III Some Methodological Issues in the Analysis of the Impact of Adjustment Programs on Poverty

International Monetary Fund
Published Date:
May 1988
  • ShareShare
Show Summary Details

Basic Methodology of the Study

The methodology underlying the case studies for this paper was eclectic. The analysis highlights the channels of impact of adjustment policy measures on poverty groups in their capacities as consumers of goods and services (i.e., uses of income), as owners of assets, and as factors of production (i.e., sources of income). It addresses both short-term and longer-term considerations. Where particular measures significantly affected other important income groups or the distribution of income, these effects were noted.

The approach is eclectic in that it incorporates both partial and general equilibrium considerations. It is an operationally feasible methodology, tailored both to the limitations of the data and to the inadequacy of the economic tools currently available to model the interaction of macroeconomic policy and poverty, particularly for the type of disequilibrium experienced by countries with Fund-supported adjustment programs.10

The methodological approach can be summarized as follows. First, the socioeconomic characteristics of poverty groups were identified, based on whatever quantitative data and qualitative evidence were available. These characteristics included the locus of their production activities, their consumption patterns, holdings of land and other assets, accessibility to government services and to formal markets, and other relevant information.11 Second, the sources of macroeconomic disequilibrium that led to the implementation of a Fund-supported adjustment program were analyzed. Third, while focusing on the measures specifically included in the program, all measures taken during the course of the program (including earlier relevant actions) were also analyzed.

Fourth, on the basis of the broad analytical framework summarized in the next subsection, the impact of particular measures was evaluated. The emphasis was on the short-term impact. In the case of tax or expenditure policy instruments, fairly strong inferences about the likely immediate impact on different income poverty groups could often be drawn. Exchange rate actions were examined mainly in terms of the likely short-term impact of increased prices of consumer goods and tradable commodities in production. The latter may significantly affect the income of some groups of the poor. Although it often proved difficult to trace the immediate effect of other macroeconomic measures, the indirect effects of the likely impact on both the uses and the sources side (wage rates, employment, and interest income) were assessed.

Fifth, the likely impact of policy measures on the earning capacity of the poor over the medium to long term was appraised wherever relevant. Particular emphasis was placed on the effects of a structural change in relative prices and the restoration of external equilibrium. These inferences are more speculative. In some cases, an assessment was made of the likely evolution of the economic status of poverty groups as a result of the measures, relative to the preprogram situation. This is a “before-after” approach, speculating on how the poor fared during the course of the adjustment process, but without attempting to impute whether any change in their status was due to the effects of the adjustment program or to the unsustainable preprogram situation. In other cases, the argumentation includes a comparison of the likely evolution of the situation with the adjustment program measures, relative to a probable counterfactual situation, namely, relative to what might have happened without such measures.

An important role is served by both the before-after and the counterfactual approaches. The latter approach recognizes that an adjustment effort is the consequence of an unsustainable disequilibrium, whether caused by inappropriate macroeconomic policies or by unavoidable exogenous shocks, and that a worsening of the situation for the poor over the medium term may reflect the inevitable consequences of this disequilibrium rather than the policy measures used to achieve a sustainable level of absorption. The prolonged world recession, the sharp deterioration of the terms of trade, and the limited access to external financing were already causing severe poverty problems in many developing countries before the adoption of Fund-supported programs.

Yet to rely on a counterfactual approach is to rely on a very elusive, and possibly misleading, yardstick for comparison. It is very difficult to impart an empirical content to the counterfactual argument—to assess the full effects of an unsustainable macroeconomic situation on the status of poverty groups. Besides, it would be unrealistic to expect poverty groups to imagine what might have happened without the adoption of an adjustment program. Their primary concern is any immediate adverse effect on the sources or uses of their income. The before-after approach avoids the issue of blame, calls attention to how particular groups fare during the course of adjustment, and highlights the potential need for compensatory policies to cushion the impact on the poor.

Three additional points should be made. First, while the approach adopted in this study is empirical, it is not always quantitative. It is empirical insofar as the analysis is based on the specific policy measures and the structure of poverty in countries with Fund-supported adjustment programs. The conclusions, however, are drawn together more through deductive reasoning and less from actual quantitative information on what actually happened to poverty during the program. Generally, lack of data precluded the quantification of the theoretical propositions embodied in the analysis.

Second, the presence of an informal sector may obscure the channels through which policies influenced the poor and thus may raise questions about the validity of inferences based only on official statistics.12 For example, the effect of the large devaluation in Ghana on real incomes of cocoa farmers can only be assessed with some knowledge of the effective price received before devaluation through informal marketing channels. Although little can be done to resolve this type of problem, it implies additional caution in the interpretation of poverty-related statistical information.

Finally, this study is concerned primarily with absolute poverty rather than with the relative distribution of income.13 Absolute poverty may have increased during the course of an adjustment program, even if the overall distribution of income has remained unchanged (as measured by an overall index of inequality, such as the Gini coefficient). Lack of data precludes rigorous analysis of the impact on poverty, which, in a comprehensive framework, should also include non-economic criteria such as nutrition and general health.

The Impact of Policy Measures on Poverty: An Overview of Channels

A primary objective of Fund-supported adjustment programs is to achieve a sustainable external balance. The approach taken to achieve this objective depends on the specific circumstances of the country and the sources of the external imbalance (e.g., an expansionary macroeconomic policy, a deterioration of the external terms of trade, or a natural disaster). In general, however, an adjustment program aims at restraining aggregate domestic demand, promoting supply, and, most important, improving economic efficiency.

A number of policy instruments may be used in the context of an adjustment program in the spheres of money and credit, fiscal policy, pricing, labor markets, and external sector policy. It is inevitable that these policies will affect poverty groups, because they influence not only aggregate demand, supply, and the overall price level but also the composition of demand and supply and, therefore, relative prices.

The demand-restraining side of adjustment measures is likely to have adverse implications for the growth of real output and employment in the short run. However, these adverse implications should be weighed against the longer-term benefits of stabilization: price stability, external creditworthiness, and the consequent favorable environment for sustainable growth. As discussed below, specific adjustment policies aim not only at demand restraint but also at supply promotion. Moreover, inflation and severe import controls often associated with unsustainable policies might have more adverse implications for poverty groups, even in the short run, than a well-designed set of adjustment policies.

Money and credit policy generally relies on restraining the expansion of domestic credit from the banking system to stabilize domestic prices and reduce the external imbalance, thus creating an environment in which sustained growth may be achieved. The implications of such policies are favorable for the poverty groups, since these groups often have less means to protect the real values of their assets and incomes from inflation than the wealthier elements of the population. Credit policies are often supported by a tight fiscal stance designed to reduce the expansion of credit to the public sector. In this way, the adverse impact of the overall domestic credit restraint on the supply of credit to the private sector is minimized.

Pricing policy aims at raising the relative prices of tradables to world market levels. Such a policy may have both negative and positive implications for poverty groups. The negative effects may be substantial, particularly in urban areas if these groups rely heavily on imported necessities, such as food, medicine, and fuel. But the actual negative impact can be diluted by a well-designed set of compensatory measures to protect poverty groups. Moreover, in the longer run, the gains from more efficient use of scarce tradables could offset the short-run adverse impact. Increases in the relative prices of exportables can also positively affect the incomes of smallholders and landless farm workers engaged in producing exportables, although the size of the benefit would depend on the extent of concentration in production and the degree of access of smallholders to markets. In the longer run, regional or sectoral migration can also limit the benefit to the rural poor.

Labor market policies invariably focus on restraining the real wages of workers in the formal sector, including the public sector. Such policies often have no direct implications for poverty groups in the rural sector or in the urban informal sector, although there may be adverse effects when urban-rural remittances are significant. Indirectly, wage restraint in the formal sector could even have favorable implications for some poverty groups (such as unemployed workers and workers in the informal sector) by increasing employment in the formal sector.

Exchange rate policy, combined with domestic pricing policy, typically aims at maintaining appropriate relative prices for tradables. The short-run implications of a devaluation are adverse for those relying on imported necessities but generally positive for those engaged in producing exportables or import substitutes. The longer-run implications of a devaluation for poverty groups depend to a large extent on how successful it is in restoring the sustainability of the external imbalance. In turn, this result depends on a number of factors, including the elasticities of the supply of exportables and of the consumption of importables, as well as on how money and credit policy is conducted. An appropriate exchange rate is important for achieving external viability. But it is crucial that the devaluation does not trigger a spiral of inflation and devaluation; in such circumstances the most vulnerable would be the poverty groups.

Another important objective of adjustment programs is to improve the efficiency with which scarce resources are used. Virtually all adjustment programs include provisions to discourage price and import controls and to improve reliance on the price mechanism in allocation, production, and distribution. The implications of such policies for the poverty groups have two aspects. First, changes in relative prices have the implications for poverty groups discussed above. Second, controlling domestic prices or imports may not only distort the allocative system but also benefit the relatively wealthy importers or domestic distributors of such goods, at the expense of reduced government revenues and smaller incomes for domestic producers, who often are poor farmers or farm workers. A market-oriented pricing system would help reduce such distortions.

Most adjustment programs attempt to raise government revenue and reduce expenditure. On the revenue side, the short-run instruments often consist of excises and other indirect taxes; in the longer run, these programs attempt to broaden the domestic tax base and to reduce tax evasion. The narrowness of the domestic tax base in developing countries often forces governments to increase indirect taxes on tradables as a means of raising government revenue in the short run. But such measures often have regressive effects on the urban consumers of importables (e.g., food, fuel, liquor, and cigarettes) and on the rural producers of exportables. The extent would depend on how these tax increases are implemented; many programs exempt goods consumed by core poverty groups. Broadening of the tax base is achieved through a longer-term tax reform, which has a gradual, but nevertheless progressive, distributional impact.

On the expenditure side, not only the level, but also the composition, of expenditure can be changed. Reduced expenditures, particularly on health, nutrition, and education, could have far-reaching implications for the poverty groups, negatively affecting their human capital. But the incidence of these expenditures is a critical issue. Expenditures on higher education or tertiary medical care often benefit only the relatively wealthy. Fiscal adjustment efforts often lead to reduced capital expenditure, which may limit employment opportunities and slow down future economic growth. In many cases, however, government capital outlays create inefficient assets that could have a negative economic impact over the lifetime of such assets; the longer-term adverse implications of reducing these capital outlays may not be large.

Some Conceptual and Empirical Questions

Several methodological issues merit further discussion because they shed more light on problems encountered in assessing the distributional impact of adjustment programs.14 These issues include (a) the choice of time frame for the analysis; (b) the limitations of the data base; (c) the analytical problems arising from a significant informal sector; and (d) the attribution of policies under a Fund-supported program.

Heuristically, therefore, the most constructive approach to the assessment of the social impact of a Fund-supported program appears to be the axiomatic acceptance of a J curve in the economy’s trajectory, without necessarily an explicit distinction between structural and stabilization measures. Although the principal focus of this study is on the costs of adjustment, that is, on the poverty implications of a Fund-supported program during the initial period of declining growth and perhaps reduced output, the analysis also examines how the adjustment process may affect the productive and income-earning capacity of the poor over a period that extends beyond the duration of the program. In general, therefore, a distinction is made between the short-term, aggregate demand implications of an adjustment effort, and the longer-term, supply-side effects of that process.

Choice of Time Frame: Stabilization Versus Structural Adjustment

The time frame chosen for the analysis has critical implications for the appropriate methodology to be used in a study of this kind and is intimately related to the distinction between stabilization and structural adjustment programs. Usually, stabilization is identified with Fund-supported programs, which place emphasis on demand management, whereas structural adjustment is identified with World Bank loans, with a focus on supply-side effects. But such a distinction is loose and not always valid. Although, formally speaking, the vast majority of Fund-supported programs still fall under the former category, in an increasing number of cases they include structural measures. Recent experience from academic discussions on this issue has confirmed the well-known difficulties in defining those two concepts.15 To identify, for example, changes in relative prices exclusively with structural programs, as some researchers have suggested, would conspicuously ignore the important role of exchange rate policies in the majority of Fund-supported stabilization programs.

Heuristically, therefore, the most constructive approach to the assessment of the social impact of a Fund-supported program appears to be the axiomatic acceptance of a J curve in the economy’s trajectory, without necessarily an explicit distinction between structural and stabilization measures. Although the principal focus of this study is on the costs of adjustment, that is, on the poverty implications of a Fund-supported program during the initial period of declining growth and perhaps reduced output, the analysis also examines how the adjustment process may affect the productive and income-earning capacity of the poor over a period that extends beyond the duration of the program. In general, therefore, a distinction is made between the short-term, aggregate demand implications of an adjustment effort, and the longer-term, supply-side effects of that process.

Limitations of the Data Base

Occasional Paper No. 46 (International Monetary Fund (1986)) pointed out the enormous statistical barriers to a quantification of the effects of past Fund-supported programs on income distribution. To all those data requirements should be added, in this instance, the information related to the more specialized focus of a study on poverty.

Indeed, a narrow definition of poverty, cast only in terms of income, could be accommodated through a data base by income group, coupled with the appropriate definition of the poverty line.16 A comprehensive definition of poverty, however, is a much more complex issue, as it would involve specialized indicators on expenditures, health, nutrition, and other basic needs. An important question is whether inequality in general and poverty in particular can be adequately measured by income statistics alone, or whether more specific criteria should be utilized. For the general question of inequality, both theoretical and empirical arguments suggest that expenditure statistics may be more reliable than income statistics as an index of welfare. For poverty in particular, specialized indicators on health and nutrition may be necessary, which obviously go beyond the routinely collected economic statistics.17

There is a serious dearth of information on the socioeconomic characteristics of the poor, since international organizations and national authorities have only recently focused on the plight of the poor within the context of adjustment programs. It is therefore highly unlikely that any country would possess adequate quantitative information for a complete analysis of the socioeconomic position of the poor and, thus, even more unlikely, data on the status of the poor before and after an adjustment program. The analysis would become even more difficult if the counterfactual case—the hypothetical situation in the absence of an adjustment program—has to be addressed empirically. The only realistic alternative is to survey the links between policy measures and the real incomes and consumption patterns of the population, and to draw inferences based on the best identifiable poverty characteristics.18

The Analytical Implications of a Significant Informal Sector

The measurement of the informal sector may have important consequences for the analytical inferences to be drawn about the distributional impact of a Fund-supported program. The principal question is whether inferences based on official statistics alone adequately and accurately reflect all the economic channels of adjustment policies. An example is offered by an economy with a dual nature, that is, a well-defined rural-urban dichotomy, in the face of a significant devaluation where the rural sector primarily comprises the producers of a primary export commodity. In the pre-adjustment period, official statistics typically will show stagnant or declining production. It is usually asserted that a devaluation has beneficial distributional consequences, since the exchange rate adjustment appears to lead to expanded supply and, presumably, higher incomes.

Such reasoning ignores the important role that may be played by informal transactions within the rural sector. In a country producing and exporting a primary commodity, for example, the official statistics will capture the sales of the commodity to the official marketing board but not the rural population’s informal transactions, which, in the presence of an overvalued currency, may include smuggling the commodity to neighboring countries. Effectively, therefore, the incomes of the rural sector during the pre-adjustment period may be seriously underestimated by the extent of informal exports of the commodity. As a result, conclusions based on official statistics may seriously overestimate the beneficial distributional consequences of the devaluation for the rural sector.

Attribution of Policies Under a Fund-Supported Program

An issue that arises in evaluating the poverty impact of a Fund-supported adjustment program is how to define which policies should be considered part of the program. In addressing this issue in the case studies, efforts were made to identify measures specifically included in the letter of intent of a program. However, in the interests of assessing the consequences of adjustment policies on poverty, an attempt was made to examine the full set of measures adopted, subject to data limitations and without excessive concern for the issue of attribution. This process reflects two considerations.

First, the focus of a Fund-supported program is typically on the broad macroeconomic aggregates. As a matter of policy, the Fund has traditionally attempted to avoid intruding on issues that bear on social and political priorities. In recent years, as the Fund has increasingly stressed the contribution of growth in facilitating the adjustment process, it has engaged in discussions with authorities on the efficiency of alternative approaches to meet aggregate budget and macroeconomic targets, and this has at times led to negotiations bearing on a limited number of fiscal and structural policy measures. Often, particular policy measures with important distributional consequences, such as tax policy, the promotion of private sector activities, and the liberalization of the price system, are the subject of active discussion with the Fund. Yet many government policies, particularly those relating to disaggregated expenditure decisions, are not discussed with the Fund staff and reflect decisions by the authorities on ways to meet their aggregate program targets.

Second, during the course of a program, countries are likely to compensate for unexpected slippages from program targets. To the extent that such policies secure the achievement of the aggregate financial targets, they rarely compromise the program as a whole. Yet it is difficult to distinguish between measures originally agreed as elements of the program and those subsequently adopted by the authorities as it evolves.

Methodology on Using a Poverty Typology to Assess Adjustment Programs

A comprehensive investigation of the poverty implications of adjustment programs would require the development of a full-fledged poverty typology for each country, on the basis of which principal policy measures could be linked with the sources and uses of income of the key poverty groups. This subsection discusses some of the underlying issues in developing such a typology for future analyses.

The basic conceptual framework of the analysis requires the classification of the poor across economically meaningful socioeconomic groups and, within that classification, some identification of key socioeconomic characteristics that are likely to be particularly sensitive to adjustment policy measures. The classification of socioeconomic groups is meant to capture their principal locus of production, thereby facilitating analysis of the impact of demand and supply factors on employment and incomes of the poor in key sectors. A typology framework recognizes that the poor are not a homogeneous group and that one must seek to understand the nature of their economic situation in order to gauge how they are affected by different adjustment policies.

To illustrate, a particular economy could be characterized by the distribution of the poor among (1) landless rural labor, (2) agricultural smallholders, (3) the urban informal sector, and (4) the urban formal sector. Refinements to this classification are obviously possible, depending on the trade-off between data requirements and the marginal improvement in the expected analytical inferences. For example, the agricultural sector could be further broken down into producers of tradable and nontradable commodities; urban dwellers could be classified according to employment in the government, the formal private sector, the informal private sector, or the unemployed.19 The finer the breakdown, the more sensitive is the analysis of the impact of particular measures on different groups.

In terms of policy instruments, the analysis would then distinguish between the economic consequences for the poor in terms of the sources and uses of their income. The effect on the uses side is primarily associated with the short-run impact of macroeconomic policy adjustments (perhaps reflecting a cutback in government services to the poor) and policies to change relative prices on the real cost to the poor of their consumption. The effect on the sources side has both a short- and a long-run dimension. Macroeconomic policy adjustments may lead to contraction in certain sectors in the short run, thus worsening income-earning opportunities for some groups of the poor. On the structural side, shifts in relative prices will improve the short- and long-run positions of other groups of the poor, for example, those involved in the production of tradables. The longer-term effects of adjustment programs, within a general equilibrium framework, could be more beneficial for the incomes of the poverty groups adversely affected in the short run.

Thus, a comprehensive analysis of the poverty implications of adjustment programs would attempt to investigate the distributional effects, not only through the reduction in real incomes, but also through the potential, if longer-term, increase in the productive capacity of the poor. Chart 1 provides a schematic representation of the conceptual framework suggested here.

Chart 1.Assessment Matrix of Impact on Poverty of Adjustment Programs

The assessment matrix must be understood to underlie every adjustment measure identified as a principal policy instrument. By linking the dominant policy instruments of a given adjustment program with the basic features of the poverty typology for a particular country, it is possible to articulate the probable net effects of adjustment programs on poverty groups in that country.

The data needed for such an approach are obviously more detailed than those needed for a distributional investigation based on income and expenditure statistics by income class. The specific informational requirements will depend on the desirable breakdown of socioeconomic groups, which, in turn, will depend on the identifiable poverty typology of a given country. For example, it may be desirable to analyze geographical differences or to separate men and women in certain subgroups.

In a pioneering study on the impact of adjustment programs on the poor in Cote d‘Ivoire, a World Bank study20 identified the poverty groups by location, occupation, type of employer for the head of household, crops grown by different groups, budget shares of staple goods in household consumption, school attendance, and health practitioners contacted by the poor. Such a detailed classification was made possible through the results of the Living Standards Measurement Survey (LSMS), which provided the data base for a thorough investigation.

The same study attempted to devise shortcuts for identifying the poor by using traits that are commonly associated with poverty. The results of this effort, however, were disappointing and suggest that in the absence of a specific survey (such as an LSMS) data requirements may present an insurmountable obstacle in the short run for the proper identification and targeting of the poor.

    Other Resources Citing This Publication