Share This
Artwork from March Cover

Search Finance & Development

Advanced Search
About F&D


Back Issues

Write Us

Copyright Information

Use the free Adobe Acrobat Reader to view a pdf file of this article

Free Email Notification

Receive emails when we post new items of interest to you.

Subscribe or Modify your profile

Finance & Development
A quarterly magazine of the IMF
March 2000, Volume 37, Number 1

Poverty Reduction
Nora Lustig and Omar Arias

Although Latin America and the Caribbean reduced the incidence of poverty during the 1990s, population increases and the greater income inequalities that had developed during the 1980s stymied the region's efforts to reduce the number of poor people. How can its policymakers fight poverty most effectively and better protect the poor during economic crises?

The incidence of poverty in Latin America was approximately 3 percent higher, and roughly 70 million more of its people were living in poverty, in 1997 than in 1980 (see table). The limited progress made to date in combating poverty is partly due to the impact of the 1980s debt crisis. Although economic growth resumed during the 1990s (averaging 3.3 percent a year for the region during 1990-98), it was not enough to produce notable progress in poverty reduction. This was partly due to the increasing inequality of income during the 1980s (see table), which was not reversed during the 1990s: with greater inequality, a given growth rate brought about a slower rate of poverty reduction.

All in all, the lack of substantial progress in reducing poverty can be linked to recurrent economic downturns and the increase in earnings inequality.

Table: Poverty and inequality trends in Latin America, 1980-96

Crises, inequality, and poverty

Macroeconomic crises, which have been a recurrent phenomenon in Latin America and the Caribbean during the past twenty years, are perhaps the most important cause of large increases in poverty in the region. Because of the pervasiveness of such crises, the 1980s came to be known as the region's "lost decade." Although the region's experience during the 1990s was somewhat better, 24 countries experienced at least one year of decrease in per capita income. Altogether, between 1980 and 1998, there were more than forty episodes in which annual per capita GDP fell by 4 percent or more. These numbers will increase once results for 1999 are included.

In all crises for which data are available, the incidence of poverty increased at the onset of the crisis, and in all cases it reached levels—after between one and five years, depending on the country, had passed—that were higher than before the resulting recession had begun. Crises have frequently been accompanied by increasing income inequality as well. Inequality rose at the onset of the crisis in 5 out of 8 episodes for which data are available; and in 15 out of 20 episodes, inequality was greater after the onset of the crisis than before.

Fields (1991) has estimated that, on average, for every percentage point decline in growth, poverty rises by 2 percent. Others have found that had Latin America reached the levels of macroeconomic stability achieved by industrial economies, roughly 25 percent of its poor people would have been lifted out of poverty. Because crises in Latin America and the Caribbean tend to be accompanied by increasing inequality, economic contraction leads to greater-than-proportional reversals of previous gains in poverty reduction: each 1 percent decrease in per capita income during a recession in the 1980s wiped out the reductions in poverty that had been brought about by increases in per capita income of 3.7 percent in urban areas, and 2 percent in rural areas, during the 1970s. Also, crises ratchet up inequality: subsequent recoveries tend not to eliminate the greater inequality generated during a severe economic downturn.

That crises result in relatively high levels of transient poverty is uncontroversial, but they can also be a cause of persistent or chronic poverty, because of the irreversible impact that income downturns may have on the physical assets and human capital of the poor. Although social indicators such as infant mortality rates continued to improve in Latin America during the crisis years of the 1980s, they did so more slowly than in the previous decade. Health indicators that were more sensitive to consumption or income downturns worsened. In Chile, the data on low-birth-weight infants and undernourished children follow the trends in economic conditions during the 1980s, following a systematic improvement in both indicators during the 1970s. In Mexico, infant and preschool mortality caused by nutritional deficiency rose during the 1980s, reversing the trend of the previous decade. In Argentina, daily per capita intake of protein decreased by 3.8 percent in 1995, and in Venezuela it decreased by 2.9 percent in 1994.

Recent research has found a link between macroeconomic downturns and changes in education indicators. For example, the average increase in years of schooling for 18 Latin American countries slowed from 1.9 years during the 1950s and 1960s to 1.2 years during the 1970s and 1980s. More specifically, improvements in schooling attainment start to decline for children born between 1960 and 1970, and decline further for those who entered the school system between 1975 and 1986—the period that roughly coincides with the region's debt crisis. Worsening macroeconomic conditions (short-term GDP shocks, volatility, and adverse trade shocks) explain 80 percent of the decline in the rate of improvement in schooling attainment.

Because crises affect investment in schooling, nutrition, and health, potentially reducing the human capital (work-related knowledge, information, and skills) of the poor, they can hinder their ability to grow out of poverty. Furthermore, an irreversible impact on the human capital of the poor is not only bad for the poor but also can weaken the overall performance of the economy in the medium run. This is especially true when nutrition and educational attainments suffer during recessions. Avoiding such an adverse development is another important part of the economic rationale for publicly funded safety nets.

The evidence presented here should establish that crisis avoidance and adequate crisis response should be high priorities in any country's antipoverty strategy. Responses to macroeconomic crises when they do occur can be more, or less, sensitive to the plight of the poor. A sensitive response should help the poor by providing for adequate consumption levels, ensuring that they continue to have access to basic social services, preventing irreversible impacts on their human capital, and encouraging alternatives to dysfunctional behavior, such as prostitution, other criminal activities, or the abuse of child labor.

The optimal policy response to a shock combines the necessary balance of payments adjustment with the smallest possible decline in output, consistent with initial conditions in the economy. Would macroeconomic responses to crises that are optimal for the economy as a whole differ from macroeconomic policies that are optimal for the poor? Perhaps. Conflicts can emerge between the interests of the poor and the nonpoor, and within the poor (between the urban and rural poor, for example), when different policy combinations result in different distributions of income.

Even if everybody's income falls by the same percentage—an outcome that would appear to be as equitable as possible—the interests of poor people may still not be well served. Consider that a country could choose between several adjustment policies, with the main trade-off being between a sharper decline in output in the short run with a higher level in the medium run, or a more gradual decline in the short run with a lower level in the medium run (with everybody's income changing in the same proportion). The welfare ranking of different adjustment packages for the poor (whose ability to smooth consumption is clearly limited) and for the economy as a whole may differ. The poor may prefer a more gradual adjustment, even at the expense of a slower recovery.

These observations should not be seen as implying that pro-poor policymakers should necessarily adopt the path that is optimal for the poor (although this may be warranted in some circumstances). If policymakers are worried about the welfare of the poor, they should introduce safety nets to compensate the poor (at least in part) for the costs imposed on them by policymakers' choice of the optimal path for the economy as a whole. Loans and grants from multilateral institutions and donors can be used to provide such compensation during an adjustment program. The multilateral organizations can induce a country to choose the optimal adjustment path but should also make sure that it undertakes the proper compensatory policies. Safety nets should not be an afterthought.

At present, most Latin American and Caribbean countries need to improve their mechanisms to protect the poor from the brunt of economic crises. While there is a widespread perception that social investment funds were put in place for precisely that purpose, a closer examination reveals that most of them were more effective at building small-scale social infrastructure than at creating employment opportunities for those hurt by the emergency. In fact, most countries in the region lack effective consumption-smoothing safety nets that could protect the poor against the output, employment, and price risks associated with systemic adverse shocks.

A recurring problem is that because the institutional mechanisms to protect the poor are not already in place, responses to crises frequently have to be improvised or make use of programs designed for other purposes. Policymakers must often formulate emergency responses to crises without taking time to do the technical analysis that is needed both to clarify the socioeconomic profile of groups most vulnerable to the adverse shocks and to evaluate the cost-effectiveness of different social-protection options.

Examples can be found, inside and outside Latin America, of safety nets that work well. Ideally, safety nets should simultaneously provide a consumption floor and protect the human capital accumulation of the poor or contribute to expanding the social and physical infrastructure for the poor. Targeted human development programs are one example of an effective safety-net program that transfers income, in cash or in kind, to poor households with children and condition the transfers on the household's investment in the human capital of their children (such as school attendance and health-care visits). The income-support component reduces current poverty, and improving nutrition and health, as well as the education of children, increases poor people's future earning capacity.

Workfare programs can also function as effective safety nets. By offering wages in exchange for work, these programs aim to transfer resources to unemployed and, in most cases, unskilled workers while minimizing the perverse incentives that discourage work. It is important to note that if the wage rate such a program offers is lower than market wages for unskilled workers, the program will appeal to only those workers who have few alternative employment opportunities. Such programs can provide unemployment protection for poor workers in response to aggregate, regional, sectoral, and idiosyncratic shocks. They can be even more valuable if they provide training for unskilled and poor workers and improve the social and physical infrastructure in poor areas.

Trends in unemployment and wages

In addition to the region's recurring crises, the limited progress it has made in reducing poverty has resulted from its economies' failure to create enough job opportunities for its low-skilled workers. During the 1990s, economic growth coexisted with high unemployment rates and a widening of earnings disparities: while the labor force grew at an average annual rate of 3.2 percent, the number of jobs increased at an annual rate of 2.9 percent. As a result, open, or reported, unemployment has soared and underemployment has also expanded. According to the United Nations Economic Commission for Latin America and the Caribbean, the open unemployment rate reached 8.4 percent (this does not include Caribbean countries) in 1998, up from 5.8 percent in 1991. In addition, the share of informal employment continued to rise, increasing from 51.8 percent in 1990 to 57.7 percent in 1997.

Wage disparities in Latin America, already among the largest worldwide, became twice as big as those in developed countries during the decade. These growing wage disparities have often been associated with uneven earnings growth owing to differences in human capital (for example, in education) and in characteristics of jobs (for example, managers versus blue-collar workers, modern sectors versus traditional ones, informal versus formal employment). In effect, the earnings gap between professionals and technical workers, and less skilled workers (especially those in low-productivity jobs) widened, on average, by more than 4 percent a year in 10 out of 14 countries.

The role of increasing compensation for general and specific skills in bringing about increasing wage disparities has been documented in a number of studies. Bourguignon, Ferreira, and Lustig (forthcoming), for example, have found evidence of the returns to education increasing more for the better educated in Brazil, Chile, Colombia, Mexico, and Venezuela. Other studies report increasing wage differentials between workers with primary education and those with higher education in 7 out of 10 countries (Bolivia, Brazil, Chile, Mexico, Peru, Uruguay, and Venezuela) for the first half of the decade. The increasing returns to higher education are seen as a key driver of poverty perpetuation and earnings inequality, because low-income workers receive lower relative, and sometimes absolute, returns on the few assets (for example, primary education) that they are able to accumulate. One proposed explanation for the increasing returns to higher-level skills is that technological change leads to capital deepening, which increases the demand for skilled labor. Others have found that trade liberalization partly explains the rising wage inequality. The patterns found, however, are not homogeneous. Arias (2000) found that the wage growth of specific groups with a given level of education varies with their age: for example, the young reaped most of the gains made by college-educated workers in Argentina and Costa Rica.

Recent studies for both developed and developing countries, indeed, find that the returns to education tend to be higher for workers in high-wage jobs. Thus, the apparently high wage premium for having graduated from college may not be available to everyone. The studies suggest that other differences among workers (for example, in family background, ethnic origin, and quality of education) significantly affect their labor market performance and thus affect the incidence of poverty and earnings inequality within a country through their effects on both asset accumulation and the returns on those assets.

The results discussed here call for policy initiatives on different fronts. On the demand side of the labor market, it is essential to correct credit market failures and eliminate obstacles to the development of micro, small, and medium-sized enterprises. Labor regulations that create dual labor markets, consisting of well-paid "insiders" and poorly paid "outsiders," should also be eliminated. On the supply side of the labor market, a set of policies is needed to increase the schooling (general skills) of the poor. Since the poor are likely to have limited access to credit, scholarship programs for children from low-income families should be introduced. However, this may not be enough. What may be needed is to give the poor access to high-quality education and training in specific skills. In addition, if part of the wage premium is due to discriminatory practices, changes in the law—and, in particular, its application—may be required. If access to high-quality jobs is determined by geographic closeness of applicants, then a greater effort at disseminating information on available employment opportunities could promote equality of opportunity.

Omar Arias, 2000, "Are All Men Benefiting from the New Economy? Male Economic Marginalization in Latin America" (unpublished; Washington: Latin America and Caribbean Regional Office, World Bank).

Fran´┐Żois Bourguignon, Francisco Ferreira, and Nora Lustig, eds., forthcoming, The Dynamics of Income Distribution in East Asia and Latin America.

Alain De Janvry and Elisabeth Sadoulet, forthcoming, "Growth, Poverty and Inequality in Latin America: A Causal Analysis," in Shielding the Poor: Social Protection in the Developing World, ed. by Nora Lustig (Washington: Brookings Institution and Inter-American Development Bank).

Gary S. Fields, 1991, "Growth and Income Distribution," in Essays on Poverty, Equity, and Growth, ed. by George Psacharopoulos (Oxford, United Kingdom: Pergamon Press).

Nora Lustig, 1999, "Crises and the Poor: Socially Responsible Macroeconomics," Presidential Address, Fourth Annual Meeting of the Latin America and Caribbean Association, Santiago, Chile, October 22.

Nora Lustig is Senior Advisor on Poverty and Inequality at the Inter-American Development Bank. When this article was written, she was on a temporary assignment with the World Bank as Deputy Director of the team preparing the World Development Report 2000/1: Attacking Poverty.

Omar Arias is a Junior Professional in the Poverty and Inequality Advisory Unit of the Inter-American Development Bank.