Nominal Anchors in the CIS
Author: Peter M. Keller and Mr. Thomas J Richardson

Contributor Notes

pkeller@imf.org, trichardson@imf.org
Publication Date:
01 Sep 2003
ISBN:
9781451859058
Language:
English
Keywords:
WP; CIS country; exchange market intervention; CIS policymaker; money demand; balance of payments; managed float; underdeveloped exchange market; foreign currency
Download PDF (625 KB)

Monetary policy has become increasingly important in the countries of the Commonwealth of Independent States (CIS) as fiscal adjustment and structural reforms have taken root. Inflation has been brought down to relatively low levels in almost all of these countries, raising the question of what should be the appropriate nominal anchor at this stage. Formally, almost all CIS countries have floating exchange rate regimes, yet in practice they manage their exchange rates very heavily, perhaps because of high levels of dollarization (i.e., they suffer from "fear of floating"). This paper explores the issues underlying the choice of a nominal anchor in CIS countries and seeks to assess whether the present mixed regime will prove durable.

Abstract

Monetary policy has become increasingly important in the countries of the Commonwealth of Independent States (CIS) as fiscal adjustment and structural reforms have taken root. Inflation has been brought down to relatively low levels in almost all of these countries, raising the question of what should be the appropriate nominal anchor at this stage. Formally, almost all CIS countries have floating exchange rate regimes, yet in practice they manage their exchange rates very heavily, perhaps because of high levels of dollarization (i.e., they suffer from "fear of floating"). This paper explores the issues underlying the choice of a nominal anchor in CIS countries and seeks to assess whether the present mixed regime will prove durable.

I. Introduction and Motivation

Much has changed over the course of approximately a decade of Commonwealth of Independent States (CIS) central banks gaining experience with managing independent national currencies. Monetary policy is beginning to play an increasingly important role. The main reason is the reduction of fiscal dominance as budget deficits have, by and large, been brought under control or even turned into surpluses (e.g., Kazakhstan, Russia). Moreover, most central banks are generally much less engaged than in earlier years with quasi-fiscal operations, such as providing directed credits at below-market interest rates to struggling public sector enterprises.2 In most countries, politicians have now accepted the link between the money supply and inflation. After the disruptions created by the 1998 Russian crisis, confidence in domestic banking systems appears to be growing in most CIS countries. Financial deepening is proceeding—in several countries at a very fast pace—and bank deposits have become freely convertible internally (and often also externally), with some notable exceptions. Although ratios of the monetary and credit aggregates to GDP are still well below those observed in the Baltic or Eastern European countries, they are rising rapidly in many CIS countries (Figure 1). These elements suggest that money and monetary policy are finally beginning to matter in the CIS context and their importance will certainly grow over the medium term.

Figure 1.
Figure 1.

Broad Money/GDP 1/

(End of 2002)

Citation: IMF Working Papers 2003, 179; 10.5089/9781451859058.001.A001

Source: IMF, European II Department centralized database.1/ ARM–Armenia; AZE–Azerbaijan; BLR–Belarus; GEO–Georgia; KAZ–Kazakhstan; KGZ–Kyrgyz Republic; MDA–Moldova; RUS–Russia; TJK–Tajikistan; TKM–Turkmenistan; UKR–Ukraine; UZB–Uzbekistan.

Section II takes stock of the monetary policies pursued in recent years in individual CIS countries, including the policy objectives targeted. With very few exceptions, inflation has been brought down from near hyperinflation to single digit annual rates. Exchange rates—while formally floating—are, in effect, heavily managed, and in some cases the outcome comes close to a fixed peg. Exchange market interventions are typically one-sided and aimed at preventing or at least slowing the real appreciation of the currency. While a variety of monetary instruments has emerged, in most instances the main influence on the money supply still comes from exchange market interventions, with the extent of sterilization constituting the major policy decision of central banks. Interest rate policies are only now beginning to play an important role in some CIS countries. Dollarization remains high, and this has a bearing on the degrees of freedom of monetary policy—although it has not undermined the very substantial progress made in reducing inflation. Money demand remains difficult to predict, and there are spurts of unforeseen financial deepening. Frequent corrections to policy targets have, however, enabled policymakers (including in the context of IMF-supported programs) to cope with this problem. In most CIS countries, monetary policy appears to be aimed mainly at achieving an inflation objective rather than a balance of payments outcome. This emphasis, despite the traditional focus of financial programs on external adjustment, probably stemmed from the following: (i) balance of payments objectives did not figure prominently among the concerns of policymakers, at least, in part, because Russia had taken responsibility for the Soviet Union’s external debt; (ii) with exchange rates more fixed than floating, fiscal policy is the instrument of choice for balance of payments correction while monetary policy is more effective for achieving the inflation objective (Mundell’s assignment problem (Mundell, 1962)); and (iii) given the low level of monetization, monetary policy would have been relatively ineffective in achieving a balance of payments objective.

Figure 2.
Figure 2.

Foreign Currency Deposits/Total Deposits 1/

(End of 2002)

Citation: IMF Working Papers 2003, 179; 10.5089/9781451859058.001.A001

Source: IMF, European II Department centralized database.1/ ARM–Armenia; AZE–Azerbaijan; BLR–Belarus; GEO–Georgia; KAZ–Kazakhstan; KGZ–Kyrgyz Republic; MDA–Moldova; RUS–Russia; TJK–Tajikistan; TKM–Turkmenistan; UKR–Ukraine; UZB–Uzbekistan.

Section III discusses the potential risks entailed in the current approach and explores the various options for anchors to guide monetary policy—namely money supply rules, exchange rate pegs, and direct inflation targeting—in a formal framework based on an empirical model. It also discusses the scope for the current approach to evolve. Section IV covers related issues—namely, what the inflation target should be; whether there are targets other than inflation; and what, if anything, should or can be done to reduce dollarization so as to create greater scope for exchange rate flexibility. Section V concludes, and Appendix I offers a cross-country progress report on mechanisms of monetary policy formation in the CIS.

II. Recent Developments in Monetary Policy

A. Background

Inflation has come down in virtually all of the CIS countries, and has reached very low levels in a few. Macroeconomic policies, including in particular deep fiscal adjustment, have shown themselves capable of achieving low to moderate inflation on a durable basis. Over the past four years, sustained disinflation has been achieved in the Caucasus, Kazakhstan, the Kyrgyz Republic, Moldova, and Ukraine, and, to a lesser extent, in Russia and Tajikistan. Relatively high levels of inflation have persisted in Uzbekistan and particularly in Belarus.

Most CIS countries have free or managed floating exchange rate regimes—at least on paper.3 The official rate in Turkmenistan seems to be pegged and differs by 400 percent from the curb market rate, while Belarus has a crawling band. Four CIS countries—Armenia, Georgia, Moldova, and Tajikistan—report freely floating exchange rate regimes, while the remainder report managed floats.

However, many of the CIS countries seem to manage the exchange rate heavily, and thus it would be difficult to consider them floaters in practice. In part this observation stems from interventions to rebuild reserves following the Russia crisis. Ukraine has seen very little nominal exchange rate movement vis-à-vis the U.S. dollar since early 2000, and at least through the first half of 2003, Kazakhstan’s tenge tracked the Russian ruble closely. Both Georgia and the Kyrgyz Republic have also witnessed relatively modest exchange rate volatility during the past three years. In most cases, the authorities point to the tight and rapid link between inflation and the exchange rate as a key reason why they prefer relatively stable exchange rates, although in a number of countries concern about the impact of an appreciation on economic performance plays a role as well.

Figure 3.
Figure 3.

CIS: CPI Inflation

(12-month percent change)

Citation: IMF Working Papers 2003, 179; 10.5089/9781451859058.001.A001

Figure 4.
Figure 4.

CIS: CPI Inflation

Citation: IMF Working Papers 2003, 179; 10.5089/9781451859058.001.A001

B. Fear of Floating

One other explanation for the divergence between de jure and de facto exchange rate regimes may stem from “fear of floating.”4 This type of implicit resistance to floating is pervasive in many parts of the world, including in the CIS, and a number of possible explanations have been offered for it, including high levels of liability dollarization, potential output costs if exchange rates are highly variable, and constraints on borrowing in external markets. Table 2, which replicates a key result of Calvo and Reinhart, shows the probability that monthly changes in nominal exchange rates and gross reserves are within a ±2.5 percent band.5 Over the period from January 1999 to end-2002, the volatility of reserves was about twice that of the nominal exchange rate. This result is even stronger if one excludes countries where reforms are at a relatively early stage—Belarus, Tajikistan, Turkmenistan, and Uzbekistan. If 1999—the post-Russia crisis year—were excluded, the volatility of the exchange rate would in most cases be still less.

Table 1.

CIS Countries: De Jure Exchange Rate Regimes

article image
Source: IMF, Annual Report on Exchange Arrangements and Exchange Restrictions, 2003.

De facto peg; de jure regime is unknown.

Table 2.

Volatility of Selected indicators, January 1999–December 2002

article image
Source: IMF, European II Department centralized database and International Financial Statistics.

Three month T-bill rale.

De facto fixed peg; data on reserves extend to the end of 2000; interest rate data from January 2000 to January 2001.

Fear of floating in the CIS is at least as strong as in more widely studied emerging market countries. Table 2 also shows a sample of Central European transition and Latin American countries, equally divided into managed and free floaters. Exchange rate stability in these countries still exceeds that of reserves, but by noticeably less than in the CIS.

Table 2 also shows that most CIS countries do not seem to rely on interest rate policy in the conduct of monetary policy and to smooth movements in the exchange rate. The volatility of nominal interest rates in the eight more advanced countries is only about one quarter of that in the relatively slower reformers, where inflation and exchange rate volatility remain high. (Interest rate volatility is still lower in the non-CIS countries in Table 2.)

The foregoing points to the existence of a number of unresolved problems affecting monetary policy formation in the CIS. In the early years of transition, monetary and fiscal policies were often directed toward support of failing state enterprises. In marked contrast, price stability has now become a core goal of the monetary authorities in almost all of them. At the same time, in a few countries price stability is in conflict with, even subordinated to, a fear of the loss of competitiveness associated with real appreciation. Although the financial sectors in CIS countries are growing, they remain small in absolute terms, and are heavily dollarized. Moreover, the nonbank financial sectors are just emerging. Central banks are usually independent, at least on paper, but many of them are quite politicized, and directed credits remain a problem in a few countries. The relationship between monetary aggregates, inflation, and output is not subject to reliable econometric specification (despite the gradual abatement of barter transactions).6 These factors heavily influence the choice of a nominal anchor for the conduct of monetary policy in the post transition period.

III. The Choice of a Nominal Anchor

A. Characteristics of the Current Approach

Although the precise relationships between monetary aggregates and inflation outcomes remain difficult to predict, the very short time lags in the CIS between changes in policy instruments and in inflation outcomes make error correction relatively easy. While in the larger industrial countries such lags have been estimated at 11/2 to 21/2 years, lags in CIS countries are 3–6 months or less for a nearly complete pass-through of a monetary impulse. In the countries with the least developed financial markets, much of the effects of a change in the money supply on inflation can be felt within a matter of weeks and on the exchange rate even within days. In effect, in recent years, CIS policymakers (supported by IMF-programs) were often quite successful in achieving inflation objectives (in the absence of fiscal dominance). The incomplete understanding of the rapidly changing behavioral relationships (such as money demand or money multiplier) and the underlying lag structure, was often successfully compensated for by frequent adjustments of the monetary instruments. Stone (2002) labeled this approach as “inflation targeting lite,” which, however, conveys a sense of greater precision and sophistication than may actually be the case. Mussa and Savastano (1999), in discussing the success of IMF-supported programs in general, stress that the “usefulness of financial programming depends not so much on the accuracy of its forecast, as on the flexibility for revising the main numerical targets as new information becomes available.” They characterize this iterative process as “open loop.” In any case, this approach has worked well particularly where the objective was to reduce inflation from still relatively high rates. Lissovolik (2003) supports the conclusion that this approach may not be as effective when the goal is to fine-tune inflation outcomes.

The main concerns with the current approach relate to the heavy focus by most, if not all, CIS countries on the exchange rate in addition to the inflation objective. While some “fear of floating” is understandable given the high degree of dollarization of the financial systems in most CIS countries, excessively stable nominal exchange rates (i) reinforce dollarization, (ii) reduce the effectiveness of monetary instruments, (iii) make it difficult to maintain control over monetary aggregates in the face of increasingly large capital flows, and (iv) could, down the road, result in serious misalignment of exchange rates.

However, the present approach has so far been very successful in lowering inflation in most CIS countries. Switching to a different approach or to different approaches (depending on the circumstances of the country) should, therefore, not be rushed without careful consideration of all available options and recognizing that introducing a new anchor entails additional risks during the switch of regimes as the choice of the regime is not de novo.

B. What Alternative Anchors Are Available?

We will examine in turn the advisability of (i) the “classical” money supply rules, (ii) the even older exchange rate pegs, and (iii) the “fashionable” direct inflation targeting framework as possible anchors.

Money Supply Rules

The IMF’s approach to monetary programming was developed in the late 1950s and 1960s (e.g., Polak, 1957) when economists were heavily under the influence of Anna Schwartz and Milton Friedman’s empirical work demonstrating that changes in nominal income were irrefutably linked, if not proportional, to changes in the money supply. The early “Polak” model was based on the assumption of a constant income velocity of money, and Polak (1997) argues that the strict quantity theory was in general a reasonable approximation. While the model does not provide a breakdown between changes in real GDP and in the price level, in high inflation situations this is not an issue as any change in output is dwarfed by changes in the price level.

Conceptually, money-supply rules are not suitable where the demand for money is very difficult to predict including because of uncertainties related to the remonetization process. The relationship between monetary aggregates and nominal GDP growth does not necessarily remain stable even in industrial countries.7 The instability and unpredictability of this relationship appears to be, however, particularly high in the CIS and compounded by a paucity of reliable data. This said, IMF-supported programs generally continue to set credit ceilings and NIR targets with a view to bringing about a certain rate of monetary expansion.8 Is this situation perhaps akin to that of the proverbial drunk, who looks for his lost keys under the street light not because he lost the keys there, but because there is light? A more benign interpretation would suggest that—as argued by Mussa and Savastano (1999)—forecast errors are typically successfully dealt with through IMF program reviews, automatic adjusters, and waivers or modification of performance criteria. Thus, money-supply targets are often shortlived and, therefore, do not “anchor” monetary policy in contrast to, for example, in Germany before the euro-area’s creation or the United States in the 1960s and 1970s.

Why do CIS policymakers state that they target monetary aggregates when they do not, but instead attach greater importance to relatively stable exchange rates? Part of the explanation may be the bureaucratic requirement formally to target monetary or credit aggregates in the context of Fund-supported programs. Despite obvious shortcomings, quantitative monetary targets figure prominently in a program context, as they are suitable for precise and timely testing. Moreover, it has been argued in the literature that meeting a declared money supply target could allow policymakers to establish credibility quickly, although this argument may not be relevant in the CIS. While a money-supply rule is easily understood by CIS central bankers, it is far less understood by government officials and perhaps not at all by the general public. Also, the theoretical linkage between money supply and inflation is still not fully understood outside central banks and narrow financial circles, although most senior policymakers now accept that rapid monetary expansion is inconsistent with price stability. By assuming responsibility for meeting a money-supply target rather than a specific inflation outcome, CIS central banks can safely point to a myriad of factors that influence price level changes such as harvest outcomes and tax rate changes without assuming responsibility for unfavorable inflation outcomes. Directly targeting a specific inflation outcome instead would, of course, require a more active response of monetary policy to such events and more difficult judgments.

Pursuing the wrong money-supply target because money demand has been misjudged is not cost free. There are examples in the CIS context (and elsewhere) of periods of unexpectedly rapid monetization where inflation targets were met with unexpectedly high rates of monetary expansion and where the observance of money-supply targets could have led to massive recession (e.g., Russia in 2001 and Ukraine in 2001/2002). As noted below, overly tight monetary policies may have been behind the economic downturns in the Czech Republic and Poland.

Exchange Rate Anchors

As discussed earlier, most, if not nearly all, CIS countries rely de facto on the nominal exchange rate as an anchor, but typically not as the exclusive anchor for monetary policy. Will this “fear of floating” continue? The answer is most likely affirmative for the next few years. For one, at least in the smaller and poorer CIS countries, financial markets are likely to remain very thin, and even in the more advanced economies, foreign exchange and domestic money markets have yet to develop depth and liquidity. Fears that exchange rates would overshoot9—whether justified or not—are likely to remain for some time an obstacle to freer floating. Also, in the heavily dollarized economies of the CIS, particularly in those where banks have extended dollar-denominated credits to borrowers without an earning base in foreign currencies, retaining some fear of floating would not be irrational given the likely impact of large exchange rate changes on financial sector stability. Moreover, the exchange rate to the U.S. dollar will more than likely continue to be widely followed by the population. Domestic currency prices of imported (and import competing) goods follow closely exchange rate movements and this has a substantial impact on the price index and perhaps even more importantly on inflation expectations. Stabilizing the exchange rate vis-à-vis the dollar, or keeping a downward slide of the exchange rate within narrow limits, will thus continue to provide immediate confirmation about the relative success of monetary policy in limiting inflation. And there is little doubt that, at least in the Baltics, fixed exchange rates have reinforced fiscal and financial discipline.

Should CIS countries take the final step and move to hard exchange pegs (or full dollarization) as monetary anchors as advocated in some of the recent literature for dollarized economies?10 Hard exchange rate pegs have been very successful in the Baltics in anchoring macroeconomic policies after a period of high inflation. Also, Austria and the Netherlands had successfully linked their currencies to the deutsche mark for long periods before the introduction of the euro.11 There is also little doubt that managing successfully an independently floating currency is not an easy or inexpensive task, and perhaps not cost effective for small countries (Buiter, 2002). Looking back over the relatively short history of CIS currencies, one might well argue that the freedom to pursue an independent monetary policy—and letting the exchange rate go—has generally not been put to good use. Particularly in the early transition periods, loose monetary policies may well have reduced economic growth as they delayed structural reform through validating soft budget constraints and because high inflation rates severely distorted resource allocation.12 And there is no guarantee that inflation objectives will not at the time of some future crisis become again subordinate to budget financing and quasi-fiscal operations.

With inflation rates much reduced in the CIS, a move toward hard pegs would, in general, not be motivated primarily by a need for rapid progress on inflation although gaining policy credibility could be an important consideration for some countries. Several CIS countries aiming at closer trade (and political) integration with neighbors are already considering linking their currencies, although the energy exporters (Russia, Kazakhstan) and energy importers (Belarus, Ukraine) in this group face very different external shocks. In the absence of large intergovernmental support arrangements, exchange rate pegs or links would, however, need to be to currencies of countries that are exposed to symmetric rather than asymmetric shocks. This said, being exposed to identical shocks is in itself an insufficient reason for a monetary union. Specifically, why would a country link its currency to that of a neighbor with much higher inflation or to the currency of a country with much stronger pressures for real appreciation?

Direct Inflation Target Framework13

As described earlier, all CIS countries are pursuing price stability and the progress made has in most instances been remarkable. However, none of the CIS countries is targeting inflation within a formal direct inflation targeting framework derived from empirical models linking policy action to inflation outcomes. Only in the case of Russia has a possible shift to direct inflation targeting been extensively discussed, while Kazakhstan and some other CIS countries have expressed an interest in moving in this direction.

What are the reasons why formal direct inflation targeting the way it is understood, for example, in Canada, New Zealand, South Africa, Brazil, Chile and some Central and Eastern European Countries, is not currently pursued by CIS countries? And can it become a viable option for the near term? First, at this stage, the relationship between changes in policy instruments and inflation outcomes remains in all CIS countries insufficiently predictable or stable as the rapid transformation of their economies and their financial sectors continues. A premature introduction of direct inflation-targeting would therefore entail substantial risks, including of discrediting this approach altogether. For example, in reviewing the experience with direct inflation targeting by the more advanced East European transition countries (the Czech Republic, Poland, and Hungary), Jonas and Mishkin (2003) note that inflation targets were often missed by wide margins. They argue that undershoots of inflation targets in the Czech Republic and Poland have resulted in serious economic downturns and undermined central banks. In their view, the main reason was the increased uncertainty prevailing in transition economies, which makes it particularly difficult to predict inflation sufficiently well as required by the forward-looking nature of formal direct inflation targeting.

For now, there are no reliable empirical models for CIS countries that can firmly link policy actions to inflation results. Reliable empirical models are, however, indispensable requirements for full-fledged inflation targeting. It is difficult to predict whether within the next few years such models could be estimated with any degree of confidence in one or two of the CIS countries. Second, successful introduction of direct inflation targeting requires both credibility of the policymakers and transparency in the policy-decision-making process. While important progress has been made in this regard over the last decade, substantially more remains to be done.14 Third, the concept of core inflation would need to be developed. In particular, food price fluctuations linked to the results of the harvest, administrative price adjustment, Balassa-Samuelson effects, and, possibly, price shocks from exchange rate changes would need to be separated out. While primarily a technical issue, the definition of an appropriate inflation measurement is also intimately linked to questions of credibility and transparency.

C. No Single Currency Regime Is Right for All Countries or at All Times15

In an open economy, exchange rate policy and monetary policy are essentially two sides of the same coin. As discussed above, choosing a formal, direct inflation-targeting framework requires a freely floating exchange rate regime, while, by definition, the use of a hard exchange rate peg or full dollarization precludes exchange rate adjustments, except in extreme situations. It would, therefore be important to consider the choice of anchor for monetary policy together with the question of broader suitability of the concomitant exchange regime choice. Fleming (1962) and Mundell (1961) showed that under high capital mobility16 fixed and floating exchange rate regimes have starkly different implications for the choice of policy instruments (monetary policy, fiscal policy) for the achievement of the targets of domestic and external balance. Reversing this argumentation and reinterpreting monetary and fiscal policies as nominal and real shocks (Guide and Keller, 2002), fixed exchange rates provide greater protection for output and/or price level in the face of nominal shocks, while floating exchange rates offer better insulation against real shocks.17 A pegged exchange rate regime will, therefore, be preferable to a floating rate regime when the central bank has an incentive to generate surprise inflation—and is unable to credibly pre-commit to not doing so—or when unanticipated monetary shocks predominate. However, when real shocks predominate and establishing policy credibility is less of an issue, a floating exchange rate regime may be preferable.18

Countries are not unconstrained in regime choice. For one, countries may simply not be in a position to credibly peg their rates because of an insufficient level of foreign exchange reserves. A second consideration is that depending on the degree of capital mobility, there is little or no scope for monetary policy under a fixed exchange rate, and the full burden of macroeconomic policies falls on fiscal policy. The latter may perhaps not be flexible enough to do the job and is constrained in any case by the risk of high budget deficits or public sector debt undermining the credibility of the exchange rate peg.

D. Modifying the Current Approach: Dropping the Anchor?

Should CIS countries set aside their fear of floating and permit greater and perhaps much greater exchange rate flexibility despite the risks brought on by a high degree of dollarization? Or should they move towards the other extreme and actually forgo monetary policy altogether and formally peg their currencies, and, if so, to what currency? Or is there a viable middle ground where it is sufficient to only keep an eye on the exchange rate when conducting monetary policy?

After years of lip service to floating exchange rates while heavily managing exchange rates, it appears that CIS countries remain to be convinced that conducting monetary policy in a setting of more freely floating exchange rates would provide a superior approach. Also, intermediate solutions have gained substantial acceptance in some of the recent literature. (See Williamson (2000)) for a spirited defense of intermediate exchange rate regimes and also Frankel (1999).) And there are not many examples of freely floating exchange rates in the world, setting aside Canada, the United States, Euro zone, Japan and some other countries which are not particularly relevant examples. Moreover, the main reasons economists advocated floating exchange rates was a fear of rigid or managed exchange rates becoming increasingly overvalued, eventually leading to balance of payments crises and recession (Fischer, 2001). Balance of payments pressures, over-valued exchange rates and recession are not—at least not currently—major concerns for most CIS countries. On the other hand, given the progress already achieved in lowering inflation, the incentives for moving to hard pegs, crawling pegs, or preannounced devaluations are also not very strong.

Switching to one of the corner solutions of either a formal inflation targeting framework or hard peg as anchor would be a radical approach, while the middle ground has proven viable in many countries. In the CIS context, the latter approach would entail policymakers eschewing the corner solutions, while adjusting monetary policy frequently to maintain acceptable levels of price and exchange rate stability. Yet some observers judge that adoption of a formal inflation targeting framework could be a useful ultimate goal of monetary policymakers in some CIS countries, even if their institution are not now adequate to the task. Barring large crises that would make radical changes in the monetary regime acceptable, an evolutionary approach may instead be called for. In current circumstances, it would seem desirable to encourage more limited exchange market interventions and the acceptance of (gradually) increased exchange rate flexibility to discourage even greater dollarization and gain better control over the money supply. Moreover, it will be important to demonstrate that lending or borrowing in foreign currency does not entail a one-way bet on the exchange rate. There are divergent views as to how much the high degree of dollarization and underdeveloped exchange markets in the CIS limit the exchange rate flexibility to less than would be needed in a direct inflation targeting framework. In our view, a substantial reduction in the dollarization of the financial sectors in the economies would be required before the option of conducting monetary policy in the context of freely floating exchanges could be considered.

IV. Related Issues

A. What Should Be the Inflation Target?

Now that inflation rates have come down from very high levels in nearly all cases, and into the single digit range in many CIS countries, it will no longer be a case of just aiming at lowering inflation, per se, but at targeting a specific outcome or range. Studies of Central and Eastern European experience suggest that a reduction in inflation from the double- and triple-digit range was associated with very substantial output gains (Stavrev, 2003). Stavrev also finds that lowering inflation further to the low single-digit range would most likely be associated with some modest, short-term output losses. A target of near-zero inflation is not desirable, mainly because nominal interest rates cannot become negative and monetary policy would become unable to stem deflationary trends (e.g., Japan).

In recent years, transition economies in Eastern Europe have typically recorded inflation rates in the single digits, but substantially above the rate in the United States or the EU. With nominal wages and prices downward rigid—although to varying degrees and perhaps less than in industrial countries—it may be difficult for CIS countries to achieve both a fast transformation of the economy, entailing a large-scale reallocation of factor inputs, and a high degree of price stability. If the authorities are prepared to accept nominal appreciations of their currencies rate, price stability can be “imported” via falling prices of traded goods that would compensate for the unavoidable increases in the absolute and relative prices of nontraded goods (i.e., the service sector where productivity gains are slower) in a growing economy. For now, the political reality in CIS countries seems to be that officials find it easier to pursue a weak currency policy so as to be seen as supporting the export sector and import competing industries. Moreover, these industries seem to be able to generate more focused political pressure than the diffuse groups and broad population affected by higher inflation. In other words, there is not—or not yet—a strong enough lobby for price stability that would encourage politicians to allow inflation objectives to override exchange objectives.19

Gradualism in lowering inflation may be more desirable than a quick adjustment that cannot be sustained. Moreover, care needs to be taken to define the inflation target appropriately and realistically so as not to create obstacles to much-needed adjustments of prices, such as for example, utility tariffs. However, setting an inflation target net of administrative price adjustments could mean a less transparent target than headline inflation.

B. Should Price Stability Be the Only Objective?

Central Banks generally have pursued financial stability with bank regulations and supervision, maintenance of a secure and efficient payment system, and creating sound macroeconomic fundamentals.20 Central banks provide liquidity to banks, including as lenders of last resort. Even “inflation targeters” (such as the Bank of England and the Reserve Bank of New Zealand) have financial stability as an explicit objective. In addition, some central banks have a mandate also to promote economic growth or to sustain output.

In the CIS countries, there is so far little evidence of a business cycle or, at least, not one as it is understood in western economies. It may well be that business cycles have so far been masked by the more massive impact of transformation on CIS economies. For now, it would seem appropriate to use monetary policy not for anticyclical purposes, but instead to focus on the primary objectives of price and financial sector stability.21

C. The Issue of Dollarization

Dollarization is a reality in the CIS countries. Dollarization reflects a choice by household and firms to protect themselves from a recurrence of high inflation by diversifying their assets. It also reflects the opening up of CIS economies to trade with the rest of the world. Dollarization of assets is particularly widespread, be it in the form of dollar-denominated bank deposits or dollar bills under the proverbial mattress. Liabilities have also been dollarized. In most CIS countries, more than half of bank loans or mortgages are typically extended in dollars, or indexed to the U.S. dollar, with dollarization highest for liabilities with longer maturities. Payments dollarization is prevalent for large household transactions involving apartments or cars or illegal and grey-market transactions (Oomes, 2003). Also intra-CIS trade is mostly invoiced, and paid for in U.S. dollars, including pipeline fees and similar charges. “Real dollarization,” defined as setting prices and wages in U.S. dollars also exist even when payments dollarization does not (legally) exist, e.g., prices for appliances, hotel services and some restaurants are often calculated in U.S. dollars, but settlement is typically required in local currency. If this were not required, it would perhaps be more convenient and entail lower transaction costs to settle directly in dollars. There is anecdotal evidence that enterprises are also frequently pricing in U.S. dollars when dealing with each other and that the private sector wages are at times set with an eye to the exchange rate to the dollar.

The origins of dollarization are twofold: First, in the Soviet Union days, dollars gave access to goods and services that were not otherwise available (e.g., dollar shops, travel to the west, imports of spare parts and machinery from outside the CIS) and the ruble was not easily converted into dollars. In the early years of transition, high inflation rates gave strong encouragement to the use of the dollars, including cash dollars. In addition, dollarization allowed the diversification of portfolios, including by investments into assets with longer maturities and into stocks, which were not available in domestic currency-denominated assets. More than a decade later, with inflation now under control in most CIS countries and currencies increasingly convertible, the portfolio choice argument still applies and the fear still lingers that governments and central banks may not be able or willing to pursue appropriately tight monetary policies if faced with adverse shocks. Even a remote possibility of a return to high inflation is sufficient to discourage holding of longer term financial assets denominated in domestic currencies. (The hysteresis of dollarization is a well-documented phenomenon also in developing countries.) This has led to a large share of time deposits being denominated in U.S. dollars and governments issuing domestic debt denominated in U.S. dollars that affords private savers and pension funds some protection and opportunities for diversification. Protection from inflation and exchange rate risk may, however, be an illusion in the case of very severe shocks, as dollarization increases the default risk for banks and the vulnerability of the financial sector and the economy as a whole.

Substantial denomination in dollars (or indexation to dollars) of banks’ assets and liabilities makes exchange rate adjustments more costly. Knowing that the authorities might therefore be hesitant to adjust the exchange rate can create a moral hazard and encourage firms to borrow excessively and cheaply in foreign currencies rather than in domestic currency, adding to the fragility of the economy.

D. Should Central Banks Fight Dollarization?

With the expected accumulation of financial assets over time by households, pension funds, and other institutions, dollarization may become a more important issue in the CIS (as is already the case in Latin America). Reducing, or limiting the increase in, dollarization through appropriately tight monetary and fiscal policies is therefore desirable so as to reduce the future vulnerability of the economy and the banking system. Differentiated reserve requirements applied to stocks or increases in dollarized deposits could also be considered. Reduced dollarization would also create the necessary degrees of freedom for greater exchange rate flexibility and the option to pursue at some later stage direct inflation targeting, if so desired. This said, all these measures are unlikely to yield substantial progress in the short term, and there are only few examples of countries, such as Israel and Poland, succeeding in substantially lowering dollarization.

Taking recourse to more heavy-handed measures, such as outlawing foreign currency deposits in banks, or imposing surrender requirements on exporters, would, however, most likely be counterproductive. Experience has shown that business and wealthy households can quickly find ways to build-up financial assets outside the country (even now payments are made between CIS firms through foreign banks that entirely bypass the domestic banking system). The result would be increased demonetization of the domestic banking system, capital flight, and tax losses.

V. Conclusions

While there are important differences among CIS countries as regards financial sector developments, there are also very important similarities:

  • Financial deepening is progressing, but its often substantial speed varies a great deal between countries and across time and remains difficult to predict. The size of the banking sector remains small compared with the Baltic and Eastern European countries. The nonbank financial sector is just emerging, even in the most advanced countries.

  • Fiscal dominance has subsided in most countries, and this is permitting monetary policy to begin to play a more important role. Financial markets, however, remain seriously underdeveloped and interest rates or more sophisticated policy instruments play at best a secondary role. The principal instrument of monetary policy remains unsterilized exchange market interventions, even in the more advanced countries.

  • Dollarization of banks’ assets and liabilities (and other financial assets) is very high and not easily reversed. While dollarization was not an obstacle to the impressive reduction in inflation rates, it has increased the vulnerability of the financial system and the economy to large exchange rate changes and contributed to a fear of floating.

  • Although protection against high inflation was the original motive for dollarization, portfolio diversification in the context of underdevelopment financial markets and relatively stable exchange rates have further reinforced dollarization.

  • Exchange rates have fluctuated much less than international reserves, and recently a number of CIS countries have tried to prevent or slow a real appreciation of their currencies, potentially putting their inflation objectives at risk.

  • All CIS countries target price stability—at times in conjunction with other objectives. As stable money-demand functions generally do not exist, monetary policy at this stage cannot easily be guided by the traditional financial programming tools or by formal, model-based and direct inflation targeting.

  • Short time lags between changes in monetary and credit aggregates and inflation, the exchange rate or net international reserves, however, make “error correction” faster and easier than in countries with deeper financial markets and correspondingly longer time lags.

  • Monetary policy and the frequent adjustments of intermediate targets typically take account of a number of considerations, which at times may be at odds with the inflation objective. Increasing the transparency of this decision-making process would foster trust in the central banks.

  • The debate about how flexible exchange rates should be, given the high degree of dollarization, is unresolved. Greater exchange rate flexibility would certainly reduce the incentive for further dollarization, including by underscoring the risk of borrowing in foreign currencies. At the same time, dollarization brings significant risks if exchange rates were to become highly volatile. This said, the “fear of floating” does not justify overly rigid exchange rates, and in most countries the scope for greater exchange rate flexibility is probably not exhausted. In general, CIS countries should be urged to intervene less in foreign exchange markets, particularly if it were to conflict with the control over the monetary aggregates and the inflation or balance of payments objectives.

  • Much has changed over the decade since the introduction of national currencies in the CIS. With continued development of short-term financial markets, sustained price stability, and accumulation of financial assets, monetary policy will become more important in the future and its impact on the economy more predictable. In perhaps another decade or so, several CIS countries may have reached a state of financial sector development close to the current situation in Eastern European countries.

  • The choice of the relevant anchor for monetary policy will, however, remain an issue. Should some of the smaller CIS countries follow the example of the Baltic countries and EU accession countries—or the earlier examples of Austria and the Netherlands—and surrender their independent monetary and exchange rate policies by joining a larger currency bloc? Or will they develop the necessary skills and reduce the degree of dollarization sufficiently to make model-based direct inflation targeting a viable option? Or is there scope for adapting the current monetary policy approach based on open-loop-feedback processes without moving toward one of the corner solutions for the concomitant exchange rate choice?

  • On the one hand, with inflation already much reduced, there does not seem to be a strong case for anchoring monetary policy with hard pegs by means of a currency union or currency board. Moreover, hard pegs to currencies of non-CIS countries, such as the euro and dollar, would not correspond to trade patterns, while a currency union among CIS countries would most likely be unworkable because of a lack of political will to subordinate national interests. On the other hand, a formal, empirical-model-based, and direct inflation-targeting framework as anchor would seem likely to be developed in the more distant future, as it, inter alia, would require more stable money-demand functions, deeper financial markets, a stable fiscal framework, and freely floating exchange rates. Concerns about financial vulnerabilities in the heavily dollarized economies and fears about competitiveness are serious obstacles to floating rates in most CIS countries. It may, therefore, be best to gradually modify the current approach of relying on heavily managed exchange rates as anchor or co-anchor by moving toward greater exchange rate flexibility and keeping less of an eye on the exchange rate objective so as to gain better control over the monetary aggregates. The high frequency open-loop-feedback mechanism would then become increasingly the sole anchor. Barring a major crisis, CIS countries are likely to be more supportive of such evolution than of a revolutionary shift to new anchors.

Same Series

Other IMF Content

Other Publishers

Cover IMF Working Papers
Nominal Anchors in the CIS
Author: Mr. Peter M Keller and Mr. Thomas J Richardson
Volume/Issue:
Volume 2003: Issue 179
Publisher:
International Monetary Fund
ISBN:
9781451859058
ISSN:
1018-5941
Pages:
42
DOI:
https://doi.org/10.5089/9781451859058.001

Broad Money/GDP 1/

(End of 2002)

Foreign Currency Deposits/Total Deposits 1/

(End of 2002)

CIS: CPI Inflation

(12-month percent change)

CIS: CPI Inflation