In formulating an economic stabilization program, the use of the exchange rate as a policy instrument is frequently a source of considerable controversy and debate. It has been argued that, in certain circumstances, depreciating the currency can be a direct and powerful method of reducing real domestic expenditure and relative prices so as to divert economic resources toward the external sector, thereby improving the balance of payments, reducing inflationary pressures over the longer run, and laying the ground for sustained economic growth. It has also been suggested, however, that the behavioral responses to relative price changes may be slow and uncertain in many countries. At the same time, the potentially disruptive effects of abrupt changes in relative prices may be significant; in particular, the once-and-for-all rise in the cost of living resulting from a depreciation, and the consequent implications for the distribution of income, may have serious social and political repercussions.
By now, a considerable body of literature exists dealing with various theoretical and empirical considerations underlying these issues. This paper is not intended to add to this general body of knowledge, but rather to focus specifically on evidence provided by a selected set of stabilization programs supported by upper credit tranche stand-by arrangements in which exchange rate action played a major role. In designing these programs, it was felt that the positive results expected from a depreciation outweighed some of the potential shortcomings mentioned above. The paper’s purpose is to examine, for the programs in question, the actual outcome in relation to the intended effects for certain key variables, namely, real growth in exports and imports, growth in gross domestic product (GDP), and the rate of domestic inflation.
The experience with Fund-supported programs of balance of payments adjustment has been reviewed by Fund staff on a number of occasions. (See Reichmann and Stillson (1978) and Johnson and Reichmann (1978).) These reviews generally assessed the overall objectives, design, and outcome of a group of Fund-supported programs within different historical periods. The approach in the present paper is similar to that of Bhagwat and Onitsuka (1974), who examined actual ex post movements in exports and imports for three groups of countries that adopted different exchange rate policies. Their results indicated that significant differences in outcome were present between countries that carried out independent devaluations and those that did not. The present investigation follows the Bhagwat and Onitsuka approach by focusing on specific effects associated with exchange rate action. However, it differs in that, first, the analysis is confined to the experience of a selected set of Fund-supported programs and, second, the empirical analysis is extended to consider, in addition to movements in exports and imports, trends in real GDP growth and inflation.
It should be recognized that the analysis undertaken here needs to be interpreted with caution because “special factors” (such as adverse weather, to cite one example) may influence actual outcomes in ways unrelated to exchange rate policy or the other policy measures contemplated in the programs. To reduce the possible biases emanating from this source, first, a reasonably large group of programs (12) was chosen, so that the average outcomes measured for the entire group would reflect a mix of special factors working in different (and to some extent offsetting) directions. Second, to control for the effects of exogenous events, such as the world recession and the oil crisis of the 1970s, countries’ performances in respect of key variables were assessed not only in relation to their own past performance but also vis-à-vis average trends in those same variables for similar groups of countries during the corresponding periods. Finally, to allow for the effect of short-run distorting shocks, comparisons were undertaken both for the program period itself in relation to the immediately preceding period and for three-year periods preceding and following the program period. Longer-run comparisons of the latter type are also of particular relevance in assessing the lag with which any effects resulting from exchange rate action may become felt.
The paper is organized as follows. Section I provides a brief overview of the theoretical and empirical considerations underlying the use of exchange rate policy. Section II discusses the choice of the programs to be investigated and the main characteristics of the exchange rate action involved. Also, the principal features of the design of the various programs are summarized, including the nature of the economic problem and the evidence indicating that the exchange rate was “out of line,” the intended effects of the exchange rate action, and the principal supporting policy measures contemplated. Section III analyzes the actual outcome for the volume of exports and imports, inflation, and real growth in GDP. The comparisons in this section are undertaken both for short-run and longer-run periods and in relation to average “world” performance. Also, in some instances, the performance of certain subgroups of program countries is analyzed separately. Section IV contains a summary of the principal conclusions.
Bhagwat, Avinash, and Yusuke Onitsuka, “Export-Import Response to Devaluation—Experience in the 1960s of the Nonindustrial Countries,” Staff Papers, Vol. 21 (July 1974), pp. 414–62.
Brillembourg, Arturo, “Specification Bias in the Demand for Imports: The Case of the Grancolombian Countries” (unpublished, International Monetary Fund, April 14, 1975).
Elson, R. Anthony, “Exchange Rate Policy and the Performance of Traditional Experts in Argentina (1946–70)” (unpublished, International Monetary Fund, December 7, 1973).
Goldstein, Morris, “Have Flexible Exchange Rates Made Macroeconomic Policy More Difficult: A Survey of the Issues and the Evidence” (unpublished, International Monetary Fund, February 14, 1980).
Johnson, G. G., and Thomas M. Reichmann, “Experience with Stabilization Programs Supported by Stand-By Arrangements in the Upper Credit Tranches, 1973–75” (unpublished, International Monetary Fund, February 28, 1978).
Lanyi, Anthony, “External Economic Problems of Developing Countries: Recent Research by the Fund Staff” (unpublished, International Monetary Fund, February 21, 1980).
Okonkwo, Ubadigbo, “Export Taxes on Primary Products in Developing Countries: The Taxation of Cocoa Exports in West Africa” (unpublished, International Monetary Fund, November 29, 1978).
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. 293–309.
Singh, Anoop, “Sri Lanka: The Response of Paddy Producers to Price Stimuli” (unpublished, International Monetary Fund, September 2, 1975).
Teigeiro, José D., and R. Anthony Elson, “The Export Promotion System and the Growth of Minor Exports in Colombia,” Staff Papers, Vol. 20 (July 1973), pp. 419–70.
Mr. Donovan, economist in the Stand-By Programs 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.
“Goods” in this discussion also includes services.
For example, in many non-oil developing countries, a “successful” devaluation may well have an income-redistributive effect away from the politically sensitive urban population to the rural areas.
During the period 1970–76, of 62 upper credit tranche stand-by arrangements approved by the Fund, 29 involved exchange rate action. Among the latter group, 13 programs involved a commitment to flexible exchange rate policy. Among the remaining 16, the program for Korea contained a balance of payments test, while it was decided to exclude the 3 programs undertaken by Indonesia (in 1970, 1971, and 1973) on the grounds (a) that a rapid succession of programs made it extremely difficult to analyze coherently preprogram and postprogram performance and (b) that the oil boom in 1973 significantly distorted the postdepreciation outcomes for that country. The extended arrangement approved for Mexico (in 1976) also contained exchange rate action; however, this case was not considered, owing to the large impact of additional oil revenues on the program’s outcome.
The series for real effective exchange rates were calculated in a uniform manner for all countries using an average of end-period monthly import-weighted and export-weighted nominal exchange rates deflated by relative consumer price indices; the trade weights were based on data referring to the eight largest trading partners in each case for the year immediately prior to the depreciation.
This group also included at least one country (Bolivia) where, although relative competitiveness had not deteriorated in the three-year period immediately prior to the depreciation, during earlier periods the economy had experienced inflation rates in excess of those of its trading partners without any exchange rate adjustment taking place.
“Average world performance” was taken to be that of the average non-oil developing countries, except for Israel and South Africa, where the comparator is the average performance of industrial countries.
For most countries, export volume changes were calculated on the basis of data for export value and average export unit value given in International Financial Statistics (IFS). For countries for which such data were not available, volume indices were constructed directly from disaggregated export data contained in papers prepared by the Fund staff.
The “relative” comparison was undertaken by comparing, for each program, performance in the postdepreciation period with average world performance during that same period.
Pakistan is excluded, owing to the unavailability of sufficient data for the predepreciation period.
For Bolivia, real exports also declined in the first postdepreciation year, while for Jamaica, a decline in the growth rate was recorded (although it still remained positive). Exports rose sharply in Zambia, however, during the first year, before declining thereafter.
Bangladesh, Burma, Pakistan, and Sudan.
Afghanistan, Bangladesh, Bolivia, Burma, Ecuador, Jamaica, Pakistan, Sudan, and Zambia.
As with exports, for most countries, import volume data were based on IFS data. However, in some instances, where an import unit-value index was not available, the average import unit-value index of all non-oil developing countries was used to deflate the import value series.
Use of another possible measure of behavioral trends in imports, namely, the import/GDP ratio, suffers from the drawback that it includes the price effect of higher import costs, particularly of oil.
One indication of this aspect of the program might be obtained by examining the outcome for the overall balance of payments for the countries in this group. For Bangladesh, Burma, and Sudan (comparable data for Pakistan are not available), the overall balance deteriorated in each instance comparing three-year predepreciation and postdepreciation periods. However, it should be borne in mind that for each of these countries, data for the postdepreciation period include any adverse impact of the oil price rise.
Bolivia, Ecuador, Israel, Jamaica, South Africa, and Yugoslavia.
The preceding results are similar to those cited by Bhagwat and Onitsuka (1974), who found that nonindustrial countries that carried out independent devaluations during the 1960s exhibited a higher export growth rate than those that did not. For many of these countries, import growth also increased at the same time.
This figure is a rough estimate calculated by multiplying, for each program country, the size of the depreciation (in local-currency terms) by the share of imports in GDP during the year immediately preceding the depreciation.
In assessing the outcome for program countries’ inflation rates, note that the growth in the money supply rose sharply on average for those countries in the first year of the program (from 14 per cent to 21 per cent). At the same time, the rate of net domestic credit expansion slowed down on average (from 30 per cent to 24 per cent). Thus, it would appear that on average the reduction in credit expansion was insufficient to sterilize the liquidity impact of the balance of payments outcome during the program period. In this regard, the overall balance of payments (measured in foreign-currency terms) registered an improvement in the first year of the program in 5 of the 11 countries for which data are available. Further interpretation of these figures, however, would require a detailed analysis of the behavior of net domestic assets, net foreign assets, and the money supply in the program countries, which goes beyond the scope of the present study.
No data are available for Afghanistan.
In Zambia, however, the real effective exchange rate appreciated during most of the postdepreciation period and was reversed only after a further (delayed) exchange rate action.
Unfortunately, data are not available for most of the programs reviewed to permit a separate analysis of movements in real domestic expenditure (as opposed to real GDP).
However, within the import-restraint subgroup, the examples of Bolivia and Jamaica need to be distinguished separately, since in those countries economic growth declined (especially in Jamaica) and there was no improvement in the trade balance variables.