Macroeconomic Performance and Adjustment Under Fund-Supported Programs: The Experience of the Seventies

In recent years, the conditionality associated with upper credit tranche adjustment programs for developing countries supported by the use of Fund resources has been a source of considerable discussion and debate. While the broad aim of Fund-supported programs—namely, “the restoration and maintenance of viability to the balance of payments in an environment of price stability and sustainable rates of economic growth” (Guitián (1982), p. 93)—is widely accepted, there is less agreement as regards the appropriateness of policies embodied in Fund conditionality as a means of achieving those objectives. It has been argued, for example, that even though what is thought of as a “typical Fund program” may achieve some improvement in the balance of payments—a conclusion that is by no means always accepted—such an outcome can often involve an excessive cost in terms of other objectives of particular importance to developing countries. This line of reasoning is frequently based on the belief that policies aimed at curbing demand pressures retard economic growth and also place a severe burden on the already low consumption and living standards of the population.1

Abstract

In recent years, the conditionality associated with upper credit tranche adjustment programs for developing countries supported by the use of Fund resources has been a source of considerable discussion and debate. While the broad aim of Fund-supported programs—namely, “the restoration and maintenance of viability to the balance of payments in an environment of price stability and sustainable rates of economic growth” (Guitián (1982), p. 93)—is widely accepted, there is less agreement as regards the appropriateness of policies embodied in Fund conditionality as a means of achieving those objectives. It has been argued, for example, that even though what is thought of as a “typical Fund program” may achieve some improvement in the balance of payments—a conclusion that is by no means always accepted—such an outcome can often involve an excessive cost in terms of other objectives of particular importance to developing countries. This line of reasoning is frequently based on the belief that policies aimed at curbing demand pressures retard economic growth and also place a severe burden on the already low consumption and living standards of the population.1

In recent years, the conditionality associated with upper credit tranche adjustment programs for developing countries supported by the use of Fund resources has been a source of considerable discussion and debate. While the broad aim of Fund-supported programs—namely, “the restoration and maintenance of viability to the balance of payments in an environment of price stability and sustainable rates of economic growth” (Guitián (1982), p. 93)—is widely accepted, there is less agreement as regards the appropriateness of policies embodied in Fund conditionality as a means of achieving those objectives. It has been argued, for example, that even though what is thought of as a “typical Fund program” may achieve some improvement in the balance of payments—a conclusion that is by no means always accepted—such an outcome can often involve an excessive cost in terms of other objectives of particular importance to developing countries. This line of reasoning is frequently based on the belief that policies aimed at curbing demand pressures retard economic growth and also place a severe burden on the already low consumption and living standards of the population.1

In order to shed further light on the debate about the appropriateness of conditionality practices, an empirical examination of certain of the main issues would seem desirable. There is at present available, on a systematic and uniform basis and for long time periods, a wealth of published data on the major macro-economic variables for most developing countries. Even if it is unlikely to settle the issues conclusively, careful analysis of these data can provide some objective information regarding different aspects of the economic performance of countries undertaking adjustment programs supported by the Fund.2

The purpose of the present paper is to determine, on the basis of published data, what broad conclusions can be drawn from the experience of those developing countries that from 1971 to 1980 undertook adjustment programs supported by an upper credit tranche stand-by arrangement. Two sets of questions of direct relevance to macroeconomic performance are addressed explicitly. First, what changes did program countries experience in their balance of payments positions (as measured by movements in both their current account balances and overall balances), inflation, and economic growth—all of which play central roles in the adjustment effort? Second, what was the outcome for other key macroeconomic variables—namely, savings, investment, and real consumption? These variables are of major concern to policymakers, since they indicate how changes in the external current account balance (defined as the difference between domestic savings and investment) reflect changes in the choice between present and future consumption priorities. For the cases considered, the change in these variables is analyzed by factually comparing both the year before adoption of the program and the program period3 itself, and (as a means of evaluating longer-term trends) the three-year periods before and after the adoption of the program. In addition, to take into account exogenous events, such as movements in the terms of trade or world recessionary trends, that affected all developing countries during the decade, information on aggregate changes in the macroeconomic variables in question for all non-oil developing countries during corresponding periods is also presented.

While the analysis described in this paper can provide useful evidence concerning the actual economic performance of the countries under review, several caveats seem in order. First, any general inferences drawn from the study cannot be interpreted as necessarily applying to a particular country or group of countries because of the highly aggregative nature of the analysis and the diversity of the countries considered. Second, no attempt has been made to distinguish between countries that implemented the policy actions envisaged in the Fund-supported programs and those that did not; any such distinction would of necessity entail elements of judgment and would reduce the factual and objective nature of the inquiry. Third, the analysis deals with the outcome in terms of overall macroeconomic performance and thus does not address the particular workings of specific policy instruments.4

More generally, caution should be exercised in interpreting the paper’s results as representing the average degree of “effectiveness” of Fund-supported programs. In the first place, the review includes programs that may have been only partially implemented. Also, as both Guitián (1982) and Williamson (1981) have observed, a rigorous examination of this issue would involve a comparison of the actual outcome with the hypothetical outcome associated with the alternative policy actions (or lack of action) that might have been pursued by the authorities in the absence of Fund assistance. Such an approach, though appealing theoretically, would involve considerable difficulties in practice, if only because it would require major elements of judgment, something this paper seeks to avoid.

The plan of the paper is as follows. Section I describes the methodology used in the statistical analysis. Section II presents empirical evidence regarding the behavior of the balance of payments, inflation, and growth, both for the countries under review and for all non-oil developing countries. Section III undertakes a similar analysis for savings, investment, and real consumption. Finally, Section IV summarizes and evaluates the main empirical findings of the paper and also discusses some issues deserving of further empirical investigation. The Appendix describes the countries and program periods that provided the basis for the analysis.

I. Methodology

The basic aim of the analysis is to assess the extent to which the performance of each of the macroeconomic variables considered varied systematically before, during, and after each program period.5 For this purpose, performance was measured by calculating the change in the variable in question during periods of comparison extending over “one year” and “three (or two) years.” The shorter period reflects the change in the variable between the period immediately before the program began and the program period itself. The longer period involves a comparison between the annual average level (or rate of change, where relevant) of the variable during the three-year period before the program period began and the three-year or two-year period beginning with the adoption of the program. Two-year comparisons were used for all 1979 programs and some 1978 programs. The longer-run, three-year or two-year comparison, it was felt, would take into account the possible lagged effects of policies undertaken during the program period itself as well as the effects of policies pursued after the program ended.

In viewing the evolution of macroeconomic performance in adjustment programs, it is important to try to take into account those exogenous factors that may have played a key role throughout the period. While this is difficult to do directly, a useful approach is to compare the economic performance of countries that undertook adjustment programs with the performance of all non-oil developing countries (NODCs). Accordingly, data on the change in each variable under analysis are presented for all NODCs together with the change in the program country variable.6 Joint examination of these two series provides some evidence on whether program countries’ performance differed systematically from that of developing countries in the aggregate.7

The disaggregated information for each variable and program period calculated in the manner described previously both for program countries and for all NODCs is then presented in aggregative form for each year during the decade and for the entire 1971-80 period. As a first standard of comparison, all programs are divided into those in which, for example, there was an increase in the variable in question and those in which there was a decrease; this distinction is applied to both absolute and relative changes (i.e., the change in the program countries’ variable measured in relation to the change for all NODCs). As a second standard of comparison, simple unweighted averages of changes in program country variables are presented together with similar averages for the corresponding time periods for all NODCs.8 While in principle these two standards of comparison might not be expected to point necessarily in the same direction, it is interesting to note that for the results reported in this paper, they provide a basis for generally similar qualitative conclusions.9

In evaluating the reported results, it is useful to consider the implications of this paper’s comparison of actual ex post economic performance as opposed to a comparison involving the hypothetical outcome in the absence of a program supported by the Fund; as was noted earlier, the latter comparison might be viewed as the theoretically ideal standard to apply in order to judge the effectiveness of programs supported by the Fund. Given the element of judgment that would be involved in the latter comparison, it is difficult to draw any general inferences. However, one possible quantitatively important aspect may be noted. If only the balance of payments indicators for a program country were considered, it might be concluded that the lack of financial assistance from the Fund (and other external assistance linked to the adoption of a program) might, in some circumstances, have led to an even smaller deficit in the current account and the overall deficit owing to the lesser availability of foreign exchange. However, such an outcome would in many cases reflect a resort to trade or exchange restrictions; although an “improvement” in the balance of payments might occur, the use of controls to limit external imbalances would in all likelihood involve some cost in terms of growth and inflation objectives. In considering, therefore, the difference between a comparison based upon the actual change in performance and a comparison based upon the hypothetical outcome without a program, it is important to take into account not only the balance of payments outcome (which by itself could, in some instances, be misleading) but also the outcomes for economic growth and the domestic inflation rate.

II. Performance in Terms of Main Objectives

This section presents quantitative evidence regarding three aspects of macroeconomic performance that are of major concern in programs supported by the Fund—namely, the balance of payments, inflation, and economic growth.

BALANCE OF PAYMENTS

As described in this paper’s opening paragraphs, the major external objective of adjustment programs is to achieve a sustainable balance of payments position. In many cases, this may involve seeking a reduction in the current account deficit by adjusting domestic demand relative to the available supply of resources. However, depending on the circumstances of the economy (and, in particular, the extent to which increased foreign savings are attracted as a result of the stabilization effort undertaken by the authorities), a widening of the current account deficit may be quite consistent with an adjustment toward a position of external equilibrium. In these latter instances, programs would normally seek to achieve some improvement in the overall balance of payments. Thus, the analysis in this paper presents information on both the current account and the overall balance indicators.10

Table 1 contains data on the percentage point change in the current account/gross national product (GNP) ratio for program countries for each program year during the 1970s, as well as the average for the entire 1971—80 period; similar data for the corresponding periods are provided for all NODCs.11 In all years except 1974, 1975, and 1980, the program countries experienced, on average, an absolute decline in the current account ratio comparing the year before the program began with the program period itself; using the three-year comparison period (thereby excluding 1980), a reduction was recorded in all years except 1973,1974, and 1975.

Table 1.

Upper Credit Tranche Stand-By Arrangements, 1971-80: Average Change in Current Account Performance Comparing Pre-Program and Post-Program Periods

(Changes in current account deficit as percentage of GNP)1

article image
Source: International Monetary Fund, International Financial Statistics, various issues.

A positive figure indicates a reduction in the deficit, a negative figure an increase in the deficit. In a very small number of cases, ratios to GDP are used. Numbers for each year are simple arithmetic averages of data for individual programs.

Equals the unweighted arithmetic average of all individual program country data for 1971-80 and, therefore, does not equal simple average of yearly columns; for the three-year comparison, calculation is made based only on 1971-79 data.

Change in deficit comparing the year before the program began and the program year.

Change in average annual deficit comparing the three-year period before the program began and the three-year period beginning with the program year. For some 1978 programs and all 1979 programs, post-program calculations are made on a two-year basis. Similarly, 1979 data for all NODCs are on a two-year basis.

It is also apparent that program countries’ deficits tended to change by considerably more than the average for all non-oil developing countries. Thus, in 15 of the 19 time periods of comparison shown in Table 1 (10 1-year comparisons and 9 3-year comparisons), the decrease (increase) in program countries’ deficit ratios was greater (lesser) than that of all NODCs. These results are mirrored by the calculation of the average change for the entire 1971-80 set of program periods. On average, program countries reduced their current account deficits by an amount equivalent to 1.5 percentage points of GNP during the program period itself and by 1.2 percentage points of GNP comparing the three-year annual average level in pre-program and post-program periods. During the same periods of comparison, the deficits of all NODCs increased by 0.4 and 0.8 percentage points of GNP, respectively.

Trends in the overall balance of payments are shown in Table 2.12 Except for 1971 (one-year comparison only), 1973 (three-year comparison only), and 1974 (both one-year and three-year comparisons), the overall balance calculated as a percentage of exports improved13 in each of the 19 periods considered.14 Viewed in relation to the change in the overall balance experienced on average by all non-oil developing countries, it can be seen that, except in 1971, a “relative” improvement occurred in every program year (including 1974, when the rise in the deficit recorded by program countries was, on average, less than that experienced by all NODCs). When the entire period was reviewed, the average improvement in the overall balances of program countries amounted to between 5 and 7 per cent of exports (depending on whether the one-year or three-year comparison period was considered); during the same comparison periods, the overall balance of all NODCs remained, on average, roughly unchanged.

Table 2.

Upper Credit Tranche Stand-By Arrangements, 1971-80: Average Change in Overall Balance of Payments Performance Comparing Pre-Program and Post-Program Periods

(Change in overall balance as percentage of exports)1

article image
Source: International Monetary Fund, International Financial Statistics, various issues.

A positive figure indicates an improvement, which is defined as either a larger overall surplus or a smaller overall deficit. A negative figure indicates a deterioration, which is defined as either a smaller overall surplus or a larger overall deficit. Numbers for each year are simple arithmetic averages of data for individual programs.

Equals unweighted arithmetic average of all individual program country data for 1971-80 and therefore does not equal simple average of yearly columns; for the three-year comparison, the calculation is made based only on 1971-79 data.

Change in overall balance comparing the year before the program and the program year.

Change in average annual overall balance comparing the three-year period before the program began and the three-year period beginning with the program year. For some 1978 programs and all 1979 programs, post-program calculations are made on a two-year basis. Similarly, 1979 data for all NODCs are on a two-year basis.

Finally, Table 3 contains information on the behavior of the current account and the overall balance considered together.15

Table 3.

Upper Credit Tranche Stand-By Arrangements, 1971-80: External Sector Performance Comparing Pre-Program and Post-Program Periods

article image
Source: International Monetary Fund, International Financial Statistics, various issues.

Change in ratio of current account balance to GNP (or, in a small number of cases, to GDP); improvement refers to a smaller deficit or a larger surplus.

Change in overall balance as a percentage of exports; improvement refers to a smaller deficit or a larger surplus.

Change in program country variable compared with change in corresponding variable for all non-oil developing countries during identical comparison periods—namely, the year before the program began and the program year, and the three-year period before the program and the three-year period beginning with the program year.

Examining performance during the year of the program itself, 72 per cent of all program countries experienced either an absolute reduction in the current account ratio and/or an improvement in the overall balance ratio; measured relative to the average change for all NODCs, the proportion of such program countries was 80 per cent. On a three-year comparative basis, the number of countries registering either a reduction in the current deficit ratio and/or an improvement in the overall balance rose further, to 87 per cent or 89 per cent, depending on whether absolute or “relative” changes are analyzed.

This evidence suggests that if changes in the current account and overall balance are taken as a standard of measurement, program countries tended, on average, to achieve some significant improvement in their balance of payments performance in both the short run and the long run. This improvement occurred both in an absolute sense and in relation to the average trends experienced by all NODCs. For the latter comparison, of course, the striking reductions recorded for program countries would be consistent with a relatively greater need on their part to undertake external adjustment.16

INFLATION

The outcome for program countries’ inflation performance is presented in Tables 4 and 5. In order to lessen statistical distortions in the averaging procedure, a small number of program countries with exceptionally high rates of inflation (defined as those in which inflation rates in excess of 35 per cent per year were recorded) were removed from the program set.17 The data indicate that in only slightly more than one third of this adjusted set of program countries were inflation rates reduced; on average, the annual rate of consumer inflation rose by about 2 percentage points following the adoption of the program, for both one-year or three-year comparison periods. It may be noted that one factor contributing to this outcome could be the effect of relative price adjustments sometimes undertaken as part of stabilization programs, such as increases in administered prices and depreciation of the exchange rate.

Table 4.

Upper Credit Tranche Stand-By Arrangements, 1971-80: Inflation Performance Comparing Pre-Program and Post-Program Periods

article image
Sources: International Monetary Fund, International Financial Statistics, various issues; International Monetary Fund (1981).

Change in program country’s rate of consumer price inflation compared with change in rate of consumer price inflation for all non-oil developing countries during identical comparison periods—namely, the year before the program began and the program year, and the three-year period before the program began and the three-year period beginning with the program year.

Table 5.

Upper Credit Tranche Stand-By Arrangements, 1971-80: Average Change in Inflation Performance Comparing Pre-Program and Post-Program Periods

(Change in the rate of increase in consumer prices)1

article image
Sources: International Monetary Fund, International Financial Statistics, various issues; International Monetary Fund (1981).

A positive figure indicates an increase in the inflation rate, a negative figure a decrease in the inflation rate. Numbers for each year are simple arithmetic averages of data for individual programs.

Equals unweighted arithmetic average of all individual program country data for 1971-80 period and therefore does not equal simple average of yearly columns; for the three-year comparison, calculation is made based only on 1971-79 data.

Change in the rate of consumer price inflation comparing year before program began and program year.

Excludes certain high-inflation program countries (i.e., those with recorded inflation rates in any one year greater than 35 per cent).

Change in the annual average rate of consumer price inflation comparing the three-year period before the program began and the three-year period beginning with the program year. For some 1978 programs and all 1979 programs, post-program calculation is on a two-year basis. Similarly, 1979 data for all NODCs are on a two-year basis.

However, a somewhat different picture emerges when one considers the behavior of the average inflation rate in all non-oil developing countries during the same periods. Such an assessment is of particular relevance, since theoretical considerations would suggest that for those small, open economies with a fixed exchange rate, the prevailing level of world inflation exerts an important influence on the domestic inflation rate, at least over the longer run. Viewed in relation to inflation trends in all NODCs, 58 per cent of program countries attained some relative reduction in inflation rates during the program period, while 42 per cent did not; the average program country inflation rate rose by 1.8 percentage points, compared with an average increase of 3.2 percentage points in NODC inflation. Over a three-year period, a relative improvement occurred in a higher proportion (75 per cent) of programs; the increase recorded for program countries averaged 2.3 percentage points, while all NODCs’ inflation rates increased by an average of 6.9 percentage point’s. This relatively better inflation performance registered by program countries over a longer time period is a not unexpected outcome, as the price adjustment measures referred to earlier would tend to have a once-and-for-all impact on recorded inflation rates in program countries over the short run (i.e., during the program period).

ECONOMIC GROWTH

The data on economic growth performance contained in Table 6 do not suggest a clear-cut pattern.18 About half (between 48 and 57 per cent) of the program countries experienced an increase in their real growth rates, and about half a decrease; this result holds irrespective of whether short-run or long-run periods are considered or whether the absolute change or the change relative to that of all NODCs is measured. A similar picture is presented in Table 7, which indicates that, on average, program countries’ growth rates declined by 0.3 percentage points in the short run but were unchanged over a longer-term comparison period. The real growth rate experienced by all NODCs fell marginally during the corresponding periods by 0.2 per cent and 0.4 per cent, respectively. Thus, in the short run, the change in the growth rates achieved on average was little different from that of all non-oil developing countries, while program countries’ growth performance improved slightly, in a relative sense, during a three-year comparison period.

Table 6.

Upper Credit Tranche Stand-By Arrangements, 1971-80: Economic Growth Performance Comparing Pre-Program and Post-Program Periods

article image
Sources: International Monetary Fund, International Financial Statistics, various issues; International Monetary Fund (1981).

Change in program country’s real growth rate compared with change in real growth rate of all non-oil developing countries during identical comparison periods—namely, the year before the program began and the program year, and the three-year period before the program began and the three-year period beginning with the program year.

Table 7.

Upper Credit Tranche Stand-By Arrangements, 1971-80: Average Change in Economic Growth Performance Comparing Pre-Program and Post-Program Periods

(Change in growth rate of real GDP)1

article image
Sources: International Monetary Fund, International Financial Statistics, various issues; International Monetary Fund (1981).

A positive figure indicates an increase in the growth rate of GDP, a negative figure a decrease in the growth rate of GDP. Numbers for each year are simple arithmetic averages of data for individual programs.

Equals unweighted arithmetic average of all individual program country data for 1971-80 period and therefore does not equal simple average of yearly columns; for the three-year comparison, the calculation is made based only on 1971-79 data.

Change in real GDP growth rate comparing the year before the program began and the program year.

Change in annual average real GDP growth rate comparing the three-year period before the program began with the three-year period beginning with the program year. For some 1978 programs and all 1979 programs, post-program calculations are made on a two-year basis. Similarly, 1979 data for all NODCs are on a two-year basis.

Overall, therefore, the aggregative evidence is consistent with the views that, in general, program countries’ economic growth performance tended not to have changed greatly during the program periods and that this outcome was not significantly different from the NODC average; however, some modest relative increases in growth rates could be said to have taken place over a longer period. In any event, the evidence analyzed here would not support the hypothesis that program countries generally experienced either sharp slowdowns or sharp rises in their economic growth rates.19

III. Developments in Other Key Variables

The preceding section sought to provide some systematic information regarding the outcome for the balance of payments, inflation, and growth—variables of major concern in Fund-supported adjustment programs. Also of considerable importance is the question of how other key macroeconomic variables might have been affected during the period of the adjustment effort. In an attempt to address some of the relevant issues, this section investigates the behavior of savings and investment separately in relation to GNP, so as to provide a basis for an interpretation of the observed change in the current account ratio. The savings ratio, in turn, together with the behavior of changes in real output, provides information on trends in real consumption—a variable often of direct political and social relevance and one that tends to be affected by policies designed to adjust domestic demand.

CHANGES IN SAVINGS AND INVESTMENT

As indicated in the previous section, the current account deficit ratio tended, on average, to decrease for program countries. Since the current account balance is equal to the difference between domestic investment and national savings, it is of interest to examine whether this outcome was the result of changes in the savings ratio, in the investment ratio, or in both ratios.

According to the evidence summarized in Table 8, in about 60 per cent of program countries, the savings ratio rose; this proportion is identical for one-year and three-year comparisons.20 The average rise recorded (Table 9) was equivalent to about ½ of 1 percentage point of GNP (again, this calculation was invariant to whether behavior in the short run or long run was analyzed). At the same time, however, slightly less than two thirds of the programs recorded decreases in the investment/GNP ratio during the program period: a decrease occurred in a slightly smaller percentage of cases (55 per cent) over the three-year comparison period.21 The average decline in the investment ratio was equivalent to between 0.6 and 1.0 percentage points of GNP.

Table 8.

Upper Credit Tranche Stand-By Arrangements, 1971-80: Savings and Investment Performance Comparing Pre-Program and Post-Program Periods

(Savings and investment as a ratio to GNP)1

article image
Source: International Monetary Fund, International Financial Statistics, various issues.

In a very small number of cases, ratios to GDP are used.

Change in ratio comparing year before program began and program year.

Change in program country variable compared with change in corresponding variable for all non-oil developing countries during identical comparison periods.

Change in ratio comparing the average levels during the three-year period before the program began and the three-year period beginning with the program year. For some 1978 programs and all 1979 programs, post-program comparison is on a two-year basis.

Table 9.

Upper Credit Tranche Stand-By Arrangements, 1971-80: Average Change in Savings and Investment Performance Comparing Pre-Program and Post-Program Periods

(Changes in savings and investment as percentages of GNP)1

article image
Source: International Monetary Fund, International Financial Statistics, various issues.

A positive figure indicates an increase in the savings ratio or the investment ratio, a negative figure a decrease in one of these ratios. Numbers for each year are simple arithmetic averages of data for individual programs. In a very small number of cases, ratios to GDP are used. The change in the (savings-investment) ratio shown in this table is identical to the change in the current account deficit ratio shown in Table 1, except for minor rounding errors present for some individual years.

Equals the unweighted arithmetic average of all individual program country data for 1971-80 and, therefore, does not equal simple average of yearly columns; for the three-year comparison, calculations are made based only on 1971-79 data.

Change in ratio comparing year before the program began and the program year.

Change in ratio comparing three-year period before the program began and three-year period beginning with the program year. For some 1978 programs and all 1979 programs, post-program calculation is on a two-year basis. Similarly, 1979 data for all NODCs are on a two-year basis.

Thus, while the results do not appear to exhibit a clear-cut pattern, it nevertheless seems that there was some tendency for the reduction experienced, on average, in the current account deficits of program countries to be reflected in a small increase in the savings ratio accompanied by a somewhat larger decline (in the longer run) in the investment ratio. As noted in the preceding section, since economic growth rates were, by and large, unaffected on average in program countries, this outcome would be consistent with some improvement in the productivity of investment expenditures. For NODCs in general, the savings/ investment performance exhibited was different; since their average investment ratio rose, while the national savings ratio either remained unchanged (in the short run) or rose by less than the increase in the investment ratio (in the longer run), their current account deficits tended to rise. This difference in behavior could reflect the fact that the external imbalances in many program countries may have required some adjustment in investment, after taking into account the potential for realizing increased domestic savings. For many developing countries not in the program set, however, a widening current account deficit (partly reflecting an increase in investment) may well have been consistent with the maintenance of a viable external position.

CHANGES IN REAL CONSUMPTION

In discussing the effectiveness of adjustment policies, it is generally recognized that the ability of many governments to implement a stabilization effort successfully and on a sustained basis will be heavily influenced by the effects of their policy measures on the living standards of the population and the length of time during which such effects may be present. Although the concept of living standards cannot be measured directly, some inferences can be drawn by examining trends in real consumption.

As was noted previously, the savings/GNP ratio tended to rise on average for program countries, which implies that there was a corresponding decline in the consumption/GNP ratio. However, the real growth rates of program countries did not exhibit any marked tendency to fall. Thus, provided the real growth rate was positive and sufficiently large, an increase in the savings rate would be consistent with some rise in real consumption, as part of the increment to the supply of resources could be used to improve the external position while also permitting some increase in the volume of consumption expenditures.22

According to the evidence summarized in Table 10, 42 per cent of program countries experienced an increase in the rate of real consumption growth23 during the program period.24 This proportion rises slightly over a long-term period (to 48 per cent), an outcome consistent with the view that the adjustment measures might exercise a relatively larger expenditure-restraining effect at the outset of the adjustment effort than later on. At the same time, between 52 and 58 per cent of program countries experienced a reduction in the growth rate of real consumption. However, Table 10 also indicates that a large majority (ranging from 73 to 85 per cent) of programs maintained positive growth rates during both one-year and three-year comparison periods. Thus, while there was some tendency for the rate of increase in real consumption to slow down, real consumption nonetheless generally continued to increase.

Table 10.

Upper Credit Tranche Stand-By Arrangements, 1971-80: Changes in Real Consumption Comparing Pre-Program and Post-Program Periods1

article image
Source: International Monetary Fund, International Financial Statistics, various issues.

Change in the sum of government and private consumption, deflated by the consumer price index.

Includes programs in which a reduction in the rate of decline of real consumption occurred.

In terms of the actual changes recorded, Table 11 indicates that, on average, program countries experienced a fall of 0.2-0.3 percentage points in their annual average growth rates of real consumption. However, the real consumption growth rate remained, on average, positive in the period following the adoption of the program (3.6 per cent or 4.3 per cent, depending on the time period examined). While no precise population data are available, these increases would in all likelihood be consistent with a modest rise in per capita real consumption. Finally, the growth rates experienced by program countries tended, on average, to be lower (by about one percentage point) than those of all NODCs in the periods following the adoption of the programs; at the same time, the reduction in the growth rate was generally less than that experienced by all NODCs.

Table 11.

Upper Credit Tranche Stand-By Arrangements, 1971-80: Average Change in Real Consumption Comparing Pre-Program and Post-Program Periods1

(In per cent)

article image
Source: International Monetary Fund, International Financial Statistics, various issues.

Change in the sum of government and private consumption, deflated by the change in consumer prices. Numbers for each year are simple arithmetic averages of data for individual programs.

Equals unweighted arithmetic average of all individual program country data for 1971-80 and therefore does not equal simple average of yearly columns. For the three-year comparison, calculations were made based only on 1971-79 data.

IV. Summary and Conclusions

The purpose of this paper has been to provide systematic quantitative evidence regarding the macroeconomic performance of developing countries that in the 1970s undertook adjustment programs supported by the use of Fund resources under stand-by arrangements in the upper credit tranches. Since the approach taken is a highly aggregative one, the overall findings conceal a wide diversity of outcomes and thus should not be taken as applying to any particular country or group of countries. Also, the results should not be necessarily interpreted as a direct gauge of the “effectiveness” of Fund-supported programs, particularly since no attempt was made either to relate results to actual policy implementation or to assess what might have occurred in the absence of a Fund-supported program. Nevertheless, by investigating the experience with a large number of programs25 over both the short run and the long run, together with the performance of all non-oil developing countries, some objective evidence can be obtained as to the extent to which program countries’ average performance differed from that of non-oil developing countries in general. Such an approach seems especially relevant to an examination of some of the important issues raised in the discussions of the Fund’s conditionality practices. Specifically, two sets of questions were addressed. First, to what extent did program countries experience, on average, improvements in their balance of payments, inflation, and growth performances (all of which are objectives of Fund-supported programs)? Second, what can be said about other key variables that may also be affected by adjustment efforts; in particular, how were changes in the current account decomposed into changes in savings and investment, and what were the effects on real consumption?

As regards the first set of questions, the statistical evidence presented strongly points to the conclusion that if changes in the current account and the overall balance of payments are viewed as indicators, the external sector generally exhibited a significant improvement in program countries. The changes that occurred in these indicators in program countries were generally larger than those recorded by all NODCs, an outcome consistent with the former group’s greater need to undertake external adjustment. In the case of inflation, the evidence indicates that considerably less progress was achieved in terms of actually reducing rates of price increase. Nevertheless, program countries tended to experience a decrease in their average domestic inflation rates relative to the inflation rates in all NODCs, particularly over the longer-run period of comparison. As regards the rate of economic growth, the evidence does not support the hypothesis that there was any significant downward bias in the performance of program countries. By and large, programs seemed overall to be associated with a broadly neutral outcome, insofar as the economic growth variable was concerned.

Turning to the second set of questions raised, the absence of a downward bias in the growth rate is to be noted, since the improvement in the current account deficit ratio was generally brought about partly by a decline in the average investment/GNP ratio; such an outcome would be consistent with some increase in the productivity of investment. The external improvement also resulted, on average, from a rise in the national savings ratio. However, largely because rates of economic growth tended to be maintained, this increased savings did not appear to result in a major compression of real consumption. Although for more than half of the program countries, annual real consumption growth rates declined slightly, in a large majority of cases, they remained positive at an average level of around 4 per cent, which, in all likelihood, would be consistent with a small increase in per capita real consumption levels.

The central finding of the paper, therefore, is that in broad terms, program countries recorded significant reductions in their external deficits while they exhibited only marginal changes in their growth rates of real GDP and consumption—changes that were not significantly different from those experienced by non-oil developing countries in general. Thus, considering the group of program countries in the aggregate, the costs associated with the external adjustment effort appear to have been less severe than has sometimes been suggested by participants in the controversy on Fund conditionality. These results could be viewed as consistent with the purpose of Fund conditionality—namely, the provision of external financial assistance conditional on the adoption of appropriate policy measures so as to promote balance of payments adjustment while minimizing the disruptive effect on economic activity or economic welfare.

The previously mentioned finding—namely, that for program countries balance of payments variables were the most affected, while inflation and (especially) economic growth were less affected—could be explained by a number of general considerations. One factor is that unlike inflation and growth, which are in principle unbounded in either direction, the balance of payments represents, in a sense, a budget constraint on the economy; even in the absence of appropriate measures, there is a limit on the size of the external deficit an economy can incur—a limit that generally tends to narrow over time. Also, since the balance of payments variables represent the difference between the demand for, and the supply of, resources, a combined use of demand management and supply policies may have a relatively powerful impact on the size of external deficits. By contrast, the rate of economic growth is less susceptible in the short run to policy action than other target variables, as it ultimately reflects supply conditions; this is particularly true for economies where the production of primary commodities determines, to a large extent, the growth rate. As regards the inflation objective, in many instances, adjustments in administered prices introduced as part of a stabilization effort aimed at improving the structure of relative prices will tend to impart an upward bias to recorded rates of inflation, especially if these adjustments are undertaken at relatively infrequent intervals; also, for those small, open economies with a fixed exchange rate, the imported inflation rate can be an important factor influencing the domestic inflation rate.

Finally, the paper’s conclusions reflect the results obtained from an aggregative analysis of certain indicators of the macro-economic performance of program countries on average. Several areas suggest themselves for further study. For example, there is scope for a more complex statistical investigation of the data involving, for instance, correlation and other statistical significance tests. In addition, in view of the wide diversity of individual experiences subsumed in the aggregative analysis, it would be useful to explore the extent to which the results vary between different subgroups of program countries—for example, by considering those countries with real per capita incomes below a certain level. In the same vein, it could be useful to distinguish economies that exhibit a broadly similar degree of openness to foreign trade or that have a comparable export structure.

APPENDIX: Definitions and Coverage of Data

This appendix describes the methodology underlying the choice of program countries and program periods and also lists the countries and periods used for each variable as well as the corresponding NODC series.

CHOICE OF PROGRAM COUNTRIES AND PERIODS

The analysis includes only upper credit tranche stand-by arrangements concluded during the period 1971–80.26 Extended arrangements were excluded from the analysis, partly because most of these arrangements were entered into during the latter part of the decade and there were relatively few cases where the program was fully completed and data were available for the post-program period. Also, since the economic content of extended arrangements, in terms of both objectives and policies, should, in principle, be different from that of shorter-term stand-by arrangements, they are deserving of separate analysis.

The data for most of the variables examined were generally available only on a calendar year or (in some cases) a fiscal year basis. For many countries, however, the period of the stand-by arrangement often spanned more than one calendar year or fiscal year. In order to deal with this problem, the concept of one or more “program periods” was used in all cases as the basic time period of analysis for stand-by arrangements. In practice, this implies that a one-year stand-by arrangement that came into effect midway through a period covered by one data series (i.e., a calendar year or fiscal year) is viewed as corresponding to two distinct program periods. As a result, the number of program periods or data points covered in the analysis is greater than the number of stand-by arrangements approved during the period. To a certain extent, the use of this technique may tend to understate the actual changes in variables that occurred during the period of the stand-by arrangement.27

The comparative statistical analysis required observations for the variable in question both for the period before the program began and for the program period itself, as well as for the three-year period preceding the program and the three-(or two-) year period beginning with the program period. For some countries, depending on the variable in question, data were either nonexistent or were available only for a limited number of years. In such instances, these program periods were removed from the analysis as necessary, and, as a result, the number of observations differs depending on the variable being analyzed (see the following subsection). For this reason, the data for both 1979 and (especially) 1980 do not include the outcomes of a significant number of stand-by arrangements concluded during these years.

LISTING OF PROGRAM PERIODS USED FOR EACH VARIABLE

The following are the program periods included for each variable; a single asterik (*) indicates that the calculation was available only for the one-year comparison period, while a double asterisk (**) indicates that the post-program calculation referred to a two-year period (this was the case for all 1979 programs and some 1978 programs). In some instances, the listing indicates that the data were on a fiscal year basis (corresponding to the series shown in IFS).

Current account balance, savings ratio, and investment ratio

Argentina (1976, 1977, 1978*); Bolivia (1973); Burma (1974/75, 1977/78, 1978/79); Chile (1974, 1975); Colombia (1971, 1972); Costa Rica (1980*); Ecuador (1972, 1973); Egypt (1977); Gabon (1978**); Guyana (1978, 1979**); Haiti (1975, 1976, 1977); Honduras (1971, 1972, 1973); Israel (1975, 1976, 1977); Jamaica (1973, 1977); Kenya (1979**, 1980*); Korea (1971, 1975, 1976, 1980*);

Liberia (1980*); Malawi (1979**, 1980*); Mauritius (1979*); Morocco (1971); Nicaragua (1972); Pakistan (1972/73, 1973/74, 1974/75, 1977/78); Panama (1971, 1978**, 1979*); Peru (1978, 1979**, 1980*); Philippines (1971, 1972, 1973, 1980*); Portugal (1978, 1979**); South Africa (1976/77); Sri Lanka (1971, 1974, 1978); Sudan (1972/73; 1973/74; 1974/75); Tanzania (1975, 1976); Togo (1979*); Uruguay (1972, 1973); Yugoslavia (1971, 1972); Zaïre (1977, 1979**, 1980*); Zambia (1976, 1977, 1978**, 1979*).

Overall balance of payments

Argentina (1976, 1977, 1978); Bangladesh (1976); Bolivia (1973, 1980*); Burma (1974, 1977, 1978, 1979**); Chile (1974, 1975); Colombia (1971, 1972); Congo (1979**); Costa Rica (1980*); Ecuador (1972, 1973); Egypt (1977); Gabon (1978); Ghana (1979**); Guyana (1978, 1979**); Haiti (1975, 1976, 1977); Honduras (1971, 1972, 1973); Israel (1975, 1976, 1977); Jamaica (1973, 1977); Kenya (1979**, 1980*); Korea (1971, 1975, 1976, 1980*); Malawi (1979, 1980); Mali (1971); Mauritius (1979**, 1980*); Morocco (1971); Nicaragua (1972); Pakistan (1973, 1974, 1977, 1978); Panama (1971, 1978, 1979*); Peru (1978, 1979*); Philippines (1971, 1972, 1973, 1980*); Portugal (1978, 1979*); Somalia (1980**); South Africa (1977); Sri Lanka (1971, 1974, 1978**); Sudan (1973, 1974, 1975); Tanzania (1975, 1976, 1980*); Turkey (1978, 1979**, 1980*); Uruguay (1972, 1973); Yugoslavia (1971, 1972); Zambia (1976, 1977, 1978, 1979**).

Inflation

Bangladesh (1976*); Burma (1974, 1977, 1978); Colombia (1971, 1972); Congo (1979**); Costa Rica (1980*); Ecuador (1972, 1973); Egypt (1977); Gabon (1978); Guyana (1978, 1979**); Haiti (1975, 1976, 1977); Honduras (1971, 1972, 1973); Jamaica (1973, 1977*); Kenya (1979**, 1980*); Korea (1971, 1975, 1976, 1980*); Liberia (1980*); Malawi (1979**, 1980*); Mauritania (1980*); Morocco (1971); Pakistan (1972, 1973, 1974, 1975, 1977, 1978); Panama (1971, 1978, 1979**, 1980*); Philippines (1971, 1972, 1973, 1980*); Portugal (1978, 1979**); Sierra Leone (1979**, 1980*); South Africa (1976, 1977); Sri Lanka (1971, 1974, 1978); Sudan (1972, 1973, 1974, 1975); Tanzania (1975, 1976, 1980*); Togo (1979**, 1980*); Western Samoa (1979**, 1980*); Yugoslavia (1971, 1972); Zambia (1976, 1977, 1978, 1979**).

Change in real GDP

Argentina (1976, 1977, 1978); Bolivia (1973, 1980*); Burma (1974/75, 1977/78, 1978/79); Chile (1974, 1975); Colombia (1971, 1972); Costa Rica (1980*); Ecuador (1972, 1973); Egypt (1977); Haiti (1975, 1976, 1977); Honduras (1971, 1972, 1973); Israel, (1975, 1976, 1977); Jamaica (1973, 1977); Kenya (1979**, 1980*); Korea (1971, 1975, 1976, 1980*); Liberia (1980*); Malawi (1979**, 1980*); Morocco (1971); Nicaragua (1972); Pakistan (1972/73, 1973/74, 1974/75, 1977/78); Panama (1971, 1978, 1979**, 1980*); Peru (1978, 1979**, 1980*); Philippines (1971, 1972, 1973, 1980*); Portugal (1978, 1979**); South Africa (1976/77); Sri Lanka (1971, 1974, 1978); Tanzania (1975, 1976); Turkey (1978, 1979**, 1980*); Uruguay (1972, 1973); Yugoslavia (1971, 1972); Zambia (1976, 1977, 1978, 1979*); Zaïre (1977, 1979**, 1980*).

Change in real consumption

Argentina (1976, 1977, 1978*); Bolivia (1973, 1980*); Burma (1974/75, 1977/78, 1978/79); Chile (1974, 1975); Colombia (1971, 1972); Costa Rica (1980*); Ecuador (1972, 1973); Egypt (1977); Guyana (1978, 1979); Haiti (1975, 1976, 1977); Honduras (1971, 1972, 1973); Israel (1975, 1976, 1977); Jamaica (1973, 1977); Kenya (1979**, 1980*); Korea (1971, 1975, 1976); Liberia (1980*); Malawi (1979**, 1980*); Morocco (1971); Pakistan (1972/73, 1973/74, 1974/75, 1977/78); Panama (1971, 1978**, 1979*); Peru (1978, 1979**, 1980*); Philippines (1971, 1972, 1973); Portugal (1978, 1979); Sri Lanka (1971, 1974, 1978); South Africa (1976/77); Sudan (1972/73, 1973/74, 1974/75*); Tanzania (1975, 1976); Togo (1979*); Uruguay (1972, 1973); Yugoslavia (1971, 1972); Zaïre (1977, 1979*); Zambia (1976, 1977, 1978**, 1979*).

SERIES FOR ALL NON-OIL DEVELOPING COUNTRIES

These series are provided in Table 12.

Table 12.

Series for All Non-Oil Developing Countries, 1968–80

article image
Sources: International Monetary Fund, International Financial Statistics, various issues, and Annual Report, 1981 (Washington, 1981). For explanations of the derivation of each series, see the text.

REFERENCES

  • Beveridge, W. A.,Fiscal Adjustment in Financial Programs Supported by Stand-By Arrangements in the Upper Credit Tranches, 1978–79” (unpublished, International Monetary Fund, July 1, 1981).

    • Search Google Scholar
    • Export Citation
  • Beveridge, W. A., and Margaret R. Kelly,Fiscal Content of Financial Programs Supported by Stand-By Arrangements in the Upper Credit Tranches, 1969–78,Staff Papers, Vol. 27 (June 1980), pp. 20549.

    • Search Google Scholar
    • Export Citation
  • Bird, Graham R.,The IMF and the Developing Countries: Evolving Relations, Use of Resources and the Debate Over Conditionality” (unpublished, Overseas Development Institute, March 1981).

    • Search Google Scholar
    • Export Citation
  • Connors, Thomas A.,The Apparent Effects of Recent IMF Stabilization Programs,International Finance Discussion Papers, No. 135 (Board of Governors of the Federal Reserve System, April 1979).

    • Search Google Scholar
    • Export Citation
  • Dell, Sidney S., On Being Grandmotherly: The Evolution of IMF Conditionality, Essays in International Finance, No. 144, International Finance Section, Department of Economics, Princeton University (1981).

    • Search Google Scholar
    • Export Citation
  • Donovan, Donal J.,Real Responses Associated with Exchange Rate Action in Selected Upper Credit Tranche Stabilization Programs,Staff Papers, Vol. 28 (December 1981), pp. 698727.

    • Search Google Scholar
    • Export Citation
  • Guitián, Manuel,Economic Management and International Monetary Fund Conditionality,in Adjustment and Financing in the Developing World: The Role of the International Monetary Fund, ed. by Tony Killick (Washington, 1982), pp. 73104.

    • Search Google Scholar
    • Export Citation
  • Independent Commission on International Development Issues (“Brandt Commission”), North-South: A Program for Survival (Cambridge, Massachusetts, MIT Press, 1980).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, Annual Report of the Executive Board for the Financial Year Ended April 30, 1981 (Washington, 1981).

  • Kincaid, G. Russell,Conditionality and the Use of Fund Resources: Jamaica,Finance & Development, Vol. 18 (June 1981), pp. 1821.

    • Search Google Scholar
    • Export Citation
  • Nowzad, Bahram, The IMF and its Critics, Essays in International Finance, No. 146, International Finance Section, Department of Economics, Princeton University (1981).

    • Search Google Scholar
    • Export Citation
  • Reichmann, Thomas M.,The Fund’s Conditional Assistance and the Problems of Adjustment, 1973–75,Finance & Development, Vol. 15 (December 1978), pp. 3841.

    • Search Google Scholar
    • Export Citation
  • Reichmann, Thomas M., and Richard T. Stillson,Experience with Programs of Balance of Payments Adjustment: Stand-By Arrangements in the Higher Tranches, 1963–72,Staff Papers, Vol. 25 (June 1978), pp. 293309.

    • Search Google Scholar
    • Export Citation
  • Schmitt, Hans O., Economic Stabilization and Growth in Portugal, IMF Occasional Paper, No. 2 (Washington, 1981).

  • United Nations Conference on Trade and Development, The Balance of Payments Adjustment Process in Developing Countries: Report to the Group of Twenty-Four, UNDP/UNCTAD Project INT/75/015 (January 1979).

    • Search Google Scholar
    • Export Citation
  • Williamson, John,On Judging the Success of IMF Policy Advice” (unpublished, Institute for International Economics, December 1981).

    • Search Google Scholar
    • Export Citation
*

Mr. Donovan, economist in the Stand-By Policies Division of the Exchange and Trade Relations Department, is a graduate of Trinity College, Dublin. He received his doctorate from the University of British Columbia.

1

For a survey of many of the issues involved, see, for example, Bird (1981), pp. 23-33) and the references cited therein: see also Independent Commission on International Development Issues (“Brandt Commission”) (1980), pp. 215-16), Dell (1981), Nowzad (1981), and UNCTAD (1979). A comprehensive analysis of the evolution of Fund conditionality is contained in Guitián (1982).

2

Published analyses of programs on a global basis have been made by Fund staff members (see Reichmann (1978) and Reichmann and Stillson (1978)). The recent programs of Jamaica and Portugal have also been studied (see Kincaid (1981) and Schmitt (1981), respectively).

3

See Section I for a discussion of the concept of the program period.

4

For an empirical examination of the fiscal content of upper credit tranche stabilization programs supported by the Fund, see Beveridge and Kelly (1980) and Beveridge (1981). The responses of selected macroeconomic variables to exchange rate actions undertaken as part of Fund-supported stabilization programs during 1970-76 are analyzed in Donovan (1981).

5

The concept of the program period used for this analysis is explained in the Appendix.

6

The data for all NODCs includes data for the program countries themselves, and thus the results tend to underestimate the actual differences between program and nonprogram countries. However, the underestimation involved is quite small, as for each period considered, program countries generally represented less than 10 per cent of the countries included in the aggregate NODC series.

7

However, care needs to be taken in interpreting the results of a relative comparison of this kind, since it is quite likely that by definition, program countries were, on average, faced with a greater need to adjust than non-oil developing countries in general. In particular, the external position of many nonprogram developing countries might not have required adjustment to restore a sustainable balance of payments position, while this was not generally the case in program countries.

8

It should be noted that for individual countries within both the group of program countries and the group of all NODCs, the interpretation of a given change in a variable during any time period would depend upon the level of the variable at the beginning of the period.

9

In a recent work by Connors (1979), a nonparametric rank test was used to assess whether during the period 1973-77, the behavior of certain macrovariables measured in absolute terms (viz., economic growth, inflation, the current account balance, and the fiscal deficit/gross domestic product (GDP) ratio) exhibited a statistically significant difference in the one-year periods before and after initiation of Fund-supported programs. The coverage of the present study differs from that of Connors, since programs during a longer time period (1970-80) are included and comparisons extending over a three-year time horizon and taking into account the comparative behavior of all NODCs are examined. Although the paper by Connors does not present data indicating the actual average changes in the variables recorded by program countries or the proportion of program countries that experienced increases or decreases in the variables in question, some of his qualitative conclusions are discussed in Section II.

10

It may be noted that in the case of medium-term programs—for example, those supported by the extended fund facility (which are not included in the present analysis)—some widening of the overall deficit might be allowed for in the initial stage of the program, again depending on the country’s particular circumstances.

11

The current account ratio of the balance of payments is defined as the difference between gross domestic investment and gross national savings, expressed as a ratio to GNP; all series were taken from the Fund’s International Financial Statistics (IFS) national accounts series. In a small number of cases where GNP series were not available, a series based on gross domestic product (GDP) was used instead. The current account defined thus is equivalent to net exports of goods and services plus net factor income from abroad. National accounts data sources (rather than balance of payments data in foreign currencies) were used to avoid complications involved in converting to local currency and to provide consistency between movements in the current account balance and changes in the difference between savings and investment (see Section III). The corresponding aggregate NODC series was calculated as the unweighted average of all individual NODC data derived in the same manner as data for program countries (see Appendix).

12

The overall balance of payments is defined as the total change in reserves (expressed in dollars) plus changes in payments arrears (where available), both of which were taken from the IFS balance of payments series. The series for the change in reserves equals the change in gross official reserves adjusted for changes in the use of Fund credit. It does not include changes in foreign assets and liabilities of commercial banks and, therefore, corresponds (except for changes in non-Fund official liabilities) to the change in net foreign assets of the monetary authorities adjusted for any accumulation/reduction in payments arrears. The adjustment for arrears was possible to undertake in most cases for those countries in the program set known to have arrears (i.e., the Congo, Ghana, Guyana, Jamaica, Zaïre, and Zambia); however, arrears data were not available for Turkey and Sudan. For all non-oil developing countries, the overall balance series was derived as the aggregate change in gross official reserves adjusted for net use of Fund credit (both series taken from IFS).

13

Improvement in this context refers to either a larger overall surplus or a smaller overall deficit.

14

Exports, rather than GDP, was employed as the scaling factor to facilitate aggregate comparisons of trends in the overall deficit, as the use of GDP would involve complications in choosing the appropriate exchange rate to use for conversion purposes.

15

Only program countries for which data are available for both indicators during the same time periods are included; thus, certain countries whose national accounts data and balance of payments data refer to different time periods are excluded. As a result, the number of programs considered is less than in Tables 1 or 2.

16

The results described (for the one-year absolute comparison of the current account balance) may be compared with those reported by Connors (1979). Although Connors used the absolute level of the current account deficit (rather than the deficit/GNP ratio)—which would tend to understate the possible improvement in an inflationary period—he found that, on average (and consistent with this paper’s results), an improvement took place. Connors also reported that the difference was statistically insignificant at the 95 per cent confidence level; in other words, there existed at least a 5 per cent chance that the “after” and “before” series were from the same “population.” However, in order properly to evaluate the importance of the latter result, it would be helpful to know how close the test statistics were to the 95 per cent level—that is, how much greater than 5 per cent the probability was that there was no difference in the pre-program and post-program behavior of the series.

17

See the Appendix for details. The inflation variable was taken to be the average change in consumer prices (from IFS). While the GDP deflator might for some purposes be a better indication of the underlying inflation rate, data for this variable were not as readily available for many program countries. The NODC series was taken from International Monetary Fund (1981).

18

Economic growth was calculated as the change in GDP at constant prices (taken from International Monetary Fund, International Financial Statistics, various issues). The non-oil developing countries series was taken from International Monetary Fund (1981).

19

This conclusion is consistent with the results reported by other authors. See Connors (1979), Donovan (1981), and Reichmann and Stillson (1978).

20

The national savings ratio was calculated from the national account series as follows: GNP minus the sum of government consumption and private consumption, expressed as a fraction of GNP. As was done with the current account balance, the domestic savings ratio was employed in a few instances rather than the national savings ratio.

21

The investment ratio equalled the sum of gross fixed capital formation plus the change in stocks, expressed as a fraction of GNP.

22

More precisely, (C*/Pc)=(C*/Y)+y*+Py*Pc* where C and Y denote nominal consumption and income, y denotes real income, Pc and Py denote indexes of consumer prices and the GDP deflator, respectively, and an asterisk indicates a percentage rate of change.

23

The change in real consumption was defined as the percentage change in total consumption, adjusted for the change in the consumer price index; total consumption equalled the sum of government consumption and private consumption, with both variables taken from the IFS national accounts series. The NODC series was derived as the simple average of series calculated in the same manner for all non-oil developing countries.

24

This includes cases where a reduction in the rate of decline of real consumption took place.

25

Between 54 and 78, depending upon the variables examined and the comparison period.

26

The study includes two-year stand-by arrangements, which were treated as equivalent to two distinct one-year stand-by arrangements; the number of such arrangements actually considered was relatively small, however, as they occurred near the end of the decade and, therefore, the necessary data were often not available (discussed later). Similarly, a succession of stand-by arrangements was treated as a series of independent arrangements.

27

As an illustration, for a stand-by arrangement in effect from midyear to midyear, the actual change in the variable during the one-year arrangement period is distributed among the two adjacent years that are analyzed as separate program periods.