From inflation targeting to the euro-peg

Economic Systems 25 (2001) 233–251

From inflation targeting to the euro-peg A model of monetary convergence for transition economies Lucjan T. Orlowski∗ Sacred Heart University, 5151 Park Avenue, Fairfield, CT 06432, USA

Abstract This study proposes a sequence of monetary convergence to the eurozone, based on autonomous monetary policy rather than on an early application of the euro-peg. The gradual adjustment process begins with a relatively strict variant of inflation targeting, followed by flexible inflation targeting, and ends with exchange rate targeting. A model outlining the optimal mode of policy adjustment is presented. The analysis warns against a premature peg to the euro, which may instigate real currency appreciation, large capital inflows and their costly sterilization. The euro-peg can be introduced only when the candidates’ monetary authorities reach a certain degree of “foundational credibility”. The model of monetary convergence is followed by the empirical assessment of inflation targeting in the Czech Republic and Poland. © 2001 Elsevier Science B.V. All rights reserved. JEL classification: E32; E52; P33 Keywords: Inflation targeting; Monetary convergence to the eurozone; Transition economies; Monetary policy credibility

1. Introduction The main objective of this study is to design a proper sequence of adjustments in the monetary systems of the candidates for accession to the European Union (EU) and, at a later stage, to the European Monetary Union (EMU). A key assumption for the presented analysis is that the candidates will be well advised to achieve monetary convergence by applying autonomous monetary policies based on direct inflation targeting (DIT). Such an approach will enable them to accomplish price stability and to enhance policy credibility by internal means. I argue that a premature peg to the euro ∗ Tel.: +1-203-371-7858; fax: +1-203-230-9860. E-mail addresses: [email protected], [email protected] (L.T. Orlowski).

0939-3625/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 3 9 - 3 6 2 5 ( 0 1 ) 0 0 0 2 0 - 6

234

L.T. Orlowski / Economic Systems 25 (2001) 233–251

would cause more damage than benefit for the central European transition economies (TEs) that wish to enter the EU and EMU. The candidates for the EU/EMU accession need to achieve price stability prior to pegging their currencies to the euro. Otherwise, the euro-peg accompanied by inflation persistently higher than that of the eurozone would trigger real appreciation of their currencies and aggravate their balance of payments position. The recommended sequence of monetary convergence begins with a relatively strict inflation targeting regime, which will allow central banks to focus on reducing inflation to the pre-determined trajectory of disinflation. Ultimately, this trajectory is geared to the degree of price stability that is indispensable for accession to the eurozone. At the same time, central banks are expected to focus on lowering inflation variability. After achieving initial gains in price stability, the candidates are advised to switch to a more flexible variant of inflation targeting, which will gradually permit them to increase the emphasis on exchange rate stability. At the final stage of the monetary convergence process, inflation targeting will be replaced with exchange rate targeting — the euro-peg — that will secure a smooth transition to the eurozone. The analysis begins with the presentation of key underlying policy assumptions in Section 2. On their basis, a dynamic model of policy adjustments from direct inflation targeting–exchange rate targeting (DIT–ERT model) is advanced in Section 3. The theoretical framework is followed by an examination of the empirical record of monetary policy adjustments in the Czech Republic and in Poland, which is presented and reviewed in Section 4. The concluding Section 5 identifies selected institutional and structural conditions for the successful future monetary convergence.

2. Policy assumptions In preparation for accession to the EU and, at a later time, to the EMU, central and eastern European TEs need to design an effective program of monetary convergence. They need to intensify a final attack on inflation, and to develop immunity against external price shocks and contagion effects of world financial crises. Before proposing a model of monetary convergence to the EU/EMU, it is essential to make several assumptions reflecting the conditions and the institutional advancement of monetary policies in TEs. A critical assumption for this analysis is that the EU/EMU candidates will be well advised to rely on internal means of achieving price stability. In other words, they will be better off pursuing disinflation through autonomous monetary policies with relatively flexible exchange rates rather than by rushing to the euro-peg. 1 They need to overcome internal sources of inflation, such as nominal indexation of wages and prices, as well as to reduce structural deficiencies associated with insufficient competition in the private sector. If the candidates apply an euro-peg prematurely, the prolonged high inflation will instigate real currency appreciation that will ultimately lead to growing current account deficits and to an unfavorable risk structure of capital inflows (the advantage of short-term over long-term capital inflows). A strong real appreciation stemming from a premature euro-peg may 1 A support for monetary convergence that begins with the introduction of flexible exchange rates can also be found in Masson (1999) and in Hölscher and Vinhas de Souza (2000).

L.T. Orlowski / Economic Systems 25 (2001) 233–251

235

question viability and sustainability of the peg. This is because the EU is unlikely to allow any deviations from free mobility of capital, which will make exchange rate pegs fragile after the EU entry. The subsequent risk of currency devaluation may induce investors to shorten maturity of financial investment in these economies, and thus, contribute to an unfavorable risk structure of capital inflows and to a more pronounced vulnerability to financial crises. These dangers are perhaps the main reason for flexible exchange rates to appear superior to fixed exchange rates at the present stage of economic transition in more advanced central European economies. In essence, the advanced stage of economic transition is fundamentally different from the initial stage in the early 1990s, in which the currency peg strategy was justifiable and perhaps indispensable. The rationality of fixed exchange rates at the initial stage of economic transition stemmed from the necessity to “borrow monetary credibility” from abroad in order to mitigate corrective inflation, which was instigated by price liberalization and by the subsequent elimination of the monetary overhang (the excess supply of money). 2 Since the early years of transition, most of the advanced reformers in central and eastern Europe have developed independent, autonomous monetary policies. Concurrently, they have departed from fixed exchange rates by applying various “exit strategies” at different times with a different intensity. By contrast, in preparations for accession to the EU and, later, to the eurozone, they will have to design a prudent strategy of “returning to a currency peg”. However, they will be well advised to switch to the euro-peg gradually, and only after establishing price stability by using internal means, that is, through autonomous monetary policy. In other words, the EU/EMU candidates need to establish a “foundational credibility” of monetary policy, which is understood as a fundamental trust for the strategies of the central bank and for tactics based on a proven, successful policy record. If such credibility is not established, a premature declaration of the euro-peg is likely to lead to speculative attacks on domestic currency since financial markets will question sustainability of the monetary regime. Another critical assumption for this analysis is that the final attack on inflation in TEs can best be facilitated by a forward-looking and, at the same time, more transparent monetary policy, rather than by a backward-looking and highly discretionary policy. Central banks in TEs need to engage in a close and transparent dialogue about policy strategies and tactics with the private sector, in particular with financial institutions, so that the private sector expectations are consistent with the official policy goals. They also need to consult their policy with the legislature and with the fiscal authorities in order to gain their attention and support for price stability. In addition, they need to demonstrate convincing and realistic targets of future inflation that are verifiable with credible inflation forecasts, so that private sector expectations and the mechanisms of nominal indexation are geared to lower expected inflation. This assumption reveals attractiveness of direct inflation targeting as a particularly useful monetary regime for more advanced economies in transition, since it is a forward-looking policy, which is predominantly aimed at reducing inflation expectations. As defined, among others, by Svensson (1999a) and Orlowski (2000a), DIT is a policy framework that is based on a pre-announced inflation target, on transparent strategies and 2 A comprehensive description of the early reform and the exchange rate strategies in former communist countries can be found, for instance, in Bruno (1992) and Sachs (1996).

236

L.T. Orlowski / Economic Systems 25 (2001) 233–251

tactics leading to its accomplishment, and on the policy implementation responding to periodic deviations between inflation forecasts and the inflation target. 3 Therefore, in my view, the operational framework of DIT is based on inflation forecast targeting, rather than on responses to historical data on monetary variables. DIT is believed to play a critical role in diminishing the scope of nominal indexation in TEs (Orlowski, 2000a). The message of gaining a “foundational credibility” through an autonomous, forwardlooking monetary regime is also consistent with the argument advanced by von Hagen (1999). He proves that the ability to move from a discretionary to a rules-based monetary regime reflects advanced maturity of the central bank and of the domestic monetary system. This statement allows for the argument that a premature declaration of the euro-peg by TEs could be interpreted as an official confession of the failure to reach price stability through autonomous monetary policy. This perception would not promote stability, and, moreover, it could hinder institutional development of financial markets. A successful implementation of a forward-looking monetary policy depends on advanced, in-depth knowledge of the policy environment and mechanisms (Mishkin, 2000). Thereby, a forward-looking DIT regime is not feasible in TEs at an early stage of transition due to the existence of numerous technical obstacles to gaining such knowledge. It is imperative that monetary authorities have the necessary knowledge about the relative importance and the prevailing directions of monetary transmission channels. More specifically, a forward-looking monetary policy relies strongly on the rational expectations channel of monetary transmission and less on the aggregate demand and on the exchange rate channels. Therefore, central banks need to become familiarized with optimal policy lags and causal interactions between monetary variables so that they can develop a reliable inflation forecast for the targeted period. Preferably, the official inflation forecast shall be conditional upon stable interest rates or stable exchange rates depending on the relative impact of these two monetary variables on inflation. Having a reliable forecast will allow for accurate monitoring of actual deviations between the inflation forecast and the inflation target. Continuously, observed and revised deviations between the forecast and the target of inflation become a key monitoring device of a forward-looking monetary policy. They are the basis for determining future policy bias by a central bank. In essence, the policy bias becomes a key instrument of DIT and it enables the central bank to gear inflation expectations to the intermediate inflation target in a more effective way than the traditionally applied adjustments in target interest rates. Under the alternative policy regimes, particularly those based on monetary or interest rate targeting, central banks use operational interest rate targets that are altered in response to the recently observed inflation data and other monetary variables. In TEs, a continuation of the traditional, highly discretionary monetary policy accompanied by frequent, mostly unannounced changes in interest rates is likely to be less effective for lowering inflation expectations and for stemming nominal indexation than forward-looking regimes are. Using the policy bias is effective for suppressing temporary external price shocks, such as those affecting central European economies during the 1999/2000 period. Under such conditions, the tightening bias announcement is 3

According to the approach by Svensson (1999a), inflation targeting is based on inflation forecast targeting mechanism. Therefore, this definition of DIT is more restrictive than a broader, more general approach by Bernanke et al. (1999).

L.T. Orlowski / Economic Systems 25 (2001) 233–251

237

likely to translate, with a relatively short impact lag, onto lower inflation expectations and, subsequently, it is likely to curtail nominal indexation. In this sense, an announcement of the future tightening bias may relieve the central bank from raising interest rates. Therefore, it is likely to result in lower output losses and, hence, in lower social costs of disinflation. However, a future policy bias becomes a superior substitute for the actual interest rate increase only if monetary policy is credible, that is, if the “foundational credibility” has already been established. Otherwise, under low credibility conditions, a tightening bias announcement may question reliability of future policy intentions, thereby leading to higher instead of lower inflation expectations within the private sector. In order to make the announcements of future policy actions credible, the central bank needs to develop a transparent system of monetary policy communication. A well-conducted forward-looking monetary policy requires that the policy goals, strategies and tactics are fully explained to the private sector, particularly to financial institutions, and also to the legislature and to the public at large. A fully communicated policy helps rationalize the expectations channel of monetary transmission. It will ultimately result in a better public perception of possible, temporary price shocks and failures to reach inflation targets. The public will accept the central bank missing the target as long as it understands the temporary nature of inflation shocks and as long as it continues to believe that the long-term target is still achievable. The public trust can only be developed if the policy is transparent and adequately communicated. In sum, it is assumed that the monetary authorities of the candidate countries begin preparations for the EU/EMU integration by applying a forward-looking, transparent and autonomous monetary regime with flexible exchange rates. Such an approach will enable them to facilitate the necessary monetary convergence to the eurozone, thereby establishing the environment for a low, sustainable and predictable inflation. By contrast, a premature application of the euro-peg or a full “euroization” of their national economies would essentially indicate a failure of the domestic monetary system to achieve the required degree of convergence. In my judgement accomplishing price stability and building “foundational credibility” through internal means is likely to facilitate the process of monetary convergence more effectively than “borrowing credibility” from the eurozone through the euro-peg. Moreover, an internal pursuit of convergence through a forward-looking monetary policy with flexible exchange rates is likely to foster the institutional advancement of central banking in TEs as well. Autonomous central banks will have to develop sophisticated analytical tools of economic environment and to apply indirect instruments of policy implementation. A premature currency peg in the presence of high inflation stemming from internal structural and systemic deficiencies will inadvertently lead to real currency appreciation and will ultimately aggravate balance of payments disequilibria. Taking into consideration the advantages of a forward-looking monetary regime for TEs, I envision the following sequence of systemic changes in preparation for the EU/EMU accession. At the initial stage, the candidates apply a DIT monetary regime. This will underpin their commitment to disinflation and it will intensify their final attack on domestic inflation stemming from nominal indexation. A sophisticated, forward-looking DIT regime can be applied when the central banks of the candidates have established a foundational credibility of monetary policy. They also need to develop sufficiently deep and efficient financial markets which will permit effective policy implementation and provide signals about the

238

L.T. Orlowski / Economic Systems 25 (2001) 233–251

policy effects (Mishkin, 2000). When inflation falls to a low, stable and predictable level at a later stage of monetary convergence, central banks of the candidates may consider gradually switching to the euro-peg. Incidentally, the announcement of the euro-peg may coincide with the actual EU accession. This will enable the new EU members to enter the ERM II mechanism. However, if the satisfactory monetary convergence is achieved earlier, the accession candidates may still apply a shadow peg to the euro. Preferably, the ERM II or the shadow peg will have, at least initially, a wide band of permitted exchange rate fluctuations, which will allow the candidates to preserve monetary independence and to focus on the last moment adjustments to monetary convergence. Exchange rate targeting will be gradually introduced at the final stage of active preparations for the EMU entry, by narrowing down the band, thus, diminishing permitted currency fluctuations in euro terms and by simultaneously withdrawing from inflation targeting. 4 Nevertheless, the euro-peg needs to be applied in the final period of monetary convergence to the EMU. Such an approach would demonstrate a strong commitment of the candidates to get ready for accession to the eurozone. It is also likely to reduce financial instability in the final period. By comparison, an abrupt switching from inflation targeting to the full euroization might lead to volatile capital flows and to instability of financial markets since the radical change in the policy regime would question feasibility of entry to the eurozone.

3. DIT–ERT model In essence, the assumed scenario of monetary policy adjustments favors approaching the euro-peg through the initial application of an independent, forward-looking DIT regime. I am not in favor of a radical leap to euroization without a proper institutional preparation of the domestic monetary system. At the initial stage of the prescribed adjustment process, central banks need to focus on a proper design of the DIT system. Useful suggestions for outlining a prudent DIT regime are provided by King (1996), who suggests that policy makers need to pay attention to two factors: (1) the optimal speed of disinflation; and (2) the speed of reactions offsetting temporary shocks to inflation. Both of these factors are relevant and important for the economies in transition. Designing the optimal speed or “trajectory” of disinflation means choosing the long-term goal of inflation and the corresponding intermediate midpoint targets. A program of the EU/EMU monetary convergence may play a pivotal role in this respect. The determination of an optimal trajectory of disinflation should be consistent with the monetary convergence to the eurozone, which in turn needs to be synchronized with the institutional advancement of the domestic monetary system. The timing of the long-term target is critical for the trajectory of disinflation to be consistent with private sector inflation expectations corresponding to the change in the policy regime. Specifically, if the candidate country chooses a steep trajectory of disinflation, e.g. a short period of monetary convergence, without a corresponding institutional and structural advancement of the monetary system, the program of monetary 4 A formal model of these dynamic monetary policy adjustments of the EU/EMU candidates is presented by Orlowski (1999).

L.T. Orlowski / Economic Systems 25 (2001) 233–251

239

convergence may suffer from the lack of credibility. Moreover, the speed of disinflation is strongly influenced by fiscal discipline. A rapid disinflation enforced with high interest rates in the presence of large budget deficits is likely to result in high social costs of disinflation and, ultimately, in the erosion of public support for the monetary convergence. Therefore, the timing of monetary convergence should be realistic. It should not be too long as it would, thus, lose its function as a tool of fiscal and monetary discipline. Neither should it be too short, since it needs to be synchronized with structural and institutional adjustments affecting domestic inflation. In fact, the speed of disinflation needs to be consistent with dynamic structural adjustments in TEs, including their increasing competitiveness and flexibility of labor and capital markets. In summary, a steep trajectory of disinflation accompanied by serious structural and institutional deficiencies may seriously undermine credibility of monetary convergence to the EU/EMU. The task of determining an optimal speed for reactions offsetting temporary shocks to inflation, and also shocks to exchange rates, is equally complex. Diminishing volatility of inflation and exchange rates is critical for achieving a smooth convergence to the eurozone. As suggested in the previous section, TEs that are moving from moderate to low inflation will be well advised to focus on lower volatility of inflation first. This will allow them subsequently to shift the focus to diminishing exchange rate volatility at a later stage of the convergence process. 5 The speed at which temporary inflation shocks are accommodated is directly related to the width of the tolerance band around the midpoint target of inflation. Therefore, it determines the inflation targeting regime’s flexibility. The adequate inflation tolerance band is particularly relevant for TEs because their small size and their high degree of openness make them vulnerable to external price shocks, as well as to the contagion effects of financial crises and speculative currency attacks. A large exposure of these countries to external shocks calls for a wider tolerance band. In addition, the band needs to be wide enough to allow for a sufficient policy decision lag. Central banks need to have enough time to decide whether the actually observed shocks are self-correcting or more permanent. They are likely to respond with policy tightening if the observed shocks are believed to cause permanent danger to economic stability. In particular, the central banks in TEs will appreciate having a longer policy decision lag because of their limited knowledge about the nature of inflation shocks and about the actual timing of monetary policy lags. By contrast, a narrower band imposes rigid constraints on the degree of discretion in policy decisions. It offers some important advantages at the early stage of DIT because it reaffirms commitment to reducing volatility of inflation and it strengthens the policy discipline. However, a wider band is more beneficial when monetary policy in a transition country becomes more credible and transparent, and the long-term price stability becomes predictable, because it allows for expanding the policy decision lags and its overall flexibility. A wider band also provides more leverage against external price shocks and possible speculative attacks on domestic currencies. The policy emphasis on stabilizing inflation as opposite to stabilizing the exchange rate is demonstrated by my DIT–ERT model in Fig. 1. It represents a specific central bank loss 5 A theoretical framework and empirical investigation of problems encountered by countries moving from moderate to low inflation can be found, for instance, in Fischer et al. (2000).

240

L.T. Orlowski / Economic Systems 25 (2001) 233–251

Fig. 1. DIT–ERT model.

function that reflects policy choices for the economies pursuing monetary convergence to a common currency system, or in the case of central and eastern European TEs, to the eurozone. 6 The DIT–ERT model shows a policy choice between diminishing inflation variability and exchange rate variability. The inflation variability Itπ is understood as a difference between the inflation forecast Πt+T /t (et−1 ) for T-periods ahead formulated at time t, which is conditional upon recently observed exchange rate stability, and the inflation target π ∗ : 7 Itπ = Πt+T /t (et−1 ) − π ∗

(1)

The inflation variability indicator Itπ is a crucial policy monitoring devise in the DIT system. Its high positive value calls for an actual policy tightening or, alternatively, for an announcement of a future tightening bias, while its negative value gives the central bank an opportunity to lower the interest rates. However, central banks may choose to maintain relatively high interest rates even if Itπ is low or negative in order to underscore their anti-inflationary policy stance and to reinforce commitment to long-term price stability. This approach will also provide an effective cushion for absorption of unexpected external price shocks. The alternative policy option within the DIT–ERT framework is the one that minimizes the exchange rate variability et viewed as a difference between the target exchange rate E∗ (π ∗ ) consistent with the inflation target π ∗ and the actual nominal exchange rate Et : et = E ∗ (π ∗ ) − Et

(2) E∗ (π ∗ )

does not have to be consistent with the It can be noted that the target exchange rate ultimate nominal exchange rate convergence target E¯ that will be determined for a candidate 6

An illuminating examination of alternative central bank reaction functions is presented by Svensson (1999b). This approach to inflation variability is similar to that of Svensson (1999b). In my analysis, the inflation variability is conditional upon exchange rate stability, as this choice is more suitable for policy dilemmas faced by the candidates preparing for the eurozone accession. 7

L.T. Orlowski / Economic Systems 25 (2001) 233–251

241

acceding to the eurozone. The accession target E¯ is likely to be affected by a broader range of factors than the purchasing-power-parity exchange rate consistent with the desirable price stability. These additional considerations may include balance of payments conditions and a perceived institutional integrity of the financial sector. The target exchange rate for the eurozone entry will likely be negotiated between the EU and the candidates and set at a level that will not damage domestic financial markets by causing large capital outflows. Therefore, there might be a discrepancy between the long-term goal of price stability and ¯ the EMU accession exchange rate target E. The trade-off between the inflation variability and the exchange rate variability allows for establishing a central bank’s loss function: Lt = 21 [(Itπ )2 + ϕ(et )2 ]

(3)

where ϕ is the preference parameter for the exchange rate stability. In the examined scenario, ϕ is treated in dynamic terms as ϕ > 0, which is a smoothing objective indicating a gradual shift of the policy emphasis from inflation stability to exchange rate stability. Considering the dynamic scenario outlined in the previous section, a central bank is expected to begin the monetary convergence process from paying more attention to the inflation variability. Therefore, it will initially apply a strict inflation targeting (SIT) regime at the early stage of the convergence process, which is reflected by point A in Fig. 1. Within the SIT framework, a central bank focuses strictly on minimizing the observed differences between the inflation forecast and the inflation target over time and it permits the exchange rate to fluctuate indefinitely, thus, et2 = ∞ and ϕ = 0. Consequently, the SIT approach is based on a narrow inflation tolerance band with a fully floating exchange rate. Such a strict adherence to inflation targeting is aimed at enhancing the policy credibility or, in other words, at building a foundational credibility over time. The initial policy credibility is assumed to be very low, as reflected by the most distant from the origin (the worst) credibility locus L2 . The ultimate goal of monetary convergence to the eurozone is consistent with achieving the highest credibility scenario L0 that will allow for achieving a well synchronized, simultaneous price stability and exchange rate stability, fully comparable to the monetary conditions prevailing in the eurozone. The process of moving from inflation targeting to exchange rate targeting is reflected by the path leading from point A to any point between 0 and B, subject to the achieved level of price stability and policy credibility. In the absence of credibility gains, a gradual departure from SIT to exchange rate targeting (ERT) will follow the linear locus L2 . 8 In this case, the central bank’s reaction (loss) function will follow a linear relationship and the alternative costs of moving between both policy regimes will be very high. But realistically, the DIT regime is likely to generate some credibility improvement, at least to the level L1 . If this happens, the monetary authority may consider switching to a flexible inflation targeting (FIT) regime. The initial policy adjustment will follow the path from A to C . At the starting point A, the central bank applies the SIT regime. This results in a high 8 This scenario is consistent with the analysis presented by De Grauwe (1996), who argues that inflation targeting and exchange rate targeting regimes are comparable in terms of credibility improvements. However, he views inflation targeting as a superior policy since exchange rate pegging invites speculative attacks on domestic currency that make it less sustainable than inflation targeting.

242

L.T. Orlowski / Economic Systems 25 (2001) 233–251

exchange rate risk premium that is build at the cost of sacrificing exchange rate stability. The strong emphasis on disinflation improves credibility to the level L1 . Once the credibility improves, the central bank moves to FIT and applies a mix of inflation and exchange rate targeting at point C as an intermediate step to introducing a currency peg. Moving to any point between 0 and B, subject to further credibility gains, reflects a gradual introduction of exchange rate targeting. Point A marks the ideal monetary convergence that qualifies the country for joining a stable common currency area (the eurozone), while point B implies that the monetary convergence has failed to improve credibility relative to the initial conditions at A. The failure to improve credibility may stem from high sterilization costs of capital inflows resulting from a currency peg at point B. This indicates a premature move to the peg, particularly in the presence of costly sterilization of large capital inflows that are predominant during the period of active structural adjustments and “catching up” effects of economic growth in transition economies. 9 Thereby, a successful monetary convergence to point 0, essential for a smooth entry to the eurozone, directly depends on solid gains of monetary policy credibility. In their absence, monetary policy in a transition country will shift from a high exchange rate risk premium at A to equally unfavorable high sterilization costs at B. A critical issue in the prescribed transition process is the timing of the introduction of FIT. This policy alternative has considerable advantages over SIT. In practical terms, it is characterized by a wider tolerance band around the inflation target, which allows for more attention to be devoted to the exchange rate stability. As suggested before, a wider inflation tolerance band entails a longer decision lag of policy reactions to the observed inflation shocks, thus, making monetary policy less discretionary, more long-term oriented, based on dynamic rules. Ultimately, it elongates the time horizon of the policy operational targets. Therefore, FIT, being a more advanced and complex system, demonstrates a higher degree of monetary policy maturity. Within the FIT framework, ϕ assumes a positive value, which indicates that a central bank is ready to adopt a supporting exchange rate target in addition to the main inflation target π ∗ . By applying a mix of inflation and exchange rate targeting, it is willing to address possible balance of payments problems and it demonstrates readiness for a “return-to-peg” strategy in the context of final preparations for accession to the eurozone. A comprehensive, well-designed FIT permits to launch active preparations for monetary convergence to the EMU, assuming that it is preceded by establishing a foundational credibility at the level L1 . The timing of the FIT introduction may be consistent with the official entry to the ERM II, however, it may also be applied earlier in the presence of a shadow peg to the euro. Under FIT conditions and in the presence of credibility gains, the actual combination of inflation and exchange rate variability ends up at point C . Moving beyond C — preferably to point 0 rather than to B — denotes paying more attention to the exchange rate pegging rather than to price stability. This means the end of a DIT regime. It becomes apparent in the DIT–ERT scenario that the efficiency of the final stage of convergence to the eurozone depends on achieving gains in monetary credibility at first. These gains are unequivocally consistent with the institutional development of the country’s financial system and are directly related to the advancement of transparency of both monetary and fiscal policies. These factors, being critical to the overall success 9

For a comprehensive analysis of “catching up” effects, see Warner (2001).

L.T. Orlowski / Economic Systems 25 (2001) 233–251

243

of monetary convergence to the eurozone, may well be included among the criteria of macroeconomic stability and monetary convergence prescribed for the EMU accession candidates.

4. The Czech and the Polish DIT regimes: preliminary empirical evidence Among the official candidates for the EU accession, only the Czech Republic and Poland have begun a process of monetary convergence that seems to follow our DIT–ERT scheme. In contrast, Hungary hopes to achieve monetary convergence to the EU/EMU based exclusively on exchange rate stability. The Hungarian monetary policy is centered on stabilizing the Hungarian Forint (HUF) in EUR terms within a 2.25% band of permitted fluctuations around the central parity rate. Hungary’s emphasis on the exchange rate stability results in a strong reliance on the exchange rate channel of monetary transmission, which is prevalent in the Hungarian monetary system. The exchange rate channel is rather disabled in Poland and in the Czech Republic, as proven empirically by Orlowski (2000b). The prevailing causal interactions between exchange rates and inflation in these two countries are reversed, in the sense that changes in inflation have a strong impact on exchange rates, contrary to the interactions in the Hungarian system. 10 This is consistent with the present monetary and exchange rate regimes in the Czech Republic and in Poland. The Czech National Bank (CNB) has adopted a net inflation targeting system accompanied by a managed float since January 1998, and the National Bank of Poland (NBP) has been following a CPI inflation targeting scheme since January 1999. The NBP allows the PLN to fluctuate freely after abandoning a wide band of permitted currency fluctuations in April 2000. In essence, the present monetary regimes of the Czech Republic and Poland roughly match the conditions of an early stage of policy adjustment following our DIT–ERT scheme. A summary of monetary regimes of these countries is presented in Table 1. The Czech DIT system emphasizes targeting net inflation based on changes in unregulated prices of goods and services that currently constitute 82% of the Czech CPI basket. However, adjustments in governmentally regulated prices (e.g. energy, transportation, etc.) continue to have strong feedback effects on the overall costs of production and on net inflation. Recognizing their importance, as of 2001, the CNB has decided to simultaneously emphasize CPI targeting. The Czech trajectory of disinflation can be interpreted as very cautious, taking into consideration the ultimate inflation goal and the actual levels of net inflation. The long-term net inflation target has been set at 2% for the year 2005, with a tolerance band of 1% on either side of the midpoint target (while the actual net inflation at the end of 1999 reached 1.5% already). The long-term inflation goal is perceived as a target level of monetary convergence to the eurozone, assuming that the Czech Republic will join the EU in 2003 and the EMU 2 years later. The CNB began the DIT regime with a SIT variant by applying a narrow 0.5% inflation tolerance band around the midpoint target. However, the actual implementation of the new monetary regime seems to be asymmetric. In each of the 3 years (1998–2000) the actual 10 An exhaustive econometric examination of the exchange rate channel of monetary transmission in Hungary is presented by Golinelli and Rovelli (2000).

244

L.T. Orlowski / Economic Systems 25 (2001) 233–251

L.T. Orlowski / Economic Systems 25 (2001) 233–251

245

net inflation was consistently well below the lower bound of the inflation tolerance range. This implies that the lower bound is not as important as the upper one. By assuming such an approach, the CNB underpins its determination to fight excessive inflation by all means. At the same time, it is comfortable with showing a readiness to cushion temporary inflation shocks in case actual net inflation falls below the target range. Its determination to bring net inflation down to the assumed long-term trajectory was very pronounced in 1998, when net inflation was running above the target for most of that year. The CNB responded with a very tight monetary policy that generated high real interest rates and ultimately contributed to the economic recession of 1998/1999. Although this was meant to manifest the CNB’s strong determination to fight inflation, the magnitude of monetary contraction was excessive (unfortunately, Poland has faced similar monetary conditions to the 1998/1999 Czech Republic scenario since 2000. NBP is applying excessively high interest rates, thereby creating a setting of a similar “monetary overkill”). Sometimes temporary inflation shocks are better dealt with by de-emphasizing the importance of reaching intermediate targets while stressing the commitment to achieving the long-term target, i.e. demonstrating compliance with the long-term trajectory of disinflation. At the initial stage of DIT, the CNB has a limited capability to forecast inflation since the data series are too short and the volatility of inflation remains very high. Large fluctuations in both net and CPI inflation are caused by uneven, large adjustments in administrative prices and by the high volatility of CZK exchange rates. In an attempt to overcome these technical constraints to inflation forecasting, the CNB has established effective communication with independent forecasters representing the ten largest financial institutions in the Czech financial markets. CNB conducts periodic surveys of their independent inflation forecasts and incorporates their expectations into its own official forecast. Moreover, the Czech DIT system is very transparent. In addition to full policy disclosure and communication with financial institutions as well as with other government agencies and the public at large, the CNB has an informative website and it regularly publishes comprehensive inflation reports. Because of its well-designed and highly transparent framework and its ability to bring actual net inflation to low, fully controllable levels, the Czech DIT system may well be characterized as a “full-fledged” inflation targeting regime. The Polish DIT regime has been less successful and institutionally less advanced. It was introduced a year later than the Czech DIT, in January 1999. Nevertheless, unlike the CNB, the NBP has rather fortunately chosen to target CPI inflation from the beginning, and this choice appears to be well understood by the private sector. The long-term CPI inflation convergence goal has been set at 4.0% for the end of 2003. This trajectory appeared to be rather cautious originally, but now it seems somewhat “stretched”, after the year-on-year CPI inflation exceeded the December 2000 intermediate target by 1.7% (as shown in Table 2). As in the Czech case, the Polish monetary convergence also began with a SIT regime, since the initial band of permitted fluctuations was 0.6% around the midpoint target of 7.2% in 1999. The band was gradually widened to ±0.7% for the end of 2000 and 1.0% for the end of 2001. It appears that in the presence of inflation shocks prevailing in Spring and in Summer 2000, the upper bound of the official intermediate inflation target could not be reached. Under such circumstances, it will be advisable that NBP de-emphasizes the importance of reaching the intermediate target and reinforces its commitment to the long-term goal instead, considering the temporary character of the 2000 inflation shocks

246

L.T. Orlowski / Economic Systems 25 (2001) 233–251

L.T. Orlowski / Economic Systems 25 (2001) 233–251

247

stemming mostly from factors unrelated to monetary policy (e.g. higher oil prices and unfavorable weather conditions contributing to higher food prices). Poland’s DIT policy is generally well implemented. The NBP conducts monetary policy in a forward-looking manner and its decisions are sensitive to the deviations of the CPI forecast from the target. Since early 2000, the NBP has been actively using the policy bias as a forward-looking instrument, instead of actually increasing already very high interest rates. Specifically, real 3 months treasury bill rates were at 10.85% in Poland, while in the Czech Republic they were only at 1.48% at the end of 2000. High real interest rates in Poland reflect the risk premium related to the inflation bias and to the continuous instability of financial markets. Transparency of the Polish CPI targeting system has not been fully developed. The NBP publishes comprehensive and informative inflation reports, but the forecasting methodology remains rather ambiguous. Moreover, the Polish central bank has yet to develop an adequate system of policy communication and consultations with financial markets. In this respect, it will be well advised to follow the CNB example. The preliminary empirical results of DIT regimes in both countries are presented in Table 2. As evidenced by the comparison of intermediate inflation targets and the actual results, the Czech Republic has demonstrated a fully successful disinflation, which implies that the long-term inflation convergence target is feasible and perhaps even too cautious. By contrast, the Polish DIT system has yet to prove its ability to bring inflation down to the long-term trajectory by curtailing the underlying causes of inflation, mainly the factors affecting nominal indexation. Nevertheless, the Polish DIT policy deserves recognition for its ability to diminish volatility of inflation, at least as measured by the standard deviation and the coefficient of variation. Both of these measures indicate a lower volatility of the Polish CPI inflation after the inception of the DIT regime, while similar results for the Czech net inflation are not pronounced. However, these simplified variability estimates should be cautiously interpreted, since the Czech net inflation has been settled at a low range territory during the entire investigated period and was even negative for most of 1999, while the Polish CPI has been declining from the level exceeding 20% in 1996. The presented empirical results indicate that the Czech DIT system has been successful in reducing the inflationary mechanisms, although the Czech inflation is plagued by excessive volatility. Poland’s CPI targeting has yet to prove its ability to reduce the long-term path of inflation, but it has demonstrated some success in reducing inflation fluctuations. Under the present conditions, both the Czech Republic and Poland may consider increasing flexibility of their DIT systems. The band of permitted fluctuations for the Czech net inflation, determined at 1% around the midpoint targets for 2000/2001, seems to be appropriate considering the recently observed standard deviation of the Czech net inflation. However, the band around the 2001 CPI target can be widened to ±1%. Poland has applied the ±1.0% band in the determination of the 2001 target. A further investigative problem is related to the structural consistency of the inflation process. The question is whether the DIT regimes have contributed to a significant structural change in the inflation pattern in both countries. As demonstrated by Orlowski (2000a,c), the inflation process in both countries is highly non-stationary. Thereby, changes in year-on-year inflation rates pt are used here to alleviate the problem of non-stationary. These changes are stationary at 5 and 10 percent confidence levels. The CPI series for Poland and the

248

L.T. Orlowski / Economic Systems 25 (2001) 233–251

L.T. Orlowski / Economic Systems 25 (2001) 233–251

249

Czech Republic as well as the Czech net inflation follow a strong auto-regressive tendency with a high degree of persistence. This proves that the inflationary process in both countries is strongly influenced by nominal indexation and that their monetary policies are not fully credible. To demonstrate the inflation persistence, I estimate the time series trend of inflation with the first-order auto-regressive and moving average components, ARMA(p, q): pt = α + β0 t + β1 AR(+1) + β2 MA(+1) + µt

(4)

The empirical results of Eq. (4) for the Czech and Polish inflation are shown in Table 3. As shown in Table 3, all values of the sensitivity coefficients β 1 of the first-order auto-regressive correction, AR(1) are close to unity and they are statistically significant, thereby proving that inflation in both countries is highly persistent. Therefore, the inflation process is strongly affected by nominal indexation. These results prove that the monetary policies in both countries suffer from the lack of credibility. Based on the preliminary assumptions discussed in Section 2, DIT could be expected to reduce the scope of nominal indexation in TEs. To prove whether this reduction has indeed taken place, I apply a dynamic Chow–Split estimation of the time series trend of inflation with the first-order ARMA(1, 1). The empirical results of the Chow–Split estimation of Eq. (4) based on the January 1995–December 2000 data series are shown in Table 4. The above results demonstrate that the moments of the DIT introduction did not interject significant structural breaks to the inflation process in Poland and in the Czech Republic. However, the DIT strategy may be at least partially credited for strengthening the decline of the Czech net inflation in the period June–August 1998, apparently after the expiration of the lagged effects on unregulated prices of the increase in administrative prices in January 1998. A visible deceleration of the Czech CPI inflation took place in the same period, thus, showing at least partial effectiveness of the DIT strategy. A stronger disinflationary effect is detected in January 1999 when, for the first time in recent years, the Czech Government refrained from raising administratively-regulated prices. Poland’s CPI inflation in the investigated period was decisively reduced in the last quarter of 1995. But the empirical tests show that the DIT strategy has not been particularly Table 4 Chow split estimation — selected, meaningful resultsa Variable

Break points

F-statistic

Probability

Likelihood ratio

Probability

Polish CPI inflation

November 1995 January 1999 July 1999

6.261 2.154 5.061

0.000 0.084 0.001

23.764 9.100 19.783

0.000 0.059 0.001

Czech CPI inflation

January 1998 June 1998 January 1999

1.793 3.783 5.227

0.141 0.008 0.001

7.654 15.283 20.347

0.105 0.004 0.000

Czech net inflation

January 1998 June 1998 August 1998

1.445 5.668 5.277

0.230 0.001 0.001

6.223 21.826 20.518

0.182 0.000 0.000

a

Source: author’s own estimation.

250

L.T. Orlowski / Economic Systems 25 (2001) 233–251

successful in lowering inflation. On the contrary, there is a structural break in the inflation series in July 1999 when the inflationary pressures were fiercely renewed as a result of supply shocks. A positive impact of DIT on disinflation in Poland still remains out of sight. The presented empirical analysis demonstrates that both countries are still at a relatively early stage of the monetary convergence process outlined in this study. The Czech Republic has been visibly successful in lowering inflation to a stable and predictable level, while Poland is still struggling to reduce inflation to the officially adopted long-term trajectory.

5. Concluding remarks In order to advance to the next stage of the convergence process or to switch to a more flexible system of inflation targeting which allows for paying relatively more attention to exchange rate stability, the Czech and the Polish authorities have yet to overcome several obstacles. Among them, some are as follows: • Inadequate competition in the industrial sector, particularly in telecommunications, utility and heavy industries. The remaining monopoly or near-monopoly structures need to be more vigorously deregulated and privatized, particularly in Poland. • Insufficient flexibility of labor markets, which leads to high structural unemployment and which is associated with rigid labor codes, promoting wage indexation and reducing labor mobility. Removing these institutional and structural obstacles to disinflation is essential for the advancement of monetary convergence to the eurozone. In addition to the above impediments, the Czech DIT system is negatively affected by pervasive adjustments in administrativelyregulated prices, and by the currency volatility related to the high degree of openness of the Czech economy and to the considerable capital account liberalization (CNB, 2000). Nevertheless, both countries may consider moving to a more flexible inflation targeting regime as long as the institutional obstacles to this type of monetary policy are removed and as they demonstrate solid gains in price stability. The presented empirical evidence suggests that the Czech Republic appears to be ready for FIT, while Poland still needs to undergo further fundamental preparations. Even if both countries declare a wider inflation tolerance band, the gradual introduction of a full-fledged FIT will depend on the determination of asymmetry of the band. More specifically, at an early stage of moving from SIT to FIT the upper bound of the official inflation tolerance band is more binding than the lower bound. By demonstrating such an approach, the central bank underpins a strong commitment to disinflation, while ignoring possible temporary deflationary tendencies. Once the fundamental price stability is established, the weights on the upper and the lower bound should be more symmetric. This will accentuate a full transition to FIT. A gradually stronger emphasis on the lower bound is indispensable for preventing TEs from falling into a liquidity trap situation in the future. It can be further noted that both the Czech Republic and Poland do not appear to suffer from the lack of fiscal discipline that would otherwise become the most serious impediment to DIT. In 1999, the budget deficit to GDP ratios were 2.0% for Poland and 1.6% for the

L.T. Orlowski / Economic Systems 25 (2001) 233–251

251

Czech Republic, and furthermore, the future fiscal outlook for both countries is even more optimistic. Nonetheless, it is imperative that the fiscal discipline be maintained in the future to facilitate a successful monetary convergence to the eurozone in both countries. Moreover, as emphasized by the DIT–ERT analysis, the future monetary convergence will require a full synchronization between inflation and exchange rate targets, coupled with the institutional advancement of autonomous monetary policies of the candidate countries. Acknowledgements I am indebted to Hubert Gabrisch, William T. Gavin, Jürgen von Hagen, Katrin Wesche and two anonymous referees for their valuable comments and suggestions. References Bernanke, B.S., Laubach, Th., Mishkin, F.S., Posen, A.S., 1999. Inflation Targeting: Lessons from the International Experience. Princeton University Press, Princeton, NJ. Bruno, M., 1992. Stabilization and reform in eastern Europe: a preliminary evaluation. IMF staff papers 39, pp. 741–777. CNB, 2000. The Setting of the inflation target for 2001, Prague. De Grauwe, P., 1996. Inflation targeting to achieve inflation convergence in the transition towards EMU, discussion paper no. 1457. Centre for Economic Policy Research, London. Fischer, S., Sahay, R., Végh, C.A., 2000. Modern hyper- and high-inflations. In: Proceedings of the Paper to the Conference of the American Economic Association, Boston, MA. Golinelli, R., Rovelli, R., 2000. Painless disinflation? Monetary policy rules in Hungary, 1991–1999, mimeo. Department of Economics, University of Bologna. Hölscher, J., Vinhas de Souza, L., 2000. Exchange rate linkages and strategies of new EU entrants. J. Econ. Integrat. 22. King, M.A., 1996. How should central banks reduce inflation? Conceptual issues. In: Proceedings of the Symposium on New Challenges for Monetary Policy Sponsored by the Federal Reserve Bank of Kansas City. Federal Reserve Bank of Kansas City, Kansas City, pp. 53–92. Masson, P.R., 1999. Monetary and exchange rate policy of transition economies of central and eastern Europe after the launch of EMU, IMF policy discussion paper no. PDP/99/5. International Monetary Fund, Washington, DC. Mishkin, F.S., 2000. Inflation targeting in emerging-market countries. Am. Econ. Rev.: AEA Pap. Proc. 90, 105–109. Orlowski, L.T., 1999. Feasibility and conditionality of inflation targeting among central European candidates for the EU accession, Forschungsreihe no. 2. The Halle Institute for Economic Research, Halle. Orlowski, L.T., 2000a. Direct inflation targeting in central Europe. Post-Soviet Geogr. Econ. 41, 134–154. Orlowski, L.T., 2000b. Monetary policy regimes and real exchange rates in central Europe’s transition economies. Econ. Syst. 24, 145–166. Orlowski, L.T., 2000c. A dynamic approach to inflation targeting in central Europe, working paper B11/2000. Center for European Integration Studies, University of Bonn, ZEI. Sachs, J.D., 1996. Economic transition and the exchange rate regime. Am. Econ. Rev.: AEA Pap. Proc. 86, 147–152. Svensson, L.E.O., 1999a. Inflation targeting as a monetary policy rule. J. Monet. Econ. 43, 607–654. Svensson, L.E.O., 1999b. Monetary policy issues for the eurosystem. Carnegie Rochester Ser. Publ. Policy 51, 79–136. von Hagen, J., 1999. Money growth targeting by the Bundesbank. J. Monet. Econ. 43, 681–701. Warner, A.M., 2001. What are the chances of catching-up to the European Union? In: Orlowski, L.T. (Ed.), Transition and Growth in Post-Communist Countries: The 10-Year Experience. E. Elgar, Cheltenham, pp. 63–93.