Operationalizing Pro-Poor Growth. The Case of El Salvador

Transcription

1 Operationalizing Pro-Poor Growth The Case of El Salvador Prepared by José Silvério Marques For the World Bank (September 2004)

2 Table of Contents Executive Summary... vi Introduction... 1 I. Historical Context and Growth-Poverty Trends Historical context Growth, Poverty, and Inequity in the 1990s... 8 Growth trends... 8 Poverty trends Income distribution trends Asset distribution II. Growth, Distribution of Income, and Poverty Sources and Determinants of Growth Changing production and employment patterns Productivity trends, Growth determinants Correlates of Poverty Distribution and Poverty Impact of Growth Interaction of growth, distribution and poverty Poverty elasticities Pro-Poor Growth Estimates III. Factors Affecting the Participation of Poor People in Growth Macroeconomic instability Did Macro Instability Increase? Why did instability not affect growth and poverty outcomes in the early 1990s? Why do people continue to feel insecure? Pro-Poor Public Spending Forging national priorities The peace dividend Social Spending Improved Social Conditions Spending Incidence Rural Development Agricultural performance Agrarian Reform Growth in rural household income Key drivers of household income growth Remittances Labor Market Gender Women in politics i

3 Girls education Women in the labor market IV. Trade-Offs Between Growth and Pro-Poor Growth Macro Stabilization Policies Trade liberalization V. Conclusions and Recommendations For Policy Making Restore rapid growth Maintain a flexible labor market Strengthen the social protection system Act on the drivers of poor households income growth Tailor policy interventions to the poorest households Pro-poor investment agenda and financing 72 References Annexes Annex Annex Boxes Box 1.1 The Cost of the Armed Conflict... 8 Box 1.2 Official Poverty Lines Box 3.2 Pro-Poor Policies in a Post-Conflict Setting Box 3.1 Panel Data, Quintiles, and Changes in Income Tables Table 1.1 El Salvador s GDP and Sector Growth, 1960s 1990s... 4 Table 1.2 Selected Education and Health Indicators, 1989/ Table 1.3 Population in Poverty, Table 1.4 GDP Growth, Table 1.5 Principal National Account Aggregates, Table 1.6 Headcount Poverty, Table 1.7 Reduction in Poverty (Adjusted Income), Table 1.8 Changes in Extreme Poverty Lines and Consumer Price Index (CPI) Table 1.9 Robustness of Poverty Incidence in El Salvador to Changes in Prices and Poverty Lines, Table 1.10 Intensity and Severity of Poverty (Adjusted Income), Table 1.11 Distribution of (Unadjusted) Income, 1991, 1996, Table 1.12 World Bank s Estimate of the Gini Coefficient Table 1.13 Distribution of land Among Land Owners, Table 2.1 Index of Sectoral Production, Table 2.2 Foreign Exchange Earning Table 2.3 Sources of Growth, ii

4 Table 2.4 Estimates of Total Factor Productivity Growth, Table 2.5 Regression Coefficients Table 2.6 Explained Changes in Growth Rates During the 1990s Table 2.7 El Salvador- Determinants of Potential Growth, Table 2.8 Age Profile of Poverty, Table 2.9 El Salvador s Growth and Inequality Elasticities of Poverty, Table 2.10El Salvador s Growth Elasticities of Poverty, Table 2.11 Ravallion and Chen Pro-Poor Growth Rate, Total Poverty, Table 3.1 Macroeconomic Volatility, 1960s-90s Table 3.2 Volatility of Private Consumption, Table 3.3 Volatility of Growth in Real Wages, Employment and Underemployment Table 3.4 Agriculture and Manufacturing Value Added and Employment Growth, Table 3.5 Central Government Expenditures By Management Area, Table 3.6 Investment in Basic Infrastructure, 1990, Table 3.7 Selected Education and Health Indicators Table 3.8 Real Prices of Selected Agricultural Products, Table 3.9 Changes in Rural Household Income, Table 3.10 Changing Composition of Rural Household Income, Table 3.11 Factors that Affect the Level and the Changes in Rural Household Income, 1995/ Table 3.12 Decomposition of Changes in Rural Household Income, by Quintile, 1995/ Table 3.13 Changes in Occupation and Labor Income, 1995 and Table 3.14 Distribution of Annual Hours Worked, 1995 and Table 3.15 Distribution of Land by Quintile of Income, 1995, Table 3.16 Land Tenure in 1995 and Table 3.17 Monthly Average Remittances Received by Household, by Poverty Level, 1998 and Table 3.18 Impact of Remittances on Poverty, Table 3.19 Rural and Urban Unemployment by Poverty Level, Table 3.20 Participation Rates, Table 3.21 Workers Occupied, by Occupational Category, 1992, Table 3.22 Indicators of Gender, 1991, 1996, Table 3.23 Ratio of Average Women s Salaries to Men s, by Years of Schooling and Region Table 4.1 Growth Elasticities of Real Wages, Employment and Underemployment, Table 4.2 Value Added and Employment in Traded and Non-traded Sectors, Table 4.3 Trade Penetration Elasticities of Real Wages, Table 4.4 Average Monthly Wage, by Years of Schooling, 1991, 1996, Table 5.1 Estimated Costs of a Pro-Poor Growth Investment Package iii

5 Table 5.2 Concentration Indices Figures Figure 1.1 El Salvador: Savings and Investment, Figure 1.2 Average Years of Schooling for 15 Year Olds, 1995, Figure 2.1 Sectoral Contribution to GDP, Figure 2.2 Growth Incidence Curves, Total Poverty, National, , Figure 2.3 Growth Incidence Curves, Total Poverty, Urban, , Figure 2.4 Growth Incidence Curves, Total Poverty, Rural, , Figure 2.5 Growth Incidence Curves, Total Poverty, National, Figure 2.6 Growth Incidence Curves, Total Poverty, Urban, Figure 2.7 Growth Incidence Curves, Total Poverty, Rural, Figure 3.1 Inflation and Unemployment Rates, Figure 3.2 Military Spending as % of Total Central Government Spending and Number of Military Personnel Figure 3.3 Average Years of Schooling for Children Aged Figure 3.4 Infant Mortality Rate Per Socio-Economic Group Figure 3.5 Ratio of the 5 th Richest to the 1 st Poorest Quintiles Household Percentage Access to Services Figure 3.6 Real Effective Exchange Rate and Remittances Figure 3.7 Per capita Agriculture Production Index, Figure 3.8 Short and Long Term Unemployment, Figure 3.9 Unemployment Duration, Figure 4.1 Inflation Rate and the Nominal Exchange Rate iv

6 Acknowledgements This paper has been prepared by José Silvério Marques with the assistance of Orlando Martinez (consultant), Reyes Aterido (World Bank), and Margarita Sanfeliu (FUSADES). Louise Cord, Helena Ribe, Andy Mason, Omar Arias, Humberto Lopez, Ignacio Fiesta, Derek Byerlee and Malcom Ehrenpreis (World Bank) provided very useful comments and suggestions during two workshops held in March and June 2004 in Washington D.C. Written comments were received from the World Bank and DFID. Additional comments were received during a workshop in Frankfurt in July 2004 with the participation of representatives of all sponsoring institutions: Agence Francaise de Developpment, DFID, GTZ, KFW, and the World Bank. Any errors are my own. v

7 Executive Summary This study illustrates the case of a country which after experiencing a major conflict during the 1980s, achieved substantial poverty reduction led essentially by growth as the country recovered and began to put some of the institutions, markets and policies in place for a broad based growth pattern within a newly established democratic framework. The study explores the factors that affected the ability of the poor to participate in economic growth and to benefit from it, and seeks to identify the policies that will make growth in El Salvador more pro-poor. The pro-poor growth definition used in this study is an absolute definition that requires only that the incidence of poverty declines with growth. After analyzing the historical context and trends in poverty and inequality, the key questions addressed are: What explains El Salvador s relatively low growth during the last four decades and its continuing high levels of poverty? How have public policies and other development such as remittances affected the participation of poor people in growth? Were there trade-offs between pro-growth and pro-poor policies in the stabilization and structural adjustment policies pursued during the 1990s? What are the poverty correlates and the key drivers of income growth in the poorest households? And, what lessons can be learned from this experience to make growth more pro-poor in the future? The study has four main messages. First, growth and pro-poor spending brought about by the stabilization and structural reforms initiated in the early 1990s, the peace agreements, and the political consensus on the priority of social policies, together with increased remittances, helped to reduce poverty by one-third in the 1990s, not a small feat. Second, given El Salvador s continuing high levels of poverty and inequality and the recent deceleration in growth and in poverty reduction, further progress in reducing poverty will require restoring rapid growth. Analysis of the determinants of growth indicates that this in turn will require investment in education and infrastructure. At the same time, there is a need to keep the economy open and the labor market flexible while strengthening the country s social protection system to minimize the impact of negative shocks on poor households. Third, to make growth more pro-poor in El Salvador, there is a need to act on the drivers of income growth of the poor. Analysis of poverty correlates and rural household income growth indicates that these drivers are education and basic infrastructure. Investment in these areas improves the access of the poor to assets and services that enable them to take advantage of economic opportunities such as accessing non-agricultural employment and microenterprises. And fourth, the design of pro-poor growth policy interventions for the poorest households requires a detailed examination of the specific drivers of their income growth, because analysis of income growth by quintile indicates that the poorest households may face constraints not easily identifiable through examination of vi

8 standard household survey data, and that this poorest group may differ substantially from other households. Historical Context and Growth-Poverty Trends Poverty has remained high in El Salvador through episodes of very rapid growth and also of sharp decline in output during the last four decades. In the 1960s the country embarked on an import-substitution effort which initially led to rapid growth; as the easier opportunities for import substitution were exhausted, growth began to decelerate in the 1970s; in the 1980s, El Salvador experienced a destructive internal conflict which caused a large drop in GDP; in the 1990s, it initiated a structural adjustment reform that was initially accompanied by rapid GDP growth, but growth has slowed again since the mid-90s. In general, growth has been low over the last 40 years, averaging 3 percent. Owing to the conflict, today s real per capita income is similar to that of 30 years ago, and poverty affects about 40 percent of the population. Some policies of the late 1970s and early 1980s were costly for the country. They included the nationalization of the banking system, the institution of price controls, the maintenance of a dual exchange rate regime that penalized traditional exports, and the creation of state export marketing boards for coffee and sugar. In the early 1980s, an agrarian reform confiscated all properties over 245 ha, and transferred this land to landless farmers and new agricultural cooperatives. The agrarian reform was poorly implemented and did not yield the results anticipated, in part because of the intensification of the conflict and the drop in agricultural prices. The conflict cost the lives of over 70,000 Salvadorans and left most of the country s infrastructure destroyed. Rural areas were those most affected by the destruction and thousands of Salvadorans emigrated, mostly to the US. Over one million Salvadorans live in the US alone. By the end of the 1980s, El Salvador was experiencing an economic and social crisis. Per capita income was only 70 percent of its 1978, pre-conflict, level. In urban areas, unemployment affected 23 percent of those in the bottom income quintile. The country s economic infrastructure was shattered and social indicators lagged. The government that took office in 1989 implemented a series of reforms designed to stabilize the economy and revive growth. Price controls were dismantled, the exchange rate was left to float, the marketing boards were abolished, the banks were re-privatized, and trade was liberalized. Gradually investment responded, particularly after the signing of the Peace Agreements in 1992, with GDP increasing at an annual average rate of 6 percent during the period. The deterioration in the terms-of-trade, owing in part to a large drop in coffee prices, and the two earthquakes in 2001 contributed to slowing GDP growth to an average of 2.8 percent per year during the period. During this period, a number of so-called second generation reforms were implemented, among them the privatization of telecommunications and electricity distribution, together with reforms of the pension system and the judiciary. In 2001, the government decided to dollarize the economy. vii

9 Headcount poverty declined rapidly during the 1990s. The intensity and severity of poverty also declined during the decade. According to official estimates, poverty affected 66 percent of the population in 1991; 33 percent of the population was in extreme poverty and another 33 percent was in moderate poverty. Poverty was higher in rural areas (71 percent) than in urban areas (60 percent). By 1995, total poverty had declined to 54 percent or by 12 percentage points, with the absolute decline being higher in urban areas than in rural areas (14 vs. 7 percentage points). During the period, poverty was further reduced to 43 percent or by 11 percentage points, with the absolute decline again favoring urban areas (12 vs. 8 percentage points). During the period, total poverty declined by 23 percentage points, 26 points in urban areas and 15 percentage points in rural areas. World Bank staff have made some adjustment to the official income estimates and confirmed the decline in poverty during the period. Poverty declined by a similar percentage during the first and second part of the 1990s (22 percent and 20 percent), though GDP growth was twice as fast during the former period compared to the latter. While the decline in urban poverty was lower in the second part of the 1990s (26 percent versus 34 percent), the decline in rural poverty was twice as large during the second part of the 1990s compared to the early 1990s (15 percent versus 7 percent). The implied growth elasticities of poverty are discussed below. The distribution of income and assets remains highly unequal. Income distribution deteriorated slightly during the 1990s. Official estimates show that the distribution of income improved somewhat during the first part of the 1990s but then deteriorated during the period. World Bank estimates also show a slight deterioration in income distribution during the 1990s. Despite the agrarian reform, land assets continue to be concentrated and the poorest children still have two years less of schooling than the richest children. Growth, Distribution of Income, and Poverty El Salvador s productive structure has changed significantly since the 1960s. Traditional agricultural exports (coffee, cotton, sugar) saw their contribution to the country s value added decline. In contrast, services and industry rose in importance. The change in the composition of output was not accompanied by an increase in the overall productivity of the economy. Indeed, over the last forty years, growth in El Salvador has been a result mostly of factor accumulation rather than productivity gains. Without considering human capital adjustments, total factor productivity s annual contribution to growth during the last four decades was -0.3 percent. On the other hand, annual average contributions of capital and labor to growth were 1.5 percent and 1.8 percent, respectively, for an average growth rate of 3 percent during These results are influenced by the destruction of capital and heavy emigration during the years of the internal conflict. Adjusting for human capital does not significantly change the results. Analysis of the determinants of growth indicates that the structural reforms of the 1990s may have increased the country s growth potential. The analysis indicates that if El Salvador is to increase its long term growth potential, the country must improve its viii

10 infrastructure, invest in education, and continue to open its economy. While analysis of the correlates of poverty supports the conventional findings related to geographic location, employment and household characteristics, there is evidence that unobserved differences among and between households play an important role in the determinants of poverty (and inequality) in El Salvador. The unobserved (unmeasured) heterogeneity of households and their member in determining per capita income related, for example, to education may include labor market connections, family human capital, school quality, and/or work ethic. The elasticities of poverty to growth and inequality for El Salvador are relatively high compared to their theoretical value, which underscores the importance of growth and improved income distribution for poverty reduction. The implied growth elasticities of poverty were much higher (in absolute terms) during the second part of the 1990s than during the first part. However, this large increase results to a large extent from the fact that the official poverty lines lagged the increase in the consumer price index. Taking this lag into consideration, the elasticity is reduced to values similar to those in the earlier period. Nonetheless, growth was more regionally balanced during 1995/2000 than during The growth incidence curves and the pro-poor growth rates indicate that these contrasting outcomes are related to changes in rural household incomes. Incomes of the poorest rural households dropped in the first part of the 1990s but recovered in the second part.. Factors Affecting the Participation of Poor People in Growth Some of the key factors that influence pro-poor growth are: macro instability, public spending policies, rural development, remittances, labor market conditions, and gender policies. Macro instability leads to insecurity which, like inequality, impairs growth and poverty reduction, as it deters investment. Insecurity about future employment and income also directly and adversely affects welfare because most households and workers care not only about the level of their standard of living, but also about its security. El Salvador was subject to intense volatility in consumption, wages and employment during the first part of the 1990s, as a consequence of the adjustment process. While the volatility of macro variables declined in the second part of the 1990s, there continued to be a sense of insecurity among the population, a phenomenon which may have reflected political factors as well as the impact of a more open economy. Public spending in the social sectors increased significantly during the 1990s, in part owing to the peace dividend. In 1996, social expenditures represented 31 percent of the budget; by 2003 they represented 46 percent. These expenditures have mainly benefited the lower income groups, a fact which is reflected in improved social indicators and access to basic services by the poor. Nevertheless, El Salvador still spends too little on education compared to the average for Latin America and the Caribbean. While the annual average growth rate of agriculture declined from 2.3 percent during to 0.6 percent during , the per capita income of the poorest rural households increased in the second part of the 1990s at rates over 5 percent. Using panel ix

11 data from FUSADES for the period, we explore the factors that impacted on income growth among the poorest rural households. The analysis indicates that the key drivers are access to non-agricultural employment and the possibility of establishing micro-enterprises, both of which require improved human capital and access to basic infrastructure. Remittances are an important source of income growth for better-off households. The poorest households are not likely to receive remittances or credit. To invest in human capital or establish micro enterprises, the poorest households may be forced to sell the few assets (cattle) that they possess. Better-off households may use remittances for such purposes. Also, while the poor families sought other sources of income outside agriculture, they intensify their work in agriculture. Many poor families rented land (in) to cultivate it. An increasing dynamic land market appears to have facilitated the resignation of resources that led to the increase in incomes. To the extent that better-off families receive higher remittances than poorer families, remittances contribute to inequality; on the other hand, remittances are an important factor in poverty reduction. With respect to the labor market, we find that women s participation has increased, particularly in rural areas, most likely owing to the agricultural crisis and the need to find non-agricultural jobs such as in maquila, and that there has not been a major change in the composition of employment categories, though the relative importance of unremunerated family members has tended to decline while that of wage earners has tended to increase. This is consistent with an increase in the formalization of labor relations. On the other hand, the labor market appears to be quite flexible, with most unemployment being of short duration. Finally, gender discrimination in the education sector appears to be on the decline, but it remains a serious problem in the labor market, particularly for women with more education. Trade-Offs Between Growth and Pro-Poor Growth While growth contributed substantially to poverty reduction during the 1990s, inequality appears to have increased slightly during the same period. Some of the policies that promoted growth may have also increased inequality. Since increased inequality works against poverty reduction, there may exist some trade-off between pro-growth policies and pro-poor policies. We explore how macro stabilization and trade liberalization policies impacted poverty and inequality. While stabilization policies are necessary for growth, they have different effects on wages and unemployment and therefore on poverty and inequality. Wage flexibility may help spread the cost of the adjustment, while unemployment has a more unequal effect. In the case of El Salvador, we find that adjustment has been mostly through changes in wages rather than through unemployment. In the future with the dollarization of the economy, the brunt of future adjustments may fall on employment. Thus, it will make labor market regulations that detract from wage flexibility more costly. As for trade liberalization, though it is expected to promote growth over the longer term, it may have adverse short term effects on income distribution and therefore on poverty x

12 reduction. In the case of El Salvador, there is no evidence of such a negative short tem impact. Conclusions and Recommendations For Policy Making After experiencing a major conflict during the 1980s, El Salvador achieved substantial poverty reduction during the 1990s. The preceding analysis indicates that the following elements have contributed to this outcome: Successful implementation of the Peace Accords, the focus of the National Reconstruction Program on the poor and demobilized, and the gradual building up of democratic institutions all created a favorable environment for growth and social investment; Successful stabilization and structural adjustment reforms led to high growth rates, particularly in the first part of the 1990s, and a drop in inflation; rapid growth in the early 1990s helped to raise many of the poor closer to the poverty line, as evidenced by a decline in the poverty gap; subsequent growth, though less rapid, helped to push many of them above the poverty line; A flexible labor market helped reduce the impact of adjustment on unemployment, which usually bears most heavily on poor households; adjustment via wages helped distribute the cost of adjustment among different income groups; Higher public social/basic infrastructure expenditures were made possible by growth (creating higher fiscal revenues) and by a new political consensus on the importance of education and other social investment; Gender discrimination fell; Rural incomes rose because of non-agricultural job opportunities (micro enterprises, maquila, services), supported by infrastructure investment and rising human capital formation, which in turn made it possible for the rural poor to take advantage of opportunities that presented themselves; and Higher remittances made it possible to finance human capital formation and physical investment in micro enterprises. These findings suggest the following are the key areas for making growth in El Salvador more pro-poor in the future. Restore rapid growth. Forty-three percent of El Salvador s population still lives in poverty. Growth has slowed in recent years and this has in turn slowed the rate of poverty reduction. Therefore, the first order of business should be to restore rapid growth. Analysis of the determinants of growth indicates that this requires investment in education and infrastructure, while keeping the economy open. The need to continue to invest in education and basic infrastructure is now an accepted priority in El Salvador, though the political consensus required to mobilize the resources to finance these needs may be lacking. Keeping the economy open is a more xi

13 controversial proposition, however. To minimize the possible negative effects of trade liberalization or trade agreements in countries with very unequal distribution of income such as El Salvador, the World Bank recommends, among other things: i) Improve the distribution of assets, because assets improve households and individuals capacity to take advantage of new opportunities; and ii) strengthen social safety nets, because the process of adjustment can lead to temporary or permanent income losses. Both income and asset distribution are still highly concentrated in El Salvador. Improving the distribution of assets could be thought to require a new round of land redistribution. But the World Bank s Rural Development Study published in 1998 indicates that considering the scarcity of farm land in El Salvador these estimates underline the unreasonableness of relying primarily on land redistribution to alleviate poverty among the rural poor and the importance of non-land factors. Thus it appears that the appropriate route to better asset and income distribution is investment in human capital and basic infrastructure, which makes it possible for the poor to take advantage of economic opportunities and move out of poverty. For those who face temporary income losses or are not able to help themselves, there is a need to strengthen the social protection system as discussed below. Maintain a flexible labor market. With the dollarization of the economy it is expected that, in the future, the labor market will absorb the brunt of any required adjustment. Since inflation is now very low, any required adjustment in the future should mostly impact employment and unemployment, because nominal wages are usually inflexible downwards. To facilitate adjustment, labor market flexibility should be maintained. Wage flexibility may help spread the cost of adjustment, while unemployment has a more unequal effect. Labor market flexibility means giving employers and employees alike efficient mechanisms with which to adapt to the business cycle It is important to ensure that the mechanisms used to set minimum wages are based not only on purchasing power, but also on considerations related to productivity and employment. Strengthen the social protection system. The existence of an adequate social protection system (insurance and social assistance programs) will facilitate the adjustment process. Since in the future, most of the consequences adjustment may fall on employment, the question arises as to whether current unemployment policies are appropriate. Severance payments should be evaluated to see whether they are better than alternative mechanisms, such as individual savings accounts. These are a funded version of the severance pay program. Social assistance programs should be strengthened, particularly those directed at the most vulnerable groups such as poor children, the disabled and the elderly, while at the some time building institutional capacity and cost effective delivery models. Act on the drivers of poor household income growth. To make growth more pro-poor in El Salvador, there is a need to act on the factors that drive income growth among the poor. Analysis of poverty correlates and of rural household income growth indicates that these are education and basic infrastructure. Investment in these areas improves the access of the poor to assets and services that enable them to take advantage of economic opportunities such as accessing non-agricultural employment and micro-enterprises. xii

14 Tailor policy interventions to the poorest households. Finally, an important implication of the findings of this study is that the poorest households may face specific constraints that require detailed analysis and tailor-made policy interventions. Arias s analysis of poverty correlates for different income quintiles (Section II) concludes that household variables that are common in survey data may not fully capture variations in socioeconomic performance and in the likely impact of public interventions. Our own analysis in Section III of the drivers of income growth finds that the poorest households have income drivers that differ from other households. Thus, the design of pro-poor growth policy interventions for the poorest households requires a detailed analysis of the specific drivers of their income growth. Pro-poor investment agenda and financing. What specifically should be done? What should comprise a pro-poor growth investment program? The recent Public Expenditure Review (PER) and Poverty Assessment (PA) prepared by the World Bank indicate that in education the government should focus on measures to increase the coverage in the 3rd cycle of basic education and in secondary education, particularly for the lowest income students. At the same time, there is a need to ensure that all people have access to quality healthcare. On basic infrastructure, access to potable water and suitable sanitation facilities in rural areas is increasingly recognized as an important health input as is improved access to all weather roads in rural areas translates into reduces isolation, lowers the costs for goods and services (including accessing education and health facilities), and increases access to markets, enabling people to better take advantage of emerging economic opportunities. On social protection, it will be important to establish an institutional mechanism for coordinating programs that would support greater emphasis on high priority, high return areas (e.g., early childhood interventions), and identification and scaling up of cost-effective models. How much would cost such pro-poor growth package? The PER and PA estimate that the cost would be equivalent to between 3.2 percent and 3.6 percent of GDP. This cost will be distributed gradually over several years. How could this investment package be finance? Within the current fiscal envelop, there is very little margin to increase social spending. In 2003, the fiscal deficit was about 3 percent of GDP and the public debt, 41 percent of GDP. El Salvador has a low tax burden (12.1 percent of GDP in 2003) with most government tax revenues originating from the VAT or value added tax (53 percent of total tax revenues), income taxes (29 percent), import duties (10 percent); and excise taxes on alcoholic beverages, tobacco products, and fuels (4 percent). Any decision to increase taxes to finance a pro-poor growth package will need to take into account its incidence on the poor. A recent study on tax incidence in El Salvador found that in general the system is regressive. The only two taxes that are not regressive are the personal income tax and the gasoline tax. The most regressive taxes are the VAT on domestic goods and the tax on cigarettes. To raise the revenues required to finance the pro-poor growth package the authorities should give priority to reduce tax illusion and evasion, particularly in the income tax. Other additional tax revenue measures that the authorities may consider should ensure a net positive resource transfer to the poor. xiii

15 Introduction Poverty has remained high in El Salvador as the country has gone through episodes of both very rapid growth and sharp declines in output during the last four decades. In the 1960s the country embarked on an import-substitution effort which initially led to rapid growth; as the easier opportunities for import substitution were exhausted, growth began to decelerate in the 1970s; in the 1980s, El Salvador experienced a destructive internal conflict which caused a large drop in GDP; in the 1990s, it initiated a structural adjustment reform that was initially accompanied by rapid GDP growth, but growth has slowed since the mid-1990s. In general, over the last 40 years, growth has been low, averaging 3 percent. Owing in part to the conflict, today s real per capita income is similar to that of 30 years ago; and poverty, though it was reduced by one-third during the 1990s, still affects about 40 percent of the population. This case study on El Salvador s growth and poverty reduction experience is part of a broader effort by the World Bank and sponsoring organizations (Agence Francaise de Developpment, Department for International Development, GTZ, KFW) to better understand the factors that contribute to making growth pro-poor. The Operationalizing Pro-poor Growth work program aims to (a) propose a common methodology for defining and measuring pro-poor growth; (b) develop an analytical approach for designing propoor growth strategies; (c) provide operational guidance on key macro, sectoral and thematic policies and how they relate to poverty and growth; and (d) investigate how country context and initial conditions might affect the selection and prioritization of such policies. This study contributes to the work program by exploring the factors that affect the ability of the poor to participate in economic growth and to benefit from it, and seeks to identify the policies that will make growth in El Salvador more pro-poor. The pro-poor growth definition used in this study is an absolute definition of pro-poor growth that requires only that the incidence of poverty declines with growth. This study illustrates the case of a country which after experiencing a major conflict during the 1980s, achieved substantial poverty reduction led by growth as the country recovered and began to put some of the institutions, markets and policies in place for a broad based growth pattern within a newly established democratic framework. The key questions addressed in this study are: 1. What explains El Salvador s relatively low growth during the last four decades and continuing high levels of poverty? What are the key determinants of growth? 2. After the destructive conflict of the 1980s, poverty declined substantially. What explains the large reduction in poverty during the 1990s? What are the key poverty correlates? How have policies related to macroeconomic development, public expenditure, rural development, remittances, the labor market, and gender 1

16 affected the participation of poor people in growth? Were there trade-offs between pro- growth and pro-poor policies in the stabilization and structural adjustment policies pursued? 3. Despite the deceleration of GDP and agricultural growth, the pace of reduction of poverty in rural areas during the second part of the 1990s was twice as fast as during the first. What were the key drivers of change in rural household incomes and poverty reduction during the second part of the 1990s? And, what lessons can be learned from this experience to make growth more pro-poor in the future? To help answer these questions, the study builds on and complements several strands of ongoing work at the World Bank and in-country, including the 2003 Country Economic Memorandum, the Poverty Assessment, the Public Expenditure Review and a recently concluded assessment of progress towards the Millennium Development Goals. The empirical analysis is based largely on three different sets of data. One is the national accounts estimates prepared by the Central Reserve Bank of El Salvador. We use these data to analyze the sources of growth, including factor and sector decomposition to help answer the first question. We find that the lack of productivity growth has been a major cause of low growth, which, together with high levels of inequality have contributed to high poverty rates. The second set of data consists of the household surveys conducted every year by the Directorate General of Statistics, DIGESTYC. Although there are some problems of comparability between surveys, their methodologies are relatively homogeneous for the period. These surveys have national coverage and the estimates are statistically representative for rural and urban areas. The surveys report household income rather than expenditures. DIGESTYC computes poverty estimates based on national poverty lines for the extremely and the moderately poor, as well as for urban and rural areas. These data will help answer the second set of questions as we explore to what extent social expenditures (education, health and basic infrastructure) have been pro-poor (incidence analysis) and discuss the pro-poor growth estimates. We find that growth has helped reduce poverty substantially in the 1990s, though income inequality has increased slightly, and that public expenditures have benefited the lowest income groups relatively more than others. The household survey data also help establish the importance of remittances for poverty reduction (although they appear to increase inequality), and provide insights into the flexibility of El Salvador s labor market. The household survey earnings data further help us to examine whether trade-offs between growth and pro-poor growth were involved in the adjustment and trade liberalization policies pursued during the 1990s. We conclude that such trade-offs were not present in the case of El Salvador. The third set of information consists of panel data collected by FUSADES 1 every other year since 1995, which will help answer the third set of questions. These data are nationally representative of rural areas and provide a good basis for evaluating the 1 FUSADES is a private sector foundation created in the mid-1980s which, among other things, conducts economic and social research in its Department of Social and Economic Studies. 2

17 determinants of rural household income growth, as well as the key transmission mechanisms between growth and poverty. We find that the key drivers of income growth among the poorest rural households are related to access to non-agricultural employment and the possibility of establishing micro-enterprises, both of which require improvements in human capital and basic infrastructure. The study has four main messages. First, growth and pro-poor spending brought about by the stabilization and structural reforms initiated in the early 1990s, the peace agreements, and the political consensus on the priority of social policies, together with increased remittances, helped to reduce poverty by one-third in the 1990s, not a small feat. Second, given El Salvador s continuing high levels of poverty and inequality and the recent deceleration in growth and poverty reduction, further progress in reducing poverty will require restoring rapid growth. Analysis of the determinants of growth indicates that this in turn will require investment in education and infrastructure. At the same time, there is a need to keep the economy open and the labor market flexible while strengthening the country s social protection system to minimize the impact of negative shocks on poor households. Third, to make growth more pro-poor in El Salvador, there is a need to act on the drivers of income growth of the poor. Analysis of poverty correlates and rural household income growth indicate that these drivers are education and basic infrastructure. Investment in these areas improves the access of the poor to assets and services that enable them to take advantage of economic opportunities such as accessing non-agricultural employment and microenterprises. And fourth, the design of pro-poor growth policy interventions for the poorest households requires a detailed examination of the specific drivers of their income growth, because analysis of income growth by quintile indicates that the poorest households may face constraints not easily identifiable through examination of standard household survey data and that this poorest group may differ substantially from other households. The paper is organized as follows: Section I provides the historical context and reviews growth, poverty and income distribution trends. Section II analyzes the sources and the determinants of growth, the poverty correlates, and the interaction between growth, poverty, and income distribution. Section III identifies factors that affect the participation of the poor in growth, including those related to macroeconomic instability, public spending, rural development, remittances, the labor market, and gender policies. Section IV discusses possible trade- offs between growth and pro-poor growth in stabilization and trade liberalization policies. Finally, Section V presents the conclusions and recommendations. The paper includes two annexes that provide empirical estimates. 3

18 I. Historical Context and Growth-Poverty Trends This Section sets the stage for what follows. It gives an overview of the historical context and reviews trends in growth, poverty and income distribution. The discussion on poverty and income distribution is mostly restricted to the 1990s because of data limitations. After pursuing an import substitution strategy in the 1960s and 1970s, and experiencing strong government intervention in the economy and a very destructive conflict in the 1980s, El Salvador was in crisis at the beginning of the 1990s: poverty was widespread, social conditions lagged, the country s economic infrastructure was shattered, and economic growth was minimal. The 1990s saw a major turnaround. Poverty declined by one-third during the decade owing to the resumption of growth brought about by the stabilization and structural reforms initiated in early 1989 and the peace agreements of 1992, together with increased remittances from Salvadorans living abroad. But despite progress in poverty reduction, evidence is presented at the end of the Section that indicates that assets and incomes continued to be highly concentrated in El Salvador. 1. Historical context El Salvador is a small country with a population of 6.5 million. The population is young with 55 percent of Salvadorans aged under 24 years. Fifty-nine percent of the population lives in urban areas and 41 percent in rural areas. Per capita gross national income in 2002 was US$ 2,080. The country is prone to natural disasters earthquakes, floods, and droughts which form an ongoing threat to the country s population, particularly to those living in high-risk zones. In the last two decades, three major earthquakes hit the country; one in 1986 that mostly impacted the capital city of San Salvador; and two in early 2001 that devastated a large part of the national territory, with 1,160 dead and 8,000 injured. Hurricane Mitch brought floods and destruction in 1998 to the eastern part of the country. The atmospheric phenomenon El Niño impacts the country periodically. Recurrent natural disasters give the population a sense of insecurity that negatively affects welfare (this issue is discussed further in Section III). Table 1.1 El Salvador s GDP and Sector Growth, 1960s-1990s (simple average of annual rates) 1960s 1970s 1980s 1990s GDP Agriculture Manufacturing Services Memo: LAC s GDP Source: World Bank s World Development Indicators (WDI) In the last four decades El Salvador has undergone two major development experiments separated by a major internal conflict. In the 1960s and 1970s, it pursued an import substitution strategy. El Salvador was one of the founders of the Central American 4

19 Common Market (CACM) established in Under high tariff protection, the CACM facilitated rapid industrialization, which was financed by agriculture surpluses mainly generated by coffee, the country s principal export crop. GDP grew at 6 percent annually during the 1960s, led by manufacturing, which grew at 8.7 percent per year. GDP growth during this period was more than double the average for Latin America and the Caribbean (LAC) (Table 1.1). In the 1970s there were signs that import substitution opportunities were quickly becoming exhausted, with average annual manufacturing and GDP growth declining to 3.1 percent and 3.9 percent, respectively. In the 1980s, the Central America region, including El Salvador, became engulfed in a conflict which was part of the broader Cold War confrontation. The conflict and a series of misguided policies contributed to a decline in GDP averaging 1.9 percent per year during the 1980s, including large drops in agricultural and manufacturing output. In 1989 a new government took office. It embarked on a major stabilization and structural adjustment program and initiated negotiations with the armed opposition (FMLN), reaching a Peace Accord in early As the country began building its democratic institutions, the governments pursued an export led growth strategy. GDP grew at an annual rate of 4.9 percent led by manufacturing and services; agricultural growth lagged, increasing at an annual average rate of only 2.2 percent. The stabilization and trade liberalization efforts may have involved trade-offs between pro-growth and pro-poor polices, as discussed in Section IV. The conflict of the 1980s cost the lives of over 70,000 Salvadorans and destroyed the country s infrastructure. Rural areas were those most affected by the destruction. Thousands of Salvadorans emigrated, mainly to the US. According to the Directorate General of Statistics (DIGESTYC), annual net migration increased from 17,600 during , to 32,200 during , to 69,000 during and numbered 43,800 in In the first part of the 1990s, annual net migration is estimated at 11,400; in the second part, at 7,600. Over one million Salvadorans live in the US alone. Some policies of the late 1970s and early 1980s were costly for the country. They included the nationalization of the banking system, the institution of price controls, the maintenance of a dual exchange rate regime that penalized traditional exports, and the creation of state export marketing boards for coffee and sugar. In the early 1980s, an agrarian reform confiscated all properties over 245 ha, and transferred this land to landless farmers and new agricultural cooperatives. The agrarian reform did not yield the anticipated results, in part because of the intensification of the conflict and the drop in agricultural prices (this issue is discussed further in Section III). 2 2 World Bank (1994) El Salvador: The Challenge of Poverty Alleviation, Report No ES, Annex A, page 4. 5

20 Figure 1.1. El Salvador: Savings and Investment, % of GDP s 1970s 1980s 1990s Gross capital formation Gross domestic savings Gross national savings Source: World Bank (WDI) Investment and domestic savings dropped sharply during the 1980s. Gross capital formation (GCF) had increased from 14.4 percent of GDP in the 1960s to 19.3 percent of GDP in the 1970s, but then fell to 12.9 percent of GDP in the1980s (Figure 1.1). Gross domestic savings (GDS) followed a similar pattern: they increased from 11.9 percent of GDP in the 1960s to 16.6 percent of GDP in the 1970s and then dropped to 6.9 percent of GDP in the 1980s. In LAC as a whole, average GCF and GDS remained above 20 percent of GDP during the period. By the end of the 1980s, El Salvador was in economic and social crisis. Per capita income was only 70 percent of its 1978, pre-conflict, level. In urban areas, unemployment affected 9 percent of the labor force (23 percent of those in the bottom quintile) and underemployment affected 50 percent (72 percent of those in the bottom quintile). In rural areas, the employment situation must have been even worse. 3 The country s economic infrastructure was shattered. Table 1.2 Selected Education and Health Indicators, 1989/90 (Percentages) Adult Illiteracy Rate Gross Primary Enrollment Gross Secondary Enrollment Gross Post- Secondary Enrollment Life Expectancy at birth (years) Infant Mortality Rate (per 1000 live births) Under 5 Mortality Rate (per 1000 live births) El Salvador LAC Source: Marques, José Silvério and Iann Bannon (2003) Central America: Education Reform in a Post- Conflict Setting, Opportunities and Challeges, World Bank, Conflict Prevention and Reconstruction Unit, Working Paper 4, Tables 3 and 4 and World Bank (WDI). 3 Data for rural areas in 1985 show unemployment at 16 percent (35 percent of those in the 1 st quintile) and underemployment at 54 percent (59 percent of those in the 1 st quintile). World Bank (1994) El Salvador: The Challenge of Poverty Alleviation, Report No ES, Annex A, page 3, table A-2. 6