Introduction and Summary

Carlo Cottarelli, and Gyorgy Szapary
Published Date:
July 1998
  • ShareShare
Show Summary Details
Carlo Cottarelli and György Szapáry 

The seminar on moderate inflation in transition economies held in Budapest on June 3, 1997 was motivated by two hypotheses: that there is something unique about moderate inflation and that there is something unique about inflation in transition economies. The papers presented at the seminar, and published in this volume together with some contributions to the open floor discussion and two background papers covering related issues, bore out these hypotheses.

In fighting moderate inflation (say, below 30 percent but well above the 0–2 percent range often regarded as constituting price stability in Western Europe), policymakers must tackle a number of problems that do not emerge at higher inflation levels and that, paradoxically, could make the disinflation process harder to complete. When inflation is high, it is usually easy to obtain political consensus for the policy tightening needed to bring inflation down because the costly distortions caused by inflation are readily apparent. However, for lower inflation levels, it may be more difficult to convince the public that the long-term benefits arising from price stability exceed the short-term costs that may be associated with disinflation (Medgyessy).1

An additional problem is the relative difficulty of changing the structure of private sector contracts and transactions as required by disinflation. As Burton and Fischer note, the shortening of the length of contracts and the growing indexation to a foreign currency that mark high inflation phases make disinflation from high levels in some ways easier to achieve than from moderate levels. Finally, new macroeconomic obstacles to disinflation may emerge when inflation is moderate. In particular, Gomulka underscores that short-term capital inflows, and the problems related to these inflows, are more pronounced in countries that have already achieved a relatively advanced stage of transition and macroeconomic stabilization. Overall, it is not surprising that inflation inertia seems to become empirically more relevant when inflation is moderate (Coorey, Mecagni, and Offerdal).

Disinflation in transition economies also presents policymakers with unique challenges. One key problem is that the transition from centralized to market economies is usually characterized by sharp changes in relative prices of goods, services, inputs, and the real exchange rate. In the presence of nominal rigidities (for example, downward rigidity of nominal wages and prices), these changes can occur only through inflation. Real adjustment may initially be quite rapid, but the process may be drawn out in some countries, particularly in the presence of indexation, thus sustaining inflation at moderate levels for several years. (Burton and Fischer; Blanchard; Griffiths and Pujol; Coorey, Mecagni, and Offerdal; and Surányi and Vincze all deal with this issue more closely.)

Against this background, it is not surprising that many transition economies in Central and Eastern Europe, while reducing inflation rapidly from high levels (often exceeding 100 percent a year), have so far failed to achieve price stability. This book addresses the issue of disinflation in these countries.

Three broad policy issues were discussed during the seminar:

  • Is moderate inflation a problem or do economies adjust to it so that their growth performance and, more generally, social welfare are not much affected?

  • What policies are needed to defeat moderate inflation? More important, how can the short-run costs of disinflation be minimized? Should any factors specific to transition economies be considered in formulating a disinflation strategy?

  • What is the appropriate speed of disinflation, and what should the inflation target be during the transition and in the long run? What are the right time and sequencing of disinflation with respect to other economic policy and structural reform goals?

We now provide a brief synopsis of the views expressed by the contributors to this book.

Is Moderate Inflation a Problem?

There is broad consensus, in both academic and policymaking circles, that high inflation hampers growth and that low inflation (say, 0–2 percent) is fairly innocuous. What is more controversial is the threshold inflation rate beyond which growth starts being hampered.

Burton and Fischer briefly review the empirical evidence on this subject and find no evidence to support the view that maintaining moderate (low double-digit) rates of inflation is good for growth, but find rather that double-digit rates of inflation are bad for growth. This view was broadly endorsed (forcefully by Icard, Škreb, and Surányi and Vincze), with virtually no dissenting voice: inflation, including at moderate levels, distorts resource allocation and income distribution (particularly through the interaction with the tax system) and reduces capital accumulation and eventually productivity growth.

How Can Moderate Inflation Be Reduced and the Short-Run Costs of Disinflation Minimized?

The answer to the first part of the question is easy: what is needed is a tightening of economic policies to reduce the growth of nominal aggregate demand. To use a term frequently heard during the seminar, it is necessary to get the “fundamentals” right (Burton and Fischer; Icard; Škreb). To do so will typically require a deceleration in the growth rate of money and often a fiscal tightening.2 A fiscal tightening is important not only to cool demand pressure, but also, in many cases, to make the anti-inflationary effort sustainable and hence credible.

This brings us to the most difficult question of how to minimize the short-run costs of disinflation. Unless prices and wages adjust immediately and commensurately to the deceleration in nominal aggregate demand, output growth is also going to be affected and unemployment will rise above the level that would have prevailed in the absence of a policy tightening. Eventually, the slack in the economy will cool the growth rate of prices and wages, but in the process some output loss will be incurred. How can this output loss be minimized?

It is by now recognized that the key factor in a successful (that is, low-cost) disinflation is the credibility of the process. If private agents believe that inflation will come down, wages and prices will adjust rapidly to the new, lower inflation rate, and output growth will not be much affected. Most participants in the seminar stressed that a strong fiscal position is essential for credibility: a large deficit requires sizable seigniorage revenues, that is, large monetary creation. The relation between inflation and public deficit may be looser in the presence of a well-developed government securities market (Icard). However, if the private sector perceives that the government’s intertemporal budget constraint is unlikely to be met, a high deficit will raise the probability of its eventual monetization even in the presence of developed securities markets. This will in turn raise inflationary expectations. Thus, a requirement for a credible disinflation is that the solvency of government finances does not rely on large seigniorage revenues (Burton and Fischer).3

It is worth noting that a strong fiscal position does not necessarily imply that the fiscal stance should be tightened (that is, that a negative fiscal stimulus is needed). In some countries, the solvency of fiscal accounts may already have been reestablished, and yet the inflation process may still be sustained by self-fulfilling expectations (Blanchard regards Hungary as a case in point). What is key, however, is that the strength of the fiscal balances be perceived as long-lasting. That is, if a fiscal tightening is needed, it should be based on structural measures.4 Also, even if a tighter fiscal stance is not needed, it may be inevitable to introduce structural fiscal measures (for example, in the pension system) to strengthen the evolution of public finances in the medium term (Burton and Fischer). The reduction of a quasi-fiscal deficit (including in public enterprises and the banking sector) is equally important (Griffiths and Pujol).5 Of course, unwarranted fiscal relaxations should be avoided (Begg).

In addition to a sustainable fiscal position, the economic literature has identified several other factors that may increase the credibility of the disinflation process.6

First, the announcement of an “anchor” for expectations—that is, a visible target through which the behavior of the monetary authorities can be clearly monitored and that is sufficiently related to inflation—will permit the public to clearly identify deviations from that target. The two main candidates for this anchoring role are the exchange rate (either by pegging the exchange rate or through a crawling peg regime) and the money supply. During the seminar, the debate on the choice between these two anchors was intense. Burton and Fischer, while recognizing that different approaches to stabilization are possible, find that “most of the successful stabilizers have used an exchange rate anchor.” Surányi and Vincze also take the view that the exchange rate is key, particularly in a small open economy like Hungary. Gomulka instead argues that, contrary to a widely held view, money velocity is stable enough in many European countries to make the money supply a good anchor.

Cottarelli and Kornai caution against using a rigid exchange rate anchor to disinflate the economy in countries with a relatively weak external position (for example, an initially high external debt or a history of persistent balance of payment difficulties) because, for those countries, very tight exchange rate policies would not be fully credible. Other contributors, while not dismissing the exchange rate anchor, downplay its importance by stressing the role of fundamentals (Škreb), or pointing at the drawbacks of excessively rigid exchange rate regimes (Begg, Gomulka), or both. There are two main drawbacks. The first is that such regimes give rise to an unsustainable appreciation of the real exchange rate if inflation does not decelerate fast enough. The second is that they encourage capital inflows. If these inflows are sterilized, the fiscal costs could undermine the credibility of the fiscal adjustment. If they are not sterilized, they will raise the money supply and inflation and, eventually, lead to a real appreciation. These risks should not be overemphasized,7 but cannot be easily dismissed either, and they are often mentioned as an additional reason for fiscal discipline.8

The second factor is a strong underlying growth performance. If the growth of potential output is strong, the incentive to boost output growth temporarily through a monetary expansion is lower. Thus, growth-oriented structural reforms (privatizing, opening up the economy, downsizing the government, reforming taxes, and increasing competition) are important for maintaining price stability (Škreb and Burton and Fischer).

Third, institutional devices may make it harder to deviate from price stability. These include setting up an independent central bank or introducing legal constraints on the deficit (Cukierman and Cottarelli).9

Aside from the factors that directly affect the credibility of the disinflation process, the cost of disinflation can also be minimized through policies designed to reduce the stickiness of goods and factor prices and, more specifically, to bring about a fast deceleration of wages and interest rates.10 These include incomes policies and the deindexation of the economy (Burton and Fischer, Blanchard, and Griffiths and Pujol) or structural reforms of the labor market that raise the sensitivity of real wages to unemployment.11 As for interest rates, Surányi and Vincze note that if the financial system is not flexible and competitive enough, real interest rates, particularly on bank loans and deposits, may increase, with a restrictive effect on aggregate demand. Cottarelli agrees on this point, but shows that in some countries, such as Hungary, the speed of adjustment of bank lending rates has increased significantly in the past few years as a result of the strengthening of bank competition and the privatization of the banking system. These developments should lower the cost of disinflation.

Should Disinflation Be Gradual or Rapid?

Blanchard addresses this thorny issue by focusing on the costs of disinflation. Because of structural differences in their economies, the cost of disinflation differs across countries. Assuming that all countries are equally averse to large increases in unemployment, those that face higher output losses for each unit of inflation should prefer a more gradual approach to disinflation. Of course, the more a country is averse to jumps in unemployment, the more likely it is to adopt a gradual approach to disinflation. However, Blanchard also notes, reporting evidence in Ball (1994), that the cost of disinflation is not independent of the adopted speed of disinflation: fast disinflation seems to be associated with a lower sacrifice ratio, a powerful argument in favor of the “cold turkey” approach. A similar conclusion can be drawn from the empirical evidence on successful disinflation discussed by Burton and Fischer, who argue that “there is little evident output cost of disinflation in the stabilizations from moderate inflation.” Cukierman also favors rapid disinflation, albeit for different reasons: he points to the “disinflation fatigue” that may arise in countries facing a slow process of disinflation; he also notes that the cold turkey approach currently has good prospects in transition economies given the consensus on the need for price stability that has now developed in the rest of Europe (partly in relation to Economic and Monetary Union).

The case for fast disinflation is, however, controversial. Kornai stresses that, in transition economies, the magnitude of the output loss related to disinflation is very uncertain, partly because of the difficulty, even at the conceptual level, of defining potential output. Given this uncertainty and the sizable output loss that accompanied the transition, policymakers may be reluctant to shift the economy into recession in order to achieve a faster disinflation rate. Indeed, many participants pointed out the difficulty of carrying forward, aggressively and simultaneously, both the real adjustments that are required in transition economies and the nominal adjustment process that characterizes disinflation. Blanchard’s paper, addressing the problem of the appropriate sequencing of real and nominal adjustment, provides a useful springboard for the discussion of this issue.

In economies in transition, is it better to embark on disinflation only after “real” adjustment (including changes in key relative prices) is far enough along? Blanchard points out the difficulty for policymakers of implementing disinflation while trying to eliminate real disequilibria (in real wages, in administered prices). The argument runs as follows. The cost of disinflation can be significantly reduced if, in the context of an incomes policy agreement, workers are willing to accept a sizable slowdown in wage increases. However, if prices end up rising faster than initially expected (as needed to eliminate the existing disequilibria in real wages or administered prices), confidence in the disinflation process may be shattered. Blanchard concludes that, “other things being equal, it is better to make disinflation a coordination problem—where the main issue is trust—than a combined problem of coordination and income redistribution. This approach may mean waiting to disinflate until, for example, subsidies to consumers have been largely removed or instead waiting for the end of disinflation to remove the remaining subsidies.”

It is debatable whether the transition economies experiencing moderate inflation have reached the point at which the interaction between structural adjustment and disinflation, and the related issue of timing, can be disregarded and whether disinflation has become purely a coordination problem (as Blanchard, for example, believes now to be the case in Hungary). But, aside from this empirical issue, some participants noted that the choice presented by Blanchard—eliminate the real disequilibria first and disinflate later, or vice versa—may not be available to policymakers. Disequilibria in real wages or administered prices may lie behind high public deficits and the inflation process (Icard). Thus, eliminating these disequilibria is a necessary step to achieve disinflation.

Surányi and Vincze, however, note that the issue of sequencing has implications for the duration of disinflation. They agree that structural reform (including trade, price, and wage liberalization; financial sector reform; privatization; and fiscal reform) should come first, but stress that structural reform is politically difficult. Structural reform policies impose severe strains on the economy and seem to require some accommodative policy to secure the political and social backing for the momentum of reforms. This means that disinflation will be drawn out a little longer. Szapáry also argues that disinflation may have to be delayed until real disequilibria are removed and that, to avoid a real appreciation, countries should not anchor the nominal exchange rate until they have achieved the required adjustments, particularly in real wages. With reference to the Hungarian experience during 1995–96, he notes that surprise inflation was instrumental in bringing about the sharp decline in real wages that was needed to redress the serious imbalance in the external accounts and that would have been difficult to achieve through a deceleration in nominal wage growth.

The main problem that may arise if disinflation is postponed is that of entrenched inflationary expectations. People simply get used to higher inflation, which makes disinflation harder. Škreb, given Croatia’s successful experience, views rapid disinflation favorably. However, he notes that, because of biases in statistics existing in industrial countries but exacerbated by the transition process, an inflation rate of 4–5 percent should be regarded as equivalent to price stability in transition economies just as an inflation rate of 0–2 percent is regarded as corresponding to price stability in industrial countries.

A final point on the issue of the timing of disinflation in transition economies needs to be mentioned. Together with others, Burton and Fischer and Cukierman point to the importance of capitalizing on favorable opportunities. In the same way that inflation can be more easily triggered by an unfavorable shock, disinflation becomes easier and less costly if it occurs at a time of a favorable supply shock, such as a decline in the prices of imported products, a fall in agricultural prices, or a sharp increase in productivity. This point is developed further by Deppler, who stresses that the transition process is one massive, ongoing supply shock and that fairly strong growth of productivity should be expected. Strong productivity enables countries to achieve a fairly rapid pace of disinflation at a cost that is significantly lower than is suggested by the experience of Western European countries.


    Alesina, Alberto, and RobertoPerotti,1997, “Fiscal Adjustments in OECD Countries: Composition and Macroeconomic Effects,”Staff Papers, International Monetary Fund, Vol. 44 (June), pp. 21048.

    Ball, Laurence,1994, “What Determines the Sacrifice Ratio?” in Monetary Policy, ed. byN.G.Mankiw (Chicago: University of Chicago Press).

Note: The authors would like to thank Mark Allen for his comments.

In this introduction, all references to authors without further information refer to papers included in this book.

Given the possible instability of money demand in some countries in transition, it may be hard to predict by how much the growth rate of money should decelerate. Thus, it may not be appropriate to anchor the disinflation process to some preannounced money target. Nevertheless, it is clear that the money supply in a scenario of disinflation will probably have to be lower than in a scenario in which inflation does not come down (assuming that money velocity does not decline too much as a result of disinflation).

Assuming that the real interest rate exceeds the growth rate of the economy, a necessary and sufficient condition for the solvency of public finances (that is, for the government” s intertemporal budget constraint to be met) is that the ratio between government debt and GDP is not increasing. Burton and Fischer argue that a reduction of the government debt burden, if it is high, is also critical to disinflation because stabilizing public debt at a high level increases the likelihood that public finances enter an explosive path as a result of some negative shocks, for example, in output. Moreover, if public debt is high, the primary surplus will have to be maintained correspondingly high (for example, by keeping tax rates higher than elsewhere). This may not be credible.

Alesina and Perotti (1997) find that long-lasting fiscal adjustments are typically characterized by expenditure cuts, particularly in transfers to the population. This is consistent with Icard’s call for tighter expenditure policies. Surányi and Vincze also note that closing the deficit gap by relying on indirect taxes (which more directly affect prices) may undermine the credibility of the disinflation process.

Griffiths and Pujol also argue that the reform of the banking system is important to impose a hard budget constraint on public enterprises. They regard delays in this reform as one important factor behind the persistence of inflation in Poland. Begg also stresses the need to strengthen the banking system to reduce unwelcome credit expansion.

Most of these factors focus on devices to reduce the risk of a “surprise” inflation, that is, of an inflation rate that exceeds the objective announced by the government.

Capital inflows may not require sterilization if they are matched by a remonetization of the economy, a distinct possibility in a period of stabilization.

Burton and Fischer recognize that, reflecting the Balassa-Samuelson effect, the equilibrium real exchange rate may be appreciating; trying to peg the nominal exchange rate would, in this case, lead to capital inflows and sustain inflation (see also Coorey, Mecagni, and Offerdal for a theoretical and empirical discussion of the factors behind real appreciation in transition economies). In this case, it may be preferable to adopt a more flexible exchange regime. More generally, the pressure from capital inflows can be alleviated through a combination of fiscal tightening, sterilized intervention, a degree of exchange rate flexibility (for example, a wide exchange rate band around the crawling peg), some monetary accommodation (if money demand is expected to increase), and, as an interim solution, some price-based controls on capital inflows. Begg also favors an eclectic approach in addressing the problem of capital inflows. Gomulka stresses the importance of fiscal adjustment, noting that the Maastricht criteria are too loose in economies that face a strong surge in domestic private credit demand and capital inflows.

Given the interest that the debate on inflation targeting has raised in industrial countries in the past few years, it is somewhat surprising that this subject was not developed during the seminar. One possible reason is that an inflation targeting framework requires a high degree of transparency in information and trust in economic policymaking, which may not yet have been achieved in countries in transition. Moreover, in a transition economy, it may be difficult to judge how much of the relative price and wage adjustment has already been completed and how much remains to be done. This increases the difficulty of making reliable projections on inflation developments.

Insofar as these policies reduce the output loss related to disinflation, they also make the prospect of disinflation more likely. In this sense, they also indirectly increase the credibility of disinflation.

Blanchard notes that if nominal wages and prices were perfectly flexible, the sensitivity of real wages to unemployment would not affect the cost of disinflation. However, given the rigidities in nominal wages, the cost of disinflation is lower if a sharp adjustment in nominal and real wages can be obtained with a relatively small increase in unemployment.

    Other Resources Citing This Publication