FRBSF Economic Letter
2001-08 | March 30, 2001
How Costly Are IMF Stabilization Programs?
- IMF programs and currency crises
- Measuring the costs of participation
- Measuring performance one year after the 1997 Asian financial crises
Pacific Basin Notes. This series appears on an occasional basis. It is prepared under the auspices of the Center for Pacific Basin Monetary and Economic Studies within the FRBSF’s Economic Research Department.
After the 1997 balance of payments problems and currency crises that hit Korea, Indonesia, Thailand, and other countries, the stabilization programs supported by the International Monetary Fund (IMF) came under critical fire. At issue was the cost of these programs in terms of forgone output and employment in the countries that adopted them. Stiglitz (2000), for example, argued that “…the IMF’s economic ‘remedies’ often make things worse—turning slowdowns into recessions and recessions into depressions.” Similar statements by other leading economists have been commonplace.
Despite this skepticism, no consensus has emerged about the impact of these programs on output and employment. Most empirical studies find that IMF-supported programs do improve the balance of payments and the current account. This is not surprising since a key purpose of many IMF programs is to provide resources to facilitate balance of payments adjustment. But views on the ultimate output and employment effects of IMF programs appear to be much more divergent.
On the surface, it may seem odd that countries would choose to participate in an IMF stabilization program if it were not in their best interests. A country might feel that the benefits of a program, such as balance of payments adjustment and continued access to credit markets, do not outweigh the costs of temporarily slower growth. But such a calculation may be overshadowed by the urgency of the situation and domestic and foreign political pressures. This Economic Letter reports on research (Hutchison 2001) that estimates the output costs of participating in an IMF program immediately after balance of payments and currency crises, much like those several Asian countries experienced in 1997.
IMF programs and currency crises
Between 1995 and 1999, the IMF approved 87 short-term stabilization programs (Standby Agreements and Extended Fund Facilities)—the lowest number during any five-year period since 1970 (over 670 total were approved during 1970-1999). Compared to the preceding 20 years, the average size (in dollar amounts) of these programs was much larger because the countries involved were so large; measured as a share of GDP, however, the average size was not unusual.
Currency and balance of payments crises also were commonplace during this five-year period (Glick and Hutchison 2001), and IMF involvement was frequent but not universal. Indeed, only about a third of the severe currency and balance of payments crises coincided with or were followed by the approval of an IMF program. In the Asian financial crises of 1997, Korea, Indonesia, Thailand, and the Philippines negotiated IMF programs, while Malaysia did not. This recent experience, along with data dating to the mid-1970s, provides the basis for a study of the benefits and costs of participating in IMF programs compared to those of not participating.
Measuring the costs of participation
I investigated short-run IMF stabilization programs for 67 developing and emerging-market countries between 1975 and 1997. These programs were aimed at adjusting the short-run balance of payments rather than at structural reform, poverty reduction, and so on.
Simple averages indicated that real GDP growth was about 3% during the years that developing and emerging-market countries participated in an IMF program, and 4% in “normal” times. This is not surprising since there are systematic differences between “normal” times and the circumstances that induce countries to approach the IMF for assistance. Focusing only on countries participating in IMF programs (at some point during the sample), for example, I found that average GDP growth was 3.6% during non-program years and 2.9% during program years.
Looking at statistics for countries before and after participating in IMF programs, I find that output growth does not decline substantially when a country enters an IMF program, but it does increase significantly during the two-year period following the program. Economies typically experience slow growth before participation and grow sluggishly until the program concludes. It does not appear that participation in the IMF program directly contributed to slower growth.
More formal multivariate regression results—controlling for country fixed effects, domestic policies and external conditions—indicate that the onset of a currency crisis in a given year is followed a year later by a decline in output growth of about 1.2 percentage points. Similarly, IMF participation appears to dampen growth—even when the onset of a currency crisis is taken into account—by about 0.7 percentage points for each year the country is in a program. A closer look indicates that growth begins to slow the year prior to IMF program approval—indeed, this is one reason countries turn to the IMF in the first place—and remains depressed only during the first year of the program. The “effect” of IMF short-term stabilization programs on output growth is not noticeable in subsequent years.
I also investigated the mechanism through which the conditions attached to IMF loans may affect credit growth. I found that countries generally respond to currency crises by expanding credit growth but reduce it during an IMF program; the magnitudes of these expansions and reductions are about the same. Balance of payments crises reduce credit growth, and IMF programs reduce credit growth, by about 4 percentage points. The joint effect of a currency crisis followed by an IMF program is estimated to reduce credit growth by about 5.6 percentage points. These results suggest that IMF program participation is associated with restrictive credit growth.
Measuring performance one year after the 1997 Asian financial crises
Using the basic model described above, I analyzed output growth in 1998 for the five East Asian countries that experienced severe currency and balance of payments crises in 1997. The results (Figure 1) show the predicted values and the “forecast errors,” that is, the differences between the predicted values and the actual values. The predictions are based on 1997 values of the explanatory variables, which include domestic factors (change in budget surplus, inflation, and credit growth), external factors (external growth and real exchange rate overvaluation), other factors (previous year’s output growth and country-specific fixed effect), the currency crisis effect, and the IMF-participation effect.
Predicted output growth in 1998 for all five countries is positive. This is in stark contrast with the actual experience—the “forecast errors,” or unpredicted declines, are huge. Why does the model predict robust growth, even with the negative effects associated with currency crises and IMF stabilization programs? Because the negative 1 to 2 percentage point effect exerted by the currency crises and the negative 0.7 percentage point effect of subsequent participation in an IMF program (except for Malaysia) are entirely dominated by positive effects—mainly a history of very strong growth in the region, and a modestly supportive external growth environment.
The average unexpectedunpredicted decline in real GDP for the four countries participating in IMF programs was 12.3 percentage points, with Indonesia having the largest unexpectedunpredicted decline in real GDP (17.6 percentage points) and the Philippines having the smallest (3 percentage points). What is striking is that Malaysia, the one country among the five that did not participate in an IMF program, also had a large unexpectedunpredicted decline in real GDP—13.5 percentage points. Thus, Malaysia’s decision against participating in an IMF program did not appear to save the country from a huge drop in output. Furthermore, this suggests that, whatever the causes of the unexpectedunpredicted fall in real GDP among these Asian countries might have been, they don’t appear to have included participation in IMF programs.
I conducted a similar analysis of credit growth. Predicted credit growth was divided into component parts and the “unexpectedunpredicted” (forecast error) calculated. In every case, credit growth in 1998 is predicted to be quite strong, ranging from 15.5% (Philippines) to 24.4% (Indonesia). Participation in IMF programs lowered predicted credit growth by about 4 percentage points, and the predicted response to the currency crises increased predicted credit growth by about 4 percentage points.
A sharp and unanticipated contraction in credit growth was experienced in every country but Indonesia following the East Asian currency crises. The countries that participated in IMF programs experienced smaller unexpectedunpredicted declines (Korea -8.4%, Philippines -17.5%, Thailand -18.2%) than did Malaysia (-23.2%), and Indonesia experienced a sharp unpredicted jump in credit. The observed decline, as opposed to the negative forecast error, in credit growth was also largest in Malaysia (-2.7%). Indonesia, by contrast, experienced an 18.6% unpredicted rise in credit and an observed rise of 43%. As with the output growth prediction results, Malaysia was seemingly hurt more by the currency crisis than the four countries that participated in IMF programs.
The estimated “cost” of an IMF stabilization program, in terms of forgone output growth, averages about 0.7 percentage point during each year of participation. The decline in output growth is concentrated in the first year of the program. This follows relatively restrictive credit policies adopted by IMF program countries—credit growth is substantially slower following approval of an IMF program. There is some evidence, however, that the decline in GDP growth actually precedes participation in IMF programs and may not be attributable to program participation per se.
Currency crises also reduce output growth over a two-year period by about 2 percentage points. Participation in an IMF-supported program following a balance of payments or currency crisis, however, does not appear to mitigate or exacerbate the output loss. This is despite the fact that countries participating in IMF-programs seem to follow much tighter credit policies when facing a severe balance of payments crisis.
The huge declines in output and credit in the aftermath of the 1997 Asian currency crises were much larger than predicted by historical patterns linking GDP developments to currency crises, IMF program participation, external conditions, and policy developments. The unexpectedunpredicted declines in output and credit were largest in Malaysia, the only country that chose not to participate in an IMF stabilization program at the time. Participation in an IMF-supported program does not appear to make the adverse effect of a balance of payments crisis worse, and Malaysia might have weathered the 1997 crisis better if it had adopted an IMF program.
Professor, U.C. Santa Cruz,
and Visiting Scholar, FRBSF
Glick, Reuven, and Michael Hutchison. 2001. “Banking and Currency Crises: How Common Are Twins?” In Financial Crises in Emerging Markets, eds. R. Glick, R. Moreno, and M. Spiegel, Chapter 2. Cambridge: Cambridge University Press.
Hutchison, M. 2001. “A Cure Worse than the Disease? Currency Crises and the Output Costs of IMF-Supported Stabilization Programs.” Mimeo. U.C. Santa Cruz (March).
Stiglitz, Joseph. 2000. “What I Learned at the World Economic Crisis.” The New Republic (April 17). http://www.tnr.com/041700/stiglitz041700.html (accessed March 2001).
Pacific Basin Notes are published occasionally by the Center for Pacific Basin Studies. Opinions expressed in FRBSF Economic Letter do not necessarily reflect the views of the management of the Federal Reserve Bank of San Francisco or of the Board of Governors of the Federal Reserve System. This publication is edited by Sam Zuckerman and Anita Todd. Permission to reprint must be obtained in writing.
Please send editorial comments and requests for reprint permission to
Attn: Research publications, MS 1140
Federal Reserve Bank of San Francisco
P.O. Box 7702
San Francisco, CA 94120