IV. Monetary Policy Framework for Disinflation and EU Convergence18
74. This chapter discusses issues related to the nature of the framework in which Hungary will conduct monetary policy in coming years. Hungary’s monetary policy framework is bound to be shaped by its prospective entry into the European Union (EU) and subsequent participation in Economic and Monetary Union (EMU). An inflation performance consistent with these goals will require further disinflation from the low double–digit rate of inflation in 1998. The appropriate medium–term inflation target and the optimal pace of disinflation are discussed in sections B and C, focusing on the period leading to EU entry, rather than the subsequent period of qualifying for EMU.
75. The monetary policy framework for this disinflation will be superseded when Hungary enters EMU. A key issue for the transitional framework is the degree of exchange rate flexibility. Vujec (1996) suggests that the factors making it optimal for Hungary and other Central and East European (CEE) countries to join EMU also encourage these countries to peg to the euro earlier. However, more flexible exchange rate regimes permit more active use of the interest rate instrument, and may also assist with adjustment to shocks. This issue is discussed in section D.
76. An inflation targeting (IT) approach to monetary policy is being used by several small open economies. Section E discusses whether Hungary meets certain prerequisites for the adoption of IT, along with some aspects of the implementation of an IT policy framework. The independence, accountability, and transparency of the National Bank of Hungary (NBH) underpins the credibility of monetary policy. Section F analyzes the NBH legislation from this perspective, also noting the requirements of EMU.
B. Inflation Objectives and EU Membership
77. To prepare for their EMU participation, Finland from early 1993 began targeting the achievement of 2 percent inflation by 1995, and Spain adopted an inflation target of less than 3 percent in mid-1994. Other countries also explicitly target low inflation rates, (Table 5). Should Hungary prepare for EU entry by adopting similar inflation goals?
|Country||New Zealand||Canada||United Kingdom||Sweden||Finland||Australia||Spain|
|Date first adopted||1990 March||1991 February||1992 October||1993 January||1993 February||1993 April||1994 Summer|
|Initital inflation target||0 to 2%||1 to 3%||1 to 4%||2%||2%||2 to 3%||less than 3%|
|Expired/revised||End 1996||No||Spring 1997||No||No||No||End 1997|
|Current inflation target||0 to 3%||1 to 3%||2.5%||2%||2%||2 to 3%||2%|
|Time-frame of current inflation target||5 years to 2003||Through end-1998||From mid-1997 thereafter||1996 onwards||1996 onwards||On average over the cycle.||From end-1997 thereafter|
|Inflation measure||CPIX||CPI||RPIX||CPI||Underlying CPI||Underlying CPI||CPI|
|Components excluded from CPI||Credit charges.||None||Mortgage interest payments.||None||Mortgage interest payments, indirect taxes, subsidies, house prices.||Mortgage interest payments, indirect taxes, other volatile items.||None|
78. An inflation rate in the range indicated in Table 5 would likely be consistent with the Maastricht Treaty convergence criterion for participation in EMU, that inflation be no higher than 1.5 percentage points above the average rate in the three EU countries with the lowest inflation. However, the Maastricht criteria do not apply to EU accession (Temprano–Arroyo and Feldman, 1998). The 1993 Economic Council in Copenhagen defined the requirements for EU accession, but did not specify an acceptable inflation rate, only the general requirement that “macroeconomic stability has been achieved including adequate price stability and sustainable public finances and external accounts.” Nevertheless, adequate progress towards the Maastricht criterion for inflation is perceived to be part of the EU convergence process.
79. The path for entry to EMU helps to define the desirable progress on inflation prior to joining the EU. Once a country joins the EU it may enter the new Exchange Rate Mechanism (ERM II), and prior to EMU entry there must be no downward realignments within ERM II for at least two years. ERM II will be a relatively flexible arrangement, with the euro as the anchor currency and fluctuation bands of ±15 percent.19 While a slow depreciation within the wide ERM II bands could accommodate a modest inflation differential, in practice speculative pressures in expectation of a realignment can develop well before exchange rates reach the floor of the band, as demonstrated by the March 1995 devaluations of the peseta and escudo, Masson (1998). Therefore, at the time of entering ERM II the inflation rate in Hungary should already be close to the level consistent with avoiding significant downward deviations from the ERM II central parity. Inflation may of course be lower, creating the potential for upward realignments.
80. Even with a stable peg to the euro, Hungary can expect higher inflation than in the EMU countries, as it will have a wider gap between productivity growth in the traded and nontraded goods sectors, i.e. the Balassa–Samuelson effect will be stronger.20Simon and Kovács (1998) find that traded goods sector productivity growth in Hungary was 4.6 percent faster than in the nontraded goods sector over 1991–96. They calculate that if this rate of differential growth continued, the long–run average real appreciation on a CPI basis would be 3.4 percent if agriculture is treated as a nontradable, or 1.2 percent if it is treated as a tradable. The productivity growth differential may narrow somewhat over time, tending to moderate the Balassa–Samuelson effect by the time of ERM II entry.
81. The inflation targets in Table 5 are not centered on zero partly due to measurement bias in the CPI, estimates of which range from 0.5 percent to 1 percent in the inflation targeting countries, Debelle (1997). Such biases may affect the Hungarian CPI more strongly, for example, the “quality bias” may be higher if the quality of goods and services rises faster in Hungary than in EMU countries as real incomes catch-up. The “outlet bias” may also be high due to the rapid growth in supermarkets and shopping malls. However, this bias should diminish to more typical levels over time, and the “substitution bias” may be quite low, as the Hungarian CPI weights are updated annually, more frequently than is typical of EMU countries.
82. Allowing for the Balassa–Samuelson effect and some potential additional bias in the Hungarian CPI relative to the EMU countries, a medium-term target range for CPI inflation of 3 percent to 5 percent would seem reasonable. Achieving such an inflation target in the run-up to EU entry will demonstrate macroeconomic stability, make good progress toward satisfying the Maastricht criterion for EMU participation, and permit a credible entry into ERM II.
C. Optimal Pace of Disinflation in Hungary
83. The above target implies a total disinflation prior to EU entry of about 6 percentage points from the CPI inflation rate 10.3 percent at end 1998. The pace of this disinflation could range from a fast one or two years to a more gradual four years. Inertia in nominal wage growth is the key factor that could slow disinflation. For the total enterprise sector, wage inflation has been rather steady at about 20 percent per annum from mid-1994 to early-1998, Figure 11.21 A fast exchange rate based disinflation may not be sustainable if credibility is limited, as the initial lack of nominal wage deceleration would reduce competitiveness and undermine the trade balance. However, a gradual disinflation may fail to demonstrate commitment to low inflation, increasing disinflation costs and risking a delay to EU entry. The following outlines the arguments on rapid versus gradual disinflation.
Figure 11.Hungary: Wage Inflation, 1993–98
84. A number of factors urge that disinflation be completed early in the period leading to EU accession and ERM II membership. First, capital mobility between Hungary and EU countries will rise as accession approaches, making sterilization increasingly costly. If inflation remained above target close to EU accession, requiring high nominal interest rates, the pre-accession capital inflows would be very costly to sterilize. An earlier disinflation would allow interest rates to fall to stem these inflows. Second, disinflation may involve a temporary real exchange rate appreciation, so an earlier disinflation will reduce the risk of entering ERM II with an over-valued exchange rate. Finally, if other CEE countries in the EU accession process make more rapid progress on disinflation, and thereby achieve earlier EU admission, the economic benefits to Hungary of EU entry may be somewhat lower.
85. Blanchard (1998) concludes that conditions in Hungary provide a strong case for a speedy disinflation, of just one or two years. His analysis of studies on OECD countries finds that faster disinflations are less costly, with the sacrifice ratio—the output loss for each percentage point reduction of inflation—rising significantly with the length of the disinflation.22 In his view, disinflation in Hungary is largely a problem of coordinating a simultaneous slowing in wage and price inflation, with no need to reduce real wages.23 Conditions for achieving this coordination appeared favorable, as wage negotiations are on an annual basis rather than less frequently, without formal backward-looking indexation. Although wage negotiations had become more decentralized, Blanchard suggested that the Interest Reconciliation Council—a forum including representatives of trade unions, employers, and the government—could facilitate a “disinflation pact,” in which agreed nominal wage increases are based on the official inflation target.24
86. One objection to such a fast disinflation is the risk of a “credit crunch” that could exacerbate any weakening in activity due to firm monetary and fiscal policies. However, Hungarian enterprise sector profits are robust at 15 percent of GDP, and enterprise debt is modest at about 200 percent of profits, with only half being forint denominated, (Table 6). Even if the performance of enterprise loans was affected, banks are well capitalized with a capital adequacy ratio of 15.8 percent at June 1998, so a sharp recession due to a credit crunch is unlikely.
|(In billions of forint)|
|Profits (tax base estimate)||690||967||1280|
|Of which: Forint||773||914||1249|
|(Percent of GDP)|
|Profits (tax base estimate)||12.3||14.1||15.2|
|Of which: Forint||13.8||13.4||14.9|
|(Percent of profits)|
|Of which: Forint||112||95||98|
87. The main argument for a gradual disinflation is provided by Ministry of Finance (1997), which presents a macroeconomic framework where reaching inflation of 4 percent to 5 percent takes four years due to “difficulties tackling inflation’s large inertial component. This is because economic policy credibility is still to be strengthened, given that targets announced in previous years have not always been attained.” However, credibility now appears to have improved, as 1998 will be the second year that the official inflation targets are achieved or bettered. Further evidence of improved credibility is provided by the slowing of wage inflation in mid-1998, led by the industrial sector (Figure 11). The political consensus for joining the EU underpins the commitment to disinflation, making the conditions for a rapid low-cost disinflation particularly propitious.25 Achieving the above medium-term inflation target within a two year period (by end 2000 or early 2001) appears feasible while sustaining growth, ensuring that disinflation is completed sufficiently in advance of entry to the EU. The specification of inflation targets is further discussed in Section E.
D. Transitional Exchange Rate Regime
88. Hungary’s crawling peg exchange rate regime is described in Chapter IX. Participation in ERM II will entail a peg to the euro rather than to the basket, the elimination of the preannounced crawl, and a wider exchange rate band.26 From January 1, 2000 the Hungarian authorities plan to define the central parity of the forint in terms of the euro, and they expect to eliminate the crawl when inflation is sufficiently low. Therefore, the key exchange rate regime decision is whether to retain the narrow ±2¼ percent band until ERM II entry, or if not, when and how much to widen the band, or even whether to float the forint at some point. Other CEE countries have been shifting to more flexible exchange rate regimes,27 but the Hungarian authorities will need to weigh many factors including: 1) minimizing the cost of disinflation, 2) facilitating growth and real convergence to the EU, 3) assisting adjustment to real shocks and maintaining external balance, 4) containing exposure to disruptive capital flows, and 5) preparing for entry to ERM II.28 The following outlines the costs and benefits of the alternative regimes after describing developments in capital flows and controls in Hungary.
Capital flows and controls
89. Capital flows appear to have become increasingly sensitive to interest rate differentials during the period of the crawling narrow exchange rate band, Figure 12. This trend reflects improving financial market confidence in Hungary, the privatization of the major banks, and the partial liberalization of capital controls made in the course of joining the OECD. Chapter IX describes the remaining capital controls, which primarily apply to short-term lending and trading in short-term securities between residents and non–residents.
Figure 12.Hungary: Interest Differential and Net Foreign Assets of NBH, 1996–98
Sources: Hungarian Central Statistical office, National Bank of Hungary, and staff calculations.
90. Until mid-1998 the forint has typically been at the most appreciated edge of the band, requiring sustained foreign exchange purchases by the NBH, and therefore sterilization operations to the extent that liquidity injections exceeded growth in base money demand. By the end of 1997 the liabilities of the NBH that arise principally from its sterilization activities had risen to Ft 681 billion from Ft 18 billion at end March 1995, a stock exceeding currency held by the public, and a flow equivalent to almost US$4 billion.29 The interest sensitivity of capital inflows became particularly evident in the first quarter of 1998. Though the rate of crawl was cut from 1.0 percent to 0.9 percent per month on January 1, 1998, interest rate cuts by the NBH were small.30 Consequently, the uncovered interest differential rose from around 2 percent in late 1997 to average 3 percent in the first quarter of 1998, attracting strong capital inflows, as reflected in sterilization liabilities rising to Ft 1083 billion by end March, a flow equivalent to about US$1.9 billion.31 The cost of sterilization is given by the interest differential between the sterilization liabilities and foreign reserves, less the revaluation gains on foreign reserves, together averaging about 2.5 percent. Sterilization in the first quarter of 1998 will cost about 0.1 percent of GDP in 1998, and had this rate of sterilization been sustained through 1998 the impact on the fiscal deficit would have been substantial.
91. Most of the funding for the purchase of sterilization instruments in the first quarter of 1998 came from foreign borrowing by banks (with net liabilities up Ft 187 billion), along with the use of their foreign exchange deposits at the NBH (down by Ft 119 billion). The NBH introduced in early 1999 reserve requirements on the short–term foreign liabilities of banks, though initially with a zero rate.32 While making the requirement effective by applying a non–zero rate would raise the interest rate threshold on capital inflows through the banks, this measure may not offer lasting insulation given that liberalization of short–term capital flows will likely occur in the next few years, as discussed in Chapter IX.
Narrow band regimes: fixedand crawling pegs
92. Introducing a fixed peg to the euro, with a narrow band, could be motivated by deepening integration with the EMU area to accelerate growth, structural convergence, and EU and EMU entry. However, pegging before inflation has declined further may undermine competitiveness due to inflation inertia. This risk is exacerbated by high capital mobility, as the drop in interest rates towards EMU levels bought about by the peg would stimulate aggregate demand. Indeed, in the face of strong and potentially volatile capital flows, Vujec (1996) and Masson (1998) argue that it is not likely feasible for even the front–running CEE countries to sustain a fixed exchange rate peg with a narrow band.33
93. Crawling pegs reduce the risk of a loss of competitiveness as they allow a more gradual reduction in inflation. However, narrow band crawling pegs involve some of the drawbacks of pegged exchange rates, in particular the loss of control over domestic interest rates (short of expensive sterilization) and the resulting potential stimulus to domestic demand. Moreover, by reducing the risk of sharp exchange rate changes, such a crawling peg encourages capital mobility (with respect to a float), thus raising exposure to the risk of volatile capital flows.
94. Another drawback of pegs and crawling pegs with narrow bands is the loss of flexibility in responding to shocks. In the early phase of convergence to the EU, Hungary will be more likely to face external shocks affecting its economy differently to the rest of the euro area. However, this risk appears relatively low, as the industrial sector is highly integrated with the EU through both ownership and trade, and the agricultural sector (at 6.4 percent of GDP in 1995) is small relative to other CEE countries, and not greatly larger than in some EU countries.34 There may be greater risk of internal shocks, e.g. to the savings rate or to real wages. A narrow band would limit changes in interest rates in response to such shocks, and may also constrain the adjustment of exchange rate policy, as an open policy shift would likely undermine the credibility of future exchange rate commitments.
More flexible exchange rate regimes: the risk of overshooting
95. The above discussion highlights the potential risks associated with retaining a narrow band in the medium–term as the pressures from capital flows intensify. A more flexible exchange rate would also enable the likely pressures for an equilibrium real appreciation during the convergence to the EU to be realized, at least partly, via nominal exchange rate appreciation rather than through higher inflation. However, there are concerns that greater flexibility could be detrimental to disinflation if an exchange rate “overshooting” weakened the current account balance. As discussed by Suranyi and Vincze (1998), Hungary’s track–record has left enterprises and households with the expectation that a weak external balance will lead to a devaluation and higher inflation, undermining the credibility of disinflation. Nevertheless, the following argues that there may be less risk of a current account deterioration when disinflating under a more flexible exchange rate regime, due to enhanced interest rate flexibility.
96. Under the narrow band, monetary policy relies on the slowing rate of crawl being transmitted through the domestic price of traded goods to reduce broader inflation. The use of interest rates to restrain demand pressures on nontraded goods prices and wages is curtailed by sterilization costs, with Figure 13 showing the stability in real interest rates during the crawling peg regime. A wider band will enhance interest rate flexibility, as reflected in the much higher uncovered interest differential in Poland though 1997 and 1998, Figure 14, though Poland and Hungary have more similar interest spreads on external debt.35 Raising interest rates under a wider band (or a float) will appreciate the nominal exchange rate, and initially the real rate given wage–price inertia. While this appreciation will likely weaken the trade balance (see Chapter VII), higher real interest rates will also reduce domestic demand, mitigating the impact on the trade balance, and also slowing nontraded goods inflation more quickly and thereby unwinding the real appreciation over time. To achieve the same pace of disinflation without interest rate flexibility, the smaller reduction in nontraded goods inflation must be offset by a larger cut in traded goods inflation, and therefore eventually a larger real appreciation, risking a greater trade balance deterioration.
Figure 13.Hungary: Real Interest Rates, 1992–98 Figure 14.Hungary and Poland: Uncovered Interest Differentials, Hungary and Poland, 1995–98
Sources: Hungarian Central Statistical office, National Bank of Hungary, and staff calculations.
97. A simple model illustrates this point (variables in logs, Δ is the difference operator):
|CPI:||p = λ pNT + (1-λ) pT|
|Traded goods price:||pT = e|
|Phillips curve for non-traded goods:||ΔpNT = Δpe - α y|
|Inflation expectations:||Δpe = δΔp+ + (1-δ)Δp-1|
|Aggregate demand:||y = - β r - γ (pNT - pT) + ∈|
Where Δp* is the announced inflation target, e is the price of foreign exchange, y is output, r is the real interest rate, and pNT - pT is the real exchange rate. The disinflation over one period is given by:
A highly credible disinflation (δ close to 1) can be achieved with little need for a monetary policy induced recession, simply by making the nominal depreciation (Δe) consistent with targeted inflation. However, given some inertia in inflation expectations, a tight monetary policy that temporarily lifts the real interest rate and/or the real exchange rate is required to achieve the inflation targets. Without the flexibility to increase real interest rates, more pressure from the real exchange rate is required to achieve same inflation reduction. Interest rate flexibility also provides a tool to counter the impact of demand shocks (∈) that would otherwise result in a deviation from the targeted disinflation unless the exchange rate was more appreciated.
98. The empirical relevance of this argument depends on the responsiveness of aggregate demand to interest rates, and of nontraded goods prices to aggregate demand. Real interest rates were found to have a significant effect on household savings in IMF (1995), suggesting that at least one channel for the interest rate effect on demand is effective. Nominal wage growth—and thus nontraded goods prices—should also be responsive to weaker demand, as Kertesia and Kollo (1997) find real wages in Hungary to be as responsive to unemployment as in the United States, and significant real wage flexibility to unemployment is also reported in Chapter III.
99. Disinflation experiences also tend to support the perspective that an exchange rate regime that permits greater interest rate flexibility involves less risk of undermining the external position. Calvo and Végh (1998) analyze countries with chronic inflation (above 30 percent to 40 percent), finding that exchange rate based stabilizations are accompanied by an initial activity boom, a growing real exchange rate appreciation, and a deterioration of the trade balance. A currency crisis and economic contraction often followed, reversing the disinflation. In contrast, “money–based” stabilizations—covering a range of more flexible exchange rate regimes—also involve a real appreciation, but the rise in real interest rates is associated with an earlier recession, and no clear–cut effect on the trade balance. Inflation stabilizations in Western Europe are examined by Detragiache and Hamann (1997). It is notable that Italy reduced inflation from 22 percent in 1980 to 4 percent by 1986 with little real appreciation during a period of “weak commitment” to the exchange rate target but a restrictive monetary stance. Portugal maybe the most relevant case, as in the early 1990s it had a broadly similar fiscal position to Hungary, and CPI inflation in 1991 of 11.4 percent was cut to 6.5 percent by 1993. Despite a real appreciation in the early 1990s the current account did not deteriorate, as the Bank of Portugal kept interest rates high, and a stable primary fiscal surplus was retained, at least until the 1993 recession.
Forms of flexible exchange regimes
100. Either a managed float or a peg (fixed or crawling) with a wider band would serve to enhance interest rate flexibility. In Hungary’s case, a wider band may achieve sufficient interest rate flexibility while putting a reasonable bound on short–term swings in the exchange rate. Furthermore, abandoning the preannounced crawling band could prove unsettling to inflation expectations, suggesting that some type of band be retained.
101. The design of a band regime involves more than determining the band width, as in practice it is common to have “intra–marginal” intervention. For example, after the ERM exchange rate bands were formally widened in August 1993 from ±2¼ to ±15 percent, ERM member central banks still typically intervened according to the old narrow bands, insisting that they remained committed to the narrow bands except possibly in the short–run under strong market pressure. Bartolini and Prati (1998) find that such a “soft” inner band is significantly less vulnerable to speculative attacks than a “hard” band, consistent with the sharp abatement of speculative pressures on the ERM currencies after August 1993. However, a pattern of intra–marginal intervention will reduce perceived exchange rate risk and therefore interest rate flexibility. Therefore, at least during a period of disinflation, it may be advisable to focus intra–marginal intervention on countering exchange rate movements inconsistent with the final policy objectives, rather than keeping the exchange rate within a narrow inner band.
102. The above discussion suggests that some widening of the exchange rate band may prove advantageous.36 The liberalization of short-term capital flows that is envisaged (see Chapter IX) would likely add to the desirability of a wider band in making the transition to ERM II. If the exchange rate band is widened, the anchoring role for inflation expectations played by the crawling peg will be diminished, suggesting that the goals that the NBH will be pursuing with its enhanced policy flexibility would need to be clearly specified.
E. An Inflation Targeting Framework?
103. A number of countries have announced that monetary policy will pursue specific targets for inflation without committing to a particular intermediate target, including the Czech Republic and, most recently, Poland.37 Such a monetary policy framework allows greater instrument flexibility to respond to shocks, but this high degree of discretion may also erode the credibility of the targets. The inflation targeting (IT) framework aims to reduce this problem by making the inflation target the sole over–riding objective of the central bank, by ensuring that the central bank has sufficient instrument independence to achieve this target, and by establishing monitoring and accountability mechanisms for its performance relative to the targets. This environment provides strong incentives for the central bank to continually adjust monetary policy to ensure that its inflation projections remain in line with the targets.
104. The following discusses whether Hungary satisfies the technical pre-requisites for IT, and considers how an exchange rate band might fit into an IT framework. From Debelle (1997) and Masson et al (1997), the pre-requisites for adopting IT are:
(i) the ability to carry out an independent monetary policy, especially one not constrained by fiscal considerations;
(ii) a well developed financial system to allow the effective operation of monetary policy;
(iii) freedom from commitment to another nominal anchor, e.g., the exchange rate;
(iv) an adequate ability to produce inflation forecasts and to assess the impact of changes in monetary instruments, and;
(v) political institutions that facilitate agreement on policy goals.
105. The broad conclusion is that Hungary has moved a great deal towards meeting these pre-requisites for using an IT approach to monetary policy. Regarding (I), the NBH Act provides for the independence of the NBH, though there are issues related to the specification of its primary objective, and to a lesser extent, its instrument independence with respect to the exchange rate, see Section F. The budget has typically had little reliance on revenue from NBH transfers, and new NBH credit to the government has been virtually eliminated. Though gross public debt is 60 percent of GDP in 1998, this is much reduced from 84 percent of GDP in 1995, and a fiscal policy consistent with EU entry will be reflected in a further decline in the debt burden. All this suggests that “fiscal dominance” of monetary policy is not overbearing on a forward-looking basis.38 As to (ii), the health of the Hungarian banking system does not to constrain monetary policy, and the key financial markets are sufficiently well developed, see van Elkan (1998).
106. Regarding (iii), in the cases where IT is a more permanent monetary framework (e.g., New Zealand, Canada, U.K., Australia) there is no commitment to another nominal anchor like the exchange rate. However, Svensson (1997) suggests it is possible to have an exchange rate band within an IT framework, but only as a “temporary intermediate target” that would be altered if a conflict with the inflation target arose. The countries that maintained an IT framework in conjunction with participating in ERM, Spain and Finland, did not face such a test of target priority. However, in Israel a conflict arose between the 1994 inflation target of 8 percent and the targeted 6 percent exchange rate depreciation. The foreign exchange market was stable but demand pressures accelerated inflation through the year. A monetary tightening—that likely would have disturbed exchange rate stability—was delayed, allowing inflation to reach 15 percent by end 1994. This experience suggests that continuing with an exchange rate band risks undermining the credibility of an IT framework. However, this risk can be reduced by ensuring that the band is sufficiently wide, and by prioritizing the inflation target ahead of the exchange rate band in a clear and transparent manner, Leiderman and Bufman (1996).
Inflation target specification
107. Before discussing (iv), it is useful to review the specification of inflation targets. Such targets should be specified as simply as possible while providing a meaningful benchmark for the performance of monetary policy. There is the choice of price index used, the calculation of inflation, and the horizon of the targets, as discussed more fully in Debelle (1997).
108. Targeting a measure of consumer prices is common to all countries using IT, and it is also the measure used under the Maastricht Treaty. While the official CPI would be the most simple and transparent target, half of the countries in Table 5 target the CPI excluding certain items, called an underlying or core measure of the CPI.39 By excluding the impact of price shocks not related to aggregate demand, e.g., seasonal swings in food prices or the impact of tax changes, a core CPI provides a more reliable benchmark for evaluating the performance of monetary policy. However, if targets based on core CPI are to be relevant, the core CPI measure should be expected to have a similar trend to the official CPI. In Hungary changes in regulated prices and direct taxes had a large impact on inflation in recent years, Figure 15, suggesting that inflation targets defined in terms of nonregulated goods prices would not have been acceptable. Nevertheless, given that domestic energy prices have moved closer to world levels this may be less problematic in future, indeed, the difference between official inflation and inflation net of regulated prices narrowed significantly by end 1998.40
Figure 15.Hungary: CPI Inflation, 12-month basis, 1992–98
109. Announced inflation targets in Hungary have in the past emphasized the annual average rate of inflation, but the average inflation rate depends on the path of prices in the previous year, so it would be preferable to set the targets in terms of the 12-month inflation rate. The budget is on a calendar year basis, making it logical to have end of year targets for 12-month inflation. Ideally, if inflation targets were adopted, they would be specified for more than one year ahead, so that through each year monetary policy could increasingly focus on achieving the inflation target in the following year. The targets could take the form of a ceiling, but no lower level, until inflation has been reduced significantly, allowing the flexibility to have a faster disinflation if conditions are favorable.
Inflation projections and indicators
110. To consistently achieve inflation outcomes close to the specified targets, a central bank under an IT framework must be able to reliably evaluate the required monetary policy stance, i.e., pre-requisite (iv) from ¶104. This evaluation requires the ability to make inflation projections that take into account the time-lags between the adjustment of monetary instruments and the impact on inflation, along with the outlook for other factors bearing on inflation. Ideally these projections would be developed within an economic model capturing the key transmission mechanisms between the monetary conditions and inflation, so that the central bank can solve for the monetary policy path that will best achieve the inflation targets.41
111. Even in countries with relatively advanced economic models, central bankers in an IT framework often draw on a long “check-list” of inflation indicators when developing monetary policy. Chapter II reports on the properties of a range of potential leading indicators for CPI inflation in Hungary, finding that the PPI, foreign producer prices, the budget balance, and both the nominal and real exchange rate have significant leading information, particularly for measures of core or underlying inflation. The impulse-response analysis finds that the effects of the nominal exchange rate are felt with 1 to 7 month lag, with a peak at 3 months, suggesting that monetary transmission lags with respect to the exchange rate are comparatively short in Hungary. Fiscal policy is found to have only slightly slower effects, with the budget balance impacting on inflation from 2 to 9 months later, with a peak effect after 5 months. Some commonly used indicators like wages, money and credit aggregates, and industrial production were found to have little leading information, suggesting these indicators be used cautiously. Moreover, even the better indicators account for less that one quarter of the variance in monthly inflation, so it is not possible to rely on just a few indicators in formulating monetary policy.
112. The price and wage equations reported in Chapter III help to interpret some of these indicator results, as well as providing information on some aspects of the monetary transmission mechanism, and allowing further evaluation of inflation predictability. Allowing for regulated price adjustments, the PPI is found to be well modeled as a tradable price, depending on foreign PPIs, the price of oil, and the exchange rate to the German mark and the U.S. dollar. Thus the good leading indicator properties of the PPI likely reflect that it captures the impact of both the nominal exchange rate and foreign producer prices.
113. While a sustainable disinflation must slow labor cost inflation, the size and timing of exchange rate “passthrough” via traded goods prices to the CPI is a key parameter in the design of an IT monetary policy in a small open economy. Chapter III reports that the first round impact—holding labor costs fixed—of a nominal exchange rate depreciation of 1 percent is to raise a core CPI measure by 0.3 percent, with about three-quarters of this impact after 6 months, and virtually complete adjustment in 12 months. The exchange rate elasticity is likely higher for the official CPI, which includes energy prices that are linked to world market prices. These results confirm that the exchange rate will be a key instrument in the disinflation process, and suggest that an IT monetary policy in Hungary could use a policy horizon of at least 12 months.
114. The dynamics of wages and inflation expectations are the key factors in medium-term inflation developments, and the sustainable pace of disinflation. Chapter III finds that real wages move in line with labor productivity over the longer-run, with a significant negative response to unemployment. Nominal wage adjustment was found to be closely linked to recent core CPI inflation, so that a slowing of core CPI inflation is soon followed by a deceleration in wage inflation. Inertia in nominal wage growth is therefore less than expected, reducing the risk of a real appreciation and trade balance deterioration from a more rapid disinflation, a finding sufficiently surprising to merit further analysis.
115. Chapter III also reports that a dynamic simulation of the wage-price system over a two year horizon performed quite well in forecasting the core CPI. This result is suggestive that prerequisite (iv) can be satisfied, at least for a measure of core inflation. Nevertheless, as is the practice of all central banks using an IT approach, it would be necessary to supplement the results of any inflation forecasting models with well-informed judgement.
F. Institutional framework for monetary policy
116. Does the institutional framework of the National Bank of Hungary facilitate a credible disinflationary monetary policy? Relatedly, does it meet requirements of EMU? Temprano-Arroyo and Feldman (1998) examine the position of Hungary and the other CEE countries in relation to the EU rules in the area of EMU. The rules that bear directly on the institutional framework for monetary policy include:
(i) The term of the central bank Governor and board members must be no less than five years.
(ii) Dismissal of the Governor and board members may occur only under circumstances of serious misconduct or inability to perform their duties.
(iii) The central bank should not take any instructions from the government.
(iv) Price stability must be the primary objective of the central bank.
(v) Central bank direct financing to government must be prohibited.
117. Under the Act on the National Bank of Hungary of 1991, as amended effective January 1, 1997, much of the legislative aspects of the institutional framework for monetary policy have already adapted to the requirements for EU membership, but there are some areas where further amendments will be required. These areas will be noted in the course of analyzing the institutional framework for the NBH in terms of those features considered to be most important for monetary policy performance: a clear objective, independence with respect to instrument setting, governance with long horizons, and adequate accountability and transparency.
118. The NBH has multiple objectives, as the “basic task of the NBH is to safeguard the domestic and external purchasing power of the national currency,” Article 4(1). Meeting the requirements of EMU, by making price stability the primary objective, will support the credibility of a disinflationary monetary. The NBH also “supports the implementation of the economic policy program of the Government with monetary policy means available to it,” Article 3. There is no qualification that this support will be provided if it does not conflict with the basic task, so this article may perhaps need amendment to ensure price stability is the primary objective of monetary policy.
119. Articles 42 to 50 shape the relations of the NBH with the Government. The “NBH shall not be subject to instructions from the Government”, as is required for EMU. However, the NBH must “mutually reconcile” with the Ministry of Finance, concerning the annual budget and its financing with the annual monetary guidelines. Also, the Government “takes a stand” regarding the guidelines before they are presented to Parliament. If the processes of consultation fail, “the NBH is authorized—if it cannot assert its standpoint in any other way—to make public its opinion about economic policy decisions of the government…” Overall, these arrangements may serve to facilitate negotiation of an agreed framework for monetary and fiscal policies between the NBH and the government, therefore satisfying requirement (iv) for adopting inflation-targeting from ¶104.
120. In Article 6, the “NBH develops its monetary policy as well as the instruments serving its implementation independently within the framework of this Act”. However, independence with respect to the exchange rate is qualified, as under Article 13(2) the “order of determining and/or influencing the exchange rate is approved by the government in agreement with the NBH.” Therefore, any change in the exchange rate arrangements, at either a broad level, e.g., adoption of the crawling peg, or at a more detailed level, e.g., a change in the rate of crawl, requires approval by government. While this has not in practice been a constraint on NBH policy, it could at some point restrict the instrument independence of the NBH. Cottarelli (1994) argues that a central bank that is not responsible for exchange rate policy is not truly independent in the scope and timing of its actions, especially when capital flows are unrestricted. With the deepening liberalization of capital movements in Hungary, it may be appropriate to update Article 13(2) to preserve the instrument independence of the NBH. Nevertheless, even the ECB does not have full independence in this regard, so an acceptable formulation might require government approval for exchange rate regime changes, but not for adjustments within a regime. An alternative approach would allow the NBH to alter exchange rate policy, though subject to over-ride by the government.
121. The instrument independence of the NBH is not significantly impaired by requirements to finance the government, as the NBH may not purchase securities directly from the state, and may only extend credits on an exceptional basis. These are “liquidity loans for bridging the momentary liquidity difficulties of the Single Treasury Account—up to 2 percent of the planned budget revenue of the actual year,” Article 19(4). Such credits may be outstanding for at most 15 days of any month, and must not be outstanding at year end. This provision has been used only few times, but it violates EMU prohibitions on central bank financing to government.
122. Members of the governing body of the NBH have long-term appointments, but meeting EMU requirements may further strengthen both appointment terms and protection from dismissal. The Central Bank Council (CBC) is the key policy-making body of the NBH. It consists of the President, up to five Vice-Presidents, and other members in a number equal to the Vice-Presidents plus one. All members are appointed by the President of the Republic, with Vice-Presidents nominated by the NBH President, and other members nominated by the Prime Minister. The President and Vice-Presidents have six year terms, longer than the four year parliamentary term, but other members have three year terms. To meet the requirements of EMU entry the terms of NBH managers on the CBC will need to be extended.
123. The NBH Act allows only the President of the Republic to relieve the NBH President or a Deputy President of their position. However, the reasons for doing so are not strictly limited, being “for reasons for which [the President or Deputy President] can be blamed, does not attend to the tasks stemming from his appointment, commits a crime proven by a valid judgement-at-law, or has become unworthy of his office in any other way,” Article 58(8)b. This article may also need to be amended to met EMU requirements.
Accountability and transparency
124. To ensure the accountability of the NBH, the Act provides for a Supervisory Committee, composed of parliamentarians, a representative of the Ministry of Finance, and an expert invited by the Minister of Finance, which is “obliged to inform the Parliament and the Minister of Finance appointing them,” Article 65. A range of publications are used to provide the transparency needed to monitor the performance of monetary policy. The NBH must provide an annual report to Parliament—it typically also publishes an interim report—and Parliament may also request occasional information. The NBH President must also present the annual guidelines of monetary policy to the Parliament, which have a more forward-looking nature. Finally, the NBH has recently started to publish quarterly reports analyzing recent inflation developments, a practice typical of the inflation-targeting central banks.
BartoliniLeonardo and PratiAlessandro1998 “Soft Exchange Rate Bands and Speculative Attacks: Theory and Evidence from the ERM Since August 1993,” Federal Reserve Bank of New York Staff Report No. 43.
BallLaurence1994“What determines the Sacrifice Ratio?”inMankiwN.G.Monetary PolicyUniversity of Chicago Press.
BlackR.CassinoV. DrewA. HansenE. HuntB. RoseD. and ScottA. 1997“The Forecasting and Policy System: an introduction,”Reserve Bank of New Zealand.
BlanchardOlivier1998“Optimal Speed of Disinflation: Hungary,”inCottarelli and Szaparyeds.Moderate Inflation: The Experience of Transition EconomiesInternational Monetary Fund and National Bank of Hungary.
CalvoGuillermo and VeghCarlos1998“Inflation Stabilization and BOP Crises in Developing Countries,”draft for Taylor and Woodford Handbook of Macroeconomics.
ChristoffersenPeter and DoylePeter1998 “From inflation to growth: eight years of transition,” IMF Working Paper 98/99 (Washington: International Monetary Fund).
CottarelliCarlo1994“Should an “Independent” Central Bank Control Foreign Exchange Policy?,”inBaliño and Cottarellieds.Frameworks for Monetary Stability: Policy Issues and Country ExperiencesInternational Monetary Fund.
DebelleGuy1997 “Inflation Targeting in Practice,” IMF Working Paper 97/135 (Washington: International Monetary Fund).
DetragiacheEnrica and HamannAlfonso1997 “Exchange Rate-Based Stabilization in Western Europe: Greece, Ireland, Italy and Portugal,” IMF Working Paper 97/75 (Washington: International Monetary Fund).
DornbuschRudi and GiavazziFrancesco1998“Hard Currency and Sound Credit: A Financial Agenda for Central Europe,”onhttp://www.mit.edu/~rudi/papers.html
International Monetary Fund1995Hungary—Recent Economic Developments and Background Issues SM/95/51.
KertesiaGábor and KöllôJános1997 “The Birth of the ‘Wage Curve’ in Hungary 1989–95,” paper presented at the Labor Markets in Transition Economies ConferenceUniversity of Michigan.
LeidermanL. and BufmanG. 1996“Searching for Nominal Anchors in Shock–Prone Economies in the 1990s: Inflation targets and Exchange Rate Bands,”inHausman and Reiseneds.Securing stability and growth in Latin AmericaOECD.
MassonPaul1998 “Monetary and Exchange Rate Policy of Transition Economies of Central and Eastern Europe after the Launch of EMU,” presented to the Fourth Dubrovnik Conference on Transition EconomiesCroatiaJune24–26. 1998.
MassonPaulSavastanoMiguel and SharmaSunil1997 “The Scope for Inflation Targeting in Developing Countries,” IMF Working Paper 97/130 (Washington: International Monetary Fund).
Ministry of Finance1997“Joint Assessment of Hungary’s Medium–term Economic Policy Priorities,”with the Directorate-General for Economic and Financial AffairsEuropean Commission.
National Bank of Poland1998Monetary Policy Framework for 1999Warsaw.
PolozSteveRoseDavid and TetlowRobert1994 “The Bank of Canada’s new. Quarterly Projection Model (QPM): An introduction,” Bank of Canada Review (Autumn): pp. 23–28.
SimonAndras and KovácsMihaly1998 “Components of the Real Exchange Rate in Hungary,” Working Paper 1998/3National Bank of Hungary.
SobczakNicolas1998 “Disinflation in Spain: The Recent Experience,” IMF Working Paper 98/106 (Washington: International Monetary Fund).
SurányiGyörgy and VinczeJános1998“Inflation in Hungary, 1990–97,”inCottarelli and Szaparyeds.Moderate Inflation: The Experience of Transition EconomiesInternational Monetary Fund and National Bank of Hungary.
SvenssonLars1997 “Inflation Forecast Targeting: Implementing and Monitoring Inflation Targets” European Economic Review Vol. 41 No. 6 pp. 1111–46.
Temprano–ArroyoHeliodoro and FeldmanRobert1998 “Selected Transition and Mediterranean Countries: An Institutional Primer on EMU and EU Relations,” IMF Working Paper 98/82 (Washington: International Monetary Fund).
van ElkanRachel1998 “Financial Markets in Hungary” inCottarelliCarloet alHungary: Economic Policies for Sustainable Growth IMF Occasional Paper No. 159 (Washington: International Monetary Fund).
Prepared by Craig Beaumont.
The participating central bank has primary responsibility for the stability of their exchange rate within ERM II, as interventions by the European Central Bank (ECB) to support the currency of an ERM II member must not impinge on the ECB’s primary objective of price stability. Timely realignments of central parities will be encouraged.
The Balassa-Samuelson effect on inflation equals the share of nontraded goods multiplied by the difference between labor productivity growth in the traded and nontraded goods sectors. See Masson (1998) for a derivation.
Ball (1994) finds that sacrifice ratios for faster disinflations are lower while controlling for the output-inflation trade-off for each country in non-disinflation periods, suggesting that the result does not just reflect the ability of countries with a low sacrifice ratio to disinflate faster. Interestingly, a study of European transition economies by Christoffersen and Doyle (1998) found no evidence of significant costs for disinflation, except, however, when substantially reducing a moderate inflation rate while maintaining a pegged exchange rate.
This perspective is supported by the initiation of aggregate employment growth in 1998, at 1.2 percent in the first half compared to the first half of 1997. Strong labor productivity growth, at 3.2 percent economy-wide in 1997, and 9.3 percent in manufacturing, suggests room for continued moderate real wage increases.
Finland may provide an example of interest, as during the approach to EMU Finland included “catch-up” clauses in its incomes policies. These clauses never came into effect, but were designed to give confidence to employees that they would not take a risk of real wage losses by accepting wage rises based on the inflation targets.
Sobczak (1998) finds that in Spain, the commitment to participate in EMU, and the implementation of fiscal policy consistent with this goal, generated credibility improvements that were that main element underlying the disinflation in 1996–97.
Countries with a sufficient degree of convergence may seek narrower fluctuation bands against the euro, but this must be approved by the European Commission, the Economic and Financial Committee, the ECB, and the finance ministers of the euro-area countries.
Poland has widened its crawling band in steps to ±12.5 percent by October 1998, and it intends to float the zloty by gradually widening the band and reducing the rate of crawl, National Bank of Poland (1998), whereas the Czech Republic and Slovakia have moved to managed floats.
Eichengreen and Masson (1998) provide an overview of exchange rate regime choice.
Sterilization liabilities include short-term (one-day, one-week, and one-month) deposits (formerly reverse repos) of banks with the NBH, and one-year NBH bills.
The key one-month reverse repo rate was cut from 19.75 percent to 19.5 percent on February 6, 1998 and to 19.25 percent on March 2, in comparison to a cut in the annualized rate of crawl of 1.2 percent.
The differential is calculated against German and U.S. interest rates weighted as in the basket, using the actual rate of crawl in place over the next three months.
The reserve requirement on domestic liabilities—denominated in either forint or foreign exchange—is 12 percent, with remuneration of 10 percent in December 1998, relative to Treasury bill yields of 16.4 percent.
Masson notes that a currency board may minimize the risk of speculative attacks, potentially making a fixed peg feasible. Dornbusch and Giavazzi (1998) recommend that CEE countries adopt currency board arrangements to attain a robust fixed exchange rate. Vujec instead proposes that the ECB form an exchange rate arrangement with the CEE countries negotiating to enter the EU as a means to make a peg feasible prior to their entry. Temprano-Arroyo and Feldman (1998) note that while no such pre-EU accession arrangement is officially envisaged, it is likely to be high on the EU political agenda.
Of Hungarian goods exports, 71 percent were to the EU in 1997, of which 87 percent were manufactures.
Widening the band will not itself lead to an appreciation, as the higher exchange rate risk will increase the risk premium on forint denominated assets.
This strategy was used by Portugal in the face of large capital inflows prior to its ERM entry in April 1992, by expanding the fluctuation margins to a few percentage points around the central parity in October 1990. Renewed inflows in mid-1991 were met by a tightening of capital controls, rather than a further widening of the band, perhaps partly reflecting the relatively narrow ERM bands at the time.
A new Act on the National Bank of Poland (NBP) came into effect on January 1, 1998, making price stability the primary objective of monetary policy, while also requiring the NBP to support the government’s economic policy so long as this did not conflict with the primary objective. The NBP elaborated a policy framework centered on inflation targets in October 1998, National Bank of Poland (1998). There are no intermediate targets for money, interest rates, or the exchange rate, these being policy instruments or indicators for achieving the inflation targets. The medium-term objective is to lower inflation below 4 percent by 2003.
An operation to securitize the foreign exchange revaluation losses of the NBH effective on January 1, 1997 strengthened the financial independence of the NBH, though it retains claims on the government with below market interest rates.
Interest rates are the most frequently excluded CPI item in Table 5, but interest rates are not included in the Hungarian CPI.
The Czech National Bank has chosen to focus monetary policy on “net inflation,” which excludes changes in administered prices. The National Bank of Poland has announced that inflation performance relative to the targets will be verified using underlying inflation indicators as well as the official CPI.
Examples of models that attempt to achieve this goal are Poloz et al (1994), and Black et al (1997). Both models used calibration techniques as well as traditional estimation, due to problems experienced with estimating models with plausible economic properties. This would suggest that the structural change that Hungary has experienced need not be an insurmountable barrier to the eventual development of a model.