Why some countries succeeded while others eventually failed
In the mid-1980s, a number of countries suffering from chronic high inflation—typically in the three-digit range—adopted what are now widely known as the “heterodox” stabilization programs. The main innovation was the inclusion of income policies (that is, temporary wage and price controls). The idea was to bring down inflation quickly at the outset, without incurring high costs, in terms of unemployment and output. In all other respects, the programs were very orthodox, centered on reducing budget deficits and restraining the growth of the money supply.
Of the countries that experimented with the heterodox approach, the best-known ones are Argentina (the 1985 Austral plan), Israel (the 1985 stabilization program), and Mexico (the 1987-88 Solidarity Pact, the Pacto). Brazil’s 1986 Cruzado plan is sometimes put in this category, but it sharply differs from the others in that it bypassed from the outset the much-needed adjustment in the fiscal accounts.
Soon after these programs were implemented, the use of income policies—combined with the orthodox approach—looked to be an extremely promising alternative. Its success rate was high, and the costs, in terms of recession or unemployment, were low. With a few more years behind us, however, the verdict has changed somewhat. For some of these countries, notably Israel and Mexico, inflation has remained relatively low, at around 20 percent per year. But the costs—in terms of high interest rates, slow recovery of growth, and loss in external competitiveness—have been much higher than originally anticipated. Moreover, the removal of controls has proved difficult, there has been a loss in exchange rate flexibility, and the process of consolidating low inflation has been time-consuming (over two years in Mexico and even longer in Israel). For others, such as Argentina (and Brazil), income policies have proved counter-productive, because they were not supported by a sustained fiscal effort. Indeed, after the Austral (and Cruzado) plans failed, hyperinflation returned and with an unprecedented vengeance.
This article is based on a longer paper, “When Do Heterodox Stabilization Programs Work?,” World Bank Research Observer, January 1992.
This article reexamines the record of these programs (see “Some Lessons from ‘Heterodox’ Stabilization Programs,” by Mario I. Blejer and Adrienne Cheasty, Finance & Development, September 1988, for an earlier assessment), taking advantage of the time that has now elapsed. Our conclusion is that income policies of the type used in the heterodox programs should only rarely be applied (i.e., under very specific circumstances), and even then, with a great deal of care.
Rationale for income policies
To understand the reasons that lead policymakers to include income policies in a comprehensive disinflation program, it is useful to step back and analyze the causes of high inflation, as well as some of the specific problems that any disinflation program would face. It is clear that the main cause has been the tendency of governments to try to finance their hefty fiscal deficits through simply printing money. But complicating matters has been the fact that once high inflation becomes chronic (e.g., after three years), the economies develop numerous mechanisms, such as asset and wage indexation, that tend to perpetuate existing inflation rates.
Under these circumstances, the announcement and implementation of serious and comprehensive stabilization programs usually leads to modest reductions in inflation. But the costs are typically large, because these economies start from very high inflation rates. Moreover, the pace is usually quite slow, in large part because there is often already a long history of failed stabilization attempts, meaning that any new stabilization program—even if well-designed—is received with skepticism by the public. The problem is that if the public does not believe that the program will succeed in bringing down inflation, inflation will then display a significant rigidity, or inertia. This occurs because firms and workers have the ability to set their prices and wages in advance, based on what they expect inflation to be in the near future. To the extent that the government cannot convince the public that the program will reduce the budget deficit in a sustainable manner and hence that it will work, inflation will only come down slowly, and this, in turn, may result in a sharp decline in output and the abandonment of the program.
The inclusion of income policies is aimed at breaking the inflation inertia that arises from adverse expectations. But price and wage controls can only be justified in situations where (1) the government starts a serious stabilization program and is fully committed to bringing down inflation, (2) the government undertakes the necessary fiscal adjustment, and (3) the public remains skeptical and continues to set prices on the assumption that the program will fail.
The available evidence indicates that income policies have been extremely effective in bringing about rapid, drastic reductions in inflation. But as the chart shows, the initial successes were sustained only in those countries that undertook and persevered with a fiscal adjustment (Israel and Mexico), while they were short-lived in those where the reduction in the budget deficit was temporary (Argentina), or where there was no adjustment to speak of (Brazil).
Exactly how the controls work, however, is rather puzzling. Because in general they did not lead to shortages or marked distortions in relative prices, one could argue that the controls were not binding and thus perhaps of dubious necessity. Yet earlier programs in these countries, which were devoid of any controls, always failed to bring about a sharp, rapid reduction in inflation. In our view, the controls serve largely as a coordinating mechanism—the announcement that controls are in place is enough to reverse short-term inflationary expectations, thus allowing the economy to move permanently to a situation of low inflation. Moreover, if the fiscal adjustment is implemented, as was the case in the successful programs, there is no reason for shortages to occur, as there are no excess demand pressures. The idea that the controls worked because they “repressed” inflation is simply misleading, mainly because these governments did not have the administrative capacity to enforce them.
But the initial drop in inflation must be seen for what it is: the easy part. The real difficulties come later on, when controls have to be removed and stabilization needs to rely entirely on orthodox policies. It is during this second phase that the initial reductions in inflation were quickly reversed in those countries that did not adequately address the fiscal imbalances and were difficult to sustain in those that did address them. Moreover, even in the more successful countries, success has come at a price. It is thus necessary to take a longer time horizon to evaluate the costs of bringing down inflation through a heterodox stabilization program.
Advantages of heterodox programs
On the benefits side, experience shows that high inflation countries stand to gain the most from income policies. It is there that inflationary rigidities typically are the strongest and the fall in inflation the largest.
Low initial costs of bringing down inflation. This is perhaps the most attractive aspect of heterodox programs. In the first phase, thanks to price and wage controls, inflation falls, usually dramatically, with little or even no increases in unemployment. In fact, Argentina and Israel experienced relatively high rates of growth in 1986 (the year immediately after the stabilization programs began), while in Mexico, the Pacto did not elicit any additional recessionary effects.
Potential revenue gains on the fiscal side. It is well known that sudden increases in inflation usually lead to a fall in fiscal revenues (the so-called Olivera-Tanzi effect). But the Olivera-Tanzi effect also works in reverse following a sudden reduction in inflation. This is most promising for countries where inflation is high and the increase in revenues could be significant. Although the size of these gains are difficult to measure, in Israel it was estimated to be around 1.5 percent of GDP, and in Argentina, around 2 percent.
The gains, while potentially significant, should be treated with caution, as they are not a “true” signal of the fiscal effort. A program that bases most of the fiscal improvement simply on greater revenues from this effect is likely to fail, since the lower deficit will remain in place only if the government succeeds in keeping inflation low. Any shock that destabilizes inflation will increase the deficit, making the inflationary process self-sustaining. Further, to the extent that public sector enterprise prices are not adjusted during the period of controls, there will also be a deterioration in their finances. This means that the gains from the reversed Olivera-Tanzi effect should constitute a relatively small part of the fiscal adjustment. In Israel, for example, it is estimated that this represented a quarter of the improvement in the fiscal accounts, and in Mexico, it was not very important, but for Argentina, it played a major role, especially for the public sector enterprises.
How inflation fared under the heterodox programs
Better ability to judge the fiscal adjustment. This occurs for two reasons. First, it gives the government some breathing time to demonstrate its commitment to the fiscal adjustment. As a result, even if the public is initially skeptical, inflation will not be immediately rekindled because the freeze is in place. However, the government should move quickly to close any fiscal gap, as this period is usually short. Second, low inflation makes it easier to calculate the size of the budget deficit, thereby rendering the stabilization process more transparent. Under high inflation, this calculation is difficult, as there is more room for using accounting tricks to mask the true state of the fiscal accounts. Indeed, Israeli officials now credit the reduction in inflation with helping them keep track of the evolution of the fiscal accounts.
More momentum to stabilization. Not surprisingly, the public will be more willing to support a program that is showing positive results early on than one that just promises lower inflation some time in the future. This means that the initial fall in inflation might make it easier for governments to introduce additional measures to strengthen the program, such as cuts in government expenditures, or increases in taxes. But it could also work in the opposite direction, as occurred in Argentina under the Austral plan, when, despite widespread support, the Government did not implement additional measures to make the initial fiscal adjustment more robust.
Disadvantages of income policies
On the other side of the ledger, the drawbacks are also significant, raising the risk of not persisting with the program after the early successes.
High later costs of bringing down inflation. Many of the costs that were avoided in stage one through the imposition of controls emerge later on, when controls are relaxed. At this point, the exchange rate takes over as the main nominal anchor trying to hold down inflation, which is now market determined. The initial low costs then turn into high ones, as the real exchange rate appreciates, leading to a recession, a loss of competitiveness (which hurts the export sector), and high interest rates (which are necessary to maintain attractive rates of return on domestic assets to avoid capital flight).
But the difficulties do not stop there. One of the trickiest phases then sets in, as the overvaluation that took place during the period when controls were relaxed needs to be corrected without losing control of the main nominal anchor. While a large devaluation could restore competitiveness, the risk is a loss of confidence about the government’s willingness to sustain the stabilization program. A large devaluation could be read as an indication that the government is not prepared to pay the costs resulting from overvaluation and instead is ready to accept a rekindling of inflation. This very problem has continued to bedevil both Israel and Mexico, with neither of them able to overcome the real appreciation of their currencies. Israel has been trying to avoid further appreciations through unannounced step devaluations (mostly in the 10 percent range), while Mexico has opted for a preannounced crawling peg (of the type used in the Southern Cone of Latin America in the late 1970s).
Disregard for fiscal adjustment. The danger here is that the favorable initial results on inflation and unemployment create a tendency for complacency on the fiscal side, potentially undermining the success of the program. “Populist” programs are an extreme case, as they try to disinflate without making the needed fiscal adjustment; in fact, in some cases, fiscal policy even turns expansionary. This is what happened under the Cruzado plan, when the combination of controls and generalized excess demands rapidly led to repressed inflation, widespread shortages, and the emergence of black markets. In the end, of course, there was an explosion in prices, which had to be stopped through a new round of price and wage controls.
One way of helping to clearly signal the government’s commitment on the fiscal side is to “overadjust”—that is, tighten up in the short run more than would have been required in the long term, perhaps even to the point of running a surplus. This is useful for two reasons: (1) to show unambiguously that this is a serious heterodox program, not a populist one; and (2) to help reduce aggregate demand, which is critical if shortages are to be avoided during the period of controls.
In practice, the differences between heterodox and populist programs become apparent shortly after the launching of the programs. Key distinguishing features are the attitude toward real wages, the importance attached to fiscal (and sometimes monetary) policy, and the fact that generalized shortages do not emerge in heterodox programs. However, it has been much more difficult to distinguish between those programs that eventually succeeded (the “persistent” programs) and those that started as heterodox programs but were later abandoned (the “not-persistent” programs). In fact, the evolution of macroeconomic variables during the initial stage of the Austral plan and the Israeli program was remarkably similar, while shortages did not arise in either of them.
Distortionary impact of price controls. Another criticism is that controls lead to a misallocation of resources. While this is valid for countries that use controls on a long-term basis, it is not necessarily the case for heterodox programs, which call for only temporary controls. In addition, in most heterodox programs, controls have not been used in a rigorous manner. Argentina, for example, allowed firms to raise prices whenever it was clear that their prices were out of line.
Difficulties in removing controls. A more-often raised—and valid—criticism is that once implemented, controls are difficult to remove. While the government might be aware that controls need to be removed, it is also worried that the removal might rekindle a new inflationary spiral. The key is to remove controls gradually and selectively, as was done in Israel and Mexico, perhaps starting with sectors where it is clear that there are no pressures for price increases.
Difficulties in assessing progress. It is always hard to determine at the outset whether a stabilization program will succeed. This is true for both orthodox and heterodox programs. In Bolivia, for example, the August 1985 orthodox program, which brought hyperinflation to a halt, had been preceded by other similar stabilization efforts that had failed. And even this successful program confronted a serious reversal, albeit temporary, toward the end of the year when hyperinflation reemerged.
This difficulty in recognizing successes is compounded in heterodox programs by the fact that low inflation is maintained initially through controls. In the absence of market-determined prices, it is harder to evaluate whether the program is working because the fundamentals are in order and low inflation is there to stay, or because controls are “repressing” inflation. The answer to this question must wait until controls are removed and inflation finds its new equilibrium. In the meantime, the credibility problem is bigger in a heterodox program.
Heterodox for whom?
Given the advantages and disadvantages of heterodox programs, their use can only be recommended in a limited number of cases. On the not recommended list we have three situations. The first involves countries that have not had chronic problems with inflation but suddenly experience an inflationary outburst. In this instance, income policies are not a good idea, because inflation is perceived (correctly) as a temporary phenomenon. Examples of this are Costa Rica and the Philippines in the early 1980s, as both countries traditionally had low inflation, but suffered a short inflationary spiral following the devaluations undertaken in response to the debt crisis. The speed with which inflation was brought down using orthodox measures illustrates that this was indeed the right approach.
In the second situation—that of hyperinflation—the orthodox approach is also preferable for several reasons: (1) inflation is less rigid; (2) the cause of inflation is clear—a large budget deficit financed by money creation; and (3) hyperinflation is inherently unsustainable, meaning that the public will demand price stability.
Third, the orthodox approach is advisable in low and moderate inflation economies. In this situation, controls are not likely to generate a large reduction in inflation, because contracts are typically long and staggered, as opposed to high inflation economies, where they are short and highly synchronized. A freeze would leave great disparities in relative wages and prices, penalizing those with the oldest contracts. This problem is difficult to handle in practice, requiring a great deal of information.
Thus, the only situation in which the heterodox approach is potentially useful appears to be when countries are plagued with chronic high inflation, as was the case with those that opted for this policy in the 1980s. But there is one important caveat: Entering a heterodox program that is later abandoned holds potential dangers. Indeed, the costs of this approach could be larger than simply postponing stabilization. For Brazil and Argentina, the repeated use of controls was a major destabilizing force, responsible for the unprecedented hyperinflations of the late 1980s.
New from the International Monetary Fund
How To Measure the Fiscal Deficit: Analytical and Methodological Issues
edited by Mario Blejer and Adrienne Cheasty
Fiscal policy seeks to equilibrate the public sector’s demand for investment and a sustainable balance of payments. Correct measurement of the public sector’s net use of resources is therefore an important prerequisite for managing the macroeconomy. This volume is organized around four issues: the adequacy of summary measures of the fiscal deficit; conventional and adjusted deficits; coverage (size) of the public sector; and the public sector’s intertemporal constraint.
Available in English. 350 pp.
ISBN 1-55775-192-7. (paper) 1992.
To order, please write or call:
International Monetary Fund
700 19th Street, N.W.
Washington, D.C. 20431 U.S.A.
Telephone (202) 623-7430
Cable Address: INTERFUND
American Express, MasterCard and VISA credit cards accepted.
Miguel A. Kiguel and Nissan Liviatan