Journal Issue

Unorthodox Adjustment and Poverty in Peru

International Monetary Fund. External Relations Dept.
Published Date:
January 1992
  • ShareShare
Show Summary Details

Faced with economic stagnation and decline during the 1980s, many developing countries adopted macroeconomic and structural adjustment programs that were recommended and partially financed by the World Bank and the IMF. A few countries that were anxious to avoid the perceived economic and social costs of structural adjustment, however, experimented with alternative “unorthodox” (the so-called heterodox) programs. While designed to encourage rapid growth, along with a more equitable distribution of welfare, such alternative approaches did not always succeed and, in fact, may have hurt the poor more than the general population.

The social consequences of World Bank and IMF-supported structural adjustment programs have been widely discussed and criticized in recent years. But whether countries adopting alternative policies have been successful in protecting the poor while reviving economic growth has not been subject to the same scrutiny. The case of Peru, where Bank and IMF-supported structural adjustment policies were deliberately avoided from 1985 to 1990, offers an opportunity to undertake just such an examination. Two unusually rich household-level surveys for Peru provide unique insights into changes in living standards under an “unorthodox” program. The findings are quite sobering; not only did the economy as a whole deteriorate but the policies also failed to protect the poor.

Unorthodox approach, 1985–90

The APRA (Alianza Popular Revolucionaria Americana) government that took office in 1985 inherited an economy that had not weathered the debt crisis well. Despite two stabilization programs between 1981 and 1985, inflation had soared, with prices more than doubling each year after 1982. At the same time, real output declined by 27 percent and gross domestic investment fell by 62 percent.

These depressed conditions reflected both international and domestic factors. Certainly the bankers’ panic ignited by the 1982 Mexican default had a negative impact; although Peru had been meeting debt payments on schedule (and even early) through 1983, loans from international private capital markets were virtually unobtainable by the end of 1982. From 1982 to 1985, Peru’s net capital inflow (public and private) fell by 78 percent, with net private capital flows negative in 1984 and 1985. The lower demand for Peru’s exports, owing to the world recession, produced a 50 percent decline in their unit value between 1980 and 1985; despite an increased volume, total export revenue declined by 24 percent over that period.

The World Bank and IMF approach to adjustment in countries experiencing such deteriorating economic conditions has emphasized immediate stabilization—getting trade and budget deficits under control—followed by a restructuring of the economy to promote sustainable, long-run growth. Such growth is to be achieved by removing price and other market distortions (such as trade barriers), and by encouraging private sector investment.

In the early to mid-1980s, the short-run impact of adjustment policies on poverty was often not explicitly addressed by the Bank/IMF approach. The implicit presumption was that growth would reduce poverty in the long run, and that short-run negative consequences were an unavoidable cost of adjustment. Only in the late 1980s was poverty recognized as a central issue of the adjustment process; incorporating poverty concerns into adjustment policies then became the topic of the World Development Report 1990.

In contrast to Bank/IMF policies, Peru’s unorthodox approach taken in 1985 did not focus on immediate stabilization. Rather, policies were aimed at “jump-starting” consumption and investment through large injections of government resources into the economy. Macroeconomic stabilization, it was thought, could only be attained after an initial period of growth. The government introduced price controls, along with broad-based subsidies and transfers, with the explicit aim of protecting the poor while encouraging demand-induced growth. Further measures included a unilateral reduction in foreign debt payments, which would finance the process of “economic reactivation,” redirecting funds toward domestic expenditures, with priority given to meeting the needs of the poor through wage increases, jobs programs, and investments in education and health.

While these policies fostered rapid economic growth in the short run, their unsustainability was evident by mid-1987, when inflationary pressures and widening budget deficits appeared. Reactivation policies nonetheless continued until September 1988, when the government announced “El Paquetazo,” a severe readjustment of prices that increased consumer prices by 114 percent. Ad hoc adjustment, consisting primarily of price increases without real wage increases, characterized the last two years of the APRA administration. Annual inflation rates (December-December), compared with a base of 158 percent in 1985, rose 1,720 percent in 1988, 2,776 percent in 1989, and were still accelerating when the government changed in late July 1990. From 1987 to 1990, per capita production declined by 25 percent, and the real minimum wage by 60 percent.

Recent appraisals of the Peruvian experience highlight the severe deterioration between 1985 and 1990 in virtually all indicators of macroeconomic performance (see Table 1). Briefly, the country was virtually blacklisted by both public and private international lenders, and thus had no access to foreign credit. Domestically, the use of government funds (generated in part through the diversion of debt payments) to subsidize consumption and investment brought the government to the brink of bankruptcy by 1988. At that point, the government financed expenditures by effectively printing money, and inflation soared. However, none of these studies conclusively examined changes in inequality and poverty during that period.

Table 1A profile of Peru’s macroeconomic deterioration
Index of annual:
GDP per capita1008770
Average real minimum wage, Lima1005421
Consumer prices11003,47440,216,592
Exports (dollars)21007683
Net international reserves310089-13
Source: Glewwe and Hall, 1992.

June of each year, through June 1, 1990.

Estimated from data through September 1990.

June of each year.

Source: Glewwe and Hall, 1992.

June of each year, through June 1, 1990.

Estimated from data through September 1990.

June of each year.

Living standards

Four Living Standards Surveys (see box) have been conducted in Peru. This article compares data from the 1985–86 survey, which covered all of Peru, with the 1990 survey that was conducted in metropolitan Lima only (covering 33 percent of the country’s population). Our discussion of the survey results is, therefore, limited to Lima, the capital.

The measure of household welfare used here is the value of consumption of goods and services. Very careful procedures were used to convert nominal amounts reported in the questionnaires to real monthly values, yielding real household consumption per month. These values are then adjusted for the number and ages of individuals in each household to provide an adjusted per capita consumption figure that reflects the variation in consumption levels of adults and children.

The overall results of the two surveys reveal an average decline in consumption of over 50 percent from 1985 to 1990 (see Table 2), which is without doubt one of the most rapid and severe deteriorations in living standards ever recorded.

Table 2The poorest suffered the biggest drop in consumption
Decile1Lima 1985–86Lima 1990Percent change
All Lima7,774.43,531.7-54.6
Source: Glewwe and Hall, 1992.Note: Figures are valued in thousands of constant June 1, 1990 Intis ($1 = 50,000 Intis) for Tables 2 and 3

Ranked from poorest to richest.

Source: Glewwe and Hall, 1992.Note: Figures are valued in thousands of constant June 1, 1990 Intis ($1 = 50,000 Intis) for Tables 2 and 3

Ranked from poorest to richest.

Changes in overall distribution of consumption expenditures. To examine changes in inequality from 1985–86 to 1990, both surveys divided the population of Lima into ten expenditure groups. As shown in Table 2, decile 1 contained the poorest 10 percent of the population (defined by adjusted per capita consumption expenditures), decile 2 contained the next poorest 10 percent, and so on.

Several conclusions can be drawn from Table 2. First, all deciles experienced declines in their consumption that left them at about one half of their 1985–86 levels. Second, the poorest decile lost the most, experiencing a 62 percent decline in consumption, and the second poorest decile experienced the next largest decline, a drop of 58 percent. Thus the data from these two surveys indicate that the distribution of consumption expenditures in Lima became less equal between 1985–90, despite policy goals to the contrary.

Changes in consumption expenditures by household characteristics. One way to examine the change in consumption expenditures from 1985–86 to 1990 is to look at gender, education level, and employment characteristics of household heads. Table 3 highlights three results.

Table 3Identifying the poorest families in Lima
Percent change in per capita monthly consumption since 1985–86
Education level
Secondary general-55.1
Secondary technical-46.4
Other post-secondary-38.9
Occupation of head
White collar-54.1
All Lima-54.6
Source: Glewwe and Hall, 1992Note: Population percentages do not add to 100 due to missing information for 0.3 percent of observations in 1985–86.
Source: Glewwe and Hall, 1992Note: Population percentages do not add to 100 due to missing information for 0.3 percent of observations in 1985–86.

(1) The position of female-headed households relative to male-headed households in Lima changed very little from 1985–86 to 1990.

(2) In Lima, households headed by someone with no education or only primary education experienced the largest declines in expenditures relative to those with better educated heads. The percentage declines in expenditures of these two groups were larger than for any other population group classified by household head characteristic (with the sole exception of the unemployed). Only households headed by those with secondary technical or other post secondary education exhibit declines in consumption significantly smaller than the average for all of Lima. This suggests that technical education helps reduce vulnerability to severe welfare loss. The fact that those with higher education and technical training also began the period with relatively higher consumption is consistent with the finding in Table 2 that the distribution of expenditures became less equal from 1985–86 to 1990.

(3) Persons living in households where the head was unemployed experienced by far the largest drop in consumption, and expanded from 2.9 percent to 5.1 percent of the population. Further, absolute consumption levels for these households in 1990 were lower than for all employment categories. This contrasts sharply with the economic status of the unemployed in Lima in 1985–86, when households headed by the unemployed had higher consumption levels than all other occupational categories except white collar employees. Thus, unemployment rose substantially between 1985–86 and 1990, and the bulk of the newly unemployed were among the poorest.

Other indicators of living standards. Examining changes in the physical living conditions, based on housing conditions and the provision of social services, is also a good indication of how the poor fared. The quality of homes deteriorated significantly in Lima over 1985–86 and 1990 for households in the poorest 20 percent of the population, where the incidence of housing made of improvised materials (straw, cane, and mud shacks) qua- drupled between 1985–86 and 1990. Access to public services, such as public water and sewage systems, also deteriorated most significantly among the poorest. This decline was particularly evident in the impoverished areas of Lima—Conos Norte, Este, and Sur—whose residents also suffered the greatest declines in consumption.

Measuring the change in poverty

Our findings also show significant changes in poverty in Lima from 1985–86 to 1990, although quantifying such changes is somewhat difficult. We use the standard “head-count” definition of aggregate poverty, that is, the fraction of the total population whose consumption is below the given poverty line. There is no official or widely accepted poverty line in Lima. But it is clear, as will be seen below, that poverty in Lima increased substantially between 1985–86 and 1990. This implies that poverty indexes that take into account the depth of poverty (the “gap” between the poverty line and each household’s actual expenditures), as well as distributionally sensitive indexes (those that account for inequality among the poor), would show large increases in poverty regardless of how the poverty line is defined.

To assess the extent of poverty in Lima in both years, we used two poverty lines based on the cost of a minimum basket of food for a family of six in Lima (see Table 4). The daily cost of this basket in June 1, 1990 Intis was 154,416, which implies a cost of 4,632,480 Intis per month. Adjusting the resulting per capita value of the food basket for household composition implies a poverty line of 1,477,650 Intis per “adult equivalent.” If one defines as poor any household for whom per capita adjusted consumption is less than this figure, so that even if all household income money were spent on food it could not purchase the minimum basket (Poverty Line 1), only 0.5 percent of the population of Lima was poor in 1985–86. But if one defines as poor all households where per capita food expenditures were below this amount (Poverty Line 2), 12.7 percent of Lima’s population was poor; in 1985–86.

Table 4The extent of poverty in Lima
Value of poverty line,Percent of population below poverty line in Lima
Poverty line1June 1, 1990 Intis1985–861990
1. Poverty line 11,447,6500.517.3
(based on total expenditures)
2. Poverty line 21,447,65012.754.7
(based on food expenditures)
Source; Glewwe and Hall, 1992.

Based on monthly real adjusted food and total expenditures.

Source; Glewwe and Hall, 1992.

Based on monthly real adjusted food and total expenditures.

The percent of the population with expenditure levels below the amount required to purchase the basket according to Poverty Line 1 jumped from 0.5 percent in 1985–86 to 17.3 percent in 1990—an unprecedented increase. Alternatively, according to Poverty Line 2, the incidence of poverty quadruples between 1985–86 and 1990; in 1985–86, one out of every eight residents of Lima were poor; by 1990 more than half were poor.

Some lessons

The Peruvian Government opted for a policy of “unorthodox” adjustment during 1985–86 to 1990 in an attempt to encourage rapid growth and improve living standards. While it is very difficult, and beyond the scope of this article, to prove that the increase in poverty was caused by the specific macroeconomic policies adopted between 1985–90, the findings do provide a stark example of the difficulty of using unorthodox policies to simultaneously encourage growth and reduce poverty. Our research shows that the hoped for improvement in growth rates and living standards did not materialize. In fact, the incidence of poverty in Lima increased dramatically and the distribution of consumption became more unequal.

Further investigation provides a clearer picture of what happened between 1985–90.

• Households headed by individuals with relatively low levels of education experienced greater declines in consumption level than the better educated. Thus the poor, who have lower education levels, suffered disproportionately because the returns to human capital were more stable at higher levels of education. In terms of identifying the poorest, the level of education is probably the best indicator. This suggests that long-run social investment plans would do well to give priority to education of the poor.

• Unemployment rose significantly and became a distinct characteristic of the poor. Thus social expenditure programs designed to increase low-wage employment opportunities may be an efficient method to target the poor.

• The provision of public services, particularly potable water and sewage services, deteriorated across the city, but most severely in the three poorest regions—Conos Este, Norte, and Sur. Investment in public water and sanitation services directed at these three areas would significantly improve living standards.

Perhaps the most important lesson to be drawn from these conclusions is that there are no quick fix macroeconomic solutions when addressing the twin problems of economic stagnation and poverty in developing countries. “Unorthodox” alternatives to structural adjustment, while designed to encourage rapid growth along with a more equitable distribution of welfare, may not succeed in this endeavor and may, in fact, hurt the poor more than the general population. The orthodox policies that were rejected in Peru in 1985 were adopted in 1990 after a change of government. Peru now faces the task of reviving economic growth with much higher poverty rates that those prevailing in 1985.

The Living Standards Measurement Study (LSMS) began in 1979. Its purpose is to generate high quality, policy-relevant household survey data in developing countries. While the data can be used for a variety of sectoral and economic analyses, the LSMS mandate centers on welfare, income distribution, and poverty. In particular, survey results are useful in assessing the impact of development policies and programs on different socioeconomic groups. In addition to Peru, surveys have been completed in Bolivia, Côte d’Ivoire, Ghana, Jamaica, Mauritania, Morocco, Pakistan, and Venezuela. New LSMS projects are underway in China, Guyana, Nicaragua, and Viet Nam.

For an in-depth analysis, see “Poverty and Inequality during Unorthodox Adjustment: The Case of Peru, 1985–90” by the authors, Living Standards Measurement Study Working Paper No. 86, The World Bank, 1992, available from the World Bank Publication Services.

3 NEW Studies Pinpoint Urban Issues, Provide Frameworks, Guidelines for Progress

• Urban Public Finance in Developing Countries

Roy W. Bahl and Johannes F. Linn

Provides a framework for urban public finance in developing countries. It shows ways to create higher tax revenues to strengthen urban infrastructures and meet fiscal demands to dries as urban population growth continues to explode.

Published for the World Bank by Oxford University Press

565 pages / ISBN 0-19-520805-6/


• Urban Property Tax Reform Guidelines and Recommendations

William Dillinger

Shows developing countries how to increase tax contributions in urban areas and improve the delivery of municipal services.

56 pages / ISBN 0-8213-2065-3 / $6.95

• Reforming Urban Land Policies and Institutions in Developing Countries

Catherine Farvacque and Patrick McAuslan

Outlines specific reforms in urban land management designed to decentralize and make equitable the processes that control land use, ownership, registration, and development.

133 pages / ISBN 0-8213-2092-0 / $9.95

Coupon orders accepted only from the U.S.A and France. To order from other countries, request a list of World Bank Publications Distributors. Prices vary by country.

In the U.S. return the coupon to:

World Bank Publications

P.O. Box 7247-8619

Philadelphia, PA 19170-8619

In France return the coupon to:

World Bank Publications

66, avenue d’léna

75116 Paris, France

Other Resources Citing This Publication