Workers' benefits from Bolivia's emergency social fund

LSM .77 .5.ITT.S FEB. 1991 Living Standards Measurement Study Working Paper No. 77 Workers' Benefits from Bolivia's Emergency Social Fund  LSMS Working Papers No. 6 Household Survey Experience in Africa No. 7 Measurement of Welfare: Theory and Practical Guidelines No. 8 Employment Data for the Measurement of Living Standards No. 9 Income and Expenditure Surveys in Developing Countries: Sample Design and Execution No.10 Reflections on the LSMS Group Meeting No. 11 Three Essays on a Sri Lanka Household Survey No.12 The ECIEL Study of Household Income and Consumption in Urban Latin America: An Analytical History No.13 Nutrition and Health Status Indicators: Suggestions for Surveys of the Standard of Living in Developing Countries No.14 Child Schooling and the Measurement of Living Standards No. 15 Measuring Health as a Component of Living Standards No.16 Procedures for Collecting and Analyzing Mortality Data in LSMS No.17 The Labor Market and Social Accounting: A Framework of Data Presentation No.18 Time Use Data and the Living Standards Measurement Study No. 19 The Conceptual Basis of Measures of Household Welfare and Their Implied Survey Data Requirements No. 20 Statistical Experimentation for Household Surveys: Two Case Studies of Hong Kong No.21 The Collection of Price Data for the Measurenment of Living Standards No. 22 Household Expenditure Surveys: Some Methodological Issues No. 23 Collecting Panel Data in Developing Countries: Does It Make Sense? No.24 Measuring and Analyzing Levels of Living in Developing Countries: An Annotated Questionnaire No. 25 The Demand for Urban Housing in the Iwry Coast No.26 The COte d'Ivoire Living Standards Survey: Design and Implementation No. 27 The Role of Employment and Earnings in Analyzing Levels of Living: A General Methodology with Applications to Malaysia and Thailand No. 28 Analysis of Household Expenditures No.29 The Distribution of Welfare in COte d'Ivoire in 1985 No. 30 Quality, Quantity, and Spatial Variation of Price: Estimating Price Elasticities from Cross-Sectional Data No. 31 Financing the Health Sector in Peru No. 32 Informal Sector, Labor Markets, and Returns to Education in Peru No. 33 Wage Determinants in COte d'Ivoire No. 34 Guidelines for Adapting the LSMS Living Standards Questionnaires to Local Conditions No. 35 The Denmand for Medical Care in Developing Countries: Quantity Rationing in Rural COte d'Ivoire No. 36 Labor Market Activity in COte d'Iwire and Peru No. 37 Health Care Financing and the Denmand for Medical Care No. 38 Wage Determinants and School Attainntent among Men in Peru No. 39 The Allocation of Goods within the Household: Adults, Children, and Gender No.40 The Effects of Household and Community Characteristics on the Nutrition of Preschool Children: Evidencefrom Rural COte d'Ivoire No. 41 Public-Private Sector Wage Differentials in Peru, 1985-86 (List continues on the inside back cover)  Workers' Benefits from Bolivia's Emergency Social Fund  The Living Standards Measurement Study The Living Standards Measurement Study (Lsms) was established by the World Bank in 1980 to explore ways of improving the type and quality of house- hold data collected by statistical offices in developing countries. Its goal is to foster incrased use of household data as a basis for policy decisionmaldng. Specifcally, the L16 is working to develop new methods to monitor progress in raising levels of living, to identify the consequences for households of past and proposed gov- ernment polices, and to improve communications between survey statisticians, an- alysts, and poliqmakers. The Lsms Working Paper series was started to disseminate intermediate prod- ucts from the lsMS. Publications in the series indude critical surveys covering dif- ferent aspects of the Isms data collection program and reports on improved methodologies for using living Standards Survey (ISS) data. More recent publica- tions recommend specific survey, questionnaire, and data processing designs, and demonstrate the breadth of policy analysis that can be carried out using LSS data.  LSMS Worldng Paper Number 77 Workers' Benefits from Bolivia's Emergency Social Fund JohnNewman Steen jorgensen Meno Pradhan The World Bank Washington, D.C.  Copyright 0 1991 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing February 1991 To present the results of the Living Standards Measurement Study with the least possible delay, the typescript of this paper has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. Any maps that accompany the text have been prepared solely for the convenience of readers; the designations and presentation of material in them do not imply the expression of any opinion whatsoever on the part of the World Bank, its affiliates, or its Board or member countries concerning the legal status of any country, territory, city, or area or of the authorities thereof or concerning the delimitation of its boundaries or its national affiliation. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to Director, Publications Department, at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to photocopy portions for classroom use is not required, though notification of such use having been made will be appreciated. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabetical title list (with full ordering information) and indexes of subjects, authors, and countries and regions. The latest edition is available free of charge from the Publications Sales Unit, Department F, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publications, The World Bank, 66, avenue d'Iena, 75116 Paris, France. ISSN: 0253-4517 John Newman is an economist in the Population and Human Resources Department, Steen Jorgensen is an economist in the Southern Africa Department, and Menno Pradhan is a consultant in the Population and Human Resources Department, all at the World Bank. Library of Congress Cataloging-in-Publication Data Newman, John L., 1955- Workers' benefits from Bolivia's Emergency Social Fund / John Newman, Steen Jorgensen, Menno Pradhan. p. cm-(LSMS working paper, ISSN 0253-4517; no. 77) Includes bibliographical references. ISBN 0-8213-1765-2 1. Fondo Social de Emergencia (Bolivia) 2. Job creation-Bolivia. 3. Temporary employment-Bolivia I. Jorgensen, Steen, 1958- I. Pradhan, Menno, 1965- . 1II. Title. IV. Series. HD5751.A6F666 1991 331.13'77'0984.-dc2O 91-7607 CIIP  ABSTRACT Bolivia's Emergency Social Fund (ESF) was established to cushion the adverse effects of its economic crisis and subsequent stabilization program on the poor and to facilitate transition through the phases of structural adjustment. The ESF program represented one of the first World Bank funded efforts to address the social costs of adjustment by including a separate compensatory program, rather than by modifying the implementation of a structural adjustment program in light of the expected social costs. The ESF was established explicitly as a temporary financial institution outside of the normal bureaucratic structure of the government. While its primary emphasis was to provide temporary employment opportunities, it differed from more typical government works projects in being demand-driven. The management team of the ESF approved or rejected funding requests for small-scale, labor intensive projects that came from local governmental and non-governmental agencies, but did not propose any projects themselves. The projects were executed and the workers hired by private subcontractors working under the supervision of the local agency and the central management team of the ESF. This paper is concerned with measuring the effect of the ESF program on employment and income of workers in the ESF projects. The analysis is based on the results of a survey administered to workers in ESF infrastructure projects and to the population at large by the Instituto Nacional de Estadistica (INE) of Bolivia. In the first part of the paper we identify the characteristics of the workers in the ESF infrastructure projects and compare these workers with the population in general and workers in the construction sector in particular. In the second part of the paper we perform a counterfactual simulation, asking what would have been the position of the workers in the absence of the ESF program. Based on this simulation, we infer what the employment and income effect of the ESF program has been on its beneficiaries. The average ESF worker experienced a 12.8 percent increase in wages, an increase of 9.5 hours of work a week, and a 32 percent increase in weekly earnings over what he would have earned if he were working in the absence of the ESF. Taking into account the probability that the individual may not have worked in the absence of the ESF leads to larger gains. The unconditional comparison indicates that the average ESF worker receives a 33 percent increase in wages, an increase of 15.5 hours of work a week, and a 51 percent increase in earnings. The greatest benefits from participating in the program were received by those who would have been least well-off in the absence of the program. -v-  ACKNOWLEDGMENTS We would like to thank Marcelo Mercado and Alejandro Mercado of the Instituto Nacional de Estadistica of Bolivia and Fernando Campero of the Fondo Social de Emergencia of Bolivia for their assistance in this proejct. - vi -  TABLE OF CONTENTS 1. Introduction ...................... . 1 2. Background to the ESF . . . . . . . . . . . . . . . . . . . 5 3. The Data . . . . . . . . . . . . . . . . . . . . . . . . . 7 4. Background of ESF Workers ................ . 9 5. Earnings of ESF Workers . . . . . . . . . . . . . . . . . 11 6. Hours Worked and Hourly Wages . . . . . . . . . . . . . . 11 7. Position of the Family ...... .. .. .. .. . . . 15 8. Impact of the ESF Program ...... . .. . .. . .. . 17 A. The Econometric Specification . . . . . . . . . . . 19 B. The Variables .................. . 22 9. Results .............. . 23 10. Conclusions . . . . . . . . . . . . . . . . . . . . . . . 33 References ..37 Appendix A . . . . . . . . . . . . . . . . . . . . . . . 39 TABLES AND FIGURES Table 1 Sample Distribution of Workers in ESF Projects and Distribution of Funding of Urban Projects . . . . . . . . 9 Table 2 Education Levels ... . . . . ..... . . . . . . . . 10 Table 3 Total Family Earnings by Family Food Expenditure Quartiles Percent of ESF families in Cell . . . . . . . 17 Table 4 Descriptive Statistics of Males Aged 18 to 55 . . . . . . 25 Table 5 Wage and Hours of Males 18-55 in the General Population . 26 Table 6 Wage and Hours of Males 18-55 in the ESF Sample . . . . . 28 Table 7 Expected Wages, Hours, and Earnings . . . . . . . . . . . 29 Figure 1 How the Earnings of ESF Workers Compare to Earnings in Construction .12 Figure 2 How Earnings of ESF Workers Compare to those of all prime age males .13 Figure 3 How Family Income of ESF Workers Compares to the General Population .14 Figure 4 Distribution of Hours of Work . . . . . . . . . . . . . . 16 Figure 5 Comparison of Predicted Earnings ESF Workers with and without ESF Program .................. 32 - vii -  1. INTRODUCTION Bolivia's Emergency Social Fund (ESF) was established to cushion the adverse effects of its economic crisis and subsequent stabilization program on the poor and to facilitate transition through the phases of structural adjustment. The ESF program represented one of the first World Bank funded efforts to address the social costs of adjustment by including a separate compensatory program, rather than by modifying the implementation of a structural adjustment program in light of the expected social costs. While the primary emphasis of this program was to provide temporary employment opportunities, it differed from more typical government works projects in several important ways. First, the program was intended to be demand-driven. The management team of the ESF approved or rejected funding requests for small-scale, labor intensive projects that came from local governmental and non-governmental agencies, but did not propose any projects themselves. Secondly, the ESF did not directly employ the workers. Rather, the projects were executed and the workers hired by private subcontractors working under the supervision of the local agency and the central management team of the ESF. Thirdly, the ESF was established explicitly as a temporary financial institution outside of the normal bureaucratic structure of the government. There are several avenues through which a compensatory program such as the ESF can affect the transition through the structural adjustment process. It may contribute political support to carry out the adjustment policies, create infrastructure that reduces the cost of producing tradeable goods, and alter workers' incentives to move towards favored sectors. As in any program designed to ameliorate the social costs of the adjustment process, there is a tension between ameliorating the adverse impact and counteracting the necessary reforms. Because the ESF's economic infrastructure projects generate employment in the non-traded goods sector, they do not directly contribute towards shifting the structure of production towards traded goods. Indeed, they raise the return to employment in a nontraded goods sector. If the ESF is to aid in the transition, what it does contribute towards easing the transition (the economic infrastructure and the income maintenance), must outweigh whatever negative effect it would have in counteracting the reforms.  Quite apart from the ESF's role in the adjustment process, there also has been interest expressed in Bolivia's experience with the demand-driven mechanism used to distribute the funds. Under the auspices of the ESF, a substantial amount of external funds were generated and channelled to a large group of smaller governmental and non- governmental organizations. The ESF's success in this regard has motivated efforts to establish similar social investment funds in other countries (notably in Guatemala, Haiti, and Jordan), either with or without a connection to a structural acdjustment program. As different groups have been interested in different aspects of the ESF, there is no single criterion for success of the program and, consequently, no single source of information on which to judge the program. At the risk of oversimplification, one can identify four different views of the ESF. Depending on the perspective of the observer, the ESF was a) a Keynesian aggregate demand policy designed to keep the economy from further contraction during the structural adjustment process; b) an institution for channelling externally-generated funds down to local groups that previously had no recourse to this type of funding; c) a program designed to compensate those who suffered the greatest loss due to adjustment; d) a program designed to protect; the poor and vulnerable during adjustment. Those in group (d) are not necessarily the same as those in group (c). In the early days of the ESF, the predominant view of ESF's role was that it should generate external funds and quickly inject them into the economy. External funds were required to finance the ESF since financing the project out of domestic funds would have made it that much harder to achieve the objective of the structural adjustment program of reducing the budget deficit. To judge how well the ESF did in this regard, one need only consult the administrative record. By the end of 1988, the ESF had raised more than $100 million in foreign currency, an amount equal to about one-fourth of the current account deficit in 1988. Most of the foreign financing received has been in the form of grants. There is reason to believe that foreign aid inflows would have been considerably lower without the ESF. Several donors have begun or resumed programs in Bolivia because ESF was able to ensure efficient and speedy implementation. The ESF's administrative costs amounted to no more than 3 percent of its budget. The average lag between date of commitment and disbursement of all funds is less than one year in the ESF, shorter - 2 -  than even fast-disbursing balance of payments support. Rapid disbursements were facilitated by the ESF's use of the market, relying on local groups to propose projects and mainly private subcontractors to execute them. The ESF was effective in making resources available to municipalities, community organizations, and nongovernmental organizations, which, by nature of their size, do not typically receive funding from international organizations. By the end of 1988, the ESF had committed about $20 million or 20 percent of its total commitments to non-governmental organizations (NGOs). Of course, the viability of social investment funds and the benefits of such a program as the ESF can not be judged solely on the basis of the administrative record. The administrative record provides a measure only of the inputs (the expenditure of the investment fund) and not the outputs of the investment fund. Moreover, it remains an open question whether the success the ESF has had as a temporary institution operating to a large extent independently of the line ministries and outside of many normal constraints can be replicated in a permanent institutional framework. The experience that will come with the successor to the ESF in Bolivia (the Social Investment Fund) and with efforts to establish other social investment funds elsewhere will be a strong test of the institutional vaibility of a permanent social investment fund. Certainly, the experience of the ESF provides useful lessons for the establishment of those funds. To judge the effectiveness of ESF's institutional structure, its role in compensating those adversely affected by the structural adjustment, and its contribution towards protecting the poor during adjustment, requires information on the outputs of the program - on the value of the infrastructure created and of the impact of the program on beneficiaries. This paper is concerned with measuring the effect of the ESF program on employment and income of workers in the ESF projects. This is an important element in an evaluation of the ESF, but does not, by itself, constitute a complete evaluation of the ESF.1 Two aspects, in particular, that are not addressed in this paper are worth mentioning. Given the relatively small size of the ESF program, we do not estimate spillover effects of the program on the urban labor market in general. How one might estimate such effects is discussed in section 8. Secondly, this paper is not concerned with estimating the value of the economic infrastructure created.2 Nevertheless, it is - 3 -  important to recognize the additional objectives of the program - to generate productive economic infrastructure and to generate and disburse funds quickly. These objectives imposed constraints on the type of employment generation and income maintenance that were undertaken. For example, because economic infrastructure projects were managed by subcontractors who decided whom to hire, almost no women wvere hired. This is not surprising since these were construction projects and extremely few women work in the urban construction sector in Bolivia. If the output of this employment/income maintenance program is valued highly, then it may still be desireable to fund such projects even though the mix of the workers hired does not match the desired targets. If the indirect effects of the program are judged not sufficient to improve the welfare of the targeted groups, then other facets of the ESF program would have to be directed towards the appropriate groups. Our analysis of the effect of the ESF projects on the urban labor market is based on the results of a survey administered to workers in ESF infrastructure projects and to the population at large by the Instituto Nacional de Estadistica (INE) of Bolivia. In the first part of the paper we identify the characteristics of the workers in the ESF infrastructure projects and compare these workers with the population in general and workers in the construction sector in particular. In the second part of the paper we perform a counterfactual simulation, asking what would have been the position of the workers in the absence of the ESF program. Based on this simulation, we infer what the employment and income effect of the ESF program has been on its beneficiaries. We do not attempt to analyze how these results translate into political support for the adjustment program as this is something that does not readily lend itself to empirical study. 1 Grosh (1990) describes evaluations of other facets of the ESF that have been conducted. These consist of: a) a technical audit of the technical/engineering quality of its civil works; b) an early evaluation and monitoring report conducted by a team of economists and sociologists; c) a community participation study conducted by anthropologists through open-ended but structured interviews; d) an ex-post cost-benefit analysis of a small sample of completed projects; e) a study of the macro-economic impact, comparing the results of investing through the ESF to results from investing through alternative channels; f) a study of the impact on institutions with which the ESF worked; g) a study of its impact on one particular community; and h) a study on the geographic targeting of projects. 2 Herrick (1989) presents a benefit-cost analysis of the infrastructure created using respondents' perceived increases in their property values following the completion of an ESF-funded infrastructure project as an indicator of the benefits. In an ex post evaluation of 20 street paving, sewer, and water projects, he finds 15 out of the 20 projects with a benefit cost ratio creater than 1. The lowest ratio is 0.84 and the median value is over 4. - 4 -  2. BACKGROUND TO THE ESF The Bolivian economy deteriorated rapidly and continuously after 1980, reaching a point of chaos by 1985. In retrospect, the causes are clear. In the seventies, an artificial boom, fed by capital inflows followed the discovery of substantial hydrocarbon resources. The boom lasted as long as access to financial resources was easy. When the situation changed around 1980, for both internal and external reasons, economic performance deteriorated sharply. Net foreign transfers became strongly negative, GDP fell in real terms, and capital flight accelerated. Economic policies became erratic and inconsistent. By September 1985, inflation had reached 24,000 percent a year, GDP per capita had fallen by more than a fifth since 1980, and public sector deficits were around one fourth of GDP. When the Government of President Paz-Estenssoro took office in August 1985, an orthodox, wide-ranging stabilization package was introduced. The package included freeing most prices, deregulating the trade system and the labor market, establishing a uniform exchange rate (determined through daily auctions) and cutting public sector deficits. Inflation was immediately curtailed and has remained low. Inflation was 11 percent in 1987 and 21 percent in 1988 and 1989. Public sector deficits have been brought under control and the exchange rate has remained unified. Growth has taken longer to occur, as Bolivia faced serious negative terms of trade shocks in the first year of the program. However, the economy grew, albeit slowly, in 1987, 1988, and 1989. The effects of the economic program on the public sector were especially dramatic. The Government set out deliberately to introduce a fundamentally different view of the public sector. The public sector's role was viewed as providing basic public services efficiently, engaging in only limited extraction and processing of natural resources, and otherwise providing a "level playing field" for the private sector. The state mining sector was closed for restructuring and 90 percent of the employees were fired. Public sector salary policy was left to individual entities under a fixed wage bill, and important reforms were initiated to establish a better accounting and control system in the public sector. From the outset of the adjustment program, the Bolivian Government attached - 5 -  a high priority to measures to alleviate any further deterioration in social conditions. This was due to fears over the effect of economic dislocation on the already very low living standards of the Bolivian poor, as well as concern that increased hardship would strengthen opposition to the Government's policies. Various options were considered for alleviating the social costs of adjustment before setting up the ESF. The early discussions in 1985 revolved around more traditional programs of widespread distribution of food or medicines, or direct subsidies for selected items. A number of difficulties were perceived with this approach. First, the administrative costs in directly providing or subsidizing the purchase of commodities to low income groups was felt to be high. Second, providing free food (particularly if taken from international donations) would have discouraged the production of food and increased the country's dependence on food imports. Third, permitting recipients more discretion in the use of assistance (providing money instead of food, for example) was viewed as more consistent with the reduced role of the state inherent in the adjustment program. Finally, the Government wished to avoid imposing a uniform method of assistance, but rather to respond to the demands of local groups which would be in a better position to judge the most immediate needs. Following these principles, the ESF was established in November 1986 to provide funds for small-scale, labor-intensive projects mostly in infrastructure. The projects were meant to be demand-driven, meaning that the ESF would finance projects proposed by local groups. The focus on labor-intensive, infrastructure projects was adopted to increase employment among the poor while at the same time providing services to the community that would aid development. A second feature was that the ESF was primarily a financing institution. It did not involve itself in the implementation of projects, except for some supervision and technical support. The organizations that received funding were responsible for hiring the workers in the projects and ensuring that the projects moved towards completion. ESF was designed as a temporary institution and is scheduled to disappear in December 1990. It was hoped that in four years time growth would have picked up enough to solve the employment problem and that the need for austerity measures would be less, allowing for increases in social spending as part of the regular budget. ESF's total program was planned as US$ 180 million over the life of the institution, with US$ 8 million dollars coming from the Bolivian government and the rest from - 6 -  external sources in the form of grants or loans at concessional rates. The ESF funds projects in four basic categories: economic infrastructure, social infrastructure, social assistance, and production support. Economic infrastructure (covering 37.8 percent of committed funds) encompasses infrastructure closely related to productive activities, including road maintenance and upgrading, urban improvement, irrigation, flood control, and reforestation. Social infrastructure (47.6 percent of committed funds) covers infrastructure for health and education, water and sanitation, basic housing (mostly self-construction), and some cultural projects (for example, repairs of historic buildings). Social assistance (8.8 percent of committed funds) covers recurrent costs in education and training, vaccinations, school breakfasts, and production of school materials. Production support (5.7 percent of committed funds) is mainly credit provided through NGOs to productive units which are outside of the formal financial system, such as microenterprises producing for the informal sector and small cooperatives in mining and agriculture. 3. THE DATA The management team of the ESF commissioned ex post evaluations of the economic returns of the projects and surveys of the beneficiaries of the employment generating projects. This paper analyzes the results of the 1988 household survey administered to ESF workers and to the population at large by the Instituto Nacional de Estadistica (INE) of Bolivia with the financial support of the ESF and the World Bank. Information on the general population was obtained from the 1988 Encuesta Permanente de Hogares (EPH), part of an ongoing survey effort that has been conducted yearly by the Instituto Nacional de Estadistica (INE) since 1976. The 1988 survey took place in May and covered urban households in nine capital cities of the provinces in Bolivia. Information on workers and their families in ESF funded projects was obtained from a special survey conducted for the ESF by INE. From a list of all ESF projects in execution, INE selected all those that were (a) in capital cities; (b) in operation at the time of the EPH Survey; and (c) in economic infrastructure (i.e. explicitly employment generating). Only the capital cities were chosen because INE's May 1988 household survey covers only the capitals and information on a control group -7 -  was required for the analysis. A total of 64 projects employing 3,051 workers met these requirements. INE then visited the projects to ascertain the actual number of workers working at the projects and drew a random sample of 600 workers from the list of all workers. The ESF workers were visited in their homes and were administered the same household questionnaire as that of the EPH sample. Table 1 indicates the distribution of all workers in the 64 projects and of the sample across all cities. This procedure of conducting a separate sample of ESF workers was done for two reasons. First, because at the time of the survey the size of the ESF program wvas small compared to the population economically active, the probability of finding a sufficiently large sample of workers in the ESF for analytical work from a sample of the general population was low.3 Second, previous pilot surveys indicated that many of the workers who were actually working in ESF financed projects did not know that they were doing so. This is evidenced by the fact that when INE visited the workers known to work in the projects, having obtained their names and addresses from the subcontractors, a majority of the workers had not heard of the ESF. This meant that the only way of guaranteeing that ESF workers would be interviewed was to employ a separate sample consisting only of ESF workers. The results from the May 1988 survey provide additional information to that contained in a previous survey conducted in December 1987. In the 1987 survey the ESF workers were interviewed on the job site. While this may be expected to generate reliable information on the characteristics of the ESF workers themselves, it is not the best procedure to obtain information on other family members or on family-level variables. The 1988 survey has better information on family characteristics. Despite the differences in collection procedures, many of the 1988 results confirm the picture presented from the 1987 survey and from other surveys conducted of the beneficiaries. Unless otherwise specified, the results presented in this document are taken from the 1988 survey. 3 At the time of INE's survey (May 1988), the size of the ESF program was small relative to the labor force. In all ESF-sponsored employment generating projects in the capital cities, 3,051 persons were employed. This number constitutes 0.3 percent of the economically active population and 2.7 percent of the unemployed in the capital cities. - 8 -  Table 1 -- Sample Distribution of Workers in ESF Projects and Distribution of Funding of Urban Projects CaDital Citv Total Workers in ESF Proiects Sample of Workers Sucre 173 34 La Paz 1296 255 Cochabamba 403 80 Oruro 492 96 Potosi 401 79 Tarija 118 23 Santa Cruz 82 16 Cobija 86 17 Total 3051 600 Sources: Informe de la Encuesta a Obras del Fondo Social de Emergencia, Instituto Nacional de Estadistica, Bolivia, June 1988; and internal documents of Fondo Social de Emergencia, May 1988. 4. BACKGROUND OF ESF WORKERS The workers in these ESF projects are, for the most part, prime-age married males who are heads of household. Ninety-nine percent are male and seventy-one percent are married. Ninety-one percent fall in the age category between twenty and sixty-five years old and ninety-three percent are reported as head of the household. These workers are apparently the prime income earners of their household and the earnings from the ESF projects are their primary source of income. In 62 percent of the households of ESF workers there is only one reported income earner, in 25 percent there are two and in 9 percent there are three. These results closely approximate the pattern of the 1987 survey. Only 7.2 percent of the ESF workers have a secondary job. Most of the ESF projects in the capital cities are in the construction sector and ninety-one percent of the ESF workers work in construction. This high percentage is due to the selection of only projects in economic and social infrastructure. We compare the ESF workers with other workers in the construction sector and with workers in general, using information from the EPH sample. In comparison with other workers in the construction sector ESF workers are less -9-  educated. The distribution over the different education levels is given in table 2. Table 2 -- Education Levels ESF Workers Workers in construction None 6.3 4.4 Basic 41.2 36.8 Intermediate 24.1 15.8 Middle 26.0 26.9 Technical 0.7 3.5 Normal 0.2 0.6 University 2.0 12.0 Thirty-nine percent of the sample responded that they were looking for work prior to working in the ESF, 54 percent reported that they had a job, and 7 percent had another response. Of those who had a job before joining, sixty percent worked in construction before, either as an independent or wage worker. That is, a little over 30 percent of all those working in ESF projects were working on construction jobs before working on the ESF projects. Again, of those with a previous job, five percent were vendors and 1.4 percent were miners before working in the ESF. The rest are spread out over various occupations. Taking into account both those who came directly from mining and those who had an intervening spell of unemployment, some ten percent of the 3,050 ESF workers in the capital cities were ex-miners. Given the size of the ESF employment program at that time and the approximately 23,000 miners let go when COMIBOL was closed, it is clear that the ESF did not come close to absorbing all of the ex-miners. However, that was never the intention of the program. - 10 -  5. EARNINGS OF ESF WORKERS The evidence indicates that the ESF workers are on the low skilled end of the construction market, but that they are paid according to the levels prevailing in the labor market in construction. The mean monthly earnings was 225 bolivianos per month for workers in ESF projects, 375 per month for the average prime-age male construction worker in the EPH sample, and 348 for the average prime-age male in the population at large. However, the mean monthly earnings of low skilled construction workers in the EPH sample (those with the characteristics of the ESF workers) was 196 bolivianos per month. The median monthly earnings was 194 for workers in ESF projects, 220 for prime-age males in the construction sector, and 230 bolivianos per month for workers in the population in general. Figure 1 presents the distribution of ESF workers across the deciles of monthly earnings of males between 18 and 65 in the construction industry. Figure 2 presents a similar distribution of ESF workers among all males aged 18 through 65 in the EPH sample. The height of the bars indicate the percentage of ESF workers with monthly earnings that would place them in the given decile in the EPH sample. For example, 45 percent of the ESF workers receive monthly earnings within the range of the fourth poorest decile and 31 percent receive monthly earnings withing the range of the lowest three deciles. Earnings in the fourth poorest decile range from a low of 150 to a high of 200 bolivianos a month. While earnings of all ESF workers are low relative to earnings in other countries, not all the workers in the ESF projects are low-income workers by the standards of the urban Bolivian work force. Figure 3 presents the distribution of family income of ESF workers across EPH family income deciles. It looks much the same as for the distribtution of monthly earnings. There is an overrepresentation of ESF families in the middle deciles, 3 through 6, and an underrepresentation of ESF families in the lower and upper deciles. Less than 10 percent of the ESF families are in the two highest deciles. 6. HOURS WORKED AND HOURLY WAGES On average, the workers in the ESF projects work more hours than the prime - 11 -  Figurel: How the Earnings of ESF Workers Compare to Earnings in Construction Percent in EPH Sample 25 20 15 10 5 - 0 1 2 3 4 5 6 7 8 9 10 Decile of Earnings in Primary Job Note: deciles from all constr. workers  Figure2: How Earnings of ESF Workers Compare to those of all prime age males percent in EPH sample 25 .o - 20 15 - 10 _ 5- 0 1 2 3 4 5 6 7 8 9 10 decile of earnings in primary job Note: deciles calculated from EPH sample  Fig. 3: How Family Income of ESF Workers Compares to the General Population Percent in EPH sample 25 2 0 . .. . .. . . .. . .. . . .. . .. . . .. . .. . . .. ... . . .. . .. . . .. . .. . .. . . .. . .. . . .. .. . . .. . .. . . .. . .. . . .. . .. . . .. . .. . . 1 5 . ... ... ... ... ... ... ... ... ... ... ... .... ... .. .. . ... ... ... ... ... ... ... ... ... .... ... ... ... ... ... ... ... ... ... ... ... ... 15. 1 0 - -- - -- - -- - -- - -- - -- - -- - .... . ..... ..... ......................................................... ............ 5 - _ -- _ . . . .. . .. . - --. .. ... . .. . . . . - . . . . ... . 10 1 2 3 4 5 6 7 8 9 10 Deciles of total family earnings Note: Deciles calculated from EPH sample  age males within the EPH sample. ESF workers work 50.3 hours a week compared with 46.8 for the EPH workers. Compared with other prime-age males in the construction sector and the population in general, the distribution of hours worked by ESF workers is less variable. Figure 4 illustrates the distribution of hours worked in the primary job for the two samples. The conclusion drawn from this figure is that, in general, ESF workers are employed full-time. The mean hourly wage of workers in the ESF projects was 1.16 bolivianos compared with a mean hourly wage of 2.28 of the corresponding group in the EPH sample. The median hourly wage is 0.98 bolivianos for ESF workers and 1.29 bolivianos in the EPH sample. 7. POSITION OF THE FAMILY Table 3 presents cross-tabulations of total family earnings quartiles by total family food expenditure quartiles. The quartiles were calculated using all the families in the EPH sample. In general, the ESF families appear to be worse off if judged by their position within food expenditure quartiles as oppososed to family earnings quartiles. For example, 34.6 percent of the ESF families are in the lowest food expenditure quartile, whereas only 21.5 percent of the families are in the lowest family earnings quartile. While there is always the possibility that measurement error in either food expenditure or family earnings results in a misclassification of families into the wrong cell, the observed relationship between expenditure and income is what one would expect given the fact that the jobs in the ESF were temporary. The ESF families may not be changing their expenditure patterns in response to the temporary gain in income. The majority of the ESF families (54.1 percent) are in the lowest two earnings and food expenditure quartiles. Seventeen percent of the families are in the two highest earnings quartiles, yet are in the lowest two food expenditure quartiles. Eleven percent of the faimilies in the lower two earnings quintiles are in the highest two food expenditure quintiles. The workers in this group may have come from relatively well- off households, since even receiving low earnings does not lead the household to have low food expenditure. The workers who are in both high earnings and food expenditure quartiles (18.8 percent) may have come from relatively well-off households to begin with or the high earnings may have raised the food expenditure in the household. With the - 15 -  Figure4: Distribution of Hours of Work ESF workers EPH workers 0 20 40 60 80 100 percent <= 40 41-45 46-50 L >50  limited amount of information on the household characteristics of ESF workers, we can not tell which is the more likely explanation. Table 3 -- Total Family Earnings by Family Food Expenditure Quartiles Percent of ESF families in Cell Food Expenditure Quartile 1 (low) 2 3 4 Row Total Family Earnings Quartile 1 (low) 12.9 5.5 2.4 0.7 21.5 2 16.5 19.2 7.0 1.0 43.7 3 4.3 9.6 9.6 2.9 26.4 4 0.9 2.2 2.9 2.4 8.4 Column Total 34.6 36.5 21.9 7.0 100.0 8. IMPACT OF THE ESF PROGRAM To ascertain the effect of the ESF on the labor market, we ask a counterfactual question - what would have been the position of the ESF workers and other individuals in the labor market in the absence of the program. In principle, the positions of both types of workers could be estimated by holding constant for local labor market conditions (the size and structure of demand) and estimating the incremental effect of the ESF program on wage offers. As the ESF increases the demand for labor, one would expect a measure of the size of the ESF program to have a positive effect on wage offers and, therefore, most likely an increase in the probability of working and - 17 -  hours worked. To infer the position of workers in the absence of the ESF program one would compare expected wages, probabilities of working, and hours worked with the incremental effect of the ESF program to the corresponding values obtained by setting the incremental effect to zero. Such a procedure requires an accurate indicator of the differential size of the ESF program in the different capital cities. At the time of the survey, there was substantial variation both in local labor market conditions and in the size of the ESF program. However, the variation in the ESF program was not random. The actual size of the program reflected the applications for proposals that were received from each city and the funding decision rules of ESF's management team. An attempt was made to locate projects where wages were low and unemployment and underemployment were high. Because of this nonrandom allocation, including a direct measure of the ESF (e.g. number of ESF workers or expenditure in each city) would generate biased estimates of the program's effect. The problem is that the ESF measure would be correlated with the error terms in the equations explaining wages and labor force participation because of its correlation with unobserved local labor market conditions. This is a common problem in the evaluation of government programs (Rosenzweig and Wolpin, 1986) and is typically solved by incorporating information on the program implementation to distinguish between the effects of the allocation and the impact of the program. The case of evaluating a demand-driven program such as the ESF can be easier or harder than the typical case depending on whether the applications for the projects are uncorrelated or correlated with the variables of interest. If the applications are uncorrelated with the variables of interest (as would be the case if the underlying capacity to generate proposals was randomly distributed across cities),, then information on the applications could serve as an instrument for the size of the ESF program. No additional information on the decision rules would be needed. However, if the initial applications are not randomly distributed across the cities, then even the knowledge of the management's decision rules may not overcome the problem. of a nonrandom allocation of the ESF program. Additional information would have to be available on the ability of a city to generate fundable proposals. Our attempts to measure the incremental effect of the ESF program using the dollar amount of applications in each city as an instrument for the size of the program were unsuccessful. The problem was that the ESF applications were highly correlated - 18 -  with the size and unemployment rate in the city. Consequently, it proved impossible to measure any spillover benefits of the ESF program on other workers who do not participate in the program. However, given that the total amount of workers employed in capital cities at the time of the survey amounted to just 0.3 percent of the population economically active and 2.7 percent of the total unemployed, any spillover effects would be expected to be minimal. Thus, we focus exclusively on measuring the benefits of the ESF program to the workers who participate in the program. 8.A. THE ECONOMETRIC SPECIFICATION We estimate the workers' expected gains from participating in the ESF program using two models. Both models are based on Heckman's (1974) model describing the determination of wages, the probability of working, and the hours of work.4 Our first model consists of four equations that describe: (a) wage offers in the general population; (b) the marginal value of leisure of individuals in the general population; (c) wage offers in the ESF; and (d) the marginal value of leisure of ESF workers. In this model, we assume that the errors in the equations for the general population are independent of the errors in the equations for the ESF. This implies that ESF workers differ from those in the general population only in their observed characteristics, not in unobserved characteristics and allows us to estimate the wage offer and marginal value of leisure equations separately for each group. Given the estimated wage offer and marginal value of leisure equations of the two groups, an estimate of the workers' expected gains from participating in the ESF program is obtained by comparing the expected wages, hours of work, and earnings while participating in the program with estimates of what their wages, hours of work, and earnings would have been in the absence of the program. Under the assumptions of this first model, the counterfactual wage offers and marginal value of leisures are obtained by multiplying the characteristics of the ESF workers by the coefficients of the wage offer and marginal value of leisure equations of the general population. From these estimated structural equations, one may calculate wages, hours of work, earnings, and the probability of working. 4 A succinct description of this model is given in Amemiya (1985), pp. 391-393. - 19 -  However, ESF workers may differ from individuals in the general population in ways that are not captured by measured characteristics of the survey. The information that these workers chose to participate in the ESF indicates they perceived working in the ESF to be better than their alternatives and, therefore, suggests that they might be different from the general population. Our second model takes this into consideration in estimating what the ESF workers would receive in the absence of the program. It adds an equation determining participation in the ESF and allows the error terms in the participation equation to be correlated with the errors in the wage offer and marginal value of leisure equations of ESF workers and the general population. In this section we briefly describe the estimation procedure for this simpler model. The second, more complicated model is described in Appendix A. The equation describing wage offers in the general population is: .n w