'Juaiajkp si adnponi,soyfuj (E) .......................... ood afjl paioArwf ;ou st7q uotisuajxa puD uip.iasai ;tJn;jnoq XV (j) Op...............................................................................lood at afO uo uijnponad jinnj1n9 v ol ;oddnd (n ) Ot ............................................S. IAI aS IL ocJfS aNY 1Jf1o i SV !AI NI 9MIS3A I *9 RE . ........... . suoisnl;?uoD) (f) ........ ..saotdas aivnbaptvui ap!A°od put' 1tvaM aiO s3.disIu1w uoilivonpa pu! Lt'aJ (i!) P. .pajvo011tvs.iw s! uo.tiv.npa put' ylltlaL uo ,uads si apll!i °t'qM (f) .p ........... uoilvonpa puv ,u;vay ua ds t'/vwa;tvnD (Z) 6Z.-- -... . .. . . .. . ''' 'lffdp aSin si.ojtvo!pui uoil!.,Inu put' q;ltvay 'uo1lt'onp¶V ([) 6Z ......................................V....... .W 4IIJ ) NVWf1H NI fN IS3ANa S Zz 7.................... ............................ suo.isnl,auoD (E) .z -.. -- .......A.ua!4fa asna.siou! p;nom uo!;nqu!slp alqv;,nba a.iow V (Z) ,6 ..i oqtv a tow Aoldwa pjnon uo:;nq!usip alqnl!nba alo.w V (1) t'z.---------------------.NV'I 01 SS333V S.HOOd 3H1 9NIISv3H3MI tI .Z '' ''''''''' ''' ''''.s.uo.isnpuoD (9) .1.'' '.. ''' ' 'salnJzpuadxa .iopas Ivnpos papunf dloijqndiof adnds losf dayl palidauag juawavuvt'UtW 2iuWouoda stH (S) 61 ...........................g . oqovjfo asn aAlJdnpoJd ayl padiof ,uawalt'utvw Otuaouoia sOH (t) RI................................"Iplla...;................... iio.!zqvm puv yim a;ou.ow d ;uawXua3vuwn w p7!WoUd 7p (E) 81 R................................'..........''. luaw utw .wulouo,a Jo ssaua4jl ajj0J (Z) 91.d - ......-dluvrlWofdad j.twouooao.ijnuw juaoab ([) r. IIN3N39VNWVS 31WONO33 AO SS3N3AII33A3 31H1 GNV 33NVWHOAH3d :IWONOj3O:3Vw *C ]f,[ --------------------.................... ................................. vUo a vnD u. ay ua a y, a, .s1V-------------------------------------------------- A931"IS NO11VIA3TIV AILWIAOd V Z 11 'AILIIVfl3NI g 6.dWflSUOd.i ood ayl iDnqf (0C) 8 a..O.U! .iUtaJlta uiood al' moH (6) ........................................ ..iaAod duaff' olo () 9 .. .sploqasnoq papovq alotwaff (Z) . qJliadfo sixial .aeKol puD uoilvJfnpa sajl aAVy iood ayiL (9) p.ee. v----- --- - --- ---uoips'odwto, !uq,Iq (S) ..uoiJnql.dslU (t) .,,,.,u . ..............................................a.r.. U. ul na t' ) dfi 7 (U ...................... SiJunoo daylofo ivlf OJ ai'dWuo daods, It nD IsD eP OH () .............................I'll......................................... .......... 4aAOd s! aAsuaixa oH ([) I.AIH~~~~~~~~~~~~~~~..................................l3AOd V 1... . .OHJ Al3AOd 3111.1 ....... ......A~IWV Nf1S 3AIlfl:)3Xa SIN3INO3 AREIAOJ Jo IN3NSSISSV NV :VWAa1V39 ii 7. IMPLEMENTING A POVERTY ALLEVIATION STRATEGY ................................................... 46 (1) Immediate decisions .................................................. 46 (2) Decisions in the nearfuture ................................................... 47 TABLES Table 1.1: Regional Population and Poverty Distribution .3 Table 1.2 Urban and Rural Poverty Distribution .4 Table 13: Regional Distribution of Poverty .5 Table 1.4: Headcount Poverty Rates .6 Table 1.5 Average School Completion and Illiteracy Rates ..................................,.6 Table 1.6: The Marginal Effect of Different Variables on the Probability of Being Poor .8 Table 1.7: Headcount Poverty Rates .9 Table 1.8: Sources of Income .10 Table 1.9: Distribution of Family Consumption Expenditures and Budget Share by Poverty Group 10 Table 1.10: Gini Index of Income Inequality .12 Table 1.11: Distribution of Income and Consumption .12 Table 3. 1: Nominal and Real Effective Exchange Rate .19 Table 5. 1: Water & Sanitation and Electricity Coverage, 1989 .30 Table 5.2: Regional Variation in Health Indicators andPoverty .31 Table 5.3: Regional Variations in enrollment and Poverty .32 Table 5.4: Illiteracy by Region (%) . .. 32 Table 5.5 MOE Regional Expenditures and Poverty Headcount .36 FIGURES Figure 3.1: Real Agriculture GDP per Capita .16 Figure 3.2: Real Health & Education Public Expenditures 21 Figure 5. 1: Per Capita Health Expenditure and Urbanization by Region .35 ANNEXES ANNEX I: Description of Data and the Measurement of Poverty ANNEX II: Estimated Poverty Levels from other Studies ANNEX III: Regional Poverty Differences ANNEX IV: Illness and Poverty ANNEX V: Gender and Poverty ANNEX VI: Probit Analysis ANNEX VII: Further Characteristics of the Poor ANNEX VIII: Land Distribution ANNEX IX: Main Economic Indicators ANNEX X: Comparative Data ANNEX XI: Public and Private Institutions ANNEX XII: A National Statistical System to Monitor and Evaluate Poverty Reduction Efforts ANNEX XIII: Participatory Poverty Assessment iii Priority Poverty Indicators Poverty lines (1989), Quetzales Guat. City Other Urban Rural Upper Poverty Line 98.57 77.01 66.79 Headcount 41.7 66.9 85.7 Extreme 53.62 48.75 44.68 Headcount 20.3 38.0 71.9 Short-term Income Indicator 1965 1970 1975 1980 1985 1989 1990 1991 1992 CPI (Food only,19851=00) 100.0 204.1 300.3 397.4 426.2 Social Idi cators 1965 1970 1975 1980 1985 1989 1990 1991 1992 GDP per Capita 363 605 1139 1221 940 830 988 GDP Share on Public Expenditures for Basic Social Services2 3.3 2.0 4.1 Net Primary Enroll.3 47.3 50.4 54.4 60.5 65.4 68.0 Under 5 Mortality4 94.1 66.7 50.3 Child Malnutrition(%)4 43.5 33.5 Immunization (%)4 Measels 46.0 DPT 58.0 Life Exp. Male4 54.5 56.8 59.7 Life Exp. Female4 58.4 61.3 64.4 Life Exp. Total4 52.5 58.0 63.1 63.7 64.2 Total Fertility Rate4 6.5 6.2 5.5 5.4 5.4 Maternal Mort.4 9.1 11.2 9.2 SOURCES: I World Bank, World Tables 1993, John Hopkins, 1993. 2 Psacharopoulos et at., Poverty and Income Distribution in Latin America: The Story of the 1980s. World Bank, 1992. 3 Ministerio de Educaci6n. Anuario Estadistico. 1993. 4 PAHO, An6fisis de la Situaci6n de Salud por Re2iones: Guatemala. 1992. 1992 Note: Under 5 mortality is per 1,000 births. Maternal Mortality is per 10,000 live births. GUATEMALA: AN ASSESSMENT OF POVERTY EXECUTIVE SUMMARY I. Guatemala has serious problems of poverty and inequality. Approximately 75 percent of the population is estimated to live below the poverty line, which is defined as having insufficient income to purchase a basic basket of goods and services. Almost 58 percent of the population has incomes below the extreme poverty line, which is defined as the amount needed to purchase a basic basket of food. 2. According to a recent World Bank study (Chen et al, 1993), Guatemala has the eighth highest incidence of poverty among forty-four low- to middle-income countries of the world. If the comparator group is defined in terms of countries with similar levels of purchasing-power adjusted GNP per capita ($2,000 - $4,000), Guatemala is rated as having the highest incidence of poverty. Indeed, all social and economic indicators, including patterns of malnutrition, health and education, point to extreme poverty. In the poorest part of the country, the Western Highlands, male agricultural wages in 1989 averaged $1.35 a day, and those for females $0.76 a day. More than 70 percent of the country's population lives in rural areas and approximately 85 percent of this group is classified as poor. More than 90 percent of the indigenous population live on an income which is less than the poverty line, as compared to 66 percent of the non-indigenous population. Even after controlling for other factors associated with being poor, the probability that an indigenous person is poor is estimated to be 15 percentage points higher than that of a nonindigenous person. 3. In addition to the problem of poverty, there is a high degree of inequality in land, income and consumption. The most recent agricultural census (1979) indicates that 2.5 percent of Guatemala's 5.3 million farms control 65 percent of agricultural land (with an average farm size of 200 hectares) while only 16 percent of the land is cultivated by 88 percent of the smallest farms (with an average farm size of 1.5 hectares). The Gini Index for land distribution was calculated to be 85.9 in 1979, higher than all but two Latin American countries, pre-reform Peru in 1961 and pre-reform Colombia in 1964. The pattern of land ownership that prevails in Guatemala is not only inequitable, but also most probably results in efficiency losses for the economy. This conclusion follows from the mounting international evidence that small family farms enjoy an efficiency advantage and use more labor than large scale farms operated with hired labor. 4. The unequal distribution of land is a major factor explaining why Guatemala has the third highest degree of labor income inequality (exceeded only by Brazil and Pakistan) among the 44 low and middle income countries in the aforementioned World Bank comparative study. 5. The twin problems of poverty and inequality were present long before the "lost decade of the 1980s" in Latin America and before any efforts at structural adjustment. For example, the first agricultural census in 1950 indicated that 47.6 % of all landowners owned less than 1.4 hectares each (3.6 acres). The dissatisfaction with poverty and inequality precipitated an armed conflict that has lasted, with varying degrees of intensity, for 40 years and has been estimated to vi cost the lives of approximately 100,000 people and the displacement of many more. Now, as negotiations are underway between the government and the Unidad Revolucionaria Nacional Guatemalteca (UNRG) to end the conflict, Guatemala must search for alternatives to violence and repression to deal with problems of poverty and inequality that, in many respects, are as severe as they were at the start of the conflict. In terms of the absolute number of poor, the situation is much worse as the population today (9.1 million in 1991) is so much higher than it was in 1950 (2.8 million). 6. Although the late 1980s and early 1990s have seen the beginnings of an explicit poverty reduction strategy and some advances, much remains to be done. Guatemala is the only country in Central America that has not implemented at least a nominal reform of its land tenure system. Although small farms are most probably more efficient and use more labor, large landowners seem unwilling to sell off parts of their holdings, in part because they fear this would increase the demand for land distribution or launch a move toward land invasion. At the same time, if the market price of land exceeds the present discounted value of agricultural profits, poor people without access to grants or equity will still be unable to buy land. Under these conditions, it continues to be difficult to distribute land more equitably and efficiently. 7. Poverty could also be greatly alleviated by investments in education, health and nutrition -- through their effects on labor productivity and thus on income and growth. However, households and the government alike have failed to invest enough. As a result, literacy rates remain dismal: 36.9 percent for males and 52.9 percent for females. School enrollment, although 77 percent for primary education, drops to 21 percent for secondary education and 9 percent for tertiary. In health, the situation is not much better. The infant mortality rate is 55 per 1,000 live births and the maternal mortality rate is 110 per 100,000 live births -- these indicators are even worse for the poor. 8. The causes of the low indicators are similar in both health and education. Given their incomes, the poor are likely to underinvest in the human capital of their children, even if such investments have a high social return. Years of underinvestment by the public sector, aggravated by low internal efficiency and an inequitable and inefficient allocation of expenditures biased towards urban areas and the nonpoor, have not helped the situation. For example, a provision of the 1986 Constitution earmarks 5 percent of all tax revenue to San Carlos University. In 1990, the operating expenses of the two largest hospitals in the metropolitan area exceeded the operating expenses for all health posts and clinics in the country. 9. Many of the poor believe that what they need is better infrastructure, particularly roads, and technical assistance to make them more productive and thus provide a way out of their dire poverty. But problems abound here, too, because at present the government has no sound means of financing the needed infrastructure, especially in rural areas. 10. The challenge of reducing poverty and inequality in Guatemala is enormous and, if policies continue as they have in the past, Guatemala will enjoy neither substantial increases in per capita growth rates nor significant reductions in poverty. The latest World Bank projections are that Guatemala could grow at a sustainable rate of 4.5 percent a year between 1995 and the year 2000. vii Because of the high population growth rate, GDP per capita is projected to increase only by 1.7 percent a year. If these currently projected growth rates in GDP take place in a distributionally neutral fashion, the reduction in poverty rates will be minimal. A simulation done as part of this study suggests that if the per capita income of every household in the 1989 Household Survey were to grow at a rate of 1.7 percent a year without interruption for 10 years, the poverty rate would decline from 75 percent to 69 percent. 11. Using the experience of countries that have surmounted their poverty problems in recent years as a guide, the World Bank recommends a strategy for Guatemala that has three key elements: (a) Stability in the macroeconomic environment and economic management that makes efficient use of the poor's most abundant asset - labor. (b) Greater access to land for the poor. (c) An increase in tax revenue to finance public investments that would enhance the earning potential of the poor, including: (i) Investments to raise the human capital of the poor. (ii) Infrastructure and support services to increase the productivity of the existing physical and human capital of the poor. 12. The key elements of the strategy follow the general approach recommended in the 1990 World Development Report on poverty, the subsequent recommendations made in the World Bank Policy Paper, Assistance Strategies to Reduce Poverty, and, in part, also reflect the East Asian experience with generating growth and reducing poverty. Recent analysis of successful development in East Asia suggests that pursuing sound macroeconomic policies and promoting investments that increase the earnings potential of the poor can lead to both higher rates of economic growth and lower levels of inequality and poverty. Thus the so-called inevitable tradeoff between growth and reducing inequality and poverty need not materialize. 13. The first element of the strategy is to pursue policies that generate stability in the macroeconomic environment and to manage the economy so as to make efficient use of the poor's most abundant asset - labor. Macroeconomic instability is not only bad for growth, but hurts the welfare of the poor (as illustrated by Guatemala's experience during a large part of the 1980s). However, the macroeconomic experience in Guatemala also demonstrates that stability and growth are necessary but not sufficient conditions for reduction of poverty (best illustrated by Guatemala's experience throughout the 1960s and 1970s). Even macroeconomic stability and an appropriate incentive structure may not be sufficient to reduce poverty at a socially acceptable rate if the poor do not have the ability to respond to the incentive structure (a situation that appears to be the case from the end of the 1980s to the present). 14. The second element is to increase access to land of the poor. This involves efforts not only to make it easier for the landless and land poor to acquire land, but also to strengthen land tenure security for smallholders who already possess land. Achieving a more equitable and efficient distribution of land will require coordination and cooperation among all parties viii concerned. In countries such as Nicaragua, Mexico and Honduras, where major policy issues have been resolved, one of the secrets to their success has been a process of consensus-building among all the social groups involved -- including those who were previously excluded. Furthermore, a more fruitful dialogue emerged when the key policy questions were faced before more technical issues (for example, those related to titling). 15. The third element is to structure public investment so that it enhances the earning potential of the poor. This would reinforce the increase in returns to the poor's assets brought about by changes in relative prices and incentive structures within the macroeconomic environment. There are two main ways to increase the earnings potential of the poor. Either poor families have to acquire more physical and human capital or their existing physical and human capital has to be made more productive. 16. Given the divergence between private and social returns to human capital investment and the accumulated international evidence concerning the effects of human capital investment on development, a strong case can be made that Guatemala's poverty alleviation strategy should set the public finance of human capital investments at a higher level than in the past. However, the performance of the public sector over the last twenty years in Guatemala and elsewhere does not provide a compelling case for providing more health and education services in the same way as they have in the past. In both developing and developed countries, there has been growing interest in introducing incentives that make providers more accountable for the quality of services provided. These mechanisms can include subcontracting the provision of health and education services to private companies through open competitions, operating public facilities under management contracts, and introducing performance-based budget allocations, internal markets, and greater community participation in local budget decisions. 17. How Guatemala combines these three elements - greater access to land for the poor, investments to raise the human capital of the poor, and infrastructure and support services to increase productivity -- in its poverty alleviation strategy is a political decision that must be taken by the different interest groups within Guatemala. However, if Guatemala is unable to maintain an adequate tax level for a successful poverty alleviation strategy, then the country will only be able to look forward to low rates of per capita growth, high population growth rates, continued high levels of poverty, continued social unrest and mounting instability. GUATEMALA: AN ASSESSMENT OF POVERTY 1. THE POVERTY PROFILE 1.1 To appreciate the poverty situation in Guatemala, it is necessary to (a): examine the data from household surveys and participatory assessments to determine the extent of poverty in the country; (b) compare these data with those from other countries; and (c) identify the prominent characteristics of the poor. Another telling factor to consider is the nature of inequality, both in income and land. The data and methodology used to measure poverty are discussed in detail in Annex 1. A. Poverty (1) How extensive is poverty? 1.2 Approximately 75 percent of the population of Guatemala lives below the poverty line, which is defined as having insufficient income to purchase a basic basket of goods and services. Furthermore, the average poor person must live on an income that falls 44 percent short of this line. Those in extreme poverty--that is, those who are unable to purchase a basic basket of food-- make up 58 percent of the population. 1.3 While these figures may overstate the case somewhat, because the household survey data on which they are based do not adequately take into account the value of home consumption or the seasonality of incomes there is no doubt that the country's poverty rates are extremely high. " This is confirmed by data from several other sources, including the results from the participatory poverty assessment conducted as part of this poverty study, the findings of other qualitative studies, and the country's patterns of malnutrition, health, and education, all of which are consistent with extreme poverty. 1.4 Even when varying methodologies are used, as demonstrated in some recent studies, the conclusion remains the same: namely, that poverty in Guatemala is both widespread and severe (Annex 2). Data from a 1980 survey by the United Nations Economic Commission for Latin America and the Caribbean (CEPAL) point to a rate of 71.1 percent, which the commission found had risen to 73.1 percent following its 1986/7 survey. These two surveys showed extreme poverty at levels of 39.6 percent and 48.5 percent, respectively. 1.5 The participatory poverty assessment in this study, which was based on 627 conversational interviews in seven rural regions and the metropolitan area of Guatemala City, corroborates the findings of the household surveys. A translation of the executive summary of this report is provided in Annex 13. Close to 90 percent of this group viewed poverty as not having enough to eat (which is roughly equivalent to the definition of extreme poverty commonly I/ See the diacuion on data quality in Annex 1. 2 used in poverty assessments. More than half of the 627 interviewed stated that the lack of food was their major problem in life: they cited high food prices, low salaries, and the lack of work as the main reasons they were unable to obtain the food they needed. Most of the group were eating only two meals a day. (2) How does Guatemala's poverty compare to that of other countries? 1.6 Compared with other countries, Guatemala stands out as having an inordinately high rate of poverty for its level of per capita GDP. Of 44 low- to middle- income countries worldwide, it has the eighth highest incidence of poverty (Chen, Datt, and Ravallion, 1993); and among the 18 countries in this sample with similar levels of GNP per capita ($500--$1,500 per year), it ranks third highest in this regard. Furthermore, among countries with similar levels of GNP per capita adjusted for purchasing power ($2,000--$4,000 per year), Guatemala is at the very top in incidence of poverty. (3) Wk74 are the prominent characteristics ofpoverty in Guatemala? 1.7 All the poverty studies that have been done in Guatemala indicate that the condition is especially prevalent in rural areas in the North and Northwest and occurs primarily among the poorly educated and the indigenous members of the population. In a multivariate analysis of these patterns, even after controlling for other factors associated with being poor, indigenous people were found to be 15 percent more likely to be poor. Interestingly, a simple bivariate analysis of the patterns suggests that female-headed households are not necessarily more likely to be poor. These bivariate and multivariate studies have shed considerable light on the distribution of poverty in Guatemala, the level of education and health among the country's poor, the levels of poverty in female-headed households, the main causes of poverty, the main sources of income among the poor, and the main kinds of food they consume. (4) Disibution 1.8 Since poverty is so widespread in Guatemala, it occurs in roughly the same geographical distribution as the population as a whole (see Table 1.1). However, the poor are substantially under-represented in the metropolitan area, and over-represented in the North, Southwest, and Northwest regions. 3 Table 1.1: REGIONAL POPULATION AND POVERTY DISTRIBUTION Poverty Rates by Urban and Rural Location Region Population Percent of Share Total Poor Urban Rural Total Extremely Poor Extremely Poor Extremely Poor Poor Poor Poor Metropolitan 21.3 15.4 20.3 47.1 55.4 74.6 29.3 54.2 North 8.2 9.9 46.4 66.9 91.5 96.7 83.4 91.3 North-East 9.0 8.3 38.1 61.7 55.1 71.5 51.4 69.3 South-East 8.6 9.2 43.7 67.7 70.4 82.9 64.8 79.7 Central 11.3 11.4 54.7 73.9 57.9 77.7 56.5 76.0 South-West 25.4 27.9 39.6 60.4 73.0 89.7 65.0 82.7 North-West 13.6 16.9 67.8 78.1 91.4 96.5 87.9 93.7 Peten 2.6 2.6 34.8 59.0 68.1 82.1 58.4 75.4 Total 100.0 100.0 33.7 57.2 71.9 85.7 57.9 75.2 Source: Encuesta Nacional SociodemogrAfica, 1989. 1.9 When distribution is judged by what percentage of the people in particular regions are poor, Guatemala City again stands out as being less poor than the other regions (Table 1.2). Although approximately 42 percent of the population of Guatemala City is poor, this number is well below the corresponding rates of 67 percent and 86 percent in the country's "other urban" and rural areas, respectively. A rate of 72 percent in rural areas is considered extreme, as it means that nearly three-fourths of the rural population cannot meet their daily nutritional requirements even if they were to apply 1 00 percent of their income toward food purchases. According to the poverty gap, there is considerable disparity in the level of poverty between these locations. The average poor person in Guatemala City has an income that is 19 percent below the poverty line, whereas in rural areas the income shortfall for a poor person is on average 54 percent below the poverty line. 1.10 The variability within regions is also great, as is evident from figures recorded at the departamento level (Annex 3)2' In the Northeast, for example, poverty ranges from 57 percent in Izabal to 85 percent in Chiquimula. Like the figures for the nation as a whole, those on poverty in 21 Although discussion of poverty conditions at the departamento level is useful with respect to identification and targeting in policy design, the confidence with which one can make such assessments is substantially reduced. Given the nature of the underlying household survey for this poverty Ianysis, findings at the departmental level should be interpreted with caution. For this reason, discussion at the departmental level is limited to the poverty measures and a few relevant characteristics in the poverty profile of the next section. Statistical confidence levels for departamento poverty neasure are given in Annex 3. 4 the departamentos point to a higher incidence in rural areas. Not surprisingly, the departamentos with the highest incidence also show the most serious poverty gap. However, the way that the country's 22 departamentos rank depends largely on the measure used. The headcount ratio, for example, is highest in Quiche, followed closely by the that in Huehuetenango, Alta Verapaz, and Baja Verapaz, but the poverty gap and FGT P2 index are worst for Huehuetenango. In other words, although a greater percentage of the population is poor in Quiche, the average poor person in Huehuetenango has an income that is further below the poverty line. How resources are allocated in antipoverty programs thus depend on the weight that policy planners wish to give to the poorest of the poor in their rankings. Table 1.2: URBAN AND RURAL POVERTY DISTRIBUTION Population Under Poverty Line, 1989 Population Percent of Total Po verty Ind i ces Share Poor Headcount Poverty FGT P2 Index Ratio Gap Guatemala City 18.1 10.1 41.7 18.9 9.9 (0.54) (0.27) (0.18) 'Other Urban" 20.0 17.8 66.9 35.7 23.5 Areas (0.49) (0.34) (0.28) Rural Areas 61.9 70.5 85.7 53.5 38.9 (0.20) (0.19) (0.18) Sources: Encuesta Nacional Sociodemopifica, 1989. Instituto Ncicmal Estadistica, 1990. Notes: Stard acon calculated from the formula: a,y= (*H(l -.HN)* l00 are psted in panhe. (5) Ethnic composition 1.11 Approximately 40 percent of Guatemala's population consists of indigenous people. This represents one of the highest concentrations of such people in Latin America and the Caribbean region. Almost 80 percent its indigenous population live in rural areas. More than 90 percent live on an income that falls below the poverty line, whereas this figure is 66 percent for the nonindigenous population (Table 1.3). At the regional level, the indigenous poverty rate is most extreme in the Northwest, North and Southeast--precisely the regions where the indigenous predominate in the overall population as well. Furthermore, in all regions but Guatemala City, more than two-thirds of the indigenous population live in extreme poverty. 5 Table 1.3: REGIONAL DISTRIBUTION OF POVERTY By Ethnicity Region Indigenous Non-Indigenous Total Percent Indigenous Extremely Poor Extremely Poor Extremely Poor Poor Poor Poor Metropolitan 61.2 82.2 25.8 51.1 29.3 54.2 9.7 North 88.9 94.9 50.9 70.1 83.4 91.3 84.3 North-East 82.9 91.6 48.3 67.2 51.4 69.3 8.7 South-East 86.6 95.5 64.5 79.5 64.8 79.7 71.7 Central 72.3 90.2 47.6 68.1 56.5 76.0 36.4 South-West 76.5 91.0 55.1 75.4 65.0 82.7 46.2 North-West 91.2 96.8 68.7 76.1 87.9 93.7 84.6 Peten 77.4 92.8 55.6 72.9 58.4 75.4 12.7 Total 81.3 92.6 45.2 65.8 57.9 75.2 36.3 Source: Encuesta Nacional Sociodemografica, 1989. (6) The poor have less education and lower levels of health 1.12 Close to 79 percent of household heads with no schooling can be classified as poor, compared with 48 percent of household heads who have completed primary school. The poverty rates for those who have completed secondary and higher education are 16 percent and 9 percent, respectively (Table 1.4). On average, poor individuals over the age of 18 have completed only 1.9 years of school, while the extreme poor have finished 1.6 years of school (Table 1.5). In contrast, non-poor individuals average 5.8 years of schooling, which is approximately equivalent to a full primary education. 1.13 The poor and those who live in rural areas are less likely to state that they are ill or injured than the non-poor and those who live in Guatemala City (Annex 4, Table A4. 1) 3 However, the poor tend to remain ill and inactive longer than the non-poor (Annex 4, Table A4.2). Also, the mean length of illness and inactivity is longer in rural areas than in urban areas. Although the poor are less likely than the non-poor to consult with health personnel, especially in rural, of those who are ill or injured, about 55 percent of the poor versus 71 percent of the non-poor visit health 3/ The ENSD ask whether an individual has been ill or injured during the last four weeks; if the answer is yes, a series of questions is posed. About five percent of the overall population self-identified as having being ill or injured 6 providers. In rural areas, only 52 percent of those claiming to be ill seek attention from health care professionals, whereas in Guatemala City the consultation rate is 72 percent. Table 1.4: HEADCOUNT POVERTY RATES By Level of Education Extremely Poor Non- Poor Poor Educational Attainment No Schooling 61.5 78.8 21.2 of Household Head Primary 24.3 48.1 51.9 Secondary 7.0 16.1 83.9 Higher 5.7 8.6 91.4 Source: Encuesta Nacional Sociodemogr4fica, 1989. Table 1.5: AVERAGE SCHOOL COMPLETION AND ILLITERACY RATES By Poverty Category Extremely Poor Non-Poor Poor Years of (18+ years) 1.6 1.9 5.8 Schooling Illiteracy Rate (14 + years) 54.4 49.0 18.1 Source: Encuesta Nacional Sociodemogr4fica, 1989. (7) Female headed households 1.14 Nearly one in six Guatemalan households are headed by a female (Annex 5, Table A5. 1). Of the individuals who live in these households, 70 percent are poor and 49 percent are extremely poor. In contrast, among those living in male headed households, 76 percent are poor and 59 percent are extremely poor. 1.15 Although the female heads of household possess only 0.5 years less schooling than males, they are more likely to be illiterate and to work considerable fewer hours. They also preside over smaller households and are older than their male counterparts (Annex 5, Table A5.3). In addition, they are twice as likely to be employed as professionals, managers, or office workers. A much smaller number of female household heads are farmers, but a high percentage work in the service sector workers or as vendors. More indigenous than nonindigenous households are headed by a 7 male. Because poverty is so pervasive among the indigenous population, this would increase the incidence of poverty among male-headed households. 1.16 Transfers account for approximately 17 percent of the total income for female-headed households. Although the household survey does not contain information on the origin of these transfers, they may be payments sent home by absent husbands who travel from place to place in search of work. Unfortunately such migrant behavior could not be statistically controlled for in the survey. But these transfers are extremely important for households headed by women and can be the determining factor that keeps a household from falls into poverty (Annex 5, Table A5.4). In the departamento of Sacatepequez, for instance, the marginal effect of transfers decreases the poverty rate from 60 percent to 40 percent for individuals living in female-headed households. (8) Factors affectingpoverty 1.17 Multivariate analysis suggests that poverty is closely related to schooling, age, region of residence, and ethnicity. (Table 1.6).4' This type of analysis makes it possible to measure the marginal effect of a given characteristic on the probability that a head of the household is poor, holding the other characteristics constant. The bivariate relationships discussed above, while important, cannot separate the effect of one variable--say ethnicity--from that of place of residence or schooling. Are indigenous people more likely to be poor because they are indigenous or because they live in a poor region and have less education? The bivariate relations indicate that indigenous people are more likely to live in the Northwest region and are less likely to be educated. For example, among children aged 10 to 12, only 57 percent of the indigenous as compared with 75 percent of the nonindigenous are enrolled as students. Only 8 percent of indigenous school-age individuals above the age of 15 are enrolled, as compared with 32 percent of their nonindigenous counterparts (Psacharopoulos and Patrinos, 1993: 167). According to the multivariate estimation, even after controlling for differences in education levels and other variables affecting the probability of being poor, indigenous people are roughly 15 percent more likely to be poor. 1.18 Multivariate analysis also strongly suggests that, for each additional year of schooling, the probability of being poor declines by 3 percent. It also makes clear the overrepresentation of the poor in the North and Northwest. If other factors are held constant, a resident of the North is 12 percent more likely to be poor, but the figure for a resident of the Northwest region is 18 percent. In addition, the analyses show that the probability of being poor is related to the age of the household head, with younger household heads more likely to be poor. 4 Annex 6 briefly describes the probit analysis and presents the means of the explanatory variables. 8 Table 1.6. THE MARGINAL EFFECT OF DIFFERENT VARIABLES ON THE PROBABILITY OF BEING POOR Variable Marginal Effects t-Value Family Size 6.96 24.4 Number of rooms -8.73 t 18.0 Years of schooling -3.42 t 17.4 Indigenous 15.42 t 9.6 Age" Under 20 16.47 t 3.5 21-30 19.01 t 9.0 31-45 12.26 t 6.5 46-60 0.68 0.4 Married 4.19 t 3.3 Male -2.47 1.3 Employed -1.06 0.4 Number of Hours -0.16 t 4.0 Regional Variablesb North 12.34 t 3.9 North-east -4.06 1.9 South-east 7.46 t 3.4 Central 1.49 0.7 South-west 4.91 t 2.8 North-west 18.32 t 6.6 Peten -2.87 0.7 Rural 5.90 t 4.3 Constant 0.164 N 8,956.0 Log Likelihood -3,684.7 Source: Encuesta Nacional Sociodemogr4fica, 1989. Note: All coefficients except constant represent marginal effects. t indicates statistical significance at the 95% level. Dependent variable Poor = (0,1) dummy, I if poor, 0 if otherwise. Household income per capita used in determining poverty - computed by taking total household income and dividing by total household members. Indigenous = (0,1) dummy, I if indigenous, O if otherwise; Married = (0,1) duxmmy, I if married, 0 if otherwise; Male =(0,1) dummy, I if male, 0 if female. ' Omitted age category is 61 and older. b Omitted regional category is Guatenala. (9) How the poor earn their income 1.19 Since open unemployment in Guatemala is not much higher than 1.9 percent, almost all the poor are in households with at least one worker. Table 1.7 presents the percent of the workers in a particular sector who are poor. Table 1.8 indicates the occupations of the head of the household in extremely poor, poor and nonpoor households. Approximately 60 percent of the poor receives their primary income from agriculture. The fact that 70 percent of the poor are in 9 rural areas suggests that a significant proportion of the poor in rural areas are working in nonagricultural pursuits. A 1978 study in the western highlands departments of Huehuetenango and El Quiche found that 63 percent of the income of farmers with less than 2 manzanas (1.4 hectares) came from off-farm sources, many of which were nonagricultural in nature. Similarly, a 1989 survey found that only 24 percent of the population in the western highlands is primarily employed in farming their own land. Another 27 percent rely on agricultural wage labor as their main source of income, and the remainder earn their living from commerce, artisanal production, and other nonagricultural activities (Smith, 1990). Annex 7 presents additional detail on the occupations and earnings of the indigenous and non-indigenous population, presented separately by sex and urban/rural areas. It also shows a relatively high percentage of children (10-18) who work (Annex 7, Table A). Table 1.7 HEADCOUNT POVERTY RATES By Economic Sector Industrial Sector Extremely Poor Poor Non-Poor Agriculture 71.8 85.5 14.5 Mining 33.6 69.0 31.0 Manufacturing 36.3 61.6 38.4 Utilities 11.5 32.6 67.4 Construction 35.0 61.8 38.2 Conumerce 28.0 50.0 50.0 Transportation 20.2 38.9 61.1 Financial Services 5.1 18.2 81.8 Social Services 19.9 38.8 61.2 Other 33.3 66.7 33.3 Source: Encuesta Nacional Sociodanografica. (10) What the poor consume 1.20 Some information on the consumption patterns of the poor and non-poor can be obtained from the 1979-81 National Income and Expenditure Survey, which is the latest survey to provide details on this topic. (Table 1.9). According to expenditure data, the extremely poor spend only approximately 66 percent of their total income on food, whereas the non-poor spend 47 percent. Thus, although the extremely poor are unable to purchase the ideal market -basket of food with their income, they do adjust their expenditures to enable them to continue to include a substantial portion of necessary non-food expenditures. As a result, they probably eat cheaper foods that are less interesting, less nutritious (either in terms of calories and proteins or essential vitamins and minerals). As Table 1.11 makes clear, maize plays a greater role in extremely poor as opposed to nonpoor family expenditures than do more nutritional foods, such as meats, fish, shellfish, and fruits. Extremely poor families spend nearly 50 percent more on maize than meats, fish, shellfish, and fruits, whereas nonpoor families spend four times more on meats, fish, shellfish, and fruits than on maize. The consumption patterns of the poor and extremely poor help to explain the high level of malnutrition among the children of Guatemala and the low birth weights (50 and 16 percent). As far as non-food expenditures are concerned, because the extremely poor generally 10 lack access to electricity they spend proportionately more on kerosene and firewood (9.9 percent) than do the nonpoor (2.0 percent). Table 1.8: SOURCES OF INCOME By Economic Sector Industral Sector Extremely Poor Poor Non-Poor Agriculture 70.9 60.9 20.9 Mining 0.1 0.2 0.2 Manufacturing 10.4 13.2 16.4 Utilities 0.1 0.3 0.9 Construction 3.0 3.8 5.0 Commerce 7.7 10.3 20.7 Transportation 1.1 1.5 4.7 Financial Services 0.2 0.4 3.5 Social Services 6.4 9.3 27.7 Source: Encueta Nacional Sociodemogrifica, 1989. Table 1.9: Distribution of Family Consumption Expenditures and Budget Share by Poverty Group Ext. Poor Poor Non-Poor Food 65.6 60.3 46.5 Maize 14.6 6.3 3.0 Beans 7.5 4.5 2.2 Meat, fish & shellfish 8.9 11.6 10.0 Fruits 1.2 1.8 2.0 Sugar 4.6 2.9 1.5 Milk 1.0 2.0 1.9 Non-Food 34.4 39.7 53.5 Clothes & shoes 8.6 10.7 10.5 Soap 2.7 2.0 1.2 Electricity 0.7 1.6 2.2 Water 0.3 0.4 0.4 Telephone 0.0 0.4 0.2 Kerosene 1.2 0.6 0.3 Firewood 8.7 4.2 1.7 Urban bus 0.2 0.6 0.6 Total 100.0 100.0 100.0 Source: Hichk, N. and H. Lee. "The Inpact of Govenmena Policies on the Poor in Gudemala.` Ihe World Bank. Draft. 1993. I1 B. Inequality 1.21 Another striking feature of Guatemala is the high degree of inequality in landholding, income, and consumption. The pattern of unequal distribution of land dates back to the colonial era, when the Spanish crown granted large extensions of land to colonizers in the form of encomiendas and realengas. The situation only grew worse after independence, because the liberal reforms of the 1870s expropriated church communal and municipal lands for coffee production. The first agricultural census in 1950 indicated that 47.6 percent of all landowners owned less than 1.4 hectares apiece. According to the most recent agricultural census (1979), inequality remains a serious problem: only 2.5 percent of Guatemala's 5.3 million farms control 65 percent of the agricultural land (the average farm consists of 200 hectares) whereas only 16 percent of the land is cultivated by 88 percent of the smallest farms (with an average farm size of 1.5 hectares). The Gini Index for land distribution was calculated to be 85.9 in 1979, which is higher in Guatemala than in all but two Latin American countries, pre-reformn Peru in 1961 and pre-reform Colombia in 1964. In rural areas individuals in lower-income brackets own smaller plots of land, but within each income decile the indigenous are more likely to have smaller landholdings (Annex 8). 1.22 The unequal distribution of land helps explain why Guatemala has the third highest degree of labor-income inequality (exceeded only by Brazil and Pakistan) among the 44 low- and middle- income countries mentioned earlier. The Gini coefficient for 1986/87 and 1989 was 0.579 and 0.595, respectively. A separate calculation of the Gini index by geographical location indicates inequality is greater in rural areas than in urban areas (Table 1.10). However, the national Gini coefficient is higher than either regional figure, which indicates that there is a substantial difference in the inequality between urban and rural areas. Similarly, when the Gini index is calculated on the basis of ethnicity in rural regions, indigenous people as an isolated group display relatively low levels of inequality, while nonindigenous people display relatively moderate income inequality. But when these results are combined, total inequality rises substantially because of differentials in the average income of indigenous and nonindigenous individuals. The figures below clearly reflect the pattern of extreme inequality in Guatemala during the 1980s: 12 Table 1.10: GINI INDEX OF INCOME INEQUALITY By Locatdon and Ethnicity Area Indigenous Non-Indigenous Overall Urban 0.523 0.510 0.533 (59.7) (153.2) (135.3) Rural 0.483 0.524 0.540 (26.7) (60.0) (44.7) National 0.521 0.561 0.596 (33.4) (104.0) (78.4) Encuesta Nacional Sociodemogrdfica, 1989. Note: Monthly eaniings in Quetzales appear in parenthesis. Table 1.11 Distribution of Income and Consumption Income Group 1980 1980 1986 1989 Income Consumption Income Incomes Bottom 60 percent 25.2 29.9 19.7 18.9 Decile 70-80 16.9 18.7 18.2 18.7 57.9 51.4 62.1 62.4 Source: Encuesta Nacional de Ingresos y Gastos Familiares, 1984. Psacharopoulos, et at Pwer Income Distribution in Latin America: The Story of the 1980s, 1992. Encuesta Nacional Sociodemogrdfica, 1989. " It is difficult to infer whether the declines in relative position of the poor from the beginning of the 1980s to 1986 and 1989 are real or reflect differences in methodologies across the different surveys. However, per capita agriculturl GDP did decline precipitously during the 1980s suggesting that the apparent decline could be real. 13 2. A POVERTY ALLEVIATION STRATEGY 2.1 Poverty and inequality were present long before Latin America fell victim to the "lost decade of the 1 980s" and before structural adjustment became necessary. Indeed, both were a source of a great deal of dissatisfaction among the population, which over the past 40 years has often erupted into armed conflict of varying intensity. This conflict has taken the lives of approximately 100,000 people and been responsible for the displacement of many more. Now, as negotiations are underway between the government and the Unidad Revolucionaria Nacional Guatemalteca (UNRG) to end the conflict, Guatemala must search for alternatives to violence and repression to deal with its unreleting poverty and inequality which in many respects, are as severe as they were at the start of the conflict. In terms of the absolute number of poor, the situation is actually much worse today because the population is so much higher (9.1 million in 1991) than it was in 1950 (2.8 million). 2.2 The World Bank's recommended strategy for poverty alleviation in Guatemala would pursue the following objectives: (a) Stability in the macroeconomic environment and economic management that makes efficient use of the poor's most abundant asset - labor. (b) Greater access to land for the poor. (c) An increase in tax revenue to finance public investments that would enhance the earning potential of the poor, including: (i) Investments to raise the human capital of the poor. (ii) Infrastructure and support services to increase the productivity of the existing physical and human capital of the poor. 2.3 The key elements of this strategy follow the general approach recommended in the 1990 World Development Report on poverty and the subsequent recommendations made in the World Bank Policy Paper on Assistance Strategies to reduce poverty. However, the emphasis on well- targeted transfers and social safety nets made in the Policy Paper on Assistance Strategies is less relevant for Guatemala, since poverty is so widespread there. Although the country needs some form of nutrition programs, for the near future the primary safety nets will have to continue to be provided through the family, community, and the support systems made available by the nongovernmental organizations (NGOS). The government should concentrate on financing investments that not only shown promise of increasing the earnings of the poor but also are not likely to be made by private individuals or firms. 2.4 The key elements of the strategy have been designed in part with an eye on the East Asian experience with generating growth and reducing poverty. Recent analysis of successful development in East Asia suggests that pursuing sound macroeconomic policies and promoting investments that increase the earnings potential of the poor can lead to both higher rates of economic growth and lower levels of inequality and poverty. In every one of the high-performing 14 Asian economies (HPAE), the Gini index declined between 1960 and 1980 while per capita GDP grew by more than 4 percent a year. The reductions in poverty are equally impressive. Between 1972 and 1982, for example, the percentage of people below the poverty line in Indonesia fell from 58 percent to 17 percent. 2.5 Thus the so-called inevitable tradeoff between growth and reducing inequality and poverty need not materialize. In fact, it appears that reducing inequality can lead to growth if policymakers attempt some form of "shared growth" of the kind promoted in the HPAE countries. The outcome of such efforts will greatly depend on what instruments are used to reduce inequality. As Birdsall, Ross and Sabot (1994) state: Using income transfers to reduce income inequality is unlikely to be good for growth: transfers often result in the diversion of scarce savings and investment to the subsidization of consumption; the targeted group is often not the one to benefit from transfers, reducing the effectiveness of transfers as a means of raising the standard of living, hence the savings and investment rates, of the poor; transfers tend to distort incentives and reduce both allocative and X-efficiency. 2.6 The instruments used in East Asia included land reform in Korea and Taiwan, public housing in Hong Kong and Singapore, investment in rural infrastructure in Indonesia, and, in all of these cases, high-quality basic education. (1) The Challenge in Guatemala 2.7 An enormous challenge awaits those who are preparing to tackle the problems of poverty and inequality in Guatemala. Some important advances have already been made in this direction. Unfortunately, considerably more needs to be done if Guatemala is to enjoy substantial increases in per capita growth and make significant headway against the rising tide of poverty. According to the latest World Bank projections, Guatemala could achieve a sustainable economic growth rate of 4.5 percent a year between 1995 and the year 2000. Because the high population is rapidly expanding, GDP per capita is expected to increase by only 1.7 percent a year. If these currently projected growth rates in GDP take place in a distributionally neutral fashion, poverty rates will remain much as they are. A simulation done as part of this study suggests that if the per capita income of every household in the 1989 Household Survey were to grow at a rate of 1.7 percent a year without interruption for 10 years, the poverty rate would decline from 75 percent to 69 percent. 2.8 Thus, it is vital to reach a consensus on a politically acceptable poverty alleviation strategy. The changes that are called for will undoubtedly prove to be unpopular in some parts of society. Those who have suffered because of the lack of investment in the past must be patient if the new advances do not appear to be moving quickly enough: they must maintain a pace that 15 will both achieve and sustain macroeconomic stability. Falling back into a period of macroeconomic instability will hurt the poor. At the same time, the strategy explicitly calls for an increase in tax revenues and a willingness to consider access to land. At current tax revenue levels, it is simply impossible to make a dent in poverty. 16 3. MACROECONOMIC PERFORMANCE AND THE EFFECTIVENESS OF ECONOMIC MANAGEMENT 3.1 This chapter examines Guatemala's recent macroeconomic performance and explains the importance of achieving growth through the productive use of labor. Today, the incentive structure favors labor-using growth more than it did in the 1970s or 1980s. However, declining tax revenues threaten both macroeconomic stability and the government's ability to mount an effective poverty alleviation program. (1) Recent macroeconomic performance 3.2 During the 1960s and 1970s, Guatemala enjoyed macroeconomic stability and sustained economic growth. GDP growth rates averaged 5.7 percent a year and inflation was contained below 9 percent. 3.3 This period of relative prosperity came to an end in the late 1970s when Guatemala suffered a series of external shocks. With the oil price hikes of 1979 and the decline in coffee and other commodity prices, its terms of trade (TOT) rapidly deteriorated. Export demand also declined as a result of the 1981-82 recession in the OECD countries. The situation was further aggravated by the outbreak of armed conflict in the region, which reduced the level of CACM trade by over 50 percent between 1980 and 1985. Consequently, exports as a percentage of GDP dropped from 22.2 percent in 1980 to about 13 percent in 1984. By 1984, imports as a percentage of GDP had declined to 15.5 percent, after reaching about 25 percent in 1980. As a result of these demand and supply shocks, economic activity was strongly depressed and growth contracted by an average of 0.9 percent (in real terms) during 1983-85 (Annex 9). Real agricultural GDP per capita declined by some 14 percent (Figure 3.1) during 1980-85. Figure 3.1: Real Agriculture GDP per Capita 115 110 105 a 95 - o 90 u ........ . . . a ............ .... m a ... Sources: Bank of Guatemala, SEGELAN 17 3.4 Toward the end of the 1980s, the declines in income halted. The Cerezo administration (1986-1991), the first democratically elected government in many years, stabilized the economy and reformed the exchange rate and trade policies. These reforms stimulated economic growth and diversified economic activity. By 1988, however, the government's adjustment program had begun to falter and by 1990 the Cerezo administration come to and end, leaving behind serious macroeconomic imbalances. One of the main reasons for the failure of the Cerezo structural adjustment program was the lack of any serious fiscal adjustment. Central Government (CG) current expenditures in 1989 had increased by 40 percent in real terms over their 1985 levels (pre- Cerezo) while CG investment had increased by 61 percent. However, revenues failed to keep pace with these expenditure increases, peaking in 1988 and declining thereafter. 3.5 The Serrano government that took office in 1991 immediately had to face a fiscal and balance of payments crisis. During its first two years, the government effectively stabilized the economy using expenditure reduction policies. The net financial deficit of the public sector was reduced from 2.6 percent of GDP in 1990 to a surplus of 0.4 percent in 1991 and 1992. It is important to note that the fiscal adjustment was achieved through freezes and cuts in current expenditures and not by investment cuts. Money and credit policies were significantly tightened in 1991 in order to combat the surge in inflation generated by the lax policies of the previous administration. Cumulative inflation was drastically reduced from 60 percent (Dec. to Dec.) in 1990 to 10 percent by the end of 1991, before increasing to 14 percent in 1992. The reemergence of high real interest rates along with increased public confidence stimulated a significant inflow of foreign exchange. Economic growth was not interrupted during this period of stabilization, averaging 3.8 percent and 4.8 percent during 1991 and 1992 (Annex 9). A loosening of credit policy in mid-1992, however, in conjunction with an appreciating real exchange rate allowed for a substantial widening of the current account deficit to 7.6 percent of GDP compared with 2.7 in 1991. 3.6 In May 1993, the President suspended portions of the Constitution and dissolved Congress and the Supreme Court. Strong opposition emerged to this action and he was forced from office. The political crisis immediately affected macroeconomic performance. Prices jumped 5 percent in the five weeks immediately after the crisis. Net international reserves fell by $ 80 million and fiscal revenues fell. 3.7 The macroeconomic situation was stabilized during the latter part of 1993. Inflation was 11.8 percent; international reserves increased by $180 million and real GDP growth reached 4 percent. However, the Government has been unable to reverse the deterioration in the fiscal accounts during the course of the year. Tax revenue collection declined to about 8 percent of GDP in 1993 compared with 8.6 percent in 1992 and the net fiscal public sector deficit went from surplus to a deficit of 1.6 percent of GDP. This poor fiscal performance has required the Bank of Guatemala to adopt a very tight monetary stance, resulting in high real interest rates (above 10 percent). 18 (2) Effectiveness of economic management 3.8 The effectiveness of economic management in Guatemala in recent years can be judged by whether the government (a) promoted growth and stability, (b) favored the productive use of labor, and (c) provided the fiscal support for public social sector expenditures. (3) Did economic management promote growth and stability? 3.9 For the most part, Guatemalan authorities have had considerable difficulty managing the economy since the early 1980s. The economy grew in spurts in this period, with some short periods of effective stabilization, interspersed with costly lapses that had a negative impact on both growth and the poor. 3.10 The government's exchange and trade policies, for example, proved ineffective in responding to the shocks of the early 1980s. The government could have improved the profitability of agriculture exporters (and thus saved some jobs) during 1980-83 by devaluing the quetzal. Also, by relying on import and foreign exchange restrictions, it imposed severe limitations on the use of intermediate imported inputs in the production of exports and nontradables. The resulting reduction of aggregate supply further depressed urban labor demand. When the government's decided to devalue the quetzal in 1984-85, employment began to grow again in both sectors. Thus the devaluation appears to have had a positive impact on both the rural and urban poor. 3.11 Another problem for the government at this time was that the 1987 tax reform failed to increase revenues. Unfortunately, it only made matters worse when it decided to allow the net financial public sector deficit to rise steadily--from 0.7 percent in 1986 to 3.1 percent in 1989. The unsustainability of these deficits was reflected in the large borrowing requirements imposed on the Bank of Guatemala (BOG), which weakened its ability to control base money growth. Subsequently, with the floating of the exchange rate in 1989, the BOG failed to counter actual and anticipated inflationary pressures stemming from the devaluation, with the result that real interest rates became increasingly negative. This put pressure on the exchange rate, and sucked the country into a devaluation-inflation spiral that saw inflation jump to 60 percent by 1990. 3.12 Yet another failing of the country's economic managers lay in their monetary policies, which contributed to the two large real wage reductions that occurred in 1985/86 and 1990/91. Between 1989 and 1990, real wages in agriculture and formal manufacturing declined by 27 percent and 23 percent, respectively. These real wage reductions were caused by the high inflation of 1985/86 and 1990/91. In both cases, the BOG failed to tighten monetary policy in the wake of large quetzal devaluations. 19 (4) Has economic management favored the productive use of labor? 3.13 By and large, Guatemala's economic managers have favored the productive use of labor, but only over the past five years. During the 1960s and 1970s, the incentive structure, as represented by tariffs, taxes, and the exchange rate, was biased in favor of a production structure that was capital-intensive and highly import dependent. 3.14 In the mid-1980s, the incentive structure began to change, mainly as a result of exchange rate reform and trade liberalization. Government expenditure switching and reduction policies induced a change in the structure of relative product prices such that the real exchange rate depreciated by about 40 percent over the period 1986-90 (Table 3.1). The was an important trend because it implied that over a sustained period of time (five years) the structure of prices favored the production of exports and efficient import substitutes in relation to nontradables. The tariff reform initiatives of 1986 and 1990 had the effect of reducing the average level of nominal protection and compressing the dispersion rates, thus reducing somewhat the inherent production distortions. The tariff structure, however, was still biased in favor of final consumer goods. TABLE 3.1: REAL EFFECTIVE AND NOMINAL EXCHANGE RATES (Quetzals per US$, REER Index 1980= 100) Nominal REER % Change Rate % Change 1980 100.0 1.00 1981 91.2 -8.8 1.00 0.0 1982 88.4 -3.1 1.00 0.0 1983 84.2 -4.8 1.00 0.0 1984 83.6 -0.8 1.02 2.0 1985 114.7 37.3 1.67 63.7 1986 117.7 2.6 2.11 26.3 1987 122.2 3.8 2.23 5.7 1988 132.1 8.1 2.49 11.7 1989 137.3 4.0 2.83 13.7 1990 165.2 20.3 4.50 59.0 1991 141.2 -14.5 5.03 11.8 1992 140.3 -0.6 5.18 3.0 1993 136.6 -2.6 5.64 8.9 1994 130.9 -4.2 5.78 2.5 Notes: A negative sign denotes appreciation. 1994 REER is the average January-September index. Source: International Monetary Fund. 20 3.15 There is no direct evidence on income or consumption levels to indicate what current levels of poverty are after the changes in the macroeconomic environment.7' However, indirect evidence suggests that over the last five years the changes have led to a more favorable environment for the poor. The decline in real agricultural GDP per capita has been arrested, nontraditional agricultural exports (primarily fruits and vegetables) are on the rise, and nontraditional manufacture exports, especially in the maquila sector, are also expanding. These sectors are heavily labor-intensive. 3.16 Smallholder producers have diversified production into "nontraditional" agriculture exports (primarily fruits and vegetables). Between 1986 and 1989, the value of nontraditional agricultural exports grew by 16 percent per year. The significance of this trend is that it provides an alternative source of income for the rural poor (especially in light of the weak basic grains market). Domestic producers of basic grains continued to be adversely affected by the flat, low level of international basic grain prices, which resulted in weak domestic grain prices. In addition, the rural poor benefited from increased employment opportunities during 1986-90 (about a 16 percent increase) as the renewed profitability of commercial farms increased labor demand. Unfortunately, international prices, especially for coffee, have been falling recently, and therefore the demand for labor has begun to contract. 3.17 The real income of the urban poor in Guatemala has been greatly affected by the increased competitiveness in manufactured exports, the ongoing trade liberalization process, and the resumption of economic growth. All of these factors put strong pressure on labor demand in the manufacturing sector, as reflected by a 32 percent increase in employment opportunities during 1986-90. The more labor-intensive, nontraditional export manufacturing sector has grown tremendously since the initiation of the exchange and trade reforms. The emerging maquila sector is representative of this trend. In 1986, 41 firms in garment production, the principal maquila activity, employed 5,120 workers; by 1989, there were 189 firms employing 35,000 workers. The effects of the exchange and trade reforms are also reflected in the 24 percent growth in nontraditional manufacture exports outside the CACM between 1986 and 1989. Manufactured exports in total merchandise exports increased from 22 percent to 31 percent. 3.18 Despite the changes in the macroeconomic environment, wages have not yet risen to the point where a significant number of families could be lifted out of poverty. Agricultural daily wage rates for the 1993/1994 coffee harvest period, a period of peak labor demand, were between $2 and $3. This was sufficient to generate a temporary internal labor migration of approximately 400,000 workers, which suggests that wage rates and productivity in the areas of outmigration are even lower. A recent report on the Maquila sector suggests that the average wage is around $3 a day, but with substantial variation. These continued low wages give some idea of just how difficult it will be to alleviate poverty in Guatemala and are consistent with the earlier reported simulation that 10 years of uninterrupted per capita growth in GDP at 1.7 percent would have little effect on poverty. 71 The last household survey was conducted in 1989. 21 (5) Has economic management generated the fiscal space for publicly funded social sector e-xpenditures? 3.19 Management proved unable to fund its social expenditures as well. This was mainly due to the country's low and variable tax revenue. Over the past 10 years, Guatemala's tax revenue as a percentage of GDP has averaged 8.8 percent, which is at least a standard deviation less than the mean in Latin America over the same period, namely, 19.6 percent. In addition, the tenuous state of fiscal finances has resulted in a series of "boom and bust" fiscal crises. When a financing crisis occurs and a fiscal adjustment is required, as happened in the early 1980s and again in 1989- 90, social sector investment expenditures have usually been the prime targets for cuts. 3.20 As a result, the level of social sector and infrastructure investments fell sharply during the 1980-92 period (see Figure 3.2). By 1992, the cumulative effect of these cuts was to push real health investment levels down to 80 percent below their 1980 levels. Real education investment levels were some 33 percent below their 1980 levels. Infrastructure investment has also declined considerably, particularly agriculture and transportation, which were 70 percent and 40 percent below their 1980 levels, respectively. Reductions in health and education services have adversely affected human capital formation, while a reduction in physical infrastructure has constrained the viability and productivity of producers and thus, indirectly affected labor demand. Figure 3.2: Real Health and Education Public Expenditures 140 120 .. .... .... . 8 - 2 2 m5 W3 EB Wi 8 | 8 gB i g jfi~~~~~-- - 5 80 -- Inv Fleath 10 '~~~~~ 60 -a--~~~~~~~~~~~[ Inv Educat ---Curr Hesalth !40 X-. *. Curr Eduicat Sources: Ministry of Finance, State's Accounting Office. 3.21 More recently, there has been an attempt to bring social sector investments up to their previous levels. Since 1990, health and education investment expenditures have increased in real terms by 5 percent and 64 percent respectively. Health and education current expenditures have increased by 13 percent and 30 percent respectively over the same period. However, this goal is once again threatened by declining tax revenues. 3.22 Additional fiscal space for social sector expenditures could be gained by eliminating the implicit electricity subsidy. This would make it necessary to increases electricity tariff rates. The net welfare impact of these tariff hikes on the poor is believed to be small given that only 38 percent of the poor have access to electricity and they consume only 31 percent of electrical 22 services. Moreover, if the original intent of this subsidy was to benefit the poor, then, given these incidence figures, this type of subsidy is clearly inefficient. A more efficient allocation of resources could be achieved by eliminating the subsidy and reallocating the savings generated to areas that are known to have a much more direct impact on the poor's welfare, such as health posts. (6) Conclusions 3.23 Like other countries of the world, Guatemala has discovered that macroeconomic stability and growth are necessary but not sufficient conditions for reducing poverty (best illustrated by Guatemala's experience throughout the 1960s and 1970s). Macroeconomic instability is not only bad for growth, but hurts the welfare of the poor (as it learned during a large part of the 1980s). Even macroeconomic stability and an appropriate incentive structure may not be sufficient to reduce poverty at a socially acceptable rate if the poor do not have the ability to respond to the incentive structure (as appears to have been the case since the end of the 1980s). Structuring public investment so that it enhances the earnings potential of the poor would help increase the returns to the poor's assets brought about by changes in relative prices and incentive structures within the macroeconomic environment. 3.24 In view of Guatemala's recent experience, a wise policy course would be to (a) continue with a liberalized economy, (b) avoid appreciations in the real exchange rate, and (c) increase the level of public savings. Trade liberalization is essential in order to remove the production disincentives and resource misallocations that the old import substitution tariff structure engendered. Trade liberalization would reduce general poverty by restoring incentives in favor of exports and by promoting labor-intensive industrialization. The shift in output toward exports and efficient import substitutes should promote a more labor-intensive pattern of production than in the past. Furthermore, urban and industrial growth have important backward linkages with the rural sector through the opportunities they create for migration from rural areas. The long-term effect of this rural urban migration is that it tightens rural labor markets which benefits the rural poor. 3.25 Despite the structural reforms in 1990, the quetzal appreciated in real terms by about 15 percent in 1991 and 3 percent in 1992. This signaled a loss in competitiveness of tradables in relation to nontradables and indicates how easy it is to slip back into exchange rate overvaluation. Several factors contributed to this appreciation in the real exchange rate: (a) high inflation that was not effectively brought down until the end of 1991; (b) inadequate adjustment in the exchange rate in the face of large private capital inflows; and (c) the continuing decline of international commodity prices, especially that of coffee (the international price of coffee declined by 22 percent over the 1990-92 period to its lowest level in over fifteen years). 3.26 In the past, the government has not been able to establish or maintain an adequate level of tax revenues. By establishing a sound fiscal position, it would have the capacity to accomplish two goals that are essential for poverty alleviation. First, it would allow the government to increase expenditures to social areas of higher priority. Increased public savings levels would 23 sustain increased expenditure levels to the social sectors over the long term. Secondly, it would remove the principal source of macro instability during the 1980s: unsustainable fiscal deficits. By establishing the anchor for a stable macroeconomic framework, it would create an environment conducive for economic growth and investment and also free of many of the welfare- reducing forces characteristic of macroeconomic instability, such as inflation. 24 4. INCREASING THE POOR'S ACCESS TO LAND 4.1 Guatemala is the only country in Central America that has not implemented at least a nominal reform of its land tenure system. Thus land ownership continues to be not only inequitable, but also most probably the cause of lower rates of utilization of labor and efficiency losses for the economy. (1) A more equitable distribution would employ more labor 4.2 In Guatemala, the per hectare labor use in small farms (less than 3.5 hectares) is on average almost 40 percent higher than on large farms (SEGEPLAN, 1987). However, because so many more hectares are under the control of large farms, large farms account for 74 percent of annual agricultural labor requirements. Large farms, despite the vast expansion of land under their control, have decreased their average labor requirements per hectare owing to the introduction of livestock, mechanization and less labor intensive crops (e.g. corn, fruit, sesame and sorghum).7' This decline in average labor use in large farms, combined with the initial inequitable distribution and decades of high population growth has forced rural residents to subdivide farms. As a result, peasants owning farms less than 1.4 hectares of land have seen the average size of their farms decline from .77 hectares in 1950 to .49 hectares in 1979. 4.3 A 1978 study in the Western Highlands departments of Huehuetenango and El Quiche found that on average, 63 percent of the income of farmers with less than 1.4 hectares of land came from off-farm sources, particularly nonagricultural sources (IIESO-USAC 1980). As indicated by a 1989 survey, only 24 percent of the Western Highlanders are primarily employed in farming on their own farm. Another 27 percent rely on agricultural wage labor as their main source of income and the remainder earn their living from commerce, artisanal production, and other non-agricultural activities (Smith 1990). (2) A more equitable distribution would increase efficiency 4.4 Mounting international evidence suggests that small family farms enjoy an efficiency advantage over large farms operated with hired labor. The World Bank's 1975 Land Reform Policy Paper and the empirical evidence that has been accumulated since that report strongly suggests that: (a) small farms have higher yields and higher total factor productivity than large farms; and (b) small farms use more labor and less capital per hectare cultivated than large farms. In addition, the evidence shows that small farmers' propensity to save and the productivity of the investment undertaken was at least as high as that by large farmers or government (Bevan 1993). 71 Guatemala's pattern was shared by Brazil. From 1965-1985, productivity grew at 5 percent a year due to rapid subsidized mechanization of large farms, but agricultural employment growth was less than I percent a year, far below the growth rate in the labor force. 25 4.5 What makes owner-operated small farms more efficient than large farms operated by wage labor is that the latter must spend valuable resources on supervising hired workers. Large farms, however, usually have better access to credit.8 In addition, for a few plantation crops such as bananas and sugar the scale economies associated with the processing plant are transmitted to the farm because harvesting and processing of the crops must be closely coordinated. For all other crops, the supervision cost advantage of small farmers usually exceeds the credit cost advantage of large farms, so that the economies of scale do not transcend the size of farm that can be operated by a family. 4.6 Despite efficiency advantages of small farms, there appears to be no tendency for large landowners to divide their holdings to sell to more efficient small farmers in Guatemala (Stringer and Lambert 1989). The market for land is divided between a submarket for wealthy landowners and another for sellers and buyers of smallholdings. 4.7 Such sales may not be occurring for a number of reasons. One is of a political nature: large landowners do not want to sell to small landholders because they fear that such actions could increase demands for land redistribution and land invasion. Other reasons are economic. If the market price of land exceeds the present discounted value of agricultural profits at the market interest rate, poor people without access to grants or own equity cannot buy land and repay the mortgage out of the net income generated by farm profits, even if they have access to credit at the discount market rate used to discount the future profits. Land prices may exceed the present discounted value of agricultural profits because land confers a "collateral value" of facilitating access to credit. Other reasons why land prices may exceed the present discounted value of agricultural profits include: a) preferential income tax treatment of agricultural incomes; b) macroeconomic instability which converts land into a hedge against inflation; and c) use of land as a speculative asset that is expected to increase in value near urban areas and in the presence of population growth. 4.8 While the Instituto Nacional de Transformaci6n Agraria (INTA) and several small, donor-sponsored experiments provide land to the landless and land-poor, both means are problematic. INTA is the Government agency mandated with the transformation of land tenure in Guatemala, primarily through the settlement of landless families on national or privately purchased lands. Thirty years of INTA programs have created 852 settlements with 118,663 beneficiaries, occupying 925,473 hectares of land. However, INTA farmer have collective titles until the farm is paid off. This arrangement has been a disincentive to farmers to improve and manage the land better. Furthermore, INTA lands frequently are far from roads which makes transportation difficult. INTA currently administers a relatively large fund (FONATIERRA) established for the purpose of purchasing lands for their eventual redistribution to smallholders. Among other problems, this has created substantial price speculation in the areas where INTA has expressed interest in buying lands. 8/ This is true in Guatemala. Only 16 percent of bank agricultural credit went towards basic grains production, the principal activity of most small holders, compared to the 41 percent that went for predominately large farmer export crops like coffee, cotton and sugarcane (AVANSCO, 1993). 26 4.9 The small donor-sponsored initiatives suffer from a paternalistic approach to beneficiary organization, decisionmaking, crop selection, marketing, credit, technical assistance and land titling. In an attempt to help small farmers, for example, USAID between 1984 and 1990 financed the Fundaci6n del Centavo's purchase and resale of 28 farms divided into approximately 1,400 parcels. However, this program did not provide titles to individual farmers and dictated what the farmers planted and how they should market their produce.9' Thus, the farmers in many ways had traded one landlord for another and had not developed autonomy. 4.10 The land registry process, too, makes it difficult for the poor to gain access to land. This system is antiquated, inefficient, overly centralized, complicated, until quite recently was run by corrupt officials. Furthermore, registration requires buyers and sellers to take costly actions, and in some cases they are required to be able to read and write. Most land transactions in the Highlands are not legally registered because of the high costs and inaccessibility of the land registry system. This leaves smallholders vulnerable to usurpation and acts as a barrier to their access to credit. To transfer land, a buyer and seller must reach an agreement and pay a notary or lawyer for a title-deed (Escritura Putblica). To guarantee security of ownership rights, Guatemala adheres to the principle of publicity by having a title-deed registered in the Registro General de Propiedades, with offices in Guatemala City and Quetzaltenango. If a land title holder does not properly register the title in the Registry, his or her title cannot be guaranteed. To act in good faith, a buyer is obliged to investigate the current legal status of the parcel in the Registry before purchasing. All transactions are done by hand and involve significant legal and travel costs for smallholders wishing to register their property. Moreover, there is no mechanism in the system for the provision of a field sketch (croquis) of the property to establish its characteristics and boundaries, leaving open the possibility of disputes and rendering the system useless for natural resource management purposes. The dysfunctionality of the registry has become especially apparent in recent years with the return of the refugees from Mexico, for whom available land must be identified and purchased for their resettlement. 4.11 Land ownership procedures are also biased against women. For example, in some parts of the highlands, women can receive land as an inheritance from their parents or when they are widowed. Generally, when a woman does bring land into a household, control over that land tends to revert to the male head in terms of decisions over what to plant, control over income generated, rental, sale and inheritance.'0' This pattern may vary considerably over ethnic group, degree of local land pressure, farm size and a series of other factors, but it does raise the important issue of potentially conflicting claims over property rights within the household. Women have not been beneficiaries of public sector land programs. There has never been a case of a woman receiving land in her own name under the Agrarian Transformation Law (Ley de Transformaci6n Agricola), and the 10 percent of public sector land redistribution beneficiaries who are women are a result of either widowhood or inheritance. With regard to communal lands, where these are managed locally it is generally a male mayor or committee that makes the decisions regarding the use of these lands, despite the fact that in many cases these lands serve as 9' ibid. IO/ Katz, Elizabeth G. 1992. "Intra-household Resource Allocation in the Guatemalan Central Highlands: The Impact of Non- traditional Agricultural Exports." Ph.D. Dissertation, University of Wisconsin - Madison 27 important sources of water and/or firewood, the collection of which is one of women's labor responsibilities. The establishment of communal firewood forests (bosques bioenergeticos comunales) has been identified as an intervention which would have positive impacts on both women's time utilization and on deforestation. U/ 4.12 In general, land law and its enforcement in Guatemala favors wealthy largeholders. For example, the squatter's rights law (Ley de Titulaci6n Supletoria), which authorizes titling of land to individuals or communities who have occupied and utilized for agricultural purposes an area for a period of 10 years or more, is not well enforced and is limited to farms of 46 hectares or less. This effectively protects the largest landowners from transferring their idle lands. (3) Conclusions 4.13 A more equitable and efficient distribution of land cannot take place in Guatemala unless all parties concerned endeavor to: a) increase access to land for the landless and land poor; and (b) strengthen land tenure security for smallholders who already possess land. 4.14 In countries such as Nicaragua, Mexico and Honduras, where major policy issues have been resolved, one of the secrets to their success has been a process of consensus-building among all the social groups involved -- including those who were previously excluded. Furthermore, a more fruitful dialogue emerged when the key policy questions were faced before more technical issues (for example, those related to titling). 4.15 The experience of other countries can yield many lessons. One of the first points to note is that if land is purchased for resale to the poor, policies that reduce the excess of land prices over the discounted value of agricultural profits will decrease market values of land and therefore the cost of the resale program. Second, a partial grant element has to be involved in the selling to the poor. Means-testing of such grants would improve equity and also limit the cost of a land resale program. Third, experience with centralized government land reform institutions have been universally disappointing. Successful land reforms in Taiwan, Korea and Japan were managed in a decentralized fashion by temporary district land committees without permanent staff. This is a mechanism proposed for South Africa with decentralized supervision of market assisted land reform by district land committees and a decentralized judicial system to which disputes can be appealed. Finally, historical studies and recent experience suggest that titling becomes most cost- effective and contributes to equity objectives when it is applied throughout a given area and is combined with a system of decentralized or even mobile courts through which all disputes can be l/ Vargas, Femando. 1992. Consultaria Sobre la Muier v la Producci6n de Alimentos: Tecnoloefa v Comercializaci6n. Instituto de Cooperaci6n para la Agricultura (IICA). 28 adjudicated as part of the area-based titling process. Selective titling on demand has generally increased land ownership inequality. I 29 5. INVESTING IN HUMAN CAPITAL 5.1 In Guatemala, households and the government alike have failed to invest enough in education, health, and nutrition, with adverse consequences for labor productivity and income growth. This is apparent from the education, health, and nutrition indicators. Because poor households rely on their children to generate income and are forced to devote almost all of their limited income to consumption, they tend to underinvest in the human capital of their children, even if such investments have a high social return. The government has done little to check this tendency. For many years, the public sector has also neglected to put enough effort into building the infrastructure so as to reduce the household's cost of investing in health, education, and nutrition. To make matters worse, ministries have by and large been inefficient and supported an inequitable and inefficient allocation of public expenditures, which has been biased toward urban areas and against the poor. 5.2 To alleviate poverty, Guatemala clearly needs to publicly finance higher levels of human capital investments. However, the performance of the public sector over the past twenty years in Guatemala and elsewhere does not provide a compelling case for providing more health and education services in the same way as they have been provided in the past. Guatemala should therefore give serious consideration to introducing incentives that make providers more accountable for the quality of the services they provide. (1) Education, health and nutrition indicators are deficient 5.3 Judged by almost any indicator, national education, health, and nutrition indicators are far below desirable levels.'2' Literacy rates, for example, are 36.9 percent for males and 52.9 percent for females. Gross enrollment rates are also low: 77 percent for primary education, 21 percent for secondary education, and 9 percent for tertiary education. The infant mortality rate is 55 per 1,000 live births, and the maternal mortality rate 110 per 100,000 live births. In addition, approximately 16 percent of infants suffer from low birth weight, and approximately 50 percent of children are malnourished. Water-related contagious diseases are the principal cause of death in adults over 45 years of age, and the third largest cause of death in infants under 5 years of age. Since its arrival in 1991, cholera, a clear indicator of inadequate sanitary conditions, has afflicted over 20,000 people, and the number is still increasing. 5.4 The country's poor water and sanitation conditions are directly related to the high incidence of diarrheal and infectious diseases, which are the most frequent causes of mortality in Guatemala. Only 64 percent of the country's inhabitants have access to running water, and availability in relation to poverty groups ranges from 81 percent for the non-poor to 57 percent for the poor (Table 5.1). Water coverage is particularly poor in rural areas: rural coverage in Peten and the North is only 15 percent and 3 5 percent, respectively. 12/ Annex 9 compar Guatenala's health uad education indicators to other latin Amencan countries. 30 Table 5.1: WATER & SANITATION AND ELECTRICITY COVERAGE, 1989 (%/6) Region Water EP P NP Urban Rural Total Metropolitan 56.7 66.8 89.1 87.9 46.5 78.2 North 38.7 41.9 75.9 89.7 35.0 45.6 North-East 55.8 58.9 79.8 87.9 59.6 66.5 South-East 53.6 58.1 70.6 87.4 53.6 61.3 Central 67.2 67.9 73.5 83.4 57.8 69.7 South-West 57.4 59.0 80.2 91.7 53.9 63.8 North-West 45.4 46.3 70.9 74.6 42.8 48.4 Petdn 23.1 30.0 52.2 83.4 15.3 36.9 Nationwide 52.8 57.0 81.0 87.2 49.8 64.0 Region Sanitation EP P NP Urban Rural Total Metropolitan 36.6 49.0 83.1 83.0 12.5 66.4 North 4.7 6.1 48.2 47.3 2.0 10.7 North-East 9,1 15.6 52.9 60.1 19.1 29.1 South-East 11.0 15.1 38.7 58.2 10.2 21.1 Central 23.3 26.4 51.5 57.4 14.1 34.1 South-West 10.6 14.3 54.8 69.6 7.1 23.6 North-West 2.7 3.9 47.3 36.0 1.5 7.5 Peten 2.3 3.9 19.0 23.5 1.7 8.6 Nationwide 12.4 18.8 62.9 69.4 8.5 31.7 Region Electricity EP P NP Urban Rural Total Metropolitan 55.4 68.5 92.9 95.0 35.2 80.9 North 6.8 10.0 62.7 74.0 1.8 15.8 North-East 32.7 41.6 68.5 86.1 40.1 51.3 South-East 41.6 47.2 72.1 86.3 43.9 53.6 Central 53.9 56.5 69.9 81.4 42.8 60.6 South-West 26.7 32.0 74.6 86.0 25.9 41.8 North-West 12.2 13.8 58.2 46.5 11.4 17.5 Peten 12.1 16.2 42.4 68.4 4.0 24.4 Nationwide 29.6 37.7 78.8 86.6 27.1 49.8 Source: INE, Encuesta Nacional Sociodemogrdflca, 1989. Notes: EP = Extremely Poor, P = Poor, NP = Non-Poor Water services include three source categories: exclusive, shared, and public. Sanitation services include private and shared services and exclusive toilets 31 5.5 Only 32 percent of the national population has access to sanitation services. The coverage breaks down to only 19 percent for the poor versus 69 percent for the non-poor. As with water services, there are significant differences along urban and rural lines; rural coverage is only 9 percent, as opposed to 70 percent in urban areas Table 5.2: Regional Variation in Health Indicators and Poverty Region infant Malnutrition2 Maternal Measles Poverty Mortality' Mortality3 Vaccination4 Headcount* (1988) (1986-87) % (1989) (1991) % Guatemala 51 29 7 52 54.2 North 62 38 19 42 91.3 Northeast 53 26 9 49 69.3 Southeast 66 34 10 45 79.7 Central 69 32 11 45 76.0 Southwest 72 39 11 29 82.7 Northwest 56 44 23 47 93.7 Pet6n 57 23 23 34 75.4 Notes: 1 Deaths per 1,000 live births. 2 Global measure, 0-36 months of age. I Deaths per 10,000. 4Infants under 1 year of age. Source: UNICEF, Andiisti de Situaci6n delNiRo v laMLer. Guatemala, 1991. Vaccination information from Oficina Panamericana de la Salud, An6lisis de la Situaci6n de Salud por Re2iones. Guatemala, 1992. Poverty headcount fiom Encuesta Nacional Sociodemografica, 1989. 5.6 The situation is far worse in the poorer regions and among the indigenous population (see Table 5.2). Illiteracy rates, for example, vary by as much as 43 percentage points from one region to another (Table 5.3). Overall, 60 percent of the indigenous people are illiterate compared with 24 percent of non-indigenous people. Health problems, too, are compounded in the rural regions. Thus it is not surprising that malnutrition and maternal mortality are considerably higher in the poorer Northwest region than in the Metropolitan region (Table 5.4) 32 Table 5.3: REGIONAL VARIATIONS IN ENROLLMENT AND POVERTY Region Preschool Primary Secondary Poverty Headcount gross net gross net gross net Metropolitan 57.8 39.5 103.9 88.9 51.2 34.7 54.2 North 43.0 10.7 53.8 41.1 7.3 4.0 91.3 Northeast 15.1 9.7 86.9 71.2 18.8 11.6 69.3 Southeast 10.7 8.0 92.9 78.5 14.0 9.4 79.7 Central 21.2 15.6 86.2 72.4 18.4 11.6 76.0 Southwest 27.0 13.7 85.1 69.3 15.8 9.8 82.7 Northwest 31.7 9.6 57.5 46.5 6.8 4.2 93.7 Petdn 17.9 8.2 77.1 56.5 11.6 6.7 75.4 Source: MinisteriodeEducacion,AnuarjoEstadistco. 1992. Notes: Preschool 5-6 years of age; primary 7-12 years; secondary 13-19 years. Table 5.4: ILLITERACY BY REGION (%) Region Male Female Total Metro. 12 22 18 North 51 71 61 Northeast 29 38 34 Southeast 27 39 33 Central 21 40 30 Southwest 31 52 41 Northwest 49 68 59 Peten 27 38 32 Source: ENS 1989. (2) Gusatemala spends too little on health and education 5.7 Health expenditures account for little more than I percent of GDP and expenditures on education for 2--3 percent (but have often fallen to less than 2 percent). In contrast, in Latin America and the Caribbean region as a whole, education expenditures over the 1980s averaged between 3 and 4 percent of GDP and health expenditures have averaged close to 3 percent of GDP. Because Guatamala's health and education expenditures as a percentage of total government spending are similar to those of other Latin American countries, its lower totals are clearly due to the much lower taxes raised as a percentage of GDP. 33 5.8 The magnitude of the government's expenditure shortfall (as well as an indication resource misallocation) is suggested by a comparison of the cost of providing a basic package of health services with actual per capita expenditures. The estimated cost in Guatemala of providing the type of basic health services recommended in the 1993 World Development Report is about US$ 8 per capita, or approximately US$ 82 million. This basic package includes prenatal and delivery care, management of the sick child, immunizations, and mass chemotherapy against helminthic infections. Such a package would be adequate to address about 80 percent of the burden of disease born by children in countries at similar levels of development as Guatemala. In 1992, Guatemala's Ministry of Health (MOH) spent a total of approximately US$8.11 per capita on health services, of which approximately 20 percent was devoted to primary health care."'3 This implies that per capita expenditure on primary health care in 1992 was only US$1.62 or US$ 16.6 million. This is approximately US$66 million short of the amount required to finance the basic package. 5.9 Similar shortfalls (and evidence of misallocation) are apparent in education. High estimated private and social rates of return for primary education and secondary education (34 and 20 percent, and 18 and 18 percent, respectively) suggest that high-return investments are being missed. It is unlikely that the opportunity cost of capital is much higher than this. 5.10 International evidence suggests that, for a country at Guatemala's level of development, the rates of return to education do not decline until primary school enrollment rates reach approximately 110 percent and secondary enrollment rates reach 50 percent.'4' Thus, expanding the education system until these rates are reached over a period of about ten years represents a good investment. The successful countries in East Asia reached these primary and secondary enrollment rates of about 100 and 50 percent in the early and mid-1960s. 5.11 To meet these enrollment targets, Guatemala would have to increase its expenditures in this area by 11 percent per year in real terms for the next eight years. This estimate is based on an assumption, perhaps optimistic, that the share of university, sports, and culture could be reduced by five percentage points and that the cost of enrolling an additional child would be equal to the current per student costs. An increase in the education budget of this magnitude would increase education's share of GDP by about 3 to 3.5 percent, depending on assumptions about GDP '3' World Bank, -Guatemala Social Sector Expenditure Review." White Cover, 1991. '4, A simple way to develop such rules of thumb is to ascertain, via simple regression, the relationship that exists between current rates of return and enrollment rates ten or fifteen years ago, which in turn reflect investments made fifteen to twenty years ago. At the primary and secondary levels, such rules of thumb would yield: rp = 36 -.18gerp rs = 20 - .21gers where ,p is the social rate of return to investments in primary education, gerp is the gross enrollment rate lagged 10 or 15 years, and rs and gers are the corresponding variables for secondary education. The coefficients on gerp and gers are significant at the 0.05 level 34 growth.'5' These rates and ratios, though high, are not without precedent in Guatemala. Indeed, many countries achieve education expenditure to GDP ratios around 3 to 3.5 percent. (3) What littk is spent on health and education is misallocated 5.12 The most telling evidence that public health expenditures are misallocated is that in 1990 the operating expenses of the two largest hospitals in the metropolitan area exceeded the operating expenses for all health posts and clinics in the country. In total, 72 percent of the Ministry of Health's 1990 expenditures went to hospitals, and only 19 percent of expenditures was allocated to primary health care centers. 5.13 Other evidence clearly shows that health sector expenditures are skewed toward urban areas and away from poor individuals. Figure 5.1 plots total expenditure per capita (vertical axis) against the percent of urbanization (horizontal axis) for each region and demonstrates that the greater the degree of urbanization, the greater are the resources spent on that area. Areas with the greatest poverty receive the least MOH expenditures per capita. The predominantly poor northwestern and northern regions combined received only Q9.66 (1991 US$1.92) per capita, wheeas the relatively better-off Guatemala City region received the aforementioned Q50.92 (1991 US$10.13) per capita. 5.14 Holding constant for the effects of urbanization in a multivariate regression analysis still reveals that less is spent in poor areas. The analysis suggests that for each percentage increase in poverty in a department, the per capita allocation of health resources to the department falls by US$0.50."6'. Thus a departamento in which 50 percent of the population is poor will be allocated about US$20 or 80 percent more financial resources per capita for health than a departamento that is 90 percent poor. "s' Assuming a 2 percent average gromwth rate, adequate education expenditures will eventually consume 3.5 pent of GDP. Assumiing a 4 prcent average growth rate, education expenditures will require 2.9 percent of GDP. Id' A simple way to develop such rules of thumb is to ascertain, via simple regression, the relationship that exists between cmrent rates of return and enrollment rates ten or fifteen years ago, which in turn reflect investments made fifteP to twenty years ago. At the primary and secondary levels, such rules of thumb would yield: rp = 36 -.18gerp rs = 20 -.21gers where rp is the social rate of return to investments in primary education, gerp is the gross enrollment rate lagged 10 or 15 yeas, and nv and ger are the corresponding variables for secondary education. The coefficients on gerp and gens are significant at the 0.05 level 35 Figure 5-1: PER CAPITA HEALTH EXPENDITURE AND URBANIZATION BY REGION Quetzales 70 60 Metro. 50 - 40 30 20 - 10 NW 0 0 10 20 30 40 50 60 70 80 Urban % 5.15 The most telling evidence that public educational expenditures are misallocated is that the 1986 Constitution earmarks 5 percent of all tax revenue to San Carlos University. Although social rates of return at the university level are estimated to be about one-half what they are at the primary and secondary levels, Guatemala spends about 25 percent more on tertiary education than on secondary education (16 versus 12 percent). Nevertheless, even the most expensive private university (in per student terms) has a lower cost per graduate than the public sector, since the proportion of students graduating from public universities is much lower. 5.16 The enrollment patterns are out of line with those common in countries that industrialized successfully in recent years, either before or during the beginning stages of their economic take- off. Guatemala today has about as much university enrollment, proportional to its population, as does Singapore, Malaysia, and Hong Kong; yet Guatemala has about one-half to two-thirds the secondary enrollment these countries had twenty or thirty years ago. 5.17 Other evidence shows that educational resources are skewed away from rural areas and poor individuals. As illustrated in Table 7.6, actual expenditure levels vary greatly by region."'7 Recurrent expenditures for 1991 reveal that the Metropolitan region alone spent 28 percent of total recurrent expenditures of the Ministry of Education, whereas the North and Northwest regions combined spent only 17 percent. 17/ All education sector expenditure informnation in this section is based on four key MOE programs: Mobiliario Escolar, Material Diddctico, Utiles Escolares and Refacci6n Escolar. 36 Table 5.5: MOE REGIONAL EXPENDITURES AND POVERTY HEADCOUNT Region School age Expenditure Poverty Headcount2 population' Expenditure per capita Metropolitan 708,359 10,091,406 14.3 54.2 North 327,755 2,245,790 6.9 91.3 Northeast 343,918 3,061,895 8.9 69.3 Southeast 332,829 3,270,453 9.9 79.7 Central 420,349 3,716,771 8.8 76.0 Southwest 985,982 9,023,542 9.2 82.7 Northwest 550,029 3,976,956 7.2 93.7 Peten 109,457 1,262,804 11.5 75.4 Source: MOE 1993. Notes: X estimated population ages 5-19 from: MOE, Anuario Estadistico 1992, 1993. Amounts are in 1992. Qutzals. 2ENS 1989. (4) Health and education ministries are weak andprovide inadequate services 5.18 In the health sector, the poor level of services actually delivered can best be seen in the figures on the underutilization of these services. The average occupancy rate of Ministry of Health hospitals is about 60 percent, which is significantly below the internationally recognized standard of 80 percent. The low consultation rates even in Guatemala City where health care resources are concentrated (0.8 per year compared to a World Health Organization norm of 2) suggests that the services being provided are not what is demanded by the public. 5.19 In the education sector, the inadequacy of the service that is actually delivered is reflected in the primary school repetition rate of 47 percent and the age-grade distortion rates. The age- grade distortions are highest among those who are the poorest, who are indigenous, and those who live in rural areas. 18/ Primary school age-grade distortion varies from 84.1 percent at the bottom of 20 percent of the income distribution to 49.9 at the top. Given the fact that the majority of the indigenous are poor, it is no surprise that indigenous ethnicity and over-age are also highly correlated with age-grade distortion rates for indigenous and non-indigenous of 87.6 and 64.3 percent, respectively. While 83 percent of those in primary school in rural areas are overaged, even in urban areas this rate is 55 percent. 5.20 Service delivery is inadequate not just because of the low level of resources and the misallocation of resources. Both the Health and Education Ministries also suffer from (a) Is/ Age-grade distortion is an important education indicator which reflects late enrollment, repetition, and drop-out. 37 inadequate planning, (b) poor budgetary execution, (c) overcentralization of their administration, and (c) an inappropriate personnel profile and rigidities in the personnel structure. These problems are discussed only briefly here because they are described in more detail in the World Bank's November 1993 Health Sector Report and June 1994 Education Sector Report. 5.21 Although the planning process in both these ministries sets out sector objectives, it typically lacks a detailed and implementable strategy and follow-through. For example, the 1994 annual operating plan of the Ministry of Health made almost no reference to the targets set in the 1993 annual operating plan. In the Ministry of Education, although the government announced that decentralization was to be its key policy instrument for improving the delivery of primary education services in the regions, no clear strategy has been developed to orient and support both central and regional agencies to achieve the Government's objectives. The primary education supervision system originally designed to develop both technical/pedagogical and administrative activities at the local level has not even been implemented yet. As a result, there is no systematic follow-up of primary school improvement efforts and student achievements; and there is no control of teacher attendance and the effective distribution of educational materials. 5.22 Poor budgetary execution is related to the lack of timely procedures for executing funds between the line ministries and the Ministry of Finance and, to a lesser extent, to inefficiencies within the line ministries themselves. In health, over the last four years the Government has spent roughly 20 percent of the amount programmed for investment. As in the health sector, the allocation of education resources is hampered by inefficient budget execution. In 1992, the MOE executed only 83 percent of its budget, only a slight improvement from the 77 percent execution rate found in the health sector. While the poorest execution rates are found in the metropolitan area where only 53 percent of allocated education resources were spent, the inability to execute hampers service provision in all areas of the country. 5.23 Along with overcentralization, an inappropriate personnel profile and rigidities in the personnel structure make it impossible to deliver basic health care effectively. Given the need to emphasize primary and preventive health care in rural areas, the current allocation and composition of staff will need to change. The disproportionate share of doctors to nurses and the excessive concentration of medical human resources in urban centers is in stark contrast to the health needs of the country. 19/ However, that are some almost insurmountable obstacles to the reallocation of personnel and the restructuring of the personnel profile. Rigidities imposed by the Civil Service Law and the strength of public sector unions make it difficult to use involuntary transfers or to dismiss workers with redundant skills. 5.24 A rigid salary structure and labor practices compound the problems in education. The allocation of teachers' salaries is based upon a rigid salary structure that awards teachers solely on the basis of their chronological experience. A teacher can attain the highest salary level without regard to the quality of his or her teaching, or to the nature of his or her relationship with the 19/ According to the 1991 Social Sector Expenditure Review, the ratio of nurses to doctors in 1988 was 0.8. An adequate and cost efficient ratio is considered between 2 and 4 nurses per doctor. 38 community. In addition, the teacher labor code makes it extremely difficult to fire teachers. Furthermore, the firing (like hiring) cannot be done by those to whom quality matters most, namely, those at the local level, but by a distant bureaucracy that can just as easily place an inappropriate teacher in a community as move away an appropriate one. 5.25 In the health sector, additional problems are (a) the lack of coordination between the Health Ministry and other health care providers (particularly IGSS and the private sector); and (b) inappropriate treatment strategies. Because of poor coordination, redundant services are provided in certain urban areas and there is also an excess of health infrastructure. The underutilization of health infrastructure in these areas has led to high unit costs of service provision by the involved agencies. Increased coordination between the two systems (including the private and donor sectors) is becoming of even greater importance at a time when IGSS' health care system is encountering serious financial difficulties. In the interest of reducing the inefficient use of public resources brought on by the duplicative efforts of the two agencies, serious consideration should be given to either integrating the two or clearly dividing their responsibilities. Improved efficiency in the public sector would not only increase access to all levels of health care, but the savings, accruing mainly from efficiency gains at tertiary level could be devoted to providing lower level care. 5.26 In the education sector additional problems are (a) the lack of primary school bilingual education and (b) inadequate teacher training. Compounding the problems of insufficient planning, overcentralization, and weak management are the lack of primary school bilingual education, sufficient school education materials and inadequate teacher training. Historically, language was a major reason for the limited access of the indigenous population to education opportunities in Guatemala. While national law now requires bilingual education in rural communities, often this does not occur. This problem is further exacerbated by a lack of textbooks, workbooks, and other educational aids, especially in rural schools. The absence of educational materials is particularly serious among the indigenous children in the rural communities where books provided by the schools are the only reading source. 5.27 Deficiencies in teacher training also constrain the delivery of adequate education. Teacher training programs are considered either too short or irrelevant and inadequate to attend to teacher needs. The deficiency in teacher training programs is most felt in rural communities. Typically, teachers in rural communities are those most in need of training. When training is provided, however, it is mainly offered in urban areas, thus incurring additional costs to teachers in rural communities, costs not compensated for by the training agency. (5) Conclusions 5.28 The low levels of human capital are due to the widespread poverty of the population, past inadequate levels of public financing, misallocation of what little has been invested and service delivery problems. International evidence suggests that human capital must improve for 39 Guatemala to develop. This will require more public funding, financed by increased tax revenues and, possibly, further reallocations of the budget. 5.29 However, poverty alleviation strategy cannot be simply to devote more resources to public investment in human capital. Given current and likely future financing constraints, the efficient delivery of education and health of paramount importance. There is no dearth of examples of effective alternative delivery mechanisms in education and health to traditional government-based mechanisms. Dozens of NGOs, private sector, and other non-state providers of education and health are active in Guatemala, and have been active for decades, purveying innovative projects. Instead, the problem is the public sector's ability and willingness to sustain (through time) and replicate (though space) the ideas and experience of the non-state sector, along with an almost total absence of serious efforts by donors and NGOs to evaluate prior experience and to assess the very lack of sustainability of past projects. 5.30 Even within the public sector, it is possible to improve the efficiency and accountability of public providers through performance-based budget allocations. The key element is that some part of the budget that a health area or a school will receive should be at risk. The health area or school would receive more for good performance and less for bad performance. This would introduce an element of competition (against a standard) that would substitute for the lack of competition against other providers which is present in the public institutions. 40 6. INVESTING IN INFRASTRUCTURE AND SUPPORT SERVICES 6.1 While the international evidence strongly suggests that investments in human capital are critical to promoting development, these long-term investments are not necessarily highly valued by the poor. Indeed, the Participatory Poverty Assessment conducted as part of this poverty study found that only one-fifth of the population sample stated that they valued education. Ten percent of the poor people sampled explicitly stated that they did not value education. Moreover, education did not appear to play a major part in the conversations. 6.2 The Participatory Poverty Assessment and a survey of poor communities in the Western highlands (Smith, 1990), however, does report strong demand for infrastructure and support services that would enhance the poor's earnings potential. At the depth of poverty that appears to exist in Guatemala, investments that increase earnings could be as effective as investments in educational infrastructure in increasing enrollments. With greater earnings, poor families could be expected to invest more in the human capital of their children. 6.3 Thus, the fourth part of the poverty alleviation strategy is to finance more infrastructure and support services. The greatest demands are for technical assistance in agriculture, roads and electricity. (1) Support to agricultural production of the poor 6.4 Two principal means of supporting small farmer production would be to help provide access to credit and to orient agricultural technical assistance to the poor. The allocation of agricultural credit from the formal banking system, especially the private banks, is highly skewed toward the wealthy. Evidence for this concentration is the fact that Guatemala City and the Pacific Coast account for more than half of agricultural credit from the banking system. In terms of the allocation of bank agricultural credit by crop, in 1990, only 16 percent went towards basic grains production, the principal activity of most smallholders, compared to the 41 percent that went for predominantly large farmer export crops like coffee, cotton and sugarcane. From the banking system's perspective, the information and transactions costs of dealing with large numbers of small borrowers are high; from the smallholders' point of view, collateral requirements and long administrative delays prevent them from seeking and obtaining formal sector credit. 6.5 Unfortunately, the principal public sector institution in the agricultural credit market, the National Bank for Agricultural Development (BANDESA), has not been particularly successful. Despite BANDESA's legal mandate to mobilize the savings of and deliver agricultural credit to small farmers of basic grains, the agency's outreach has been small compared with that of private banks, rural credit unions, and informal financial agents in terms of amounts of agricultural credit provided, number of clients served, and services such as deposit mobilization. At the same time, 41 20/ BANDESA has absorbed nearly half the Ministry of Agriculture's annual budget in recent years. Moreover, within its small market share, BANDESA does not appear to target its resources to the rural poor: only 15 percent of its 1993 investments are directed to the poorer North and Northwestem regions. BANDESA's portfolio also reveals a bias against rural women's economic activities: in 1991, loans for artisanal projects, an important source of rural women's income, represented less than 2.5 percent of the number of loans and the value of BANDESA's national portfolio. 6.6 Largely in response to donor concern over smallholders' lack of access to public sector credit, BANDESA had increasingly relied on trust fund mechanisms, with special conditions, lower rates and better terms than bank credit, to target financing to resource-poor clients. The widespread use of these mechanisms, however, led to increased operating costs and reduced incentives for effective loan recovery. The cost problem is expected to be resolved in part by the streamlining of administrative procedures which BANDESA is undertaking as part of its current restructuring efforts; however, the basic issues of reliance on soft external financing and lack of savings mobilization and loan recovery remain. In its Agenda para la Reactivaci6ny Modernizacion de la Agricultura, the Guatemalan Ministry of Agriculture proposes the creation of a "second tier window" (ventanilla de segundo piso) within BANDESA to carry out wholesale operations with grassroots financial intermediaries such as cooperatives and NGOs. 6.7 The feasibility of the "second tier window" has not been thoroughly examined. If BANDESA can be reformed to more closely meet its original objectives, then it would merit continued support. Failing that, an alternative way to help provide access to credit would be to support successful agricultural cooperatives and credit unions. The emphasis should be on successful. In 1988 it was estimated that out of approximately 140 rural credit unions serving over 100,000 families, only 20 percent of these were considered to be solvent and well functioning.21' One of Guatemala's most successful cooperatives, the Cooperativa 12 de Octubre, has learned this lesson by incorporating an effective, community-based borrower selection and loan repayment mechanism. The cooperative formed solidarity groups on a community level that approve loans to group members and collect payments, pressuring for repayment in the case of tardiness. The group does not receive a new loan if it has not repaid its prior obligation in full. If someone does not repay, the group pays that person's part and may then eliminate him or her from future loans. This mechanism is effective because the group chooses its own members and approves the loans (see Vogel, Robert C., Robert P. Cristen, John McGuire, Juan Carlos Protasi, Antonio M. Salas and Michael Saperstein. An Assessment of Rural Financial Markets in Guatemala. Report to USAID-Guatemala). 6.8 Another recent study of three rural credit unions found that these filled a major gap in local markets by mobilizing the savings of and making loans to households engaged in small to 20/ BANDESA's share of agricultural public sector spending has been reduced from 46% in 1992 to 35% in 1993. 21/ Vogel, Robert C., Robert P. Cristen, John McGuire, Juan Carlos Protasi, Antonio M. Salas and Michael Saperstein. An Assessment of Rural Financial Markets in Guatemala. Report to USAID-Guatemala. 42 medium-scale agricultural and commercial enterprises, providing a scale and scope of financial intermediation to this sector of the population that would otherwise go unserved by other public and private institutions.22 The study further concludes that credit unions are important to local economic development insofar as they are the only organizations that are effectively channeling local savings from the lower and middle ends of the wealth spectrum back into productive loans in the community. (2) Agricultural research and extension has not favored the poor 6.9 Agricultural research in Guatemala has historically been focused on the development of improved seed varieties with little attention to natural resource management, small animal husbandry, agroforestry, or the development of appropriate technologies for rural women. In addition, extension activities have been largely divorced from research, non-participatory, and ethnic group and gender-stereotyped. In the participatory poverty assessment, close to half of the 487 farmers interviewed stated that they had received no training. 6.10 The national agricultural research institute (ICTA) was created in 1973 to conduct research on basic grains and promote the resulting new technologies among small landholders. While the institute has expanded its research scope from basic grains to include fruits, vegetables, oilseed crops and livestock, due to a decline in financial and human resources, ICTA has been unable to respond to the challenges of new agricultural policies and the changing production systems of rural Guatemala. Specifically, in order for agricultural research to serve the needs of the rural poor, it needs to incorporate smallholder participation into its decision-making structure and research methodologies and to diversify its research agenda to respond to the needs of the labor market and to include areas such as integrated soil, water and pest management, small animal husbandry, agroforestry, and appropriate domestic technologies. For example, while it has been estimated that women represent 28.4 percent of the economically active population in agriculture, there is no program to investigate technologies that might address some of the most pressing needs facing rural women: access to fuel and water, increasing their productivity in the agricultural tasks in which they participate, and improving their small livestock care techniques.23' Research has indicated that rural women's priority objectives with regard to technology generation are reducing production-related and health risks, reducing costs, and reducing labor- time. Nevertheless, ICTA's field diagn6sticos and sondeos, which almost exclusively utilize male agronomists, focus on increasing productivity of male farmers. 6.11 Extension activities are limited in scope and have been largely divorced from research, nonparticipatory, and are ethnic and gender-stereotyped. Public rural extension services are ' Barham, Bradford and Stephen Boucher. 1993. The Imoact of Credit Unions on the Savinas and Credit Patterns of Small and Medium Scale Producers in Three Regions of Guatemala. Department of Agricultural Economics, University of Wisconsin - Madison. 2 de Le6n E. 1992. 43 provided by three separate agencies: DIGESA (agriculture), DIGEBOS (forestry), and DIGESEPE (livestock). However, it is estimated that the three extension agencies together serve only 7 percent of Guatemalan smallholders. In addition, the effectiveness of extension agents in the transfer of appropriate technologies is questionable, with the notable exception of those areas participating in the IDB/IFAD- financed Technology Generation and Transfer and Seed Production Program (PROGETTAPS). Also, lacking in both PROGETTAPS and non- PROGETAPPS areas, however, are the use of participatory extension methodologies and effective collaboration with NGOs, agricultural cooperatives, and other grassroots rural organizations. Furthermore, the extension agents are primarily latino, they often do not speak the local languages, and they give little consideration to the support and/or improvement of traditional mayan technologies. 6.12 Extension activities have also limited their attention to women in largely the more traditional female activities of "home economics"--such as cooking and sewing lessons, personal hygiene, infant care. While the types of work that women do in the rural areas vary considerably by region, ethnicity, socioeconomic strata, age, and marital status, their principal areas of concentration can be characterized as (a) domestic labor (food purchases, milling corn, cooking, child care, weaving, washing, etc.); (b) collecting and transporting water and firewood; (c) maintaining family gardens; (d) working together with male household members in the most labor- intensive agriculture tasks, such as planting, weeding, and harvesting; (e) postharvest management, including processing, classification, storage, transport and retail marketing of many crops: (f) raising small animals (pigs, chickens, sheep, goats); and (g) other income-generating activities such as artisanal production, taking in laundry, day agricultural labor, petty marketing, etc. Despite that type of work that women do, it is widely considered that the home educators (educadoras del hogar) employed by DIGESA are engaged in mostly unproductive and inappropriate activities, such as teaching rural women how to make cakes and cook on gas stoves. They commonly make use of technologies and inputs not normally available to the community, and fail to take into account local knowledge. Educadoras del hogar receive no agronomic training whatsoever, and are therefore unqualified to work with women beneficiaries in this area. The male promotores agricolas, meanwhile, rarely work with women to improve their agricultural skills, despite the fact that women are generally more in need of outside training and more open to learning new techniques. Little attention is given to increasing women's productivity by (a) alleviating the burdens of water and firewood collection and corn milling with sanitary installations and appropriate technologies, (b) promoting viable income-generating activities at both the individual and collective levels, and (c) training women in the agricultural tasks in which they participate. 44 (3) Infrastructure is deficient, especially in rural areas 6.13 Guatemala's public economic and social infrastructure is inadequate mainly because the govermment invests so little in this area and because it is difficult and costly to provide infrastructure in the many remote areas. The lack of investment is apparent in what is perhaps the most important economic infrastructure in the country---roads. A recent assessment of the road network (Perrone, 1992) indicated that as a result of aging, unsuccessful efforts in controlling vehicle-axle-load and neglected maintenance, what had been a reasonably well developed primary and secondary road network had reached an advanced state of deterioration. In addition, many of the bridges included under the secondary and tertiary networks are wooden structures built close to 20 years age to low standards and are now unsafe and should be replaced. 6.14 Half of the population has access to electricity and there is a large disparity between poor and non-poor and between urban and rural areas. Only 38 percent of the poor versus 79 percent of the non-poor and 27 percent of the rural residents versus 87 percent of rural residents have electricity. Electricity provides not only consumption benefits, but also makes nonagricultural pursuits more productive. Even in rural areas, non-agriculture pursuits are gaining in importance as a result of continued population growth and difficulties in acquiring land. (4) Conclusions 6.15 Guatemala's infrastructure and support services have suffered from insufficient investment (roads), misdirected investment (agricultural extension) and poorly functioning public institutions (agricultural credit). These problems need to be corrected. 6.16 On a more promising note, there have been three relatively recent developments that present promising approaches to financing infrastructure and support to production in the rural areas. Guatemala has (a) introduced a constitutionally mandated 8% transfer of tax revenues to municipalities (subsequently raised to 10% in the 1994 referendum); (b) established the Fondo de Inversion Social (FIS), a financial intermediary supported by the World Bank that finances but does not execute projects proposed by local groups; and (c) established the Fondo Nacional para la Paz (FONAPAZ). Both funds offer assistance in the areas of social services, economic and social infrastructure, and institutional strengthening. The difference between the two lies not so much in the services they provide but in their target population. While they are both designed to provide services to the poor, FONAPAZ targets those affected by armed conflict. 6.17 Both funds should help generate a pattem of investment that closely reflects local needs. There is an increasing acceptance of the view that successful interventions require the active sustained involvement of local groups. This involvement is more likely to take place if investments are proposed by local groups than if imposed from above. Providing the mechanisms for communities to become involved in investment decisions is especially attractive in Guatemala 45 because community organizations already exist at the grassroots level in most of the country's 330 municipalities and thousands of small rural settlements.24' 6.18 While some municipalities have already taken advantage of the 8 percent transfers, many other municipalities do not have the administrative expertise to formulate and manage infrastructure projects. However, the administrative procedures developed as part of the FIS (and in FONAPAZ) should help eliminate this obstacle. Municipalities can follow the examples of the FIS in contracting private contractors to carry out the infrastructure investments. Because they are private autonomous organizations that finance rather than implement projects, they avoid the build-up of bureaucracy, circumvent public sector inefficiencies and can take advantage of private sector participation. 6.19 Municipalities would be expected to share in the costs of the projects with the FIS. In addition, monetary and/or in-kind contributions are expected from the beneficiaries. Insisting on counterpart financing not only guarantees that the investments are valued by local groups, but also enable limited public and international donor funds to help finance more infrastructure. The analysis of the proceeding chapters suggests that those who have worked as agricultural laborers planting and harvesting traditional agriculture products such as coffee, sugar and cotton have been hurt by falling prices. There is no indication from an analysis of the funds and programs that this group has been specifically targeted. 24/ Amwex 11 identifies and briefly describes the inportaft public and private sctor ititutions. 46 7. IMPLEMENTING A POVERTY ALLEVIATION STRATEGY 7.1 As the preceding pages indicate, a strategy for alleviating poverty in Guatemala must aim at achieving a stable macroeconomic environment, must try to manage the economy in a way that will make efficient use of labor, and must increase public investment to raise the earnings potential of the poor. These objectives cannot be achieved unless the government is able to raise taxes and to make good use of the funds it does receive. The private sector is understandably reluctant to give resources to a public sector that it views as inefficient at best and corrupt at worst. Yet the public sector does not have the power to enforce tax collection and finds its revenues subject to substantial fluctuations around a low level. These fluctuations in revenue, combined with the government's antiquated administrative mechanisms, create wide swings in public investment and put the public sector unions at odds with the administration in their effort to capture new funds when the budget constraint is relaxed. The resulting poor and uncertain quality of public providers and highly visible work stoppages further erode the public's confidence in the public sector to provide any services. (1) Immediate decisions 7.2 Recent World Bank projections suggest a strong possibility that revenues during 1994 could fall to 6.4 percent of GDP. If the administration is not able to increase tax revenues, its priorities should be to pursue an extremely limited program aimed at maintaining macroeconomic stability. An effort should be made to maintain macroeconomic stability even at the cost of needed publicly financed investment, for it is costly to recover from an unstable environment. 7.3 While new initiatives and sector-wide reforms are probably not feasible in the current political climate, the administration could take advantage of existing mechanisms to pilot-test promising approaches to improving the delivery of publicly financed investments. For example, it could promote links between the municipalities, which have access to funds but no expertise, and the Social Investment Fund, which is developing mechanisms which allow for greater participation of the private sector and NGOs. The experience with these mechanisms should be documented so that there will be some Guatemalan-specific experience to draw on when the political climate would permit debate over alternative poverty alleviation strategies. 7.4 The current administration could also initiate efforts to establish a base line for poverty. It has been five years since the last household survey and many policies have changed in the interim. There have also been improvements in methodologies for collecting poverty-related data.25' 251 Annex 10 discusses a possible strategy for developing a national statistical system to monitor and evaluate poverty reduction efforts. 47 (2) Decisions in the nearfuture 7.5 There is an ongoing and an upcoming political event during which the elements of a poverty alleviation strategy are likely to be debated. The ongoing event is the peace process. The upcoming event is the next presidential election. 7.6 The peace process presents an opportunity for a wide debate among all sectors of society since such topics as the recognition and rights of the indigenous community and the socioeconomic aspects and the agrarian situation are on the agenda. The peace process also contains risks in that a poverty alleviation strategy that is politically acceptable may be financially infeasible. Any poverty alleviation strategy that calls for levels of investments far in excess of the level of taxes that the private sector is willing to pay and what is available from international financing is internally inconsistent and doomed to failure. 7.7 If a consensus is not developed on the appropriate poverty alleviation strategy as part of the peace process, then this strategy will be confronted in the next presidential election. The absolute amount Guatemala is willing to invest to enhance the earnings potential of the poor is likely to affect not only the levels of poverty, but also the rate of growth the country can expect to enjoy in the future. 7.8 The appropriate combination of the three recommended interventions (increasing the poor's access to land, investment in human capital, and infrastructure and support services) is a political decision that must be taken by the different interest groups within society. If Guatemala is unable to maintain an adequate tax level for a successful poverty alleviation strategy, then it can expect to face low rates of per capita growth, high population growth rates, minimal reduction in poverty, social unrest and threats to sound economic policy from populists on the left and threats to democracy from the right. 49 BIBLIOGRAPHY Arends, M. (1992). "Female Labor Force Participation and Earnings in Guatemala." In G. Psacharopoulos and Z. Tzannatos, eds., Case Studies on Women's Employment and Pay in Latin America. Washington, D.C.: The World Bank. Berham, B. L., Carter M., and Sigelko, W. (1993). Agro-Export Production and Peasant Land Access: Examining the Dynamic between Adoption and Accumulation. University of Wisconsin - Madison. Barham, B. and Boucher, S. (1993). The Impact of Credit Unions on the Savings and Credit Patterns of Small and Medium Scale Producers in Three Regions of Guatema- la. Department of Agricultural Economics, Unversity of Wisconsin - Madison. Binswanger, H.P., and Deininger, K. (1994). World Bank Land Policy: Evolution and Current Challenges. Washington, D.C.: The World Bank. (Mimeo). Birdsall, N., Ross, D., and Sabot, R. (1994) Inequality and Growth Reconsidered. Washington, D.C. (Mimeo). CELADE. (1992). Demographic Bulletin 50 (July). CEPAL. (1992). Latin American Poverty Profiles for the Early 1990s. CEPAL. (1991). Magnitud de la Pobreza en America Latina en los Anlos Ochenta. Estudios e Informes de la CEPAL 81. Santiago de Chile: Naciones Unidas. Chen, S., Datt G., and Ravallion, M. (1993). Is Poverty Increasing in the Developing World? (WPS 1146). Washington, D.C.: The World Bank. De la Cadena, M. and Strochlic, R. (1990). The Penny Foundation and INTA Case Studies: Findings and Recommendations. Madison: Land Tenure Center, University of Wisconsin. de Le6n Escribano, C.R. (1992). La Mujer Productora de Alimentos en Guatemala: Diagn6stico y Polfticas. Programa de Analisis de la Politica del Sector AgropecuaTio Frente a la Mujer Productora de Alimentos en Centroamerica y Panama. IICA- Guatemala. Direcci6n General de Estadistica. (1984). Encuesta Nacional de Ingresos y Gastos Familia- re, 1979-81, Volume II Ministerio de Economfa. Guatemala. 50 Episcopado Guatemalteco. (1988). El Clamor por la Tierra. Carta Pastoral Colectiva del Episcopado Guatemalteco. Garcfa H. (1993). Guatemala: Revisi6n de los Programas de Enfrentamiento a la Pobreza. Informe de Misi6n. World Bank, Washington, D.C. Hintermeister, A. (1984). Rural Poverty and Export Farming in Guatemala. World Employment Programme Research. Huppi, M. (1991). Guatemala Social Sector Expenditures. Washington, D.C.: The World Bank. (Mimeo). IIESO-USAC. (1980). Guatemala: Estructura Agraria del Altiplano Nor-Occidental. Centro Universitario de Occidente, Universidad de San Carlos de Guatemala. Instituto Nacional de EstadIstica (INE). Various years, Indice de Precios al Consumidor. Boletfn Anual, 1990 and 1991. Guatemala. Instituto Nacional de Estadfstica (INE). (1991). Perfil de la Pobreza en Guatemala, Volume V, July. Guatemala. Instituto Nacional de Estadfstica (INE). (1992). Perfil de la Pobreza en Guatemala. Guatemala. Katz, E. G. (1992). Intra-household Resource Allocation in the Guatemalan Central Highlands: The Impact of Non-traditional Agricultural Exports. Ph.D. Dissertation, University of Wisconsin - Madison. Lee, H. (1993). The Ethnic Dimension of Poverty and Income Distribution in Latin America. Background Paper for Research Project (RPO 677-68) "Indigenous People and Poverty in Latin America: An Empirical Analysis." Washington, D.C.: The World Bank. Melendreras and Cabrera. (1991). El Mapeo de la Pobreza en Guatemala. Proyecto Instituto Nacional de Salud, Ministerio de Salud Publica y Asistencia Social. Newman, J. (1993). Guatemala: Health Sector Study. Washington,D.C.: The World Bank. (Mimeo). Psacharopoulos, G. and Patrinos, H. A. (1992). Socioeconomic and Ethnic Determinants of Grade Repetition in Bolivia and Guatemala, (WPS 1028). Washington, D.C.: The World Bank. Psacharopoulos, G. and Patrinos, H. A. (1993). Indigenous People and Poverty in Latin America: An Empirical Analysis, (RPO 677-68). Washington, D.C.: The World Bank. 51 Psacharopoulos, G., Morley, S., Fiszbein, A. Lee, H. and Wood, B. (1992). Poverty and Income Distribution in Latin America: The Story of the 1980s. Latin America and the Caribbean Technical Department, Regional Studies Program, Report No. 27. Washington, D.C.: The World Bank. Psacharopoulos, G., Velez, E. and Patrinos, H. A. (1992). Poverty in Paraguay Today. Prepared for Paraguay Country Economic Memorandum (March 22, 1993). Wash- ington, D.C.: The World Bank. PAHO. (1991). Estudio Basico del Sector Salud. Ravallion, M. (1992). Poverty Comparisons: A Guide to Concepts and Methods. Living Standards Measurement Study Working Paper No. 88. Washington, D.C.: The World Bank. Ravallion, M. (1993). Poverty Comparisons, Chur, Switzerland: Fundamentals in Pure and Applied Economics, Volume 56, Harwood Academic Press. Ravallion, M. and Benu B. (1993). How Robust is a Poverty Profile?. Washington, D.C.: The World Bank Economic Review, forthcoming. Richards, M., Gates, S., and Stringer, R. (1990). Land Tenure, Land Utilization, and Household Economy of Cooperative Members Affiliated with the Federaci6n de Cooperativas Agr"colas Regionales (FECOAR). Guatemala. Guatemala:USAID. Secretarfa General del Consejo Nacional de Planificaci6n Econ6mica (SEGEPLAN). (1987). Agricultura. Poblaci6n y Empleo en Guatemala. Direcci6n de Planificaci6n Global, Departamento de Poblaci6n y Empleo. Proyecto GUA/79/PO3-OIT/FNUAP, Serie Resultados No. 5. Guatemala. Shearer, E. B., Lastarria-Cornhiel, S., and Mesbah, D. (1991). The Reform of Rural Land Markets in Latin America and the Caribbean: Research, Theory, and Policy Implica- tions. LTC Paper 141. Madison: Land Tenure Center, University of Wisconsin. Smith, C. A., with the assistance of CISMA. (1990). An Analysis of Economic Variation, Development Projects. and Development Prospects in the Highlands of Western Guatemala. Report for the InterAmerican Foundation. Stringer, R. and Lambert, V. (1989). A Profile of Land Markets in Rural Guatemala. Madison: Land Tenure Center, University of Wisconsin. van de Walle, D., and Ravallion,M. (1993). A Profile of Poverty in Morocco. Washington, D.C.: The World Bank. (Mimeo). Vargas, F. (1992). Consultorfa Sobre la MUier y la Produccion de Alimentos: Tecnologfa y Comercializaci6n. Instituto de Cooperaci6n para la Agricultura (IICA). 52 Vogel, R. C., Cristen, R., McGuire, J., Protasi, J.C., Salas, A., and Saperstein, M. a Assessment of Rural Financial Markets in Guatemala. Report to USAID-Guatemala. Williams, Robert G. (1993). States and Social Evolution: Coffe and the Rise of National Governments in Central America. Unpublished manuscript. World Bank. (1990). World Development Report 1990. New York: Oxford University Press. _. (1991). Country Economic Memorandum. Report No. 9378-GU. Latin America and the Caribbean Region, Country Department II. . (1991). Assistance Strategies to Reduce Poverty. A World Bank Policy Paper. -. (1992). Guatemala Social Investment Fund Project. Staff Appraisal Report No. 11060-GU. Country Department II, Human Resources Operations Division, Latin America and the Caribbean Regional Office, Washington, D.C. (1992). Indonesia: Public Expenditures. Prices and the Poor. Green Cover Report Nop. 11293-IND, December 1992. (1992) Economic Modernization Loan. President Report P-5858-GU LA2C2. (George Monserrat: 39416). (1993), A Poverty Profile of Guatemala. Report No. 12313-GU. Latin America and the Caribbean Regional Office, Human Resources Operations Division, Country Department II. (1993). Kingdom of Morocco: Adjustment, Gropwth and Poverty, Yellow Cover Report No. 11918-MOR, May. Annex I Page 1 of 10 Description of Data and the Measurement of Poverty A. Description of Data I. Available primary data from which to determine the incidence of poverty in Guatemala is relatively limited. The most recent reliable source of such information is the 1989 National Socio- Demographic Household Survey (ENS) and its companion module, the National Employment Survey. This survey was administered by the National Statistics Institute (INE) to 9,272 households comprising 48,909 individuals nationwide within eight regions composed of 22 departments between April 4 - July 24, 1989. Because households reporting zero income were excluded from the analysis, the resulting sample size was 8,988 households comprised of 47,850 individuals (Annex Ia, Table 2). Table 1.2 presents the list of regions and departments as they are defined in the ENS.Y' t 2. The ENS is an income-based survey which was previously conducted in 1986/87. The survey provides information on individual characteristics including age, gender, relation to household head, ethnicity, educational level, fertility, health, and morbidity. It also contains labor market information for each person such as labor force participation, industrial sector, occupation, economic sector, levels of employment, and source and amount of income earned. In addition, information on land tenure and access to public services is included. No consumption data are provided by the survey and reliable consumption information is not available from other sources. 3. Data availability therefore dictates Table A1: REGIONS AND DEPARTMENS that income be used as the fundamental Region Department basis for determining poverty levels in this 1. Metropolitan Guatemala report. This approach is not without 2. North Baja Verapaz, Alta Verapaz shortcomings. First, there is general 3. North-East El Progeo, Izu, Zacapa and agreement among researchers that Chiquim3ula consumption is a better measure of welfare than reported income. Income is not 4. South-East Santa Rosa, Jalapa and Jutiapa necessarily indicative of how household s. Central Sacatepequez, Chimaltenango and members live on a consistent basis since it Escuintla can be saved or dissaved to smooth for 6. South-West Solola, Totonicapan, fluctuations over time. Second, Quetzaklengo, Suchitepequez, fluctuations in income, caused by, among Retalhuleu and San Marcos other things, multiple harvest seasons and 7. North-West Huehuctenango and Quiche labor migration, can bias poverty estimates 8 Peten Peten depending upon when the survey is conducted (see Annex lb). Third, there is Source: Encuesta Nacional Sociodemogrdfica, 1989. considerable debate over the effects of income underreporting in household 1/ Urban areas include cities, villages, and towns (ciudad, villa and pueblo). Rural areas consist of hamlets nd farns (aldea, caserio paraje, finca, etc.). 2/ The ENS draws on the master sample established from the 1981 National Census and uses a two-stage sampling methodology. In the first stage, the sample was stratified by 18 domains consisting of cight regions and three sub-areu divided into two central urban areas, eight 'other urban' and eight rural areas. This resuled in 611 primary study sample units or census sectors (Annex la, Table 3). Each census sector was then divided into a further three sectors. In the second stagc, five or six households were randomly drawn from each census subsector and included in the final household survey ample. Anex-I Page 2 of 10 surveys. Underreporting is most prevalent at the tails of the income distribution, with the wealthy tending to underestimate rents and secondary incomes while the poor often overlook income from auto-production and in-kind transfers. Finally, income represents only one dimension of well-being (as does consumption), and it does not take directly into account other welfare attributes such as malnutrition prevalence, infant/child mortality levels, housing characteristics, and access to water, sanitation, health and education services. 4. The income concept used is per capita household income for the month prior to the survey interview. Per capita household income was derived as follows. First, income from all sources was aggregated to obtain total individual incomeY Second, total individual income was combined within each household to obtain total household income. Finally, total household income was divided by family size to obtain per capita household income. 5. This analysis did not adjust for income under-reporting. The bias that under-reporting causes is difficult to accurately identify and alleviate. Furthermore, there is also often a risk that correcting for under-reporting may introduce additional bias into the study. B: Income Fluctuation and Own-consumption Effects 6. A common problem associated with many poverty studies, especially in the developing world, is the difficulties presented by the lack of adequate data. Two issues that often challenge poverty analysis based on income data are: (a) fluctuations in the income of the poor over the year in relation to the time period in which the survey was conducted; and (b) accounting for own-consumption. 7. Since a majority of the poor do not receive a steady income, poverty estimates are sensitive to the time period in which the data is collected. This problem can be particularly pronounced in the agricultural sector where various harvest seasons in the course of the year can fluctuate income levels. Tle ENS survey, like many other surveys, is no exception. Conducted between April and July, the ENS survey, though including an April harvest peak, excludes a harvest season (coffee) that occurs from November to December. 8. Another problem all too common in poverty studies arises from the lack of data accounting for own-consumed product. Own-consumption is the consumption of goods produced by the household. A common form of own-consumption is farm output, where often a portion of farm product is consumed by the farmer. Poverty measurements based on household income may overestimate poverty if the imputed in-kind value of products produced and consumed at home is ignored. Additional problems may arise from the difficulty of imputing the monetary value of in-kind products.1 The difficulties imposed by own-consumption are particularly relevant given the April harvest is mainly grains, product often used for own-consumption. 9. Several factors, however, mitigate the potential problem of the ENS survey underestimating or overestimating poverty. Income fluctuations over the year are somewhat moderated by the geographic location of harvest seasons, agricultural diversification and by non-agricultural income. Although the larger November-December harvest season is excluded from the survey, this harvest occurs primarily in 3/ Total individual income is the aggregate income from principal, secondary, retirement, other trnsfrs and in-lind soures. F' Ravalion, M. 'Poverty Comparisons: A Guide to Concepts and Methods", LSMS Working Paper No. 88. 1992. Anmex 1 Page 3 of 10 the relatively better-off regions, though, as mentioned above, does effect the income of the poor through migrant labor. Whereas the April harvest, included in the survey, occurs mainly in the poorest regions (north, southwest and altiplano). In addition, over the last 20 years, with the introduction of mini- irrigation systems, improved farming methods and improved seeds, there has been a stabilization of seasonal fluctuations in income, due to the opportunities modernization has created for cultivating a wider variety of crops during more months of the year. Non-agricultural income also plays a substantial role among the poor. Various studies have placed the proportion of income attributed to non-agriculture activities at around 50 percent among poor populations. C. Constructing Poverty Lines Poverty Line IO.The primary focus of this report is absolute poverty, whereby certain individuals within the Table 13: POVERTY AND EXTREME POVERTY population have a standard of living which is less LINES than some absolute minimum. Following By LasPr h common practice, this standard of living is measured according to real per capita income. GOteaWa *Other Rural The level of income which is equivalent to the cAt UA defined standard of absolute poverty is known as the poverty line. In this report, a poverty line Poverty Line 93.S7 77.01 66.79 and an extreme poverty line have been developed Extreme 53.62 48.75 44.63 for three geographical subdivisions: Guatemala Poverty Line City, "other urban" areas and rural areas (Table Source: Encuesua Nk-ioul Sociawo,gr4ilca, 1939. Al. I). These poverty lines have been constructed to represent an identical level of welfare for each individual across the different geographic areas, taking into account cost of living differentials across different locations. 11. In making assessments of poverty, the choice of methodology will be dictated in part by the purpose of the exercise. In a poor country such as Guatemala, the main concern is with absolute poverty - the situation in which individuals do not attain a standard of living deemed to be a minimum below which one is deprived in some absolute sense in that society. T'here are various ways of defining that minimum - commonly known as the poverty line. A number of alternative methods can be proposed, each with strengths and drawbacks. 12. In the present case, the aim is to inform policy decisions concerning the alleviation of absolute poverty. Whether the policies chosen ultimately reduce poverty crucially depends on the consisteny of the underlying assessment across subgroups, such as sectors, geographical regions and dates.-F This is because the outcome of the poverty assessment will have strong implications for the most appropriate 1' Consistency implies that any two households with the same real standard of living are ctegorized in an identical manner: one is not deemed poor in one subgroup and one non-poor in another subgroup. This requirs that the poverty cutoff embody the same level of welfare across each subgroup. In particular, the focus in this assessment is to ensure that consiateny is maintained across geographic regions. F A more in depth and exhaustive discussion of the issues which arise in constucting poverty linea is put forth in Ravallion, 12. Annex 1 Page 4 of 10 choices in the allocation of development expenditures and the targeting of public services across geographical areas. For these reasons, getting the relativities right between sub-groups is of considerably more consequence to policy choices than the precise poverty numbers and levels. 13. Following common practice, the level of living is measured by real income. It is thus essential to adjust for spatial differences in the cost of living, although the data to do this are often far from ideal. Methodology 14. This report employs a methodology which, at a general level, is used quite commonly in poverty profiles for Latin American countries. It begins by defining a basket of basic food commodities which achieves a stipulated food energy requirement. This bundle is then costed to give the food component of the poverty line. The cost of this food component is used to denote the "extreme poverty" cutoff point. Next, this amount is divided by the food share of a "typical poor" person in order to take into account a basic minimum of non-food expenditures, such as housing, clothing, health and education, transport and so on. The combination of basic food and non-food expenditures then comprise the poverty line. In order to maintain the aforementioned consistency across geographic regions, separate poverty lines are constructed for each of three geographical subdivisions: Guatemala City, 'other urban" areas and rural areas. 15. The problems associated with this approach are not intrinsic to it, but rather, arise in how it is implemented. The present case of Guatemala provides a good illustration of several potential pitfalls associated with this method. In particular, three separate issues arise: i) choosing the basket of food goods, ii) pricing the basket, and iii) choosing the food share."' 16. The basket of basicfood goods: While the specification of the basket of basic food commodities is anchored to a fixed food intake requirement, this requirement itself provides little guidance as to what food items and quantities should be included. An endless number of food commodity combinations can satisfy the food energy requirement. And finding the minimum cost basket which fulfills the requirement is likely to result in a bundle of food items unpalatable to even the poorest households in most countries. Therefore it is desirable to have the food bundle be consistent with local food tastes and habits. 17. A common approach to address this problem is to rely on the actual consumption behavior of an appropriate group of households as identified by a household expenditures survey. By this method, a food basket that is representative of the consumption patterns of those just meeting the caloric requirements can be defined as an appropriate reference standard. Calculations to this effect were made by the Instituto Nacional de Estadfstica (INE) in Guatemala, based on the most recent national household expenditure survey.Y INE identified the per capita income interval in which households attained the 2' It should be noted that defining the food energy requirement is also subject to controversy. Agreement is compliated by the fact that energy intake needs vary across individuals according to age, gender and activity levels, as weU as by other unobservablephysical attributes such as individual metabolic rates. In the case of Guatemala, a delincated average daily caloric requirement of 2,150 per person is assumed to be given and the focus is on other aspecs of the medhodology. The choice of 2,150 calories per person per day is based on the recommendations of an FAO/WHO/UNU Consulative Group of nutrition experts. Y The household expenditure survey used as the basis for determining expenditure pattems was the Ecwusta Nacionl de Ingresos y Casios Familiares, conducted between November 1979 and August 1981. Unfortunately, household expenditure pattems have undoubtedly changed during the following decade, so that the contents and composition of the food basket should Annex 1 Page 5 of 10 caloric requirement for each of the three main geographic breakdowns. These households were found to be between the 23rd and 55th percentiles of the income distribution in Guatemala City, between the 50th and 83rd percentiles in urban areas, and between the 81st and 97th percentiles in rural areas. Next, food consumption bundles were selected based on these reference households' average consumption expenditures. 18. However the separate calculation of food consumption bundles by this method embodies the very disparities in welfare across regional subgroups which a poverty analysis must avoid. For example, since wealthier individuals tend to eat much more expensive calories, by construction they will attain their caloric requirements at higher expenditure levels. (It costs much less to reach 2,150 calories through the consumption of corn than it does through caviar.) This can yield a poverty line which is higher for richer groups -- precisely because they have higher living standards and more expensive tastes. As urban areas tend to be wealthier, the method generally produces more generous poverty lines for urban than for rural areas--even allowing for differences in prices--and hence, overstates urban relative to rural poverty. This is not a sensible thing to build into the poverty line and indeed often leads to sector rerankings and erroneous policy prescriptions. 19. A casual perusal of Table I of this annex suggests that a great deal of the diversity among the three separate food baskets appears to be attributable to variations in welfare across the regional subgroups. The most striking example is given by the tortilla and maize allowances. Tortillas account for 105 grams in the Guatemala City basket versus only 11 in the rural basket. In contrast, corn and its derivatives are much more significant in the rural basket: 284 grams versus 66 in the capital. Based on this rural food basket, nearly half (1042) of the caloric requirement is obtained from maize, versus just over one tenth in the Guatemala City basket (and a third in the other urban basket). These amounts may well accord with observed consumption habits. But, such habits are likely to be due to differences in living standards as well as differences in preferences. It has often been observed in Latin America that the poor and rural dwellers buy corn flour and make their own tortillas at home. Ready made tortillas are either unavailable where they live, or making them saves money. However, the diet is essentially the same. The difference is not so much in the corn consumed but whether the processing to form tortillas is done within the home or is paid for by the consumer. The latter costs more but saves leisure. A household who can afford to buy ready made tortillas is presumably better off than one who can not. 20. Similarly, it is difficult to believe that the carbonated beverages and bread (French and sweet) which figure more prominently in the Guatemala City basic food basket are not also primarily driven by higher living standards there. Seventy-four percent of the caloric requirement is attained from corn, sugar, beans and cooking oil in the rural basket as opposed to 57 and 36 percent in the "other urban" and capital city baskets, respectively. In general, the Guatemala City basket allows for more calories to be derived from vegetables, fruits, milk, eggs and meat -- all likely to be expensive sources of calories (relative to corn) and commodities whose consumption rises with income. Clearly, these baskets reflect living standard differences between Guatemala's geographical areas. Thus, the weights used by INE to measure differences in welfare across regions are themselves biased. Since these consumption bundles do not embody an identical standard of living, they are unlikely to yield a consistent poverty profile. 21. In order to avoid the problems associated with variations in welfare of the three food consumption bundles across geographic subgroups, the basket of basic food goods was fixed at the rural be updated to account for these changes. However, in the absence of more recent data, the rural basic food baskct listed in Table 1 of this annex is likely to be the most accurate estimatc possible of the food consumption patterns of someone just attaining the minimum daily requirement of 2,150 calories. Page 6 of 10 food basket across all regions in this analysis. 22. Costing the basket offood goods. The next step is to cost the food basket at the prevailing food prices in each of the three geographic regions. In a poor country like Guatemala with a large agriculural sector, it is common for there to exist substantial food price differentials based on geographic location. Pricing the basket was straightforward for Guatemala City and 'other urban' areas since price data for these areas are collected at regular intervals; the 1990 prices collected by INE show a 10 percent differential in favor of Guatemala city.Y But, a problem arose in pricing the rural bundle since rural prices are not collected on a consistent basis in Guatemala. The collection of these rural prices should be a high priority in the near future. In the meantime, and in the absence of any hard data, the only recourse was to try to make a reasonable estimate of the rural to urban price differential. In this analysis, rural food prices were assumed to be 20 percent lower than food prices in Guatemala City, which corresponds to rural food prices that are 10 percent lower than in "other urban" areas. 23. The extreme poverty line for each geographical region was set equal to the cost of the corresponding basket of food goods. 24. Allowance for non-food basic needs. Lastly, a poverty line also needs to allow for non-food basic needs, such as housing, clothing, health and education, transport and so on. However, there are no universal standards or guidelines for defining minimum needs in these categories. Again, the common practice is to turn to household behavior as a guide, and this is what has been done here. Estimates were made of the food shares of the poorest household income interval as identified by the 1979-1981 household expenditure survey.A' This group corresponds to households with incomes in the 0 to 1200 quetzales range who comprise approximately the bottom quartile of the income distribution (Annex B, Table 2).1" 25. The final poverty line for each geographical region was determined by dividing the cost of the food consumption bundle by the share of food in total expenditures. These are given in Table 3. i Pnces for Guatemala City and "other urban" areas are readily available from INE, Boletin Anual, 1990 and 1991. Z See footnote 4 of this annex. Il' Unfortunately, only household-level data was available for makdng these calculations. Thus houschold size could not be taen into account, with consequent variations in per capita expenditure levels wihin this refcrence group, pualculary across geographic regions. Spatial cost-of-living differentials have not been accounted for either. Thercfore, it is unlikly that dte households reflect similar welfare levels. But in the absence of more detailcd data, this approach is the most ccurae poibl. Annex 1 Page 7 of 10 Table A1.2: Basic Food Basket Composition Across Regions Basic Food Basket (grams/person/day) Food Items Guatemala City Other Urbau Rural 1. French Bread 44.959 15.730 5.206 2. Sweet Bread 40.111 24.488 12.131 3. Spaghetti 6.835 5.087 3.948 4. Rice 25.025 19.688 19.342 5. Tortillas 104.704 43.734 10.830 6. Corn 65.682 186.321 283.980 7. Other Cereals 43.231 35.341 5.920 8. Potatoes 28.717 36.002 27.929 9. Onions 35.157 16.825 13.127 10. Sugar 72.454 66.297 68.418 11. Beans 38.668 50.347 55.161 12. Other Legumes 0.429 0.328 0.255 13. Tomatoes 44.422 30.046 28.305 14. Guisquil 8.702 6.450 4.390 15. Other vegetables 35.392 21.080 16.456 16. Bananas 16.233 15.942 12.728 17. Plantain 22.273 16.553 8.991 18. Oranges 26.289 13.915 11.068 19. Other Fruits 74.482 47.517 39.847 20. Beef 53.937 50.288 45.442 21. Pork 10.023 9.707 7.151 22. Poultry 37.392 24.640 23.856 23. Other meat 6.214 3.583 2.716 24. Fresh fish 8.158 10.269 10.667 25. Milk 110.941 92.467 83.713 26. Cheese, Milk Products 33.082 31.230 22.147 27. Eggs 35.263 29.706 20.963 28. Coffee 4.846 3.647 2.784 29. Carbonated Drinks 73.365 59.272 51.923 30. Cooking Oil 14.939 12.653 10.409 31. Lard, Fatb 0.586 0.423 0.344 32. Margarine 3.426 1.140 0.465 Total Gnms 1125.9 980.7 910.6- Total calonres 2,150 2,150 2,150 Source: INE, July 1991. Annhl Page 8 of 10 Table A13. Per Capita Expenditures, Food and Population Shares By Region and Family Income Class Total Expenditures per Food Share Population Capita Frequency Annual Family Income Class 0 - 1200 Quetzales Guatemala City n.a .544 4.2 "Other Urban" n.a .633 14.2 Rural n.a .669 35.5 Source: Direccidn General de Estadfstica, 1984. Table A1.4. Alternative Poverty Lines by Region (quetzales/person/month) Poverty Lines Guatemala City 'Other Urban' Rural Alternative Poverty Line Extreme 53.62 4A.3 ' 44.68 Poverty ;. Poverty 98.57 77O~66.79 Source: [NE, July 1991 and calculations based on Direcci6n General de Estadfstica, 1984. Note: The poverty lines are in March 1989 prices. Annel Page 9 of 10 D. POVERTY MEASURES Poverty Measures 26. In this study, the three empirical measures employed to document the extent and depth of poverty in Guatemala are the poverty headcount ratio, the poverty gap index, and the Foster-Greer-Thorbecke (FG7) P2 index. The poverty headcount ratio is a measure of the percentage of the population whose income is less than the poverty line. The poverty gap index is a measure of the average shortfall in the income of a poor person from the poverty line, expressed as a percentage of the poverty line. The FGT P2 index is a distributionally sensitive measure which assesses the severity of poverty; by this index, poor individuals are given an increasing weight according to the magnitude of their income shortfall from the poverty line. The measures used to demonstrate income and land ownership inequality are the decile income distribution, the Lorenz curve and the Gini coefficient.2 Poverty Headcount Ratio The poverty headcount index (H) is the proportion of the population whose income (y) is less than the poverty line (z). Where q denotes the number of poor individuals in a population of n, the headcount ratio is: HI=q/n. This is one of the most popular measures of poverty because it is easy to understand and interpret. However it is unable to cast any light on the depth of poverty among the population. Pove r Gar) The poverty gap (PG) is a measure of the depth of poverty. It is the difference between the poverty line (z) and the mean income of the poor (u), expressed as a ratio of the poverty line (also known as the 'income gap ratio"): I 1 - HZ and multiplied by the poverty headcount ratio (H). Therefore the poverty gap can be expressed as: PG = I *H. The poverty gap is sensitive to the deficit of the poor relative to the poverty line. A limitation of the a For the decile income distribution, each decile represents ten percent of the population mnked in tennrs of pcr capita household income. The Lorenz curve shows the cumulative distribution of income from the poorest to the wealhiest individual, buae on corresponding percentage shares of population and total national income. Thus, as the sharm of the popultion included in the distribution grows larger, so does the cumulative income share of that population (Psacharopoulos et al., 1992). Te Gini coefficient is an inequality index which is defined u the ratio of the area enclosed between a 45 dege line and a Lorenz ctuve (a plot of cumulative ime distribution based on perentage shares of population and total national incomc), to the catire triangle enclosed by the 45 degree line. When a large percentage of total national incomc is conocated among a reaively small percentage of individuals, the overall Lorenz curve tends to be lower, with a sharp rise in incomc share at the top of the income distribution. Thercfore, in general, the Gini coefficient will increae when the distribution of inoome beoom morm skewed (Psacharopoulos et al., 1992). Page 10 of 10 headcount ratio and the poverty gap measures is that they are not sensitive to the distribution of income within the poor population. Foster-Greer-Thorbecke (FGTn P Index The FGT P2 index is a measure of the severity of poverty. By this index, each poor person is weighted by the size of their income shortfall from the poverty line. For example, two separate distributions which each contain two poor people may have an identical headcount ratio and poverty gap. However, if one distribution includes a very poor person and a marginally poor person, it will rank worse than an alternative distribution containing two equally moderately poor individuals. This is because the FGT P2 index is exponentially sensitive to increasing severity of poverty at the disaggregated individual level, thereby taking into account the distribution of income within the poor population. The FGT P2 index can be expressed as: PG2/H + ((H-PG)2/H) * CV 2 where CV.2 denotes the squared coefficient of variation of income amongst the poor (Ravallion 1992: 36-39). Annex 2 Page 1 of 1 This table presents estimated poverty levels from studies conducted between 1980 and 1989. Table 1. Estimated Poverty Levels from Other Studies Poverty Study Unit of |Poverty Rate Analysis 1980 1986-1987 1989 | Hintermeister (1984) '' individuals 70.8 -- -- CEPAL (1991) a' individuals 71.1 73.2 -l urban 47.0 60.3 -- rural 83.7 79.7 -- CEPAL (1991) 2/ households 65.4 67.6 -- urban 40.7 54.4 -- rural 79.2 75.4 72.03/ Melendreras and Cabrera households 63.4 83.4 89.0 (1991) 4/ Instituto Nacional de individuals -- -- 80.0 Estadistica (1991) 1' households -- -- 75.5 Psacharopoulos et al., individuals -- 66.4 70.4 (1992) 6' Sources: 1/ Hintermeister, A. 1984. "Rural Poverty and Export Farming in Guatemala." World Emplovment Programme Research. 2/ CEPAL. 1991. Magnitud de la Pobreza en America Latina en los Afios Ochenta. Estudios e Informes de la CEPAL 81. 3/ CEPAL. 1992. "Latin American Poverty Profiles for the Early 1990s." Poverty rate for rural area in 1989 is a preliminary estimate. 4/ Melendreras and Cabrera. 1991. Proyecto Instituto Nacional de Salud. "El Mapeo de la Pobreza en Guatemala." Ministerio de Salud Pudblica y Asistencia Social. 5/ Instituto Nacional de Estadistica (INE). 1992. "Perfil de la Pobreza en Guatemala." 6/ Psacharopoulos et al., 1992. Poverty and Income Distribution in Latin America: The Story of the 1980s. The World Bank. Note: Year in parentheses is the date the study was carried out. Annex 3 Page 1 of 9 The following tables report poverty differences across first regions and then departments. The departmental poverty measures are presented with their standard errors calculated using the methodology described in Kakwani, "Testing for Significance of Poverty Differences", LSMS Working Paper no. 62. The World Bank, 1990. Having the standard errors allows one to test whether any departmental poverty measure is statistically significantly different from the national average. These teste are presented in Tables A3.6 and A3.7. Table A3.1: REGIONAL DISTRIBUTION OF SELECTIVE SAMPLE CHARACTERISTICS Sample Size Percent Average Percent of Households (Individuals) Indigenous Family l Size Count % Living in Living Urban with Area Family Head 1. Metropoljian 11,596 24.2 1 0.0 4.8 74.3 19.() 2. North 4,317 9.(0 85.5 5.7 17.9 15.3 3. North-East 3,766 7.9 8.8 5.3 21.9 17.4 4. South-East 4,273 8.9 1.6- 5.20.8 17.7 5. Central 6,148 12.8 36.0 5.2 44.7 12.0 6. South-West 10,499 21.9 46.4 5.6 23.9. 15.1 7. North-West 3,955 8.3 85.0 6.0 15.0- 12.0 8. Peten 3,296 6.9 12.6 5.5 2.2 15.7 ...W.,:,......... 4.....35. Souce 100.0E 3coa 5.3 1989. Source: Encuesta Nacional Sociodemogrcifica, 1989. Annex 3 Page 2 of 9 Table A3.2: REGIONAL POVERTY RATES By Region, Urban/Rural Location and Departamento Headcount Ratio (H) Region Urban Rural Total Very Poor Poor Very Poor Poor Very Poor Poor Metropolitan Guatemala 20.3 47.1 55.4 74.6 29.3 54.2 North 46.4 66.9 91.5 96.7 83.4 91.3 Baja Verapaz 48.2 64.9 94.1 98.6 81.3 89.2 Alta Verapaz 45.3 68.2 90.7 96.1 84.1 92.1 crth-1~~~~~st 38J ~~~ 6L. 55. t 1. $14 6J Izabal 35.2 52.8 39.9 57.2 39.5 56.8 Zacapa 31.9 65.4 67.0 79.4 52.4 73.6 El Progreso 51.4 66.9 51.0 79.1 51.1 76.4 Chiquimula 38.6 60.6 90.2 96.0 74.2 85.0 WuJiI~m i--;::- 3 .- ; 7: 7: 67,7 19-7.4 82- 648 Santa Rosa 39.4 63.8 65.1 08.0 58.1 75.6 Jutiapa 32.8 62.9 72.5 83.8 67.1 81.0 Jalapa 63.3 80.1 74.8 85.9 72.0 84.5 ,..,. .... . -...-- .... ... f , 547 7357 7756 764 Sacatepequez 58.5 76.3 36.9 65.3 46.1 70.0 Escuintla 41.1 60.7 60.4 77.3 52.9 70.9 Chimaltenango 65.6 85.4 75.7 91.9 70.2 88.3 ~~oiztk-WMt 396 60;;4i ; -.Q 8-.; 6- 0 ''2"'" Quetzaltenango 39.3 61.5 67.1 83.8 54.1 73.4 Retalhuleu 30.1 54.0 55.7 82.1 51.3 77.2 Suchitepequez 42.8 63.6 64.2 90.8 57.4 82.2 Solola 51.7 63.5 79.1 88.9 74.8 84.9 Totonicapan 58.4 68.7 80.6 93.8 75.8 88.4 San Marcos 23.6 49.5 81.9 94.6 76.0 90.1 ,.N,,-. ,. 67 , 7.1 ,,, 93. Huehuetenango 59.0 73.0 94.7 96.9 87.6 92.1 Quiche 92.3 92.3 87.8 96.0 88.2 95.7 Peten 34,8 59.0 68.1 82.1 58.4 7L4: Total 33.7 57.2 71.9 85.7 57.9 75.2 Source: Encuesta Nacional Sociodemografica, 1989. Note: The headcount ratio of poverty (H) is the proportion of the population whose income (Y) is less than the poverty line (z). Hence, H =q/n, where q is the number of people who are deemed poor in a population of n. Figures under the heading "extremely poor" are the headcount ratio of poverty for the population below the extreme poverty line; figures under "poor" indicate the headcount ratio under the poverty line. Annex 3 Page 3 of 9 Table A3.3: REGIONAL POVERTY INDICES Level and Rank by Departamento (rank in parenthesis) Region Headcount Ratio Poverty Gap FGT Index Guaah 54.2 (1) 23.9 31) 13.6 (1) North Baja Verapaz 89.2 (1)6. 1) 48.0 (9 Alta Verapaz 92.1 ) 6. 4 (0 507 ( North-East El Progreso 76.4 8 9 37.2 t 20.4 ( Izabal 56.8 (2.. UIi =: 27.0 17.0(2 Zacapa 73.6 40.9 ~ 1) 28.3 (0 Chiquirnula 85.0 (4 56.7 43.1 South-East Santa Rosa 75.6 ) 39.9 ( 25.5 9 Jalapa 84.5 (1)55.0 (15)i 41.4 (5 Jutiapa 81.0 ( 49.7 35.5 Central Sacatepequez 70.0 ) 36.6 ( 24.4 ( Escuintla 70.9 (16 35.9 (4) 22.0 (12 Chimnaltenango 88.3 ()52.4 (1) 36.7 (~ South-West Solola 84.9 (1)57.4 1) 4.4.0 (1 Totonicapan 88.4 17iX 54.5 (14) 38.8 () Quetzaltenango 73.4 z) 37.7 23.4 Suchitepequez 82.2 ( 40.2 (. 24.0 ) Retalhuleu 77.2 35.3 20.4 ( San Marcos 90.1 (I)60.2 (I) 46.0 (S North-West Huehuetenango 92.1 (2)70.0 (Z~ 56.3 ) Quiche 95.7 ~ 2~ 68.8 (2) 53.8 21 Peten 75.4 (7) 43.0 (1) 29.7 (1 7S.2 43.9 30.S Source: Encuesta Nacional Sociodemogrdfica, 1989. Annex 3 Page 4 of 9 Table A3.4: REGIONAL POVERTY RATES By Ethnicity and Departamento Headcount Index (H) Region Percent Indigenous Non-Indigenous Total Percenous Extremely Poor Extremely Poor Extremely Poor poor poor poor Metropolitan Guatemala 61.2 82.2 25.8 51.1 29.3 54.2 9.7 .Noth 88.9 94.9. 50.9 70.1 83.4 91.3 84.3 Baja Verapaz 95.5 99.1 48.5 66.6 81.3 89.2 67.9 Alta Verapaz 87.2 93.9 53.6 74.1 84.1 92.1 90.0 .WNrth-Eaat 82.9 91.6 48.3 67.2 51.4 69.3 8.7 Izabal 41.0 53.8 39.4 56.8 39.5 56.8 0.2 Zacapa 0.0 100.0 52.5 73.5 52.4 73.6 2.8 El Progreso 100.0 100.0 51.0 76.4 51.1 76.4 0.0 Chiquimula 88.7 96.6 66.3 78.7 74.2 85.0 34.8 South-East 86.6 95.5 64.5 79.5 64.8 79.7 71.7 Santa Rosa 100.0 100.0 58.0 75.6 58.1 75.6 0.1 Jutiapa 100.0 100.0 67.1 80.9 67.1 81.0 7.5 Jalapa 85.7 95.2 71.0 83.7 72.0 84.5 0.1 (:eutral . 72.3 90.2 47.6 68.1 56.5 76.0 36.4 Sacatepequez 67.0 85.1 35.8 62.5 46.1 70.0 33.7 Escuintla 58.1 79.7 52.6 70.3 52.9 70.9 79.5 Chimaltenango 75.7 93.2 48.3 69.2 70.2 88.3 6.8 South-West 76.5 91.0 55.1 75.4 65.0 82.7 46.2 Quetzaltenango 66.3 85.5 43.7 63.2 54.1 73.4 83.6 Retalhuleu 59.0 86.2 48.9 74.5 51.3 77.2 90.0 Suchitepequez 67.5 92.5 47.0 71.5 57.4 82.2 46.6 Solola 83.0 92.0 27.0 43.3 74.8 84.9 51.6 Totonicapan 77.8 90.8 56.3 64.8 75.8 88.4 23.8 San Marcos 96.4 98.5 69.5 87.4 76.0 90.1 24.5 WW¢ : . : 91.2 96.8: 68.7 76.1 87.9 93.7 844 Huehuetenango 93.4 97.0 61.2 69.6 87.6 92.1 81.9 Quiche 88.8 96.6 83.3 88.7 88.2 95.7 87.7 Petesi 77.4 92.8 55.6 72.9 58.4 75.4 12.7 Total 81.3 92.6 45.2 65.8 57.9 75.2 36.3 Source: Encuesta Nacional Sociodemogrdfica, 1989. Note: The headcount ratio of poverty (H) is the proportion of the population whose income (Y) is less than the poverty line (z). Hence, H=q/n, where q is the number of people who are deemed poor in a population of n. Figures under the heading "extremely poor" are the headcount ratio of poverty for the population below the extreme poverty line; figures under "poor" indicate the headcount ratio under the poverty line. Annex 3 Page 5 of 9 Table A3.5: Inter-Department Poverty Comparison By Urban/Rural Breakdown Poverty Indices Region Department Area N Headcount Poverty Gap Foster-Greer- Index (H) Index (PG) Thorbecke Index (FGT P2) (standard errors in parentheses) 1. Metropolitan Guatemala Urban 8,618 47.1 18.9 9.9 (0.54) (0.27) (0.18) Rural 2,978 74.6 38.3 24.3 (0.80) (0.57) (0.47) 2. Nordt Baja Verapaz Urban 305 64.9 36.0 24.5 (2.74) (1.95) (1.64) Rural 790 98.6 72.7 57.1 (0.42) (0.74) (0.90) Alta Verapaz Urban 468 68.2 36.8 23.9 (2.16) (1.49) (1.25) Rural 2,754 96.1 70.2 55.3 (0.37) (0.47) (0.53) 3. North-East El Progreso Urban 148 66.9 34.0 19.4 (3.88) (2.31) (1.68) Rural 531 79.1 38.1 23.3 (1.77) (1.29) (1.08) kabal Urban 159 52.8 22.0 12.4 (3.97) (2.19) (1.52) Rural 1,472 57.2 27.6 17.5 (1.29) (0.82) (0.65) Zacapa Urban 260 65.4 26.0 14.1 (2.96) (1.68) (1.26) Rural 364 79.4 51.5 38.5 (2.12) (1.82) (1.68) Chiquimula Urban 259 60.6 28.9 18.2 (3.04) (1.95) (1.56) Rural 573 96.0 69.2 54.4 (0.82) (1.06) (1.18) Annex 3 Page 6 of 9 Poverty Indices Region Department Area N Headcourt Poverty Gap Foster-Greer- Index (H) Index (PG) Iborbecke Index 4. South-East Santa Rosa Urban 437 63.8 29.3 17.1 (2.30) (1.40) (1.04) Rural 1,165 80.0 43.9 28.7 (1.17) (0.90) (0.78) Jalapa Urban 226 80.1 50.4 37.7 (2.66) (2.35) (2.11) Rural 710 85.9 56.5 42.5 (1.31) (1.22) (1.17) Jutiapa Urban 232 62.9 27.4 16.4 (3.18) (1.96) (1.52) Rural 1,503 83.8 53.2 38.4 (0.95) (0.82) (0.77) 5. Central Sacatepequez Urban 684 76.3 44.6 31.4 (1.63) (1.30) (1.15) Rural 919 65.3 30.6 19.2 (1.57) (1.03) (0.87) Chimaltenango Urban 1,038 85.4 51.3 36.6 (1.10) (1.00) (0.93) Rural 856 91.9 53.7 36.8 (0.93) (0.97) (0.93) Escuintla Urban 1,029 60.7 29.0 17.1 (1.52) (0.92) (0.67) Rural 1,622 77.3 40.3 25.1 (1.04) (0.74) (0.61) Source: Encuesta Nacional Sociodemogr4fica, 1989. Note: Entries with "'" are significant at 5 percent level (Z.DS= 1.96), while those with "**" are significant at 1 percent level (Z.0,=2.576) for the two-tailed test. Test statistic is given by Z=(P,-Pi)/SE(P,-Pj), where Pi and P, are estimated poverty rates for sub-regions, and SESE(P,-P V(VP,)/N,+P,)/N, is the standard error of estimates of poverty difference (Nanak Kakwani, 'Testing for Significance of Poverty Differences," LSMS Working Paper No. 62. The World Bank, 1990). And the subscript i refers to the subarea to be compared, and j to the national average. Therefore, the incidence of poverty is less than national average in locations with Z-score less than zero and significant. Annex 3 Page 7 of 9 Poverty Indices Region Department Area N Headcount Poverty Gap Foster-Greer- Index (H) Index (PG) Tborbecke Index 6. South-West Solola Urban 230 63.5 40.5 29.2 (3.18) (2.36) (2.06) Rural 1,248 88.9 60.5 46.7 (0.89) (0.90) (0.94) Totonicapan Urban 166 68.7 39.4 25.2 (3.61) (2.42) (1.87) Rural 612 93.8 58.6 42.5 (0.98) (1.16) (1.21) Quetzaltenango Urban 1,113 61.6 29.5 18.0 (1.46) (0.92) (0.72) Rural 1,269 83.9 45.0 28.2 (1.03) (0.79) (0.69) Suchitepequez Urban 467 63.6 32.0 18.9 (2.23) (1.36) (1.00) Rural 1,004 90.8 44.0 26.3 (0.91) (0.83) (0.76) Retalhuleu Urban 239 54.0 25.2 15.1 (3.23) (1.92) (1.44) Rural 1,149 82.1 37.4 21.6 (1.13) (0.81) (0.67) San Marcos Urban 301 49.5 20.1 10.7 (2.89) (1.48) (1.06) Rural 2,701 94.6 64.6 49.9 (0.43) (0.55) (0.58) 7. North-West Huehuetenango Urban 437 73.0 48.4 36.9 (2.13) (1.76) (1.63) Rural 1,750 96.9 75.4 61.2 (0.41) (0.50) (0.57) Quiche Urban 156 92.3 67.9 52.8 (2.14) (2.07) (2.28) Rural 1,612 96.0 68.8 53.9 (0.49) (0.63) (0.70) 8. Peten Peten Urban 962 59.0 26.9 15.8 (1.59) (0.95) (0.71) Rural 2,334 82.1 49.7 35.4 (0.79) (0.68) (0.63) Source: Encuesta Nacional Sociodemogrdfica, 1989. Anne-x 3 Page 8 of 9 Table A3.6: Test Statistics for Significance of Poverty Difference Between Individual Regions and National Average Asymptotic z for Region Area Headcount Poverty Gap Foster-Creer- Index (H) Index (PG) ThorbedOe Index _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ .YG T P2) 1. Metropolitan Urban 49.10 4' -80.12 ** -91.25 4* Rural -0.76 -9.61 ** -13.25 4' 2. North Urban 4.90 ** -6.22 ** 6.62 ** Rural -59.50 *" -62.67 ** 52.21 * 3. North-East Urban -7.92 ^ -16.00 ** -19.07 * Rural -4.41 # -5.12 ** -4.21 4* 4. South-East Urban -4.78 * -8.97 ** -9.63 *4 Rural -11.35 ' 11.85 ** 9.79 * 5. Central Urban -1.55 4.06** Rural 3.39 ** -5.20 ** -9.07 4'e 6. South-West Urban -14.90 ** -21.86 ** -24. S Rural 36.65 '4 26.58 ** 20.63 7. North-West Urban 1.66 6.61 ** 7.48 * Rural 56.86 ** 65.37 ** 56.64 ". 8. Peten Urban -10.13 * -17.80 ** -20.55 4. Rural 8.44 *' 8.33 ** 7.16 A Annex 3 Page 9 of 9 Table A3.7: Test Statistics for Significance of Poverty Difference Between Department and National Average Asymptotic Z for Region Deportrnenit Area }Irerdcoung Poverty Gaip FotrGreerr- Index (11) Index (PG) lborbeckie lmlex (F'T P3) 1. Metropolitan Guatemala Urban -49.10" -80.12 *s -91 25 ' Rural -4.76 -9.61 se -13 25 ' 2. North Baja Verapaz Urban .3.76" -4.05 ** 3.R* *" Rural 50.64 *: 38.06 ** 2d.9# s Alta Verapaz Urban -1.27 * 4.75 ** -.5 48 *' Rural 49.97 "* 53.38 ** 44 901 * 3. North-East El Progreso Urban 2.15 4.27 e* -6.72 " Rural 2.17 ' -4.46 e* *6.h0 so Izabal Urban 5.1AaS " -10.02 ** -12.t3 "* Rural -13.82 ' -19.58 es -20."5 * Zacapa Urban i3 33 " -10.61 ** *13.15 " Rural 1 95 4.17 so 4.601 " Chiquimula Urban -4.80 " -7.65 ** b 12 * Rural 24.58 *' 23.56 ** 10.91 to 4. South-East Santa Rosa Urban 4 93 -10.42" 3.07. Rural 4.(11 so 4001 ' 65 ' Jalapa Urban 1.82 2.74 so ;.30 ' Rural q ns * 10.24 ** 10.01 ' Jutiapa Urban 3.86 " -8.41 s .9 42 so Rural 8.78" 11.08 so " 83 s 5. Central Sacatepequez Urban 0.66 ().51 11.5, Rural 428' *' -12.73s "! -i to Chimaltenango Urban 9.07 "' 7.32 ** 6 23 "* Rural 17.55 's 9.96 ** (s.43 's Escuintla Urban 'i s4 " -16.04 ** 19 '93 * Rural 96 * -4.81 ** 9.904 6. South-West Solola Urban -3.a)' es -1.45 -0 ?' Rural A4 93 *" 18.15 *e I16 87 "' Totonicapan Urban - 81 -1.85 .2.9 '' Rural I S.63 *" 12.54" 9.64 ^- Quetzaltenango Urban 9.30 ' -15.54 ** .17 36 " Rural K. lb *' 1.29 .3.69e Suchitepequez Urban -5 20 " -8.71 ** -11.76 * Rural 16 74 " 0.09 .5.75 s Retalhuleu Urban J.57 " -9.70 se -10 78 "6 Rural 5.95 " -7 88 ** 13.52 O' San Marcos Urban -8 49 ' -15.95 se -il M " Rural 40.70 "' 36.33 s* 31 9s 7. North-West Huchuetenango Urban -1.05 2.55 * 3.76 ' Rural 47 31 * 59.97 so 51.54 "0 Quiche Urban 7.94 " 11.58 ** 9.64 's Rural 39.61 " 38.15 ** 32.27 0 8. Peten Peten Urban 10.13 " -17.80 ** -20.55"* Rural Y.44 * 8.33 ?* '.16' Source: Encuesta Nacional Sociodemogrdfica, 1989. Annex 4 Page 1 of 4 The following chapter shows the differences in the effect of illness on poor and non-poor by locality, poverty group and type of facility visited. Conclusions are that mean length of days inactive is longer for poor than non-poor, possibly because the poor live in rural areas which inhibits travel to distant medical facility, are more likely to receive lower quality medical care or no medical care, and do not have the income for medical expenses. Table A4.1: PERCENT OF INDIVIDUALS WHO HAVE BEEN ILL OR IN AN ACCIDENT DURING THE PAST FOUR WEEKS By Locality and Poverty Group Extremely Poor Poor Non-Poor TotW . Guatemala City 6.2 6.9 6.9 6.9 "Other Urban" Areas 4.5 4.9 7.0 5.6 Rural Areas 4.5 4.5 6.5 Total 4.6 4.9 6.8 5.3 Source: Encuesta Nacional Sociodemografica, 1989. Table A4.2: MEAN LENGTH OF ILLNESS AND INACTIVE DAYS DUE TO ILLNESS FOR THOSE ILL By Locality and Poverty Group (in calendar days) Extremely Poor Non-Poor Total Poor . . . . Mean Lergth of Illness (A) Guatemala City 13.4 13.4 12.6 12.9 "Other Urban" Areas 14.0 141 12.9 13.6 Rural Areas 14.0 14.1 14.1 .14.1 Total 14.0 14.0 13.2 13.7 Mean Length of Days Inactive (B) Guatemala City 7.6 6.9 6.5 6.7 l "Other Urban" Areas 6.8 6.3 4.9 5.7 l Rural Areas 7.5 574 5.4 7.0 Total 7.4 7.1 5.7 6.' ource: Encuesta Naaconaf Sociodemogrdfica, 1989 Note: The relevant time period is during the past four weeks prior to the survey, and those ill include people who suffered from an accident. Annex 4 Page 2 of 4 Table A4.3: PERCENTAGE OF ILL INDIVIDUALS WHO CONSULTED HEALTH PERSONNEL By Locality and Poverty Group [ Extremely Poor Poor Non-Poor To1tal Guatemala City 59.6 67.3 75.2 71.5 "Other Urban" Areas 66.2 65.4 70.8 07.7 Rural Areas 47.7 48.9 66.0 52.2 Total 51.5 54.6 70.8 59,7 Source: Encuesta Nacional Sociodemogrdfica, 1989 Note: The relevant time period is during the past four weeks prior to the survey, and those ill include people who suffered from an accident. Table A4.4: TYPE OF CONSULTATION By Ill People and Poverty Group (in percent) Extremely Poor Non-Poor Totl Poor Type of Consultation Medical Doctor 76.6 80.5 94.7 85.8 Nurse 16.4 13.0 3.5 9.5 Traditional Healer 2.2 2.3 0.6 1.7 Mid-Wife 0.5 0.3 0.2 0.3 Others 4.3 3.9 1.1 2.8 All Types 100.0 100.0 100.0 100.0 Source: Encuesta Nacional Sociodemogrdfica, 1989 Annex 4 Page 3 of 4 Table A4.5: TYPE OF CONSULTATION By Ill People and Locality (in percent) Guatemala City 'Other Ilkban" Rural Area Total Type of Consultation Medical Doctor 97.4 92.3 76.4 85;8 Nurse 0.7 4.6 16.5 9.5 Traditional Healer 0.2 1.3 2.6 1.7 Mid-Wife 0.0 0.0 0.5 0,3 Others 1.7 l.8 4.0 2.8 All Types 100.0 l00.0 100.0 100:0 Source: Encuesta Nacional Sociodemogrdfica, 1989 Table A4.6: PLACE OF CONSULTATION By Ill People and Poverty Group (in percent) Extremely Poor. Non-Poor Total Poor Place of Consultation Hospital 22.5 210 18.1 19.9 Health Center 30.6 26.0 9.4 .:< .B Social Security Institute 4.2 7.0 9.4 7.9 Private Clinic 32.5 35.9 51.1 41.6 Nurse's House 3.0 3,2 8.5 5.2 Others 7.3 6.9 3.4 :5.-6 All Types 100.0 ;00.0 100.0 . 100.0 Source: Encuesta Nacional Sociodemogrdfica, 1989 Annex 4 Page 4 of 4 Table A4.7: PERCENTAGE DISTRIBUTION OF TIME TAKEN TO REACH CONSULTATION By Type of Facility Visited by Locality Place of Guatemala City "Other Urban" Rural Area Consultation Ext. . Non- Ext. poor Non- Ext. Poor Non- Poor Poor Poor Poor Poor Poor Hospital 1. <30 Minutes 21.7 - 40.7 52.9 $7A 56.1 16.7 19.1 25.0 2. 30 Min-I Hr. 39.1 36.5 38.9 23.9 21.5 15.4 18.9 19.8 23.0 3. 1-2 Hours 34.8 ; 3.0A8 20.4 7.0 9.0 13.9 29.0 . 29.9 18.5 4. >2 Hours 4.4 -1. X4 0.0 16.2 12.1 14.6 35.5 31. 33.5 Health Center 1. <30 Minutes 91.4 .933I 63.6 94.6 95.9 67.2 47.5 5S2.1 76.0 2. 30 Min-I Hr. 8.6 6? 36.4 5.4 4.1 32.8 16.9 16.2 17.7 3. 1-2 Hours 0.0 0.0 0.0 0.0 0.O 0.0 18.4 16.7 4.3 4. >2 Hours 0.0 00 0.0 0.0 0.0 0.0 17.2 15.0 2.0 Social Sec. Inst. 1. <30 Minutes 42.9 44.4 35.4 100.0 100.0 89.3 28.7 21.7 21.0 2. 30 Min-1 Hr. 28.6 . LA1 35.4 0.0 0.0 0.0 27.6 35.0 42.0 3. 1-2 Hours 28.6 244 27.1 0.0 0.0 0.0 25.5 22.6 24.5 4. >2 Hours 0.0 . 2.1 0.0 0.0 10.7 18.2 20.7 12.6 Private Clinic 1. <30 Minutes 53.6 729 64.6 74.4 742 78.6 26.7 29.1 37.4 2. 30 Min-I Hr. 28.6 -91 25.0 10.4 9.9 8.3 26.4 24.6 21.1 3. 1-2 Hours 17.9 . 3 9.2 7.8 10.2 5.4 26.6 27.0 17.1 4. >2 Hours 0.0 .O. 1.2 7.4 5.6 7.7 20.3 19.3 24.5 Nurse's House !. Others 1. <30 Minutes 88.9 93.1 66.7 70.5 58.1 100.0 58.1 51.7 31.5 2. 30 Min-i Hr. 0.0 .0 33.3 15.9 32.1. 0.0 20.5 23.3 32.7 3. 1-2 Hours 11.1 0 13. 39. 0.0 10.9 15.1 35.7 4. >2 Hours 0.0 0-.0. 0.0 0.0 0.0 0.0 10.5 9.9 0.0 Overall 1. <30 Minutes 53.9 62.7 55.4 74.6 75.2 76.1 35.7 37.2 44.5 2. 30 Min-i Hr. 24.7 - 30.1 12.4 12.0 10.5 20.8 21.0 21.7 3. 1-2 Hours 20.2 143 13.5 6.0 7.6 5.5 22.5 22.6 14.6 4. >2 Hours 1.1 04 1.0 7.1 5.3 7.9 21.1 19.2 19.2 Source: Encuesta Nacional Sociodemogrdfica, 1989. a/ No data was available on the time taken to reach the facility for people visiting nurses' houses for consultation. Annex 5 Page 1 of 4 The following Annex shows the differences in poverty level according to gender and locality. Male-headed households tend to be 48.5 percent poorer in rural areas than in Guatemala City while female-headed households are 46.8 percent poorer. Findings in Table A5.4 indicate that female-headed households are less likely to be poor (by department) due to the fact that they received more transfers. Table A5.1: DISTRIBUTION OF THE POOR By Gender of Household Head (in percent) [Extrndy Poor Nlon-Poor All Poor Male 86.6 85.4 80.9 84.0 Female 13.4 14.6 19.1 16.0 Total 100.0 100.0 100.0 100.0 Source: Encuesta Nacional Sociodemogrdfica, 1989. Table A5.2: HEADCOUNT POVERTY RATES By Gender and Locality Gender of Guatemala City "Other Urban" Rural Area Household Head Extremely Poor tremely Poor Extremely Poor Poor Poor Poor Male 21.7 48.5 39.5 64.8 67.2 83.0 Female 20.0 46.8 48.3 67.4 72.4 86.0 Source: Encuesta Nacional Sociodemogrdfica, 1989. Annex 5 Page 2 of 4 Table A5.3: HOUSEHOLD HEAD DIFFERENCES ACCORDING TO GENDER FenSles Males All Household Size 4.3 5.6 5.4 Age 48.9 42.8 43.7 Indigenous 28.1 37.2 35.8 Years of Schooling 2.6 3.1 23.0 Illiteracy Rate 53.9 36.9 39.6 Hours Worked/Week 19.9 45.5 41.5 Total Income (Quetzaleslmonth) 170.9 249.3 236.9 Occupation Professional 9.8 4.2 4.7 Manager 5.9 3.2 3.5 Office worker 2.4 2.2 2.3 Vendor 22.2 6.6 8.0 Farmer 18.9 56.8 53.4 Miner 0.0 0.2 0.2 Transportation 0.3 4.0 3.7 Artisans 17.3 14.9 15.1 Manual workers 1.6 4.6 4.3 Service sector 21.5 3.2 4.9 N 1,428 7,529 8,957 Source: Encuesta Nacional Sociodemogrdfica, 1989. Note: N = number in this analysis. While the original survey had a sample size of 8,988 households, only those household heads with positive income were included in this analysis. Annex 5 Page 3 of 4 Table A5.4: SOURCE OF INCOME By Poverty Group Extremely Poor Poor Non-Poor Total Males and Females Principal Income 86.2 88.2 881 88.1 Secondary 1.8 1.3 ?.3 2.1 Retirement Income 2.6 2.0 2.3 2.2 Other Transfers 4A 4.2 5.6 ; 5.4 In-Kind 5.0 4.3 1.8 2.1 Males Principal Income 904 92.2 91.3 91.4 Secondary 2.1 1.3 . 2.3 Retirement Income 1.7 1.3 .6 1.6 Other Transfers : .8 1.8 3. 3.0 In-Kind 4.0 3.4 1.5 1.8 Females Principal Income 67.7 71.2 72.9 72.6 Secondary 0.7 1.0 1.4 1.3 Retirement Income 6.4 5.2 5.4 5.5 Other Transfers 15.8 14.7 17.2 16.8 In-Kind 9.4 7.9 3 3.8 Source: Encuesta Nacional Sociodemogrdfica, 1989. Annex 5 Page 4 of 4 Table A5.5: POVERTY RATES BEFORE AND AFTER INCOME TRANSFER By Departamento Extreme overty Rate Povert te BseforeAn[er Before Arier D%p aranento Gender Transfer Transfer Transfer Transfer buatemala Al-lf 27.VT 2T ,r., 49.9 Male 27.1 25.8 50.0 49.2 Female 36.9 25.3 57.6 47.8 Progreso All 55.9 49.6 75.6 70.9 Male 53.4 50.5 73.8 70.9 Female 66.7 45.8 83.3 70.8 Sacatepequez All 45.0 42.4 66.9 63.9 Male 44.6 44.2 67.8 67.0 Female 48.6 28.6 60.0 40.0 Chimaltenango All 65.8 65.3 85.0 84.5 Male 65.6 65.3 84.7 84.1 Female 67.5 65.0 87.5 87.5 Escuintla All 46.2 44.3 63.8 62.3 Male 46.4 44.8 63.7 62.6 Female 45.2 40.3 64.5 59.7 Santa Rosa All 53.4 51.5 71.0 69.7 Male 53.9 53.2 71.0 70.6 Female 50.0 39.5 71.1 63.2 Solola All 70.3 68.3 79.5 78.3 Male 70.6 70.6 80.4 80.4 Female 68.6 54.3 74.3 65.7 Totonicapan All 74.1 72.7 87.1 84.9 Male 72.5 72.5 87.2 86.2 Female 80.0 73.3 86.7 80.0 Quetzaltenango All 49.8 48.9 7(0.3 69.2 Male 51.4 51.2 70.5 70.2 Female 40.0 35.4 69.2 63.1 Suchitepequez All 51.8 48.9 76.2 74.5 Male 51.3 51.3 75.4 75.0 Female 53.7 38.9 79.6 72.2 Retalhuleu All 49.2 45.7 72.5 71.3 Male 48.4 45.7 71.0 70.6 Female 54.1 45.9 81.1 75.7 San Marcos All 71.3 69.7 86.6 85.7 Male 72.9 72.7 87.5 87.5 Female 59.3 47.5 79.7 72.9 Huehuetenango All 86.4 84.4 90.6 89.4 Male 86.3 85.4 90.4 89.4 Female 86.8 76.3 92.1 89.5 Quiche All 85.2 84.9 94.2 94.2 Male 85.3 84.9 93.6 93.6 Female 85.0 85.0 97.5 97.5 Baja Verapaz All 77.9 75.8 86.8 86.3 Male 77.7 76.4 85.1 85.1 Female 78.6 73.8 92.9 90.5 Alta Verapaz All 81.7 80.9 90.2 89.9 Male 83.1 82.7 90.8 90.8 Female 73.0 68.9 86.5 83.8 Peten All 53.1 51.9 70.1 68.9 Male 51.7 51.7 69.7 69.5 Female 60.9 53.3 71.7 65.2 Izabal All 37.2 34.2 53.3 50.3 Male 34.8 34.4 51.0 50.2 Female 49.0 33.3 64.7 51.0 Zacapa All 50.0 45.8 71.2 66.9 Male 52.1 50.0 72.9 71.9 Female 40.9 27.3 63.6 45.5 Chiquimula All 72.2 69.0 84.8 81.0 Male 73.8 72.3 84.6 83.1 Female 64.3 53.6 85.7 71.4 Jalapa All 68.2 67.6 81.2 80.6 Male 69.4 69.4 81.9 81.9 Female 61.5 57.7 76.9 73.1 Jutiapa All 67.0 62.0 79.6 75.9 Male 66.7 65.9 78.9 78.5 Female 67.9 50.0 82.1 67.9 Source: Encuesra Nacional Sociodemogrdfica, 1989. Annex 6 Page 1 of 2 This Annex briefly describes the probit analysis used in Chapter 1 and presents information on the mean values of the explanatory variables. PROBIT ANALYSIS The probit model expresses the probability (P) of an individual being below the poverty line as a function of various characteristics (X), such as age, gender, ethnicity and so on. P=Prob (Individual is poor) =c1(EbkXk)= f e (2 21I du Where 4 is the normal cumulative distribution function. The reported marginal effects are partial derivatives indicating the change in the probability of being below the poverty line relative to a unit change in one of the independent variables, a P =_1 e(-