Economic Systems 66 (2003) 1–25

Is Hungary ready for inflation targeting? Pierre L. Siklos a,∗ , István Ábel b a

Department of Economics, Wilfrid Laurier University, 75 University Avenue, Waterloo, Ont., Canada N2L 3C5 b Advisor to Executive Director, International Monetary Fund, Washington, DC 20431, USA Received 20 March 2001; received in revised form 2 May 2002; accepted 3 May 2002

Abstract This paper considers whether the adoption of inflation targets for Hungary, an emerging market economy, is a desirable option. We consider the qualitative and quantitative pre-conditions required for the successful adoption of inflation control objectives. Although it is found that the National Bank of Hungary appears to possess a reaction function with some of the main features found in those estimated for industrial economies, there remain certain aspects of the relationship between the government and the central bank that require clarification and further evolution. © 2002 Elsevier Science B.V. All rights reserved. JEL classification: E58; E61; P20 Keywords: Inflation targeting; Central bank accountability and transparency; Exchange rate targeting

1. Introduction Following experimentation with a variety of monetary policies over the past four decades or more, several industrialized economies have opted for a form of inflation targeting (Siklos, 2002; Mahadeva and Sterne, 2000). Other monetary policy strategies, notably money supply targets and exchange rate pegging, were implemented, but mostly found wanting.1 In some countries the target is quantified and is ordinarily a joint initiative of the government and the central bank (e.g. Canada and New Zealand). In others, the goal is more implicit as ∗ Corresponding author. Tel.: +1-519-884-0710x2559; fax: +1-519-888-1015. E-mail address: [email protected] (P.L. Siklos). URL: http://www.wlu.ca/∼wwwsbe/faculty/psiklos/home.htm 1 It should be noted that one of the “pillars” of the European Central Bank’s strategy includes concern for money supply growth (see ECB, 1998), though this approach has been deemed inappropriate by at least one critic (Svensson, 1999). The debate for or against fixed exchange rates seemingly is a never ending one (e.g. see Bank of Canada, 2001).

0939-3625/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 3 9 - 3 6 2 5 ( 0 2 ) 0 0 0 5 9 - 6

2

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

in Germany or the US.2 However, there is a tacit, if not formal, understanding nowadays that central banks ought to promote a form of price stability with some numerical value possibly attached to that condition. The apparent popularity of inflation targeting is due, in large part, to the success of these policies in reducing inflation and in anchoring inflationary expectations (Bernanke et al., 1999). However, it has been noted that the disinflation of the 1990s is a worldwide phenomenon (e.g. Siklos, 1999a) and it is therefore difficult to ascertain how much of the improvement in the record of inflation is attributable by itself to the adoption of the policy.3 Nevertheless, since inflation targeting represents a “framework” for the conduct of monetary policy, there has been keen interest in the applicability of such a policy regime to emerging market economies. The IMF (Masson et al., 1997; Blejer et al., 2000; Coats, 2000) has gone so far as to explore the record of inflation targeting in industrial and some emerging market economies (including Brazil, Chile, Israel, and Mexico) in order to provide some lessons for other countries seeking to go down that route. Taylor (2000) and Mishkin (2000) have also joined the chorus of those in favor of some form of inflation targeting in emerging market economies. The present paper considers the case of Hungary, which is regarded as being among the most mature of the so-called emerging market economies.4 The country has experienced almost a decade of relative economic stability and has largely completed its transition to a market-based economy.5 Hungary neared a cross-roads of sorts by 2001, as the crawling pegged exchange rate system adopted in 1995 (see, inter alia, Siklos, 2000; Szapáry and Jakab, 2000; Neményi, 1997, and Kornai, 1997) reached the point where changes in the pre-announced exchange rate were close to zero.6 Indeed, on 4 May 2001, the fluctuation band for the Hungarian forint (HUF) was widened to ±15% from ±2.25%. On 24 August 2001, it was announced that the crawling peg would cease to exist 1 October 2001. Consequently, the exchange rate no longer served a useful function as a focal point for inflationary expectations. As one of the candidates for accession to an enlarged European Union (EU) and, someday, into the European Monetary Union (EMU), it is appropriate to ask whether the inflation targeting framework is a sensible choice for Hungary under these circumstances.7 The press 2 Hence, in the case of Germany, the inflation objective was achieved via the announcement of monetary targets. The Bundesbank, of course, ceded control over key monetary decisions in 1999 to the European Central Bank. 3 Siklos (2002), however, presents evidence for 20 OECD economies to the effect that inflation targeting countries have reduced nominal interest rates by more than non-inflation targeting countries. 4 Research for this paper was completed prior to the official announcement of imminent inflation targets in Hungary. However, as will be seen below, a number of outstanding issues remain so the question whether Hungary is ready for inflation targeting remains relevant. 5 Falcetti et al. (2000) claim that, by 1998, Hungary enjoyed 8 years of macroeconomic stability. Further, they construct a reform index ranging from 0 to 4 and report that Hungary reached the maximum index value by 1997. 6 By April 2001, the daily rate of depreciation was only 0.0066% which translates to an annual depreciation of approximately 2.4%. 7 The present paper is primarily concerned with the conditions that would ensure a successful transition to an inflation targeting regime. Krzak and Ettl (1999) outline the inflation targeting approaches used in the Czech Republic and Poland but are unable to present empirical evidence ostensibly owing to the scarcity of relevant data for these two countries. Also, see Šm´ıdkova and Hrn´ır (2000), Czech National Bank (2000), Jon´aš (2000); IMF (2000), and Christofferson and Westcott (1999).

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

3

release of 24 August 2001 (available at http://www.mnb.hu/index-n.htm) states that the next step is to target inflation, though there are relatively few publicly available details to date about the strategy (see National Bank of Hungary, [NBH] 2001a,c). The paper is organized as follows. The next section discusses the necessary conditions that ought to be satisfied for the successful introduction of inflation targeting. We also speculate about the relative importance of each condition. Section 3 explores qualitatively, as well as quantitatively, the extent to which the necessary conditions for the adoption of inflation targets have been met in Hungary. Section 4 considers the overall verdict of the exercise. In particular, we conclude that the adoption of inflation targets is both desirable and compatible with entry into the EU and EMU, and that most, but not all, of the pre-conditions have been met. The de facto adoption of a floating exchange rate regime, together with an announcement preparing the public for an eventual full inflation targeting strategy is the appropriate course of action, and largely follows the path taken earlier on by the Bank of Canada and the Reserve Bank of New Zealand in the 1990s. Section 5 concludes.

2. Necessary conditions for the adoption of inflation targets? Following Siklos (2002), Bernanke et al. (1999), Blejer et al. (2000), and others, Table 1 lists some of the institutional and economic pre-conditions necessary for the successful adoption of inflation targets. They are not listed in any order of preference nor are they ranked in any order of importance. The table is sub-divided into two parts. This reflects the need to meet both institutional and economic requirements prior to the introduction of inflation control objectives. Others (e.g. Masson et al., 1997; Taylor, 2000; Mishkin, 2000) have created comparable lists but did not adopt the institutional-economic dichotomy proposed here. 2.1. Institutional conditions Institutional pre-conditions reflect the need for the central bank to be “autonomous” in meeting the specified objectives. This requires instrument independence, as pointed out most notably by Debelle and Fischer (1994). It also necessitates the associated independence or autonomy of the monetary authorities from government instructions, while shifting ultimate responsibility for monetary policy to the government. The central bank must also have at its disposal the tools necessary to signal to financial markets, in particular, its intentions concerning the state of monetary conditions. In general, this is accomplished via an overnight rate. A central bank granted considerable autonomy to accomplish a specified task must also become accountable for its actions. In other words, it must be prepared to answer to the public, via its elected representatives, concerning the successes, failures, and remedies in case of failure to meet the agreed upon objectives. It must also be clear which individual or group is answerable to the government for accomplishing the inflation control objectives. In some countries, accountability is enforced in the legislation of the central bank (e.g. New Zealand). Elsewhere, enforcement is achieved either via the threat of a loss of credibility or, alternatively, many central banks face the possibility of government override of their decisions if there are serious disagreements with the government (e.g. as in Canada and Australia). Accountability, however, must be accompanied by the communication of

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

4

Table 1 Adopting inflation targeting: necessary institutional and economic pre-conditions Institutional requirements

Rationale

Instrument independence

The central bank needs a free hand to attain a specified inflation objective The central bank should have at least one major instrument that can inform markets about its desired stance for monetary policy and influence expectations of inflation in particular There needs to be clarity about the responsibilities of the central bank for achieving stated price stability objectives The central bank ought to communicate clearly and frequently with government, markets and the public. This condition is often referred to as the transparency objective The flexibility to achieve a quantified inflation objective must be in the context of an exchange rate regime that permits domestic monetary policy to determine monetary conditions The adoption of quantified inflation targets must be announced by government as an agreed to objective that is to be met by the central bank The smooth operation of an inflation target cannot be assured if fiscal and monetary policies are not in harmony. Fiscal policies must be designed such that they are compatible with monetary policies The adopted strategy should be seen by the public as the most appropriate one under the circumstances Financial markets should be sophisticated and liquid enough to generate a strong preference for price stability and provide useful inflationary expectations

Effective monetary policy instrument

Accountability Disclosure

A floating exchange rate

Joint responsibility

Harmony with fiscal policy

Public support Well developed financial system

Economic requirements Choice of an appropriate price index

Selection of a width and horizon, and symmetry for the inflation target Adequate knowledge of the transmission mechanism of monetary policy Provision of inflation forecasts

Adequate measurement and timely arrival of economic information

The inflation objective must be set in terms of an index that is widely understood and representative of developments in the purchasing power of money Inflation targeting can be credible and simultaneously provide needed policy flexibility. However, the degree of flexibility needs to be announced as well as the central bank’s intention to aim toward the mid-point of any target range (symmetry) Policy makers ought to be reasonably well informed about how quickly and how strongly their instrument settings are likely to affect the main aggregates of the economy Inflation targeting is primarily about ensuring that inflation expectations are well anchored within the stated inflation target band. This means that the central bank should be forward looking and react, at least in part, to forecasts or inflation projections Financial markets need to be well developed to produce useful signals in central bank decision-making. Economic information needs to be produced in a timely manner with a reasonable amount of measurement error

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

5

monetary policy decisions in a clear and regular manner to financial markets and the public. In most industrial countries this has typically taken the form of a monetary policy or inflation report (see Siklos, 1999a, 2002). Again, some countries mandate such reporting (e.g. UK, New Zealand, the US until 2000). In other countries, reporting on monetary policy decisions is voluntary. Nevertheless, all these recent developments reflect the recognition that inflation targeting requires greater disclosure of information by the central bank, that is, a demand for greater transparency.8 Indeed, the transparency question has seemingly acquired a life of its own as a desirable characteristic of central bank operations comparable to the importance allotted to the need for central bank “independence” beginning in the1990s. Perhaps more fundamentally, inflation targeting cannot adequately fulfill its accountability and transparency requirements unless the determination of monetary conditions is firmly in domestic hands. This can only be accomplished under flexible exchange rate conditions because only then is the nominal anchor of monetary policy clear. Under other exchange rate regimes the nominal anchor shifts responsibility, depending on the nature of the peg, to the monetary policy of another country.9 A central bank that is autonomous must nevertheless ultimately answer to the public via elected officials who have ultimate responsibility for the chosen monetary policy strategy while leaving the central bank with the task of how best to meet the stated inflation objectives. Consequently, there must be joint or shared responsibility following the announcement of the targets. In this fashion the government formally recognizes the desirability of a stated inflation objective with the responsibility for carrying out these objectives resting with the central bank alone. In most inflation targeting countries the inflation objective is set by the government, in consultation with the central bank. However, it is up to the central bank alone to use the necessary instruments with the intention of meeting the inflation goal. Otherwise, the central bank will not have the requisite credibility to meet the inflation target. Consequently, the doctrine of shared responsibility requires that fiscal policy does not undermine the objectives of monetary policy (viz. an inflation goal jointly agreed to with the government) or vice versa.10 It is largely for this reason, and to signal the fact that fiscal and monetary conditions were not yet in place to permit announcement of stable inflation rate via the publication of inflation control targets, that the initial inflation targets in New Zealand and Canada were in fact inflation reduction targets.11 8 For our purposes, transparency and disclosure can be thought of as the same thing. However, Siklos (2002) points out some subtle but important differences between the two terms. In addition, rules concerning accountability and transparency introduced by the IMF in 2000 (see IMF, 2000) also relate to the disclosure practices of central banks. 9 Some (e.g. Mishkin, 2000) have argued that inflation targeting may not be feasible without some importance being placed on an exchange rate objective. However, the Israeli experience (e.g. see Bufman et al., 1995, and Leiderman and Bufman, 2000) suggests that tensions between two potentially conflicting targets, one expressed in terms of inflation, the other in terms of the exchange rate, may be too great. A point underemphasized in this context is that, typically, the central bank is responsible for achieving the inflation target while the government generally sets the exchange rate policy (see Siklos, 2002). 10 While fiscal dominance was a feature of Hungarian monetary policy until the mid 1990s, it has been on the wane (see Szap´ary and Jakab, 2000, and Siklos, 2000). 11 Masson et al. (1997) also point out that inflation rates were relatively low at the outset, by the standards of most developing countries, thereby giving inflation targeting an additional credibility boost many emerging markets may not enjoy.

6

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

Inflation control targets were announced only when fiscal houses in both countries were in order. Two other institutional considerations are worthy of discussion though they are related to some of the pre-conditions discussed below.12 The monetary policy strategy must have sufficient public support and must be seen as the most appropriate one under the circumstances.13 Clearly, this is a necessary feature of any successful monetary policy strategy that cannot be divorced from the credibility requirement of any monetary policy strategy. Moreover, public support is also necessary to increase the impact on inflation expectations of any central bank decision.14 Finally, a well-developed financial sector will ensure the success of inflation targeting. Posen (1993, 1995), for example, argues that the financial sector has a vested interest in avoiding inflation. Presumably, the more sophisticated or developed the financial system is, the more likely it is that a central bank will place a relatively greater emphasis on inflation control.15 2.2. Economic conditions The list of economic pre-conditions for inflation targeting given in Table 1 reflect the many practical decisions that have to be made once the institutional conditions have largely been met. They include the choice of a price index used to define the inflation target, the degree and nature of permitted flexibility around the target, knowledge of the transmission mechanism, providing inflation forecasts or projections, timeliness and minimization of measurement errors in the required data. Their importance has been emphasized by many (e.g. Haldane, 1995; Leiderman and Svensson, 1995; Bernanke et al., 1999; Siklos, 1997a, to name a few) and they seem, for the greatest part, self-explanatory. One issue that deserves a little more discussion, however, concerns the release of inflation projections or forecasts. The subject appears controversial (e.g. see Tarkka and Mayes, 1999) not least because central banks worry that the release of such information can cloud accountability and create conflict between clarity and transparency in central bank activities. Generally, forecasts are prepared by the staff of the central bank and represent one input used to decide the appropriate level at which to set the instrument of policy.16 More importantly, if such forecasts are conditional on some scenario, such as leaving the current instrument settings unchanged, they are no longer forecasts as such but reflect a projection of the future in an unchanged monetary policy environment. Unless consumers of such projections are clear about the underlying assumptions behind them, they can be led to believe that the projection is the most likely outcome for the variable being projected. 12

We are grateful to an anonymous referee for raising these points. Hence, it is quite possible that the public does not quite see the need for low inflation when moderate inflation (e.g. 10–30%) produces adequate economic growth. There is a vast literature dealing with the relationship between inflation and growth, with conflicting results, that we cannot survey here. See, for example, Temple (2000). 14 Mahadeva and Sterne (2000) also point out the role of public support in devising a successful monetary policy strategy. 15 An implication of Posen’s argument, which we do not subscribe to nor do we take up explicitly in this paper, is that statutory measures of central bank independence need not always be a good predictor of inflation developments. 16 See Longworth and Freedman (1995) for an explanation of the role of staff forecasts in the Canadian case. The situation is no doubt similar at other central banks in the industrial world, including the young European Central Bank (see Duisenberg, 2000). 13

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

7

Instead, a forward-looking central bank will use the projection, if taken literally, to change the instrument of monetary policy precisely to prevent the projection from becoming reality, especially if the outcome is considered undesirable from the standpoint of achieving an inflation target. There is one sense in which providing inflation forecasts is particularly useful in the case of transition economies. Siklos (2002) argues that inflation targets were effective in countries with central banks that had little credibility or reputation in delivering low and stable inflation rates. The reason is that the targets were seen as a highly visible means of making the central bank accountable for its policies. However, once inflation targets are unchanged, and since the public may become less inclined to be forward looking if the forecasts are met for some time, it becomes less clear how useful published forecasts are consistently within the target ranges. The reason is that, under such conditions, expectations are unlikely to be influenced by anything other than the announced target, unless the public also routinely assesses to risks of missing the target. To be sure, in this kind of environment, a projection can be a useful way of communicating a scenario for the series of shocks that can combine to produce a bad economic outcome but only if the projection is not viewed as having the quality of a self-fulfilling prophecy. Aside from these concerns are issues concerning the risks around any projection. Different central banks have taken somewhat different approaches to this problem by providing ranges or bands around a central projection or forecast (e.g. as in the Bank of England’s fan chart; see Siklos, 2002; Tarkka and Mayes, 1999 for additional details). In what follows we are not interested in the manner in which a projection or forecast is made public but whether the forward-looking nature of monetary policy is adequately communicated to the public.

3. Has Hungary met the requisite conditions? 3.1. Institutional pre-conditions It is useful to use the institutional and economic classifications in Table 1 to consider whether Hungary has met the pre-conditions for adopting an inflation targeting policy. Tables 2 and 3 considers Hungary’s status in this connection.17 Part (A) of the table lists what is known about the principal ingredients of the planned inflation targeting regime. Inflation reduction targets are anticipated with a long-term inflation goal presumably driven by the requirements of the Maastricht Treaty.18 Part (B) of the same table reveals that by 2001 all institutional conditions for inflation targeting were met. As far as statutory independence is concerned, Hungary is among the most independent central banks in the world, at least according to some estimates. Siklos (1994) uses a version of Cukierman’s (1992) 17 A separate table (not shown) presents some information relative to the separate Inflation and Financial Stability Reports produced by the NBH. Such publications are considered to be important ingredients in the monetary policy strategy of industrial countries that target inflation. 18 Technically, the Maastricht treaty does not require 2% inflation but inflation not greater than +1.5% of the average of the three lowest inflation rates in the euro area. This is likely to give the NBH more room to achieve the required inflation objective than the notional long-term goal currently outlined.

8

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

Table 2 Basic ingredients of Hungary’s planned inflation targeting strategy Index targeted

CPI

Inflation objective Calculation period Contingencies Targeting horizon Adoption date Separate inflation report Who sets target? Publishes inflation forecasts?

7% by December 2001; 4.5% by December 2002; 2% long-run objective Ostensibly over the cycle Contemplated Yes 12 June 2001 Yes, predates inflation targeting Joint Yes

well-known index of central bank independence adapted for transitional economies and finds an index value of 0.24 to 0.46 out of a maximum of 1.19 These values are more modest than others published in the literature. Cukierman et al. (2002) also apply Cukierman’s index and also place Hungary near the top of the list of independent central banks, as do Loungani and Sheets (1997). Hochreiter (1994), Hochreiter and Riesinger (1995), Dvorsky (2000), and Radzyner and Riesinger (1997) investigate variants of basic measures of institutional independence and generally conclude that these have improved markedly in Hungary and elsewhere, especially in central Europe (see also Neyapti, 2001; Cukierman et al., 2002). Changes to the National Bank Law were passed into law in 2001 in anticipation of the eventual need to meet the Maastricht Treaty requirements, including the importance of the price stability objective (available at http://www.mnb.hu/index-n.htm). It is well-known, however, that statutory autonomy need not translate into actual independence (e.g. see Forder, 2000). As noted earlier, Hungary’s exchange rate regime is technically not of the floating variety so that it does not meet this requirement.20 In principle, however, the ±15% fluctuation band adopted in 2001 comes close enough to a floating exchange rate regime so that pre-requisite may be interpreted as having been satisfied. Since the beginning of the transition to market, arbitrarily dated at around 1989, Hungary has adopted two versions of the fixed exchange rate regime.21 Until March 1995, the government and the National Bank of Hungary followed a policy of pegging the exchange rate with irregular devaluations. In March 1995, a crawling peg was introduced. By 2000, the rate of crawl was nearing zero within a narrow band until it was widened considerably in May 2001.22 It remains to be seen whether the wider 19 Masson et al. (1997, Table 4) rank Hungary the 10th most independent central bank in the 1980s in a list of 38 industrial and developing countries, behind nine industrial countries. 20 NBH (2000a, p. 73) provides a plot showing the rate of devaluation in the HUF and in core inflation (see Siklos, 2000 for a plot of the real exchange rate and NBH (2000b, 2001b) for the schedule of exchange rate devaluations. A referee correctly drew attention to an apparent inconsistency in the manner in which the NBH (as did we in an earlier draft) contrasted the one time devaluations in the forint versus the rate of crawl. The appendix provides additional information about this issue. 21 While all remaining restrictions on capital flows were eliminated as of 18 June 2001, the NBH has not stated any new policy on foreign exchange rate intervention. The new Act stipulates “. . . the NBH shall protect and influence exchange rates on domestic and foreign exchange markets, when necessary and possible.” (Article 11 (3)(2)). 22 According to NBH (2001c, p. 49), the goal is to meet the Maastricht Treaty requirements for entry into ERM II by 2004–2005.

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

9

Table 3 Pre-conditions for inflation targeting? Institutional conditions: are they met?

Status

Comment

Instrument independence?

YES

Effective policy instrument?

YES (in principle)

Accountability?

YES

Adequate disclosure?

YES

Exchange rate regime?

YES

Joint responsibility for monetary policy?

YES

Fiscal-monetary harmony?

YES (in principle)

NBH Law provides sufficient statutory autonomy in this regard The 2 week repo rate is adequate in this connection Prior to June 2001 legislation did not require the President to testify in Parliament and appointment procedures were not conducive to accountability Inflation Report and Financial Stability Report adequately fulfill these requirements Crawling peg. The rate of crawl was close to zero by May 2001. Crawl abandoned in May 2001 in favor of a wider band Conflicts between NBH and Finance Ministry but decisions on exchange rate regime made jointly according to legislation. Recognition in 2001 that responsibility for announcing the inflation target is a joint one Improvement over 1990s but doubts linger

Economic conditions: can they be met?

Option(s)

Price index for targeting?

CPI, eventually some target for “core” inflation

Width, horizon, symmetry of target bands?

±1% around a specified mid-point of target band

Sufficient knowledge about transmission mechanism? Role of inflation forecast?

Dependent on choice of exchange rate regime

Measurement, timeliness issues?

Distribution of risks around some central tendency

Largely met

Differences in measurement and behavior as between core and headline inflation is a potential problem as is the role of regulated prices High inflation variability and poor historical inflation record make flexibility more desirable than otherwise but at the risk of credibility NBH’s own acknowledgement of considerable uncertainties in this regard is problematical Given difficulties over specification of transmission mechanism and thinness of financial markets this is problematical although private surveys of inflation already published and perform adequately based on a very small number of observations. Beginning in August 2001, new inflation forecasts and risk assessments published Hungary has met the IMF’s conditions for the Special Data Dissemination Standards

A qualitative assessment for Hungary. Sources: NBH, IMF, and Siklos (1999a).

exchange rate band will, in an environment with numerous economic shocks and significant capital flows, contribute to excessive exchange rate volatility (Eichengreen et al., 1999). A potential difficulty concerns the nature of the relationship between the central bank and the Ministry of Finance. While the NBH’s policies continued to be firmly guided by the need to generate an orderly disinflation, as we shall see below, differing views

10

Table 4 Inflation persistence and the determinants of inflation expectations, 1993–1999

Constant πt −1 Exchange crawl reg πt−1 “Bokros” effecta deterministic trend Sample N R2adj

Dependent variable: CPI inflation

Dependent variable: core inflation

Dependent variable: inflation expectations

(1)b

(1)b

(1)c

(2)b

1.78 (0.56)∗ 0.84 (0.04)∗ 0.13 (0.04)∗ 0.05 (0.02)∗ −1.96 (0.50)∗

−0.19 (0.29) 0.77 (0.05)∗ 0.09 (0.05)@ 0.12 (0.02)∗

93:01–99:12 84 0.97

95:03–99:12 58 0.99

(2)c

1.26 (0.57) 0.92 (0.03)∗ 0.17 (0.05)∗

13.38 (2.70)∗ 0.73 (0.04)∗ 0.07 (0.05)+

−1.76 (0.46)∗

−0.10 (0.02)∗

93:03–99:12 82 0.98

95:03–99:12 58 0.99

(2)d

1.60 (.84)@ 0.46 (0.13)∗ 0.67 (0.13)∗

4.50 (1.10)∗ 0.11 (0.17) 1.00 (0.19)∗

97:01–00:10 46 0.91

98:01–00:10 34 0.75

Note: data are monthly for the samples shown. (∗) Signifies statistically significant at the 1% (+5%, @ 10%) level of significance. a π reg is inflation in the regulated sector. When core inflation is the dependent variable, lagged core inflation is the relevant independent variable. CPI inflation is 100 times log CPIt − log CPIt−1 , where CPI is the Consumer Price Index published by the NBH. Data for core and regulated prices are also from the NBH. b Estimated via OLS. When included, the “Bokros” dummy is set to 0 until 1995:02 and is 1 thereafter. c Estimated via OLS estimates. A deterministic trend was included to capture the Bokros period instead of a dummy variable. d Instrumental variable estimation. The instruments are: a constant, the government’s announced inflation objective (no such announcement was made in 1997), and πt −2 .

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

Independent variablesa

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

11

between the monetary and fiscal authorities can send the wrong signals to financial markets and the public. It should also be noted that similar conflicts are a feature of the normal tensions between fiscal and monetary policies. Nevertheless, the European Union (2001), in reviewing the state of preparedness for joining the EU among the so-called accession countries, commented that progress was incomplete toward making the NBH sufficiently autonomous of the government.23 Joint responsibility over the choice and defense of the exchange rate regime (Article 11(2)) is problematical, for it means that the size and timing of any realignment must be negotiated as was the case for the rate of devaluation under the crawling peg regime. Moreover, an additional difficulty arises because foreign exchange transactions appear on the books of the NBH. Given the high rate of FDI (see Siklos, 2000) the resulting sterilization is costly for the central bank.24 In principle, some of these issues have been resolved with the central bank law of 2001 but there is as yet insufficient clarity about the scope of central bank intervention in foreign exchange markets. Nevertheless, it must be stressed that previous legal changes that enshrined central bank autonomy did not prevent conflicts from erupting from time to time. As Siklos (1999b) points out, such conflicts can represent defining moments in central bank-government relations. Additionally, there is the matter of the degree of harmony between fiscal and monetary policies. The situation has improved dramatically since the 1980s when a substantial amount of deficit financing was carried out via central bank loans to the government (see, for example, Kiss and Szapáry, 2000; Barabás et al., 1998). Nevertheless, past history suggests fragility in the relationship between the government and the central banks, an understandable outcome of the fact that Hungary has been through a decade of major reforms. Annual “guidelines for monetary policy” published by the NBH relied on estimates of a “target” for inflation that was determined by the government.25 In late summer 2001, inflation objectives were announced jointly by the government and the central bank. Hungary is ostensibly following in the footsteps of the Canadian and New Zealand examples by adopting inflation reduction targets before formally announcing inflation control objectives consistent with the Maastricht Treaty. A difficulty is that previous inflation guidelines were generally not credible as there was no formal requirement that the NBH meet a particular inflation objective. Clarity in this respect, and sanctions, if any, for failing to achieve the stipulated inflation objectives, is essential so that markets and the public do not confuse the old and 23

This view was made public prior to the June 2001 changes to the central bank law. This can be seen most notably in NBH (2000a). Szap´ary and Jakab (2000) estimate the total costs of stabilization to be around 0.41 % of GDP in the period between May 1995 and the end of 1997, which does seem modest. However, they do not consider fully the impact of sterilization on the profits of the central bank (see Siklos, 2000, regarding the treatment of foreign exchange transactions) nor the costs for the attendant loss of central bank autonomy, which are admittedly difficult to quantify when the costs of sterilization are borne by the central bank. Data in Kiss and Szap´ary (2000) reveal that the profits of the NBH, as a percent of GDP, ranged between 0 and 0.5% of GDP. 25 The annual NBH publication outlining monetary policy plans for the following year (“Monet´ aris Politika Ir´anyelvek” [Monetary Policy Guidelines]) mentions an inflation objective—based on government budgetary projections—of 12–13% for 1998, a 9% target for 1999, and 6–7% for 2000. These have never been achieved. In addition, the June 1999 issue of the NBH’s Quarterly Report on Inflationis the first to mention the goal of reducing inflation “permanently” to EU levels. The December 1999 issue of the same publication refers to a “sustainable reduction of inflation” to EU levels, the March 2000 issue expresses the desire to “reduce inflation at a sustainable pace” in order to “achieve price stability”. 24

12

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

new monetary policy strategies and the relative responsibilities of each party in the outcome of the program. Finally, the monetary authority must have at its disposal an effective instrument of policy. In 1999, Hungary changed the maturity of its policy instrument, namely, the yield on the NBH’s deposit facility, to 2 weeks (see NBH, 2000a, p. 76) in order to meet the requirements of having an effective monetary policy instrument. Nevertheless, there have been problems of liquidity that can be traced in part to the restrictions on capital flows at the short-end of the maturity structure that have since been lifted.26 Moreover, the continued large inflows of foreign direct investment, though these have abated somewhat, have required effective foreign exchange sterilization. Siklos (2000) presents empirical evidence in support of this proposition (also see NBH, 2000a). For the most part, the institutional requirements for the adoption of inflation control targets thus appear to have largely been met. 3.2. Economic pre-conditions Fig. 1 plots inflation in the CPI, core inflation,27 and inflation in regulated prices. Reports on inflation in Hungary—the same is true for Poland and the Czech Republic, two recent additions to the list of inflation targeting countries—mention the importance of regulated prices, in large part a legacy of the era of central planning and capable of distorting inflationary developments. Core and headline inflation move reasonably closely together over the sample considered.28 However, in Hungary, the distinction between core and headline concepts of inflation, much discussed and debated in industrial countries, is not yet well established in the public’s mind. The new monetary policy strategy was introduced on 12 June 2001. Targets for headline CPI, as well as a new calendar for the Quarterly Report on Inflation, were also announced. Ambiguity still remains over whether core inflation would be the more relevant indicator of inflationary pressures that the NBH ought to respond to.29 In this connection, one issue that is seemingly idiosyncratic to emerging market economies is the behavior of prices in the traded versus non-traded goods sectors (see below). Emerg26 Taylor (2000) points out that a “deep” long-term bond market, where market expectations of future inflation are formed and are influenced by central bank policies, is an important requirement. Five and 10-year government bonds were introduced in Hungary in the mid to late 1990s though the markets for these instruments remain relatively thin. 27 The NBH and the Central Statistical Office (CSO) each produce their own version of “core” inflation. See NBH (1999). Essentially, the NBH and the CSO core measures differ because the latter index excludes items such as meats, gas and electricity which are left in the NBH’s version of core inflation. The CSO’s measure of core inflation includes 80% of the CPI basket while the NBH’s version includes 91% of the same basket. In the early years of inflation targeting in New Zealand, the RBNZ also published its own measure of core inflation but the task was later transferred to Statistics New Zealand to ensure independent measurement of core inflation. Also, see Valkovszky and Vincze (2000) on problems with measures of core inflation in Hungary. Beginning in 2002 the CSO produces an official measure of core inflation. 28 It is also noteworthy that the distribution of inflation rates in CPI inflation and in core inflation do not show signs of any excess kurtosis (i.e., non-normalities). In contrast, there is excess kurtosis in movements of inflation in regulated prices. 29 While central banks in the industrial world have made a strong case for gearing monetary policy actions on the basis of developments in core inflation, doubts have begun to surface as core and headline inflation began to diverge sharply in late 2000 (see Siklos, 2002).

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

13

Fig. 1. Varieties of inflation rates in Hungary. Source: NBH. Inflation rates are defined in Table 4.

ing market economies such as Hungary are small and open and much of the benefits from policies aimed at raising living standards will originate through the traded goods sector. Productivity growth especially is expected to be far higher in the traded goods sector than in the non-traded sector. As a result, inflation is more likely to be a feature of the non-traded sector. The resulting bias (measured via relative prices in the two sectors) implies that any inflation objective, generally set in terms of some aggregate price index, may not be able to deal adequately with the phenomenon referred to as the Balassa–Samuelson effect.30 The next two requirements are related to each other since the specification of the width, horizon, and treatment of inflation movements inside the target band presuppose a reasonably stable and well-defined transmission mechanism. The role of the transmission mechanism in a small open economy has been emphasized, among others, by Longworth (2000). Although policy makers in Hungary are well aware of the importance of the transmission mechanism (e.g. see NBH, 2000a, pp. 68–78), there have been important changes in the instruments of monetary policy over the past few years. Moreover, the existing transmission mechanism is undoubtedly influenced by the current policy of a narrow band for the crawling peg that ended on 1 October 2001. The new regime might lead to a change of emphasis among the various channels through which monetary policy affects the real economy. Together with developments in the banking sector, there is sufficient evidence to warrant considerable caution concerning the smooth functioning of the transmission mechanism. It should be pointed out, however, that Golinelli and Rovelli (2002) are able to estimate a well-behaved monetary policy reaction function for the 1991–1999 period (see 30 In addition, the existence of such an effect is believed to be sufficient to justify a higher inflation target. There exist a range of estimates on the size of the Balassa-Samuelson effect and many are summarized in Deutsche Bundesbank (2001).

14

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

Fig. 2. Survey-based inflation forecasts errors. Source: NBH. Note: the forecast error is actual CPI less Reuters survey of macro-analysts expectations.

also Section 4 below). They suggest that the NBH’s policy was such that either inflation targeting or a fixed exchange rate would become feasible (see also Szapáry, 1998). Indeed, the NBH’s press release of 24 August 2001 says as much with the option of inflation targeting seemingly the preferred choice.31 If we accept the problematic nature of the state of knowledge about the transmission mechanism, inflation projections32 are then also potentially difficult to generate. Fig. 2 plots the forecast errors based on a survey of inflation forecasts presented in the NBH’s Quarterly Report on Inflation.33 As we shall see in the following section, in the sample considered, these forecasts react strongly to the rate of crawl or the target range of inflation reported in the annual monetary policy guidelines of the NBH. If the rate of crawl represents an implicit targeted inflation rate it is unclear how informative such forecasts are about the NBH’s credibility. It has been emphasized by many (e.g. see Longworth, 2000; Berg, 2000; Bernanke et al., 1999) that a successful inflation targeting strategy involves anchoring inflationary expectations to the targeted inflation rate, preferably publicly announced in a manner described earlier. With the publication of the August Quarterly Report (NBH, 2001a), the NBH has replaced the survey with a “fan” chart around the central projection. The latter is the product of input from a variety of models and other pieces of information (including surveys). 31 Though, under the Maastricht Treaty, the adoption of irrevocably fixed exchange rates against the euro will eventually be necessary if, and when, Hungary joins EMU. 32 Recall that inflation projections refer to a conditional forecast based on unchanged monetary policy instruments, the standard scenario employed by staff in presenting their outlook to senior officials of the central bank. 33 In presenting the survey the NBH “trims” the respondents forecasts to exclude extreme observations. The data are obtained from the Reuters survey of macro-analysts.

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

15

Such projections do not, however, directly address the potential role of uncertainty in the transmission mechanism. For example, the projections issued by the Bank of England in November 2001 reveal a maximum dispersion in the 2-year ahead inflation forecast of between approximately +1.25 and +3.80% (http://www.bankofengland.co.uk/inflationreport/ consrpixnov01large.gif). A similar chart from the NBH, also appearing in November 2001, reveals dispersion in the 2-year ahead inflation forecast from 0.5 to 7.25%, and reveals the difficulties of forecasting inflation in Hungary (http://www.mnb.hu/hungarian/7 sajto/ kozlem/2001/hu011105 1.pdf). The foregoing highlights not only the importance of inflation reduction targets but the manner in which the central bank communicates such breaches and the eventual return of inflation within the target band. Unfortunately, we have very little actual experience with such breaches. We do know, however, that they can have immediate and deleterious effects on central bank credibility (e.g. see Svensson, 2001; Siklos, 2002). Moreover, the NBH will need to face the types of questions that were raised in the recently published report on operations at the Bank of England (Kohn, 2001). These include: whose inflation projection (staff or Monetary Policy Council), what contingencies are to be contemplated (i.e. the caveats), and the dangers of publishing a projection that is based on the assumption, unlikely to be correct, of constant monetary variables (i.e. nominal interest rates and exchange rates). Measurement issues plague all countries, including industrialized economies, since there are many questions surrounding inherent biases in CPI data. But the difficulties are exacerbated in emerging market economies because of the sheer scale of the reforms and transformations that have taken place over the past decade.34 The measurement difficulties around key economic indicators are complicated by changes in the calculation of the currency basket to which the HUF is pegged. The bottom line is that some of the economic pre-conditions for inflation targeting may be somewhat problematical at the present time.

4. The verdict: is Hungary ready for inflation targeting? It goes without saying that any empirical assessment of the impact of a change in policy regime in the present case must confront the fact that the existing data are a reflection of major shifts in policy over the past two decades. Examination of key macroeconomic aggregates, such as CPI inflation, core inflation, inflation in regulated prices, and the real exchange rate, together with the application of a test to detect endogenous structural breaks in each series conclude that all of these series display a break in early 1995.35 More precisely, statistical breaks in the series are found in the months of January or February 1995, that is, in the months just before the so-called Bokros package was introduced in the Hungarian Parliament.36 34 A separate Table (not shown) shows numerous measurement issues that have drawn the attention of the NBH, as reported in the Quarterly Report on Inflation since 1998 alone. 35 A description of the statistical technique that form the basis for these results are not shown to conserve space. See, however, Burdekin and Siklos (1999) for details. A practical consequence of these tests is that, for example, the finding of a unit root in inflation, for the period 1990–2000, does not hold up once an endogenously estimated structural break is permitted. 36 The Bokros package is named after the Minister of Finance who introduced the package in Parliament.

16

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

The discussion in the previous section leads to the conclusion that Hungary is able, if it chooses to do so, to adopt an inflation targeting policy. Whether this strategy is a desirable one remains to be seen. First, the nature of the transitional process, and the express wish to join the EU has meant that, as in other emerging market economies, a policy of real exchange rate targeting (as in Calvo et al., 1995) has been a deliberate aim of monetary policy in several emerging market economies. Existing research has confirmed this to be true of Hungary as well (e.g. see Siklos, 1997b, 2000). Indeed, since at least 1991, it has been frequently stated that real exchange rate targeting is desirable though the objective has never been clearly enunciated (e.g. see Obláth, 1995). However, real exchange rates have certainly not been appreciated at the same rate since the crawling peg was adopted in early 1995.37 The reason, primarily, has to do with institutional rigidities that permeate transitional economies. Prices were liberalized gradually, while the pressure on exporting sectors to compete internationally conflicted with the need to prevent serious balance of payments problems due to the desire to import the goods and services needed to sustain the transition process. In this connection, it is important to note that competitiveness considerations cannot be ignored in the context of accession to the EU, and it is unclear whether such a goal is compatible with a flexible exchange rate or a target zone peg with a wide band. The bottom line, however, is that, in an inflation targeting regime, the central bank cannot simultaneously target the real exchange rate and inflation. Table 4 shows that inflation persistence has been reduced by the policy shift in 1995. Inflation persistence is measured here by the coefficient on lagged inflation.38 Both headline and core inflation is less persistent after March 1995 than before. Moreover, the rate of crawl is seen as a statistically significant contributor to core inflation (as are regulated prices) and there is a slight drop in core inflation over time as captured by the deterministic trend term. These results are consistent with the notion of enhanced monetary policy credibility since early 1995. However, as noted earlier, an important feature of inflation targeting is its role as an anchor for inflationary expectations. Table 4 also provides a little bit of evidence that the survey of inflation expectations appears to be largely driven either by the rate of crawl or the published inflation guidelines. Indeed, if we use instrumental variables, the rate of crawl is able to explain one-for-one movements in inflation forecasts. Previous forecasts then play an insignificant role. However, for reasons already stated, it is not immediately obvious that these results, even putting aside econometric and data related considerations, are consistent with an anchoring of expectations of the kind purported to have succeeded in inflation targeting countries. An important aspect of central bank behavior is whether it is seen as responding appropriately to shocks using an effective instrument of monetary policy. This is true not only for inflation targeting countries, of course. Indeed, it is an argument of this kind that led Taylor (1993) to suggest that while US Fed policy is discretionary, its actions can be viewed through as akin to a rule, since referred to as Taylor’s rule. Taylor did not “estimate” his 37 The standard deviation of first log differences in the real exchange rate is 0.01 for the 90:1–95:2 period, and 0.02 for the 95:3–99:12 sample. Throughout the paper we rely on monthly data. 38 Burdekin and Siklos (1999), and Siklos (1999a), show that an AR(1) model of inflation, suitably augmented by variables capturing “breaks” in the relationship, adequately describes the time series properties of inflation. Also, see references therein for theoretical justifications for such models, and Bleany (2001).

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

17

rule but instead calibrated the coefficients to achieve a reasonable fit with the Fed funds rate, the central bank’s instrument of policy.39 Taylor’s rule can be written Rt = ρ∗ + π¯ t + α(π¯ − π∗ ) + β(yt − y∗ ) + γRt−1

(1)

where γ measures the persistence in interest rate movements arising out of the possibility of smoothing interest rate changes. The first term consists of an “equilibrium” real interest rate (ρ∗ ) and a moving average inflation rate. Hence, this term, with γ = 0, can be interpreted as a target nominal interest rate. π∗ is a targeted or desired inflation rate, and (yt − y∗ ) is the output gap. The coefficients α and β in (1) represent the weights the central bank attaches output and inflation objectives, respectively. Taylor (1993) did not originally allow for interest rate smoothing, and a fully persuasive justification remains the subject of some controversy, with some advocating its presence (e.g. Rotenberg and Woodford, 1999) while others argue against the practice (e.g. Ball, 1999). Rudebusch (2002) argues that interest rate smoothing (i.e. persistence) is an artifact of the reliance, in most studies, on quarterly data in estimating Taylor rule coefficients. Nevertheless, several authors who have calibrated Taylor rules for US data typically assume some interest rate smoothing (e.g. γ = 0.5, as in Rudebusch and Svensson, 1999, or γ = 0.1 as in Favero and Rovelli, 2002). Taylor (1993) is able to obtain good results for the US by setting equal weights on the inflation and output gap objectives. Subsequently, however, the literature that derives or estimates an optimal rule from some structural model finds low values for β relative to α, together with a small coefficient for γ, describe the actual policy stance of major central banks quite well (e.g. Rudebusch and Svensson, 1999; Cecchetti and Ehrmann, 2000). Needless to say, in the case of a country such as Hungary, many of these parameters cannot be precisely estimated for reasons that, by now, should be apparent. In particular, there is considerable uncertainty about the output gap.40 Interested readers can consult Appendix (not shown) for estimates of a Taylor rule for Hungary using monthly data. We turn, therefore, to “calibrated” estimates that may be equally revealing about the conduct of monetary policy in Hungary over the past decade. These are shown in Figs. 3 and 4.41 We consider two cases, namely when inflation and the real exchange rate are in the reaction function, or when inflation and the output gap appear instead in the Taylor rule. The cases are considered separately for at least two reasons. First, as noted already, 39 As Taylor (2000) points out, his advice was meant to provide a normative rule for the conduct of monetary policy. However, in a kind of reversal of what has come to be called “Goodhart’s law”, namely, that once a relation has become known it ceases to hold, Taylor’s rule is increasingly being viewed as a good description of how many central banks actually behave. Policy rules have, therefore, migrated into positive uses. 40 Calculations of the output gap used here imply a growth in potential output of 4.55%. This figure compares favourably with estimates of the NBH (see Darvas and Simon, 2000) and of 4.76% published by the OECD (OECD, 2000, p. 219). In addition, we computed a measure of the output gap from estimates of a structural VAR. The resulting estimates are comparable to the ones from the H-P filter from 1995 onwards. However, the SVAR based output gap reveals either persistently more excess capacity or excess demand than does the H-P in the 1991–1994 period. 41 The calibrations are based on monthly data. Most US studies rely on the quarterly data frequency. In the case of Hungary, quarterly GDP data are not as reliable as the monthly industrial production data (and the sample availability is also much shorter). Moreover, it is likely that, given the speed with which outside shocks are transmitted through the Hungarian economy, policy makers’ horizons are likely to be shorter than the quarterly frequency.

18

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

Fig. 3. A Taylor rule for Hungary: central bank reacts to the real exchange rate and inflation gaps, the transition and post-Bokros phases. Note: the output gap is replaced by the real exchange rate gap in the Taylor rule equation. The resulting Taylor rule equations (see (1)) are: 2+ 12-month moving average of CPI inflation + 0.25 reer gapt + 0.75(π − 12)t + 0.1 Rt−1 for the full sample and 2+ 12-month moving average of CPI inflation + 0.20 reer gapt + 0.80(π − 12)t + 0.1 Rt−1 for the post-Bokros period. The reer gap is the residual of an HP filter applied to the real effective exchange rate with a smoothing parameter of 28,800.

there are good reasons to believe that the monetary authorities in Hungary in the sample considered were primarily concerned about real exchange rate and inflation developments alone. Second, difficulties surrounding the measurement of the output gap for Hungary were also discussed. Consequently, there are good reasons to believe that a real exchange rate gap acts here as a proxy for the output gap, especially as Hungary’s economy is highly sensitive to international developments (e.g. Szapáry, 1998; Neményi, 1997; Siklos, 2000,

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

19

Fig. 4. A Taylor rule for Hungary: central bank reacts to the output gap and inflation the transition and post-Bokros phase. Note: the figures show the actual T-bill rate and the rate implied by a Taylor rule of the form 2+ 12-month average of CPI inflation + 0.35ygapt + 0.65(π − 12)t + 0.1 Rt−1 for the full sample and 2+ 12-month moving average of CPI inflation + 0.30ygapt + 0.7(π − 12)t + 0.1 Rt−1 for the post Brokros period. The variability ygap is the output gap estimate by applying an HP filter to the log of industrial production with a smoothing parameter of 28,800. Also, see Fig. 3 and Appendix (not shown).

and Appendix). While some (e.g. Leitemo, 1999) have specified open economy Taylor rules containing both an output gap and a real exchange rate gap, Clarida et al. (2001) have also shown that a Taylor rule in an open economy setting can have the same form as the closed economy Taylor rule. Indeed, the closed economy form of the Taylor rule seems to depict

20

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

reasonably well interest rate developments in several small open economies (e.g. Cˆoté et al., 2001; Collins and Siklos, 2001). It must, however, be borne in mind that the simulations reflect past economic performance and so the rules cannot be reliably used to forecast the future as such. After considerable experimentation, we set γ = 0.1, that is, there is only modest interest rate smoothing, and the inflation objective is set to 12%.42 The real interest rate target is set at 2% (in line with assumptions for industrial countries) with the nominal interest rate target at ρ∗ plus a 12-month moving average for inflation (again in line with US practice of using a four-quarter moving average. See, for example, Rudebusch, 2002). When the real exchange rate gap is used instead of the output gap in the Taylor rule, the respective weights are, in the full sample, 0.25 for the real exchange rate, and 0.75 for inflation, and, for the post-Bokros period, 0.20 and 0.80, respectively. When the output gap is used in the Taylor rule, weights of 0.35 on the output gap and 0.65 on inflation were applied in the full sample, and 0.3 and 0.7, respectively, for the sample following the introduction of the Bokros package.43 When viewed through the lens of real exchange rate behavior, the NBH has been able to disinflate by placing a relatively larger weight on inflation while slightly reducing the nominal interest rate target, especially in the aftermath of the Bokros reforms. In the case where the output gap appears in the rule, the NBH also appears to have placed relatively greater emphasis on inflation with the desired result, namely, a substantial reduction in nominal interest rates over time, despite raising its nominal interest rate target temporarily around the time the Bokros package was announced (i.e. beginning early 1995). These results illustrate that without an explicit commitment and credibility in the inflation objectives of the government it was simply not possible at first to significantly reduce nominal interest rates. As in other countries with little credibility in fighting inflation, the battle against inflation is won via higher real interest rates, at least for a time. This is even more so when there is no joint commitment to an orderly disinflation strategy between the central bank and the government. The difficulty is in finding ways to influence expectations of inflation in such a way as to permit the central bank to lower the nominal interest rate target in a credible manner. As noted earlier, the crawling peg likely ceased to perform this function and so an inflation targeting policy is one strategy that can achieve the desired results. Note also that, regardless whether the real exchange rate gap or the output gap appears in the central bank’s objective function, the degree to which the NBH’s policy was too loose (i.e. the simulated interest rate is higher than the actual interest rate) is far more modest and eventually disappears following the “regime” change in March 1995. Previously, the gap is large and widens, at least until the middle of 1993. This is consistent with the OECD’s recent assessment of monetary conditions in Hungary (OECD, 2000).However, into 1999, the NBH’s monetary policy begins to be too tight in all the simulations considered. Nevertheless, the simulations also reveal that the NBH was clearly not an “inflation nutter”.44 Instead, it did appear to increase the weight placed on the inflation objective after 42 This is the coefficient used by Favero and Rovelli (2002). Interestingly, this assumption produces a better “fit” vis-a-vis the actual interest rate path than γ = 0.5 suggesting relatively less interest rate smoothing at the monthly frequency than for quarterly data. 43 Appendix (not shown) shows a plot of the implied target nominal interest rate. 44 The term used by the Governor of the Bank of England, Eddie George, to describe a hypothetical central bank that focuses only on the inflation objective to the exclusion of any concern for real economic performance.

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

21

the Bokros package of 1995 but not dramatically so.45 This is true in both the pre and post-March 1995 samples considered. Notice also from Figs. 3 and 4 that the “fit” is fairly good other than shortly after the two major shocks in the sample, namely, the collapse of the CMEA and the immediate aftermath of implementation of the Bokros package.46 We also experimented with an alternative specification (not shown) where the central bank’s inflation objective (␲∗ ) is allowed to decline in a stochastic manner toward an inflation objective of around 7–8% in 1999, as “forecasted” by the Hungarian government. If all other parameters in the simulation are unaffected the implied interest rates are higher than those shown in Figs. 3 and 4 but are roughly the same if we instead set ρ∗ to 3% or higher. In other words, these inflation objectives do not appear to have been credible. One reason such an outcome was not feasible is that fiscal policy was, for the most part, inconsistent with the objective of a lower nominal interest target rate. Another explanation involves concerns over the state of the current account balance and its implications for the real exchange rate. As noted above, these types of pressures on monetary policy, evident over a significant portion of the sample considered, made it difficult to operate an internally consistent inflation-targeting regime.

5. Conclusions The results of this paper suggest that inflation targeting is a viable option for an emerging market economy such as Hungary. A review of some of the qualitative features of the economic performance of the Hungarian economy, and the relationship between the central bank and the government does, however, reveal a few gaps. In particular, the relative responsibilities and expectations of the central bank vis-a-vis the government need clarification and elaboration. Nevertheless, most of the pre-conditions are in place or may be so in the foreseeable future. Hence, Hungary’s decision to adopt the inflation-targeting route is both defensible and generally appropriate. A quantitative examination reveals that the NBH has shifted emphasis toward placing more importance on the inflation objective in recent years. Indeed, application of a technique used to evaluate central bank performance in industrial countries, namely, the Taylor rule, suggests that it may be used to describe some of the policy choices made by the NBH since the early 1990s. Although inflation targeting is seen as a desirable policy option there are additional considerations that must be kept in mind to ensure the success of the chosen monetary policy strategy. First, implementation of monetary policy will undoubtedly be influenced by the process and requirements leading up to membership in the European Union. Second, if the policy of reducing inflation and inflation expectations via inflation targets succeeds, a hard currency strategy will eventually be required since irrevocably fixed exchange rates will be necessary to join the euro area. Inflation targeting is inherently a more flexible policy 45 Golinelli and Rovelli (2002) find some evidence consistent with interest rate setting by the NBH without regard to aggregate demand. We use the same interest rate series as they did. However, they estimate a Taylor rule from the data unlike the calibration exercise considered here. 46 The mean absolute errors in Fig. 3 are 9.42 and 8.36, respectively; in Fig. 4, the MAEs are 7.75 and 7.55, respectively.

22

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

strategy and may therefore be more credible. This might produce conditions conducive to the exchange rate conditions stipulated in the Maastricht Treaty.47 However, stringent fiscal constraints were a necessary ingredient prior to the start of EMU. Therefore, declaring inflation targets is not enough. Credible policies, or institutions, to guarantee good conduct in fiscal policy are also necessary and may be incomplete in the Hungarian case. In addition, markets for long-term debt may not yet be sufficiently liquid to provide guidance about inflationary expectations. Finally, even if fiscal concerns are waning, there remain concerns over the sustainability of current account deficits. Inflation targeting also makes it clear that markets will ultimately decide the “equilibrium” exchange rate level. Whether the irrevocable exchange value around the time Hungary is expected to join EMU is politically acceptable is, of course, another matter. Nevertheless, it should be added that some now argue (e.g. Calvo and Reinhart, 2000) that many countries that claim a “floating” exchange rate have demonstrated a “fear of floating”. What those arguments fail to point out, however, is that inflation targeting within a floating regime represents a coherent policy strategy while simply declaring that exchange rates are permitted to freely float, without announcing a nominal anchor (viz. an inflation objective), does not constitute a coherent strategy.

Acknowledgements This paper was partially prepared while the first author was a Visiting Scholar at the National Bank of Hungary and Wilfrid Laurier University Research Professor for 2000–2001. The financial support of CIDA, under a Public Sector Bridge Program, and the Social Sciences and Humanities Research Council of Canada is gratefully acknowledged. The views expressed in this paper are the authors’ and not necessarily those of CIDA, the National Bank of Hungary, or the International Monetary Fund. Both authors are grateful for the helpful comments from two anonymous referees and the Managing Editor. Some results not shown are relegated to an unpublished Appendix that can be accessed at the link provided below. References Ball, L., 1999. Policy rules for open economies. In: Taylor, J.R. (Ed.), Monetary Policy Rules. University of Chicago Press, Chicago. Bank of Canada, 2001. Revisiting the case for flexible exchange rates. In: Proceedings of a Conference Held at the Bank of Canada, November 2000. Bank of Canada, Ottawa. Barabás, G., Hamecz, I., Neményi, J., 1998. Fiscal deficit and public debt during the transition. In: Bokros, L., Dethier, J.-J. (Eds.), Public Finance Reform During the Transition: The Experience of Hungary. The World Bank, Washington, DC, pp. 59–93. Bernanke, B., Laubach, T., Mishkin, F., Adam Posen, 1999. Inflation Targeting: Lessons from the International Experience. Princeton University Press, Princeton, NJ. Bleany, M., 2001. Exchange rate regimes and inflation persistence. IMF Staff Papers 47, 387–402. Blejer, M., Ize, A., Leone, A.M. (Eds.), 2000. Inflation Targeting in Practice: Strategic and Operational Issues and Application to Emerging Market Economies. International Monetary Fund, Washington, DC. 47 The relevant Maastricht convergence criterion stipulates the “observance of normal fluctuation margins provided for by the exchange-rate mechanism of the European Monetary System, for at least two years, without devaluing against the currency of any other Member State.”

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

23

Bufman, G., Leiderman, L., Sokoler, M., 1995. Israel’s experience with explicit inflation targets: a first assessment. In: Leiderman, L., Svessnon, L.E.O. (Eds.), Inflation Targets. CEPR, London, pp. 169–191. Burdekin, R.C.K., Siklos, P.L., 1999. Exchange rate regimes and shifts in inflation persistence: does nothing else matter? J. Money, Credit and Banking 31, 235–247. Berg, C., 2000. Inflation targeting: the five dish experience. In: Blejer et al. (Eds.), Inflation Targeting in Practice: Strategic and Operational Issues and Application to Emerging Market Economies. International Monetary Fund, Washington, DC. Calvo, G.A., Reinhart, C.M., Végh, C., 1995. Targeting the real exchange rate: theory and evidence. J. Develop. Econ. 47, 97–133. Calvo, G.A., Reinhart, C.M., 2000. Fear of floating. Quart. J. Econ. 177, 379–408. Cecchetti, S., Ehrmann, M., 2000. Does Inflation Targeting Increase Output Volatility? An International Comparison of Policymakers’ Preferences and Outcomes, Working Paper 69. Central Bank of Chile, Chile. Christofferson, P.F., Westcott, R.F., 1999. Is Poland Ready for Inflation Targeting? IMF Working Paper 99/41. Clarida, R., Gali, J., Gertler, M., 2001. Optimal monetary policy in open versus closed economies: an integrated approach. Am. Econ. Rev. 91, 248–252. Coats, W. (Ed.), 2000. Inflation Targeting in Transition Economies: The Case of the Czech Republic. International Monetary Fund, Washington, DC, available at http://www.imf.org. Collins, S., Siklos, P.L., 2001. Optimal Reaction Functions, Taylor’s Rule and Inflation Targets, Working Paper. Wilfrid Laurier University, available at http://www.wlu.ca/∼wwwsbe/faculty/psiklos/home.htm. Cˆoté, D., Kuszczak, J., Lam, J.-P., Liu, Y., St.-Amant, P., 2001. What Have We Learned from This Workshop? Taylor Rule Workshop Held at the Bank of Canada, June 2001, available at http://www.bank-banquecanada.ca/workshop2001/papers.htm. Cukierman, A., 1992. Central Bank Strategy, Credibility, and Independence. MIT Press, Cambridge, MA. Cukierman, A., Miller, G.P., Neyapti, B., 2002. Central bank reform, liberalization and inflation in transition economies—an international perspective. J. Monetary Econ. 49, 235–236. Czech National Bank, 2000. Inflation targeting in the Czech Republic. In: Coats, W. (Ed.), Inflation Targeting in Transition Economies: The Case of the Czech Republic. International Monetary Fund, Washington, DC. Darvas, Z., Simon, A., 2000. Potential Output and Foreign Trade Terms in Small Open Economies. NBH Working Paper 2000/9. Debelle, G., Fischer, S., 1994. How independent should a central bank be? In: Fuhrer, J.C. (Ed.), Goals, Guidelines and Constraints Facing Monetary Policy. Federal Reserve Bank of Boston, Boston. Deutsche Bundesbank, 2001. Monetary Aspects of the Enlargement of the EU. Monthly Report 53, pp. 15–30. Duisenberg, W.F., 2000. Hearing Before the Committee on Economic and Monetary Affairs of the European Parliament, 23 November 2000. Dvorsky, S., 2000. Measuring Central Bank Independence in Selected Transition Countries, Focus on Transition 2/2000. Austrian National Bank, Vienna, pp. 77–95. Eichengreen, B., Masson, P., Savastano, M., Sharma, S., 1999. Transition Strategies and Nominal Anchors on the Road to Greater Exchange Rate Flexibility. Essays in International Finance 213, Princeton University, Princeton, NJ. European Central Bank, 1998. A Stability-Oriented Monetary Policy Strategy for the ECSB, available at http://www.ecb.int/press/pr98013 1.htm. European Union, 2001. Economic and Monetary Union: Applicant Countries and the Community, Hungary, 15 February 2001, available at http://www.europa.eu.int/scadplus/leg/en/lvb/e01103.htm. Falcetti, E., Raiser, M., Sanfey, P., 2000. Defying the Odds: Initial Conditions, Reforms and Growth in the First Decade of Transition, EBRD Working Paper 55. Favero, C., Rovelli, R., 2002. Macroeconomic stability and the preferences of the fed: a formal analysis, 1961–1998. J. Money, Credit and Banking, in press. Forder, J., 2000. Central bank independence and credibility: is there a shred of evidence? Int. Finance 3, 167–185. Golinelli, R., Rovelli, R., 2002. Painless disinflation? Monetary policy rules in Hungary, 1991–1999. Econ. Trans. 10, 55–91. Haldane, A. (Ed.), 1995. Targeting Inflation. Bank of England, London. Hochreiter, E., 1994. Central banking in economies in transition. In: Willett, T.D., Burdekin, R.C.K., Sweeney, R.J., Wihlborg, C. (Eds.), Monetary Stability in Emerging Market Economies. The Westview Press, Boulder, CO, pp. 127–144.

24

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

Hochreiter, E., Riesinger, S., 1995. Central banking in central and eastern Europe—selected institutional issues. ECU-J. 32, 17–22. International Monetary Fund, 2000. Code of Good Practices on Transparency in Monetary and Financial Policies, available at http://www.imf.org. Kiss, G.P., Szapáry, G., 2000. Fiscal adjustment in the transition process: Hungary, 1990–1999. Post-Soviet Geogr. Econ. 41, 233–264. Jonáš, J., 2000. Inflation targeting in transitions economies: some issues and experience. In: Coats, W. (Ed.), Inflation Targeting in Transition Economies: The Case of the Czech Republic. International Monetary Fund, Washington, DC. Kohn, D.L., 2001. The Kohn Report on MPC Procedures. Bank of England Quarterly Bulletin, pp. 35–49. Kornai, J., 1997. The political economy of the Hungarian stabilization and austerity program. In: Blejer, M., Skreb, M. (Eds.), Macroeconomic Stabilization in Transition Economies. Cambridge University Press, Cambridge, pp. 173–203. Krzak, M., Ettl, H., 1999. Is Direct Inflation Targeting an Alternative for Central Europe? Focus on Transition 1. Austrian National Bank, Vienna, pp. 28–59. Leiderman, L., Bufman, G., 2000. Inflation targeting under a crawling band exchange rate regime. In: Blejer, M., Ize, A., Leone, A., Werlang, S. (Eds.), Inflation Targeting in Practice: Strategic and Operational Issues and Applications to Emerging Market Economies. International Monetary Fund, Washington, DC. Leiderman, L., Svensson, L.E.O., 1995. Inflation Targets. Center for Economic Policy Research, London. Leitemo, K., 1999. Inflation Targeting Strategies in Small Open Economies. Norges Bank. Longworth, D., 2000. The Canadian monetary transmission mechanism and inflation projections. In: Blejer, M., Ize, A., Leone, A.M. (Eds.), Inflation Targeting in Practice. International Monetary Fund, Washington, DC, pp. 58–67. Longworth, D., Freedman, C., 1995. The role of the staff economic projection in conducting Canadian monetary policy. In: Haldane, A. (Ed.), Targeting Inflation. Bank of England, London, pp. 101–112. Loungani, P., Sheets, N., 1997. Central bank independence, inflation, and growth in transition economies. J. Money, Credit and Banking 29, 381–399. Mahadeva, L., G. Sterne, 2000. Monetary Policy Frameworks in a Global Context. Routledge, London. Masson, P., Savastano, M.A., Sharma, S., 1997. The Scope for Inflation Targeting in Developing Countries, IMF Working Paper WP/97/130. Mishkin, F.S., 2000. Inflation targeting for emerging market economies. Am. Econ. Rev. Pap. Proc. 90, 105–109. National Bank of Hungary, 1999. Annual Report. National Bank of Hungary, Budapest. National Bank of Hungary, 2000a. Monetary Policy in Hungary. National Bank of Hungary, Budapest, May, available at http://www.mnb.hu/english/4 public/monpol/rw-monet.pdf. National Bank of Hungary, 2000b. Annual Report. National Bank of Hungary, Budapest. National Bank of Hungary, 2001a. Quarterly Report on Inflation, August. National Bank of Hungary, 2001b. Monthly Report, July. National Bank of Hungary, 2001c. A forint Útja az Euróhoz, The Forint’s Path to the Euro, November. Neményi, J., 1997. Monetary policy in Hungary: strategies, instruments and transition mechanisms. In: Monetary Policy in Transition in East and West: Strategies, Instruments and Transition Mechanisms. Austrian National Bank, Vienna. Neyapti, B., 2001. Central bank independence and economic performance in Eastern Europe. Econ. Syst. 25, 381–399. Obláth, G., 1995. Exchange Rate Policy and Exchange Rate Regimes: A Comparison of Experiences in Hungary and Some of the Other Economies in Transition. National Bank of Hungary, Budapest. Organization for Economic Co-operation and Development, 2000. OECD Economic Surveys: Hungary, November. OECD, Paris. Posen, A., 1993. Why central bank independence does not cause low inflation: there is no institutional fix for politics. In: O’Brien, F.R. (Ed.), Finance and the International Economy. Oxford University Press, Oxford, pp. 40–65. Posen, A., 1995. Declarations are not enough: financial sector sources of central bank independence. In: Bernanke, B.S., Rotemberg, J.J. (Eds.), Macroeconomics Annual 1995. MIT Press, Cambridge, pp. 253–274. Radzyner, O., Riesinger, S., 1997. Central Bank Independence in Transition: Legislation and Reality in Central and Eastern Europe. Focus on Transition 1, Austrian National Bank, Vienna, pp. 57–84.

´ / Economic Systems 66 (2003) 1–25 P.L. Siklos, I. Abel

25

Rotenberg, J., Woodford, M., 1999. Interest-rate rules in an estimated sticky-price model. In: Taylor, J.B. (Ed.), Monetary Policy Rules. University of Chicago Press, Chicago. Rudebusch, G., 2002. Term structure evidence on interest rate smoothing and monetary policy intertia. J. Monetary Econ. 49, 1161–1187. Rudebusch, G., Svensson, L.E.O., 1999. Policy rules for inflation targeting. In: Taylor, J.B. (Ed.), Monetary Policy Rules. University of Chicago Press, Chicago, pp. 203–246. Siklos, P.L., 1994. Central bank independence in transitional economies: a preliminary investigation for Hungary, Poland, and the Czech and Slovak Republics. In: Székely, I., Bonin, J.P. (Eds.), The Development and Reform of Financial Systems in Central and Eastern Europe. Edward Elgar, London, pp. 71–98. Siklos, P.L., 1997a. Charting a future for the Bank of Canada: inflation targets and the balance between autonomy and accountability. In: Laidler, D. (Ed.), Where We Go from Here: Monetary Policy and the Future of Inflation Control in Canada. C.D. Howe Institute, Toronto, pp. 101–184. Siklos, P.L., 1997b. The connection between exchange rate regimes and credibility: International evidence. In: Exchange Rates and Monetary Policy. Bank of Canada, Ottawa, Ont., pp. 73–121. Siklos, P.L., 1999a. Inflation target design: changing inflation performance and persistence in industrial countries. Rev. Federal Reserve Bank of St. Louis 8, 47–58. Siklos, P.L., 1999b. US and Canadian central banking: the triumph of personalities over politics? In: Holtfrerich, C.-L., Reis, J., Toniolo, G. (Eds.), The Emergence of Modern Central Banking from 1918 to the Present. Ashgate Press, Aldershot, UK, pp. 231–278. Siklos, P.L., 2000. Capital flows in a transitional economy and the sterilization dilemma: the Hungarian experience 1992–1997. J. Pol. Reform 3, 373–392. Siklos, P.L., 2002. The Changing Face of Central Banking: Evolutionary Trends Since World War II. Cambridge University Press, Cambridge. Šm´ıdkova, K., Hrn´ır, M., 2000. Disinflating with inflation targeting: lessons from the Czech experience. In: Coats, W. (Ed.), Inflation Targeting in Transition Economies: The Case of the Czech Republic. International Monetary Fund, Washington, DC. Svensson, L.E.O., 1999. Monetary policy issues for the eurosystem. In: Proceedings of the Carnegie–Rochester Conference Series on Public Policy 51. North-Holland, Amsterdam, pp. 79–136. Svensson, L.E.O., 2001. Independent Review of the Operations of Monetary Policy in New Zealand: Report to the Minister of Finance. New Zealand, Wellington, available at http://www.monpolreview.govt.nz. Szapáry, G., 1998. Exchange rate policy in transition economies: the case of Hungary. J. Comparative Econ. 26, 691–717. Szapáry, G., Jakab, Z., 2000. Maastricht and the Choice of Exchange Rate Regime in Transition Countries During the Run-up to EMU, NBH Working Paper 2000/7 (Revised). Tarkka, J., Mayes, D., 1999. The Value of Publishing Official Central Bank Forecasts. Bank of Finland Working Paper 22/99. Taylor, J.B., 1993. Discretion versus policy rules in practice. In: Proceedings of Carnegie–Rochester Conference Series on Public Policy, vol. 39. pp. 195–214. Taylor, J.B., 2000. Using Monetary Policy Rules in Emerging Market Economies. Stanford University, Stanford. Temple, J., 2000. Inflation and growth: stories short and tall. J. Econ. Surveys 14, 395–426.