Occasional Papers - BNR

National Bank of Romania

Occasional Papers

No. 1 October 2002

Direct Inflation Targeting: A New Monetary Policy Strategy for Romania

The paper was prepared under co-ordination of Ph.D. Cristian Popa by: Surica Rosentuler, Elena Iorga, Wilhelm Salater, Daniela Ruxandra Sasu, Adrian Ionut Codirlasu.

ISSN 1583-3836

Note The views expressed in this paper are those of the authors and do not necessarily reflect those of the National Bank of Romania. The Research and Publications Department carried out the English version, and technical co-ordination. Reproduction of the publication is forbidden. Data may be used only by indicating the source.

Contents SUMMARY ...............................................................................................5 INTRODUCTION.....................................................................................6 1. PRECONDITIONS FOR ADOPTING THE DIRECT INFLATION TARGETING REGIME................................................7 1.1. Institutional prerequisites ............................................................7 1.2. Technical prerequisites .................................................................9 2. ASSESSING THE APPROPRIATENESS AND TIMING OF ADOPTING A NEW FRAMEWORK FOR CONDUCTING MONETARY POLICY IN ROMANIA .............. 11 3. ROMANIA AND THE PRECONDITIONS FOR ADOPTING INFLATION TARGETING .........................................14 3.1. Institutional prerequisites...........................................................14 3.2. Technical prerequisites ...............................................................17 4. ECONOMETRIC ANALYSIS OF THE LINKS BETWEEN INFLATION AND SELECTED MACROECONOMIC VARIABLES ......................................................................................19 4.1. Determinants of the CPI inflation rate .....................................21 4.2. Determinants of the CORE1 inflation rate...............................25 4.3. Determinants of the CORE2 and CORE3 inflation rate ...............29 5. CONCLUSIONS ................................................................................31 BIBLIOGRAPHY ...................................................................................33 ANNEXES

National Bank of Romania Occasional Papers No. 1/2002

SUMMARY In the 1990s, the growing concern for price stability as a prerequisite for sustainable economic growth determined the central banks in numerous countries to adopt a new monetary policy strategy, i.e. direct inflation targeting. In the “Preaccession Economic Programme” drawn up by the Government of Romania in 2001, the direct inflation targeting is mentioned as a first option for the political strategy that the National Bank of Romania takes into account for the years 2003-04, provided that the disinflation process has recorded durable gains, backed by coherent and time-consistent macroeconomic policies. The present paper aims at assessing the general macroeconomic framework adequate for adopting the inflation targeting strategy and the stage Romania currently finds itself from this perspective. Section 1 presents the criteria that the institutional and economic structures of a country must fulfil in order for this strategy to be successfully implemented and the reasons behind their setting. Section 2 analyses whether changing Romania’s monetary policy strategy is appropriate or not by considering five elements characteristic to a certain stage of economic and political development of a country. Section 3 assesses, based on the institutional and economic criteria presented in Section 1, the extent to which Romania is ready to adopt direct inflation targeting. In Section 4 the relevance of the statistical links between different inflation measures (CPI, core inflation) and real, monetary and fiscal variables is tested in order to determine whether building an econometric model for forecasting inflation is feasible. The paper concludes that for the time being Romania fulfils these criteria only partially, as: (i) the National Bank of Romania enjoys the autonomy necessary for implementing the strategy, but has not made a firm commitment to exclusively pursue the disinflation objective; (ii) the fiscal dominance, albeit on the wane, persists in the form of the Public Finance Ministry’s influence on not yet fully-fledged developed financial markets; (iii) the monetary policy transmission mechanism and the net debtor position of the NBR towards the banking system reduce the effectiveness of the monetary policy actions; (iv) progress has been made as regards a greater transparency of the monetary policy, but it is necessary to increase the clarity of the signals on the forward-looking view of the monetary authority; (v) the inflation projections of the National Bank of Romania cover at most 12 months and are not disclosed to the public. Taking into account the current stage of economic development and the forecast on the main macroeconomic indicators as shown in the Preaccession Economic Programme, we consider that Romania needs about two to three years to prepare the adequate framework for adopting the new monetary strategy, including initiation of required legislative proceedings. In this period the central bank should promote a proactive attitude by gradually introducing some elements that define the inflation targeting regime: –

consistently pursuing the inflation target and fully relieving the monetary policy of supporting other macroeconomic objectives;

–

consolidating the central bank’s credibility by avoiding adjustment of the initially announced inflation target during the year;

–

strengthening the central bank’s accountability relating to fulfilling the inflation objective by publishing an inflation report that should include the inflation forecast.

5


INTRODUCTION The monetary policy framework consists in institutional arrangements whereby the monetary policy decisions are adopted and implemented. Choosing the monetary policy framework generally depends on the financial system characteristics, the competence of the decision-makers within and outside the central bank, as well as on other institutional matters. The direct inflation targeting regime recognises the importance of the inflationary phenomenon in modern economies and, implicitly, the fact that ensuring price stability represents the most efficient way for the monetary policy to sustain the overall objective of long-term economic growth. After being widely used in industrialised countries in the 1990s, the direct inflation targeting strategy has become, in the aftermath of the Latin American and Asian crises, an attractive alternative for emerging countries as well. From the point of view of the theoretical dispute on rules versus discretion, which has dominated the dedicated literature during the last decades, inflation targeting attempts to reconcile the extremes. This monetary policy regime combines the constraints deriving from the set objective with a flexibility margin as regards the instruments used by the monetary authority. In order to achieve the targeted inflation, the central bank uses the rules as “guiding elements of the monetary policy conduct” (Svensson, 1998), without being obliged to apply them mechanically. The central bank thus keeps its capacity to react to unforeseen shocks by adjusting the level of its monetary policy instruments, provided that it does not sacrifice the final objective. The direct inflation targeting strategy is defined by the following features: –

non-equivocal commitment to price stability as the overriding objective of the monetary policy, ahead of the other traditional objectives (economic growth, increase in external competitiveness, coverage of the fiscal gaps or reduction of unemployment);

–

transparency of the monetary policy strategy by communicating the monetary policy objectives and decisions to the public;

–

increase in the central bank’s accountability for attaining the inflation target;

–

dependence on having in due time a whole set of information on relevant variables concerning the four macroeconomic areas (real, monetary, fiscal and external areas).

An inflation targeting regime allows monetary policy to focus on issues relating to the domestic financial environment and to better respond to shocks in the domestic economy. The moneyinflation relation is not the characteristic element of the direct inflation targeting, but represents a favourable prerequisite for the adequate determination of the monetary tools chosen by the monetary authority.

6


1. PRECONDITIONS FOR ADOPTING THE DIRECT INFLATION TARGETING REGIME 1.1. Institutional prerequisites Absolute priority for the inflation objective The institutional commitment towards price stability implies that monetary policy should be given a clear mandate whereby this objective is considered fundamental and thus takes priority over other objectives such as economic growth, external competitiveness or increase in employment. Instrumental independence of the central bank The direct inflation targeting regime requires an increased independence of the central bank in conducting monetary policy. In terms of institutional arrangements, a law is necessary to lay down the independence of the central bank’s board of directors (appointed on a long term and protected from arbitrary revocations) from the government and to grant the central bank full and exclusive control over the choice and implementation of the monetary policy tools. Harmonisation of monetary policy with fiscal policy The de facto independence of the central bank – defined as the discretion to use the available instruments in the way the central bank deems appropriate for attaining its objective – is constrained by fiscal dominance, which reduces the efficiency of the monetary policy measures. As long as monetary policy operates in an environment where the fiscal policy has the advantage of the first move, the room for manoeuvre of the central bank is noticeably narrowed (Niepelt, 2001). The central bank is fully independent only when the operational framework ensures that the inflation target overrides the fiscal objectives. In this case, it is fiscal policy that must adjust its conduct so as not to jeopardise the inflation target, which translates into the absence of any signs of fiscal dominance (Masson et al, 1998), i.e. (i) the public deficit financing directly by the central bank or by the banking system should be low or inexistent; (ii) the government should not depend on revenues from seigniorage; (iii) domestic financial markets should be sufficiently developed for allowing the trading of public or private debt instruments; (iv) public debt should be sustainable. Unless these conditions are fulfilled, protracted fiscal deficits will generate inflationary pressures that will undermine the effectiveness of the monetary policy in attaining any nominal target, the monetary authority being forced to adopt an accommodative stance. Masson et al. (1998) argue that in a country that recorded annual inflation rates in the range of 15-20 percent for a number of 3 to 5 years, monetary and fiscal policies are virtually inseparable, as the central bank is unable to achieve alone a significant and lasting reduction in inflation. The fiscal consolidation required for implementation of the direct inflation targeting regime depends to an overwhelming extent on the fiscal discipline. Fiscal non-compliance either in the form of non-payment of obligations towards the government budget or by allowing a lax wage 7


policy as regards the loss-making state-owned enterprises, contributes to the increase in the budget deficits beyond sustainable limits. The need to finance these deficits takes priority over the inflation objective and calls in question the main condition for the functioning of the inflation targeting regime. A well-developed financial system The underdevelopment of capital markets makes the fiscal dominance effects stronger. The economies encountering difficulties in accessing international capital markets, with low savings and a thin financial system, limit the government possibilities to borrow on the domestic market, leaving the seigniorage and other forms of financial repression1 as single options. Under such circumstances, the financial markets cannot provide information on medium-term inflationary expectations, which the central bank needs to prepare the direct inflation targeting strategy; moreover, financial intermediation institutions, which generally have a strong aversion against inflation and the inflationary policies, do not have the strength to put pressure on authorities to observe their commitments, when the implemented policies tend to deviate from the established inflation objective. The banking system fragility also represents a risk factor for the success of the direct inflation targeting strategy, as it increases the possibility that the central bank may sacrifice the inflation objective in order to avoid a systemic crisis. However, the recent experience of several developing and transitional countries proves that the implementation of inflation targeting is also compatible with a less advanced stage of financial market development. Empirical evidence shows that a full-fledged financial system facilitates the implementation of inflation targeting rather than being a sine qua non condition for it. Flexible exchange rate Another issue that may cause serious problems to the direct inflation targeting strategy is the dollarisation phenomenon, a characteristic of many emerging markets. The flexibility of the nominal exchange rate is a condition of the inflation targeting regime which stems from setting the inflation target as the overriding objective. The related risks are however not negligible in the case that sharp depreciation may induce the rise in the USD-denominated debt burden, while an appreciation of the local currency may lead to trade balance worsening. Under such circumstances, implementation of the direct inflation targeting in partially dollarised economies must be preceded by the introduction of strict prudential regulations and an effective supervision of financial institutions, which should ensure the system capacity to absorb exchange rate shocks. Transparency and accountability Monetary policy transparency is extremely useful to a central bank that has a strong antiinflationary stance. If the bank pursues a consistently anti-inflationary conduct, the public 1

Interest rate ceilings, high reserve requirements, sectoral lending policies, mandatory acquisition of public debt instruments.

8


communication of its objectives, instruments, procedures, decisions and projections enhances its credibility and its accountability towards the general public. The repeated attainment of the announced targets generates reputational gains and the public support for the central bank independence rises even in the absence of a clear, formal framework for assessing the bank performance. The explicit announcement of a numerical inflation target reduces the dynamic inconsistency risk in two ways (Mishkin, 2000): (a) firstly, the announcement of the target is equivalent to giving the central bank the mandate to use monetary policy tools to attain this target, so that both markets and the politicians are aware that the short-term objectives (economic growth, reduction in unemployment) will not take precedence over the inflation objective; (b) secondly, the public commitment to the inflation target strengthens the central bank’s capacity to resist the possible political pressures to promote an expansionary monetary policy. 1.2. Technical prerequisites Beyond the institutional prerequisites, adopting a direct inflation targeting regime implies dealing with some technical issues (Debelle, 1997): –

choosing an adequate price index (representative for the purchasing power of money and easily understandable by the public);

–

explicit setting of a quantitative target, of the accepted fluctuation band and of the time horizon over which the target is pursued;

–

building of an effective model for inflation forecasting by the central bank.

The choice of an adequate price index The choice of the price index in relation to which the inflation target is set depends on the methodological aspects of calculating the consumer price index2 and on its sensitiveness to supply-side shocks. The second criterion is the more so important in the transition countries that pursue the implementation of an inflation targeting regime as administered prices have a significant influence on the general price level. Under the circumstances, the implementation of the direct inflation targeting strategy implies either a very good communication between the monetary and the fiscal authorities as regards the timing and the dimension of the change in the administered prices or the exclusion of their effect on the target-index. In striving to isolate the influence of the non-monetary determinants, many of the countries that pursue direct inflation

2

There are three reasons why the inflation targets are set based on the consumer price index (or on some variant thereof) and not on the GDP deflator: (i) the CPI rather than the GDP deflator is familiar to the general public, (ii) it is timely, and (iii) it is not subject to revision.

9


targeting define their objective in terms of core inflation, with differing items excluded from the CPI (Annex 1). However, setting the target in terms of core inflation should be decided while taking into account that such a methodology is not easily understandable and acceptable by the general public; consequently, an information campaign aimed at ensuring the objective transparency and the public support for the new strategy should be carried out before implementing such a decision. The fluctuation band and the time horizon a) The width of the target band. The need to specify a bandwidth results from the imperfect control of monetary policy over the inflation rate. Given the long and variable lags of monetary policy transmission mechanism, and given the imperfect ability to forecast inflation, it is not possible to restrict the variability in inflation below a certain level. Bandwidth is also needed to maintain some flexibility of the monetary policy in responding to short-term shocks. The choice of bandwidth reflects a trade-off between announcing a tight band and breaching it occasionally, and announcing a wide band, that may be regarded a too loose monetary policy stance. A narrow band indicates a strong commitment of the central bank to pursue the price stability objective, but it may prove difficult to implement, thus undermining the credibility of the monetary authority. On the other hand, adopting a narrow band brings the benefit of a higher short-term accountability of the central bank for achieving the inflation target, but may induce instability in the instruments of the monetary policy – for achieving a given movement in inflation rate, the shorter the time horizon, the larger the intervention3. The width of the target band (or even its existence) represents a major difference among the inflation targeting regimes adopted in different countries (Annex 1). As concerns the Central and East European countries which adopted a direct inflation targeting strategy, most of them defined their objective in terms of a band with a width that was gradually reduced in line with the disinflation progress (for example, the Czech Republic started in 1998 with a ±0.5 percentage point band around a 6 percent target whereas in 2000 the fluctuation band was narrowed to ±0.1 percentage points around the 4.5 percent target). b) Time horizon. Setting the horizon of the targets essentially depends on the initial inflation rate and the duration of the transmission mechanism. In Canada and New Zealand, for example, authorities were allowed an 18-month period for attaining the initial target. The next targets were set at shorter time intervals (12 months), but once inflation has been reduced to the desired longterm level, both countries decided that inflation was expected to remain within the target bands at all times during a 5-year period. Similarly, Poland’s medium-term Monetary Policy Strategy sets a one-year horizon for achieving the inflation target, but specifies that this will be extended as the

3

Sudden movements in interest rates may destabilise financial markets, even though inflation continues to be within the target band. Similarly, the frequent changes in the exchange rate causes a conflict between the short-term price stability objective and maintaining the inflation target in the medium term as well as an undesired movement in domestic demand.

10


reaction lag of the economy to the central bank’s actions increases (against the background of low inflation and diversification of financial markets). Inflation forecasting capacity The central bank’s forecast of inflation is essential in an inflation targeting regime, given the forward-looking characteristic of such a monetary policy strategy4. However, making satisfactory projections depends on the existence of consistent time series that should allow the estimation of stable relations between inflation and other economic variables. It is necessary to have a sound knowledge of the monetary policy transmission mechanism and financial markets able to provide timely and reliable information to the central bank in response to its actions. In order not to become recursive, the central bank’s projections should not incorporate the expectations on its future actions that other institutions use in their forecasts. 2. ASSESSING THE APPROPRIATENESS AND TIMING OF ADOPTING A NEW FRAMEWORK FOR CONDUCTING MONETARY POLICY IN ROMANIA In the approach taken by Fry et al. (2000), a central bank should take into account at least five elements prior to adopting a new monetary policy framework: (a) history of the monetary policy in the country in question; (b) current inflation; (c) monetary policy conducted by other countries facing similar conditions; (d) main economic characteristics of the country; (e) main political and institutional characteristics of the country. The appropriateness of implementing the inflation targeting regime in Romania should be analysed by considering these five elements. (a) History of monetary policy. In the last decade, the NBR pursued a monetary targeting strategy, by using monetary base (M0) as the operational objective and broad money (M2) as the intermediate objective. It is noteworthy that the NBR’s monetary policy set quantitative targets both at intermediate and operational levels. The attempts at using exchange rate as an antiinflationary anchor failed, while interest rate was not taken into consideration to this end, as the monetary and extra-monetary conditions were adverse to this solution. The monetary strategy carried out by the NBR led to a gradual disinflation process. The inflation rate fell from extremely high, close to hyperinflation, annual levels (peaking at 295.5 percent in 1993, December/December), to the relatively moderate figure of 2001 (30.3 percent). This performance 4

The direct inflation targeting virtually comes down to ensuring that expectations stay within the target band, which means that the central bank must respond to the gap between the inflation forecast and the target before the inflationary pressures become visible.

11


is lagging far behind the results recorded by the other Central and Eastern European countries, whose annual inflation rates tend to stick to one-digit levels. Moreover, it should be mentioned that the disinflationary process in Romania lacked continuity, as it was interrupted by inflation flare-ups in 1997 (151.4 percent) and 1999 (54.8 percent). Even though the NBR’s monetary policy framework has lasted for a very long time – only Estonia, which adopted a currency board in 1992, came close to our country from this perspective – the basic principles of the monetary targeting regime have not been rigorously observed. (b) Current inflation. In Romania, the inflation rate, albeit on a downtrend, is still far away from running at a level compatible with the price stability concept. Theoretically, adopting a more discretionary regime is not a recommended choice at this point of the disinflation trend. Many central banks switched to the new monetary policy regime when inflation was already very low. The most notable exceptions to this rule were Chile and Israel, which successfully implemented inflation targeting starting from a 20-25 percent annual inflation rate. The lack of homogeneity in the empirical evidence shows that inflation targeting can be introduced on different bands of the level of the targeted variable, the extent to which inflationary expectations are low and can be controlled in the transition period between the two monetary policy regimes playing the decisive role. (c) Monetary policy of countries with similar economic conditions. In recent years, Central and Eastern European countries have shown the tendency to abandon the different variants of fixed exchange rate regime (from currency boards to adjustable pegs) in favour of more flexible arrangements, such as inflation targeting. The Czech Republic (in 1998), Poland (in 1999), and Hungary (in 2001) introduced this monetary policy strategy, giving up the solution they had opted for in the first stage of transition, i.e. use of the exchange rate as a nominal anchor. (d) Main economic characteristics of the country. Romania is a small, open economy where investments are mainly financed through the banking system and to a lower extent through the capital market. These characteristics, without making the Romanian economy an ideal candidate for inflation targeting, may be compatible with this regime, as the recent experience of other developing countries has shown. Among the countries that introduced this regime, some recorded inflation rates, budget or current account deficits close to those in Romania (Annex 2). Specifically, Chile ended 1990 with an annual inflation rate exceeding 25 percent, Israel (1991), Spain (1994), and Hungary (2000) recorded budget deficits of 4.5 percent, 6.4 percent and 3.7 percent of GDP respectively, and in 1997 the Czech Republic posted a 6.1 percent-to-GDP current account deficit. The persistence of high budget and current account deficits – above the ceilings of 3 percent5 and 5 percent6 respectively, beyond which these deficits are considered excessively high – could however preclude the effective functioning of the new strategy in Romania. These indicators are expected to fall below the above-mentioned 5

Maximum level established by Maastricht criteria Lawrence Summers, former USA Treasury Secretary, indicates this level as a ceiling beyond which any country should be concerned (The Economist, 5 January 1996, pg. 46-48). 6

12


limits by 20047, which creates much more favourable conditions for a successful implementation of inflation targeting. (e) Main political and institutional characteristics of the country. The de jure and de facto independence of the central bank represents a fundamental condition for the implementation of the inflation targeting regime. The new strategy cannot be applied unless there is a sound support of the political structures of our country for a strictly anti-inflationary monetary policy stance and for a consistent pursuance of a set of macroeconomic policies in accordance with the agreed and publicly announced numerical inflation target. From a strictly electoral perspective, a favourable timing for the change of the monetary policy framework could be at the start of the next election cycle, when, according to an optimistic scenario, inflation rate could decrease to a level comparable to that the Czech Republic, Poland and Hungary recorded when they adopted the same strategy while the twin deficits could fall below the conventional ceilings mentioned at point (d). If the change of the monetary policy framework is approached from a broader institutional perspective that also considers the nearing integration in the European Union, close to which this change would be less desirable, inflation targeting could be adopted in 2004. Change management is extremely important for the successful implementation of the new monetary policy regime. In this respect, two types of behaviour adopted by the monetary authorities related to the change in the monetary policy framework can be identified. The most frequent approach consists in a conservative attitude of full commitment to the current strategy. The central bank adopts a defensive position, trying to bring justifications to maintain the status quo and refusing de plano or delaying sine die any change in the monetary policy regime. In many countries where central banks excessively promoted this approach, the change occurred at a time of crisis and generated high economic and political costs. The most recent example is Argentina, where fostering upon the currency board long after it had stopped being beneficial led to a sweeping crisis, that could have been avoided – or at least softened – if a more flexible monetary policy framework had been adopted in due time. The other approach is forward looking, entails accepting the idea of change and lays the groundwork for its implementation. A proactive attitude of the central bank may ensure the smooth adoption of a new monetary policy regime. Poland is a good example in this respect. In the case of the NBR and of the possible shift from monetary targeting to inflation targeting, adopting the second approach is facilitated by the existence of some continuity elements, such as the disinflationary stance, flexibility of the exchange rate and the important role of the monetary aggregates. As regards the appropriate timing for initiating the direct inflation targeting regime, theoretically two alternatives can be considered: 7

In accordance with the Preaccession Economic Programme prepared by the Government of Romania in 2001.

13


–

immediate adoption;

–

implementation around 2004.

The arguments in favour of the second alternative are the following: –

the level of inflation, which will be comparable with that in other countries when they adopted the same regime;

–

the possible consolidation of NBR credibility by attaining the inflation objectives for two years in a row before introducing inflation targeting;

–

the existence of a 2-3 year time period for laying the foundation for a smooth functioning of the new monetary policy strategy (including ruling out fiscal dominance and reducing the current account deficit).

3. ROMANIA AND THE PRECONDITIONS FOR ADOPTING INFLATION TARGETING 3.1. Institutional prerequisites In Romania, the institutional criteria for adopting the direct inflation targeting strategy are partially fulfilled. In pursuit of Law No. 101 – the NBR Act, the National Bank of Romania enjoys independence of monetary and foreign exchange policy decisions, upon which politicians bear no influence. The same law limited the volume of borrowings that State Treasury may take from the central bank during one year to 7 percent of the government budget revenues in the previous year, while outstanding borrowings shall not exceed the total amount resulting from doubling the NBR’s capital and reserve fund. The central bank’s autonomy, laid down by law, is confirmed in practice as well, if the low turnover rate of the NBR governors8 is taken into consideration. Fiscal dominance fell significantly in the last years as the amount of loans granted by the NBR to the State Treasury declined, the policy of sectoral lending was abandoned and seigniorage revenues decreased. According to Masson et al. (1998) developing countries resort to this form of financing to a much greater extent than developed countries (in this case, seigniorage may reach up to 3 percent of GDP, compared with less than 1 percent of GDP in the industrial countries). In Romania, seigniorage averaged 2.5 percent of GDP in 1994-2000, but the 2001 value (1.4 percent) is below those recorded by the Czech Republic and Poland. The NBR’s fundamental objective – ensuring national currency stability to contribute to price stability – is compatible with direct inflation targeting. However, for certain periods of time this goal lost its primacy, as the insufficient co-ordination of the macroeconomic policy mix forced the NBR to support other economic segments encountering difficulties (foreign trade, the 8

Cukiermann (1992) considers that the turnover rate of governors represents one of the elements that define a central bank’s independence.

14


financial and banking system). Thus, the central bank’s attention was divided among attaining the inflation target, ensuring a controlled depreciation of the national currency, building up foreign exchange reserves and acting as a lender of last resort. Neither in 2001 was the NBR’s interest rate policy fully subordinated to the disinflationary objective, as the central bank followed the descending trend of interest rates induced by the Ministry of Public Finance in an attempt to reduce the costs relating to public debt. Table 1. Seigniorage* in Romania and in advanced Central and East European countries percent

1993 1994 1995 1996 1997 1998 1999 2000 2001 Average 1994-2001

Romania 3.4 3.9 2.0 2.9 1.1 2.3 3.1 1.9 1.4 2.3

Czech Republic … 4.8 8.7 0.1 0.0 4.2 ** 2.0 1.6 2.5 3.0

Hungary … -0.2 1.9 -0.9 2.3 1.7 2.3 1.8 0.2 ** 1.1

Poland 0.7 1.6 2.9 1.5 2.5 1.4 -0.1 ** -0.6 1.5 1.3

* calculated as the annual change in the monetary base divided by nominal GDP ** year of adoption of inflation targeting Source: NBR, IFS and central banks' publications.

As concerns the implementation of monetary policy decisions, the NBR does not yet have an effective instrument whereby to inform the markets on the monetary policy stance and, most of all, to influence the inflationary expectations. The interest rate channel, used by all central banks that pursue inflation targeting, was hampered in the last years by the net debtor position of the central bank vis-à-vis the banking system. The interest rates on the NBR’s sterilisation operations proved effective in orienting the market for the first time in 2001, but this tool needs to be further strengthened in order to play the part that inflation targeting strategy implies. The structural liquidity surplus is also manifest in the Central and Eastern European economies that have adopted inflation targeting, with the interest rate on reverse repo transactions taken as the central bank’s reference rate. Even though in 2001 the Ministry of Public Finance enjoyed a much easier access to foreign financing than in the past, the international context may become more restrictive in the future. Against the background of difficulties in strengthening financial discipline, this state of affairs may lead to the increase in government pressure on the domestic market. The Ministry of Public Finance has dominated financial markets by offering handsome returns at a low risk, thus rendering the access of economic agents to financing rather difficult. The improvement in the budget revenue collection procedures and in economic agents’ financial discipline is an imperative for the appropriate implementation of direct inflation targeting. 15


Romanian authorities succeeded in maintaining moderate budget deficits and did not resort to financial repression in the form of interest rate ceilings or mandatory investments in public debt instruments, but the shallow financial markets and the lack of a wide range of saving instruments have made the State Treasury have a strong position in the market. The banking system showed visible improvement in 2001, both in terms of capitalisation and financial performance of banks, and the NBR continued to better the prudential norms and the supervision activity. The systemic risk fell down, but time is needed for the positive results recorded last year to consolidate. Restructuring and privatisation of Banca Comerciala Româna and CEC (the Savings Bank) are important steps in this respect, as they hold a large share in total bank assets and enjoy a high potential for destabilising the banking system, in the case that depositors lose confidence in these institutions. The Bank Deposit Guarantee Fund does not yet have enough resources to deal with a bank crisis without the NBR support, and the central bank acting as a lender of last resort would conflict with the disinflation objective. Significant risks for the successful implementation of a direct inflation targeting regime may stem from the dimension of the external imbalance and from the extent of dollarisation9, especially in the context of capital flow liberalisation. Given that this type of monetary strategy requires exchange rate flexibility, fluctuations in exchange rate are unavoidable. In the case of a sharp depreciation (caused, for example, by a capital flow reversal or by the need to use the exchange rate as trade balance adjuster) the financial crisis risk would rise following the increase in the foreign-exchange-denominated debt burden and the worsening of financial standing of economic agents (enterprises and banks alike). The National Bank of Romania annually submits to the Parliament a report on the monetary policy conducted in the previous year and the guidelines for the current year, being accountable for attaining its objectives. The NBR releases monthly bulletins, annual reports and communiqués whereby it explains the general public the developments in real economy and financial markets, monetary and foreign exchange policy decisions, and the reasons why the inflation objective was not attained. As concerns the monetary policy stance, the information released refers only to the actions undertaken by the central bank in the reported period, while appropriate technical tools that could allow the development of a forward-looking approach are not yet available. The retrospective analyses are necessary, but not sufficient in case of adopting a direct inflation targeting regime, where the proactive attitude of the central bank plays an essential part. Moreover, since the central bank’s manoeuvre room and accountability for curbing inflation are not clearly specified, the impact of monetary policy decisions on the inflation rate is not stated. This information should be given in case of adopting direct inflation targeting regime, generally via an “Inflation Report” – the main communication vehicle of the central bank, whereby it takes responsibility for attaining the set target. 9

In December 2001, the share of foreign-exchange-denominated deposits in M2 stood at 42.8 percent, and foreignexchange-denominated loans accounted for 59.8 percent of non-government domestic credit and 45.3 percent of total domestic credit.

16


The central bank’s communication with the general public, with financial markets improved considerably in the last years, both through a greater openness to dialogue by the NBR management (in the form of meetings with banks, interviews, organisations of seminars, etc.) and through the transparency created by shifting from direct monetary policy instruments to indirect ones. The central bank’s efforts towards disinflation can materialise only provided that the general public trusts in the implemented strategy and adjusts its inflationary expectations in line with the announced target. So far, this occurred only to some extent, as the authorities’ credibility has been eroded by the frequent upwards adjustments in the initial inflation objective during the time horizon set for attaining that target. 3.2. Technical prerequisites If, in the case of institutional preconditions, some of them are met, when it comes to actual implementation, the National Bank of Romania does not yet have the technical tools necessary for adopting the inflation targeting strategy. The instability of the economic and legal environment, the insufficient restructuring of state-owned enterprises, the slow privatisation of loss-making companies, the depreciation of the national currency are just some of the causes that fuelled a lasting high and volatile inflation. The share of products with administered prices in the consumption basket has been significant, and the alignment of these prices to international prices has generated repeated inflation flare-ups, with knock-on effects on free prices. Many of the central banks that have implemented inflation targeting set their inflation target in terms of core inflation, by excluding from the consumption basket the goods and services whose prices are subject to strong seasonal adjustments, administrative measures or supply-side shocks. A comparison between the consumer price index (CPI) and core inflation index allows the identification of the limits of the central bank’s influence on the inflationary processes occurring in the economy, and at the same time, underlines the aspects that may endanger the anchoring of general public’s expectations to the target. The calculation of core inflation implies the direct exclusion from the CPI of some items that strongly influence the public’s perception of price developments (for example, administered prices) or the use of a complex exclusion algorithm10, which is difficult to be assimilated by the general public. Moreover, since core inflation is generally determined by the central banks, this could enhance the public distrust in the results obtained in pursuing the target. These reasons made the National Bank of Poland set its inflation target in terms of CPI and the Czech National Bank to replace (in April 2001) the net inflation index with the CPI. In the case of Romania, the CPI is used to measure inflation since the start of the transition period and it is deeply rooted in the operators’ perception; in the NBR, calculations on the level of core inflation have been done only recently and have not been made public. In view of introducing inflation targeting, the NBR should make a comparative analysis of the two price indices and decide which of them better serves the goal aimed at. In the case that core inflation is preferred, a campaign for familiarising the public with the new concept should be 10

Trimmed mean method, for example.

17


initiated, whereby the reason for the decision made and the calculation methodology should be presented. As the economic environment is further characterised by a high level of uncertainty it is difficult to build a macroeconomic model establishing the relationships between nominal and real variables precisely enough to allow the central bank to correctly measure the effectiveness of the monetary policy transmission mechanism and the effects of its decisions on inflation. The information obtained through econometric methods should be added to price trend analyses made for a period of at least 12 months in advance, based on surveys concerned with both supply (inflationary expectations of economic agents) and demand (consumers’ confidence). The lack of reliable information on inflationary expectations increases the risk that inflation projections made by the NBR will not materialise, which means making some incorrect monetary policy decisions and, finally, failing to attain the inflation target. According to Debelle (1997) and Masson et al. (1998), one of the main prerequisites of successfully adopting the inflation targeting regime is an appropriate instrument the central bank should use to forecast inflation and to assess the impact of monetary policy actions on inflation (dimension of effects, transmission lag, etc.). In order to fulfil this condition, many central banks that have implemented inflation targeting resort to econometric models for inflation – although there is not explicit requirement for using such methods – or, at least, use inflation leading indicators such as interest rate, broad money, exchange rate, industrial producer prices, wages, etc. In conclusion, adopting direct inflation targeting in Romania depends to a large extent on the strength of statistical relationships between inflation and the macroeconomic variables that can be used for its forecast. Even though the tests performed (Section 4) identified links between inflation (based on CPI and CORE1) and the exchange rate, M2, real interest rate on deposits taken by the NBR and government budget deficit, respectively, these links are not strong enough to allow for a reliable forecasting model to be built. Presently, the NBR makes only short-term inflation projections, as it does not have a methodology for medium-term forecasting, and the fact that the inflation forecast is not disclosed to the general public is yet another impediment in anchoring inflationary expectations of the economic agents.

18


4. ECONOMETRIC ANALYSIS OF THE LINKS BETWEEN INFLATION AND SELECTED MACROECONOMIC VARIABLES The present econometric analysis investigates the possibility of forecasting the inflation rate using econometric models. For this purpose, we computed the distribution of price changes for the main groups of goods and services and constructed bi- and multi-variate VAR models for identification of micro- and macro-economic variables that could be used to forecast the inflation rate as in the Christoffersen and Wescott (1999) approach. Most of the analyses regarding the inflationary process start from the hypothesis that price changes for the main components of the CPI basket are normally distributed. This normality is important for at least two reasons. First, as Ball and Mankiw (1995) and Pujol and Griffiths (1996) considered, the asymmetry coefficient or the high variability of the relative price changes can lead to an inflation propensity as a consequence of agents’ asymmetric behavior towards a price change in any direction (the inflation can rise only because of a non-normal distribution of the relative prices). Second, normality of the distribution is important because it facilitates the forecasting of the price movements. To construct the distribution of price changes in Romania, 35 subgroups of prices, included in the Consumer Price Index were used. For a period of 55 months (Jun. 1997 – Dec. 2001), 1925 observations resulted. Price changes for every subgroup were seasonally adjusted. Every price change was standardized by subtracting the average monthly inflation rate for every month in those 35 subgroups and dividing by the corresponding standard deviation. Distributions and characteristics of the unadjusted and seasonally adjusted series are presented in the Graphs 1a and 1b: Graph 1b. Seasonally adjusted series

Graph 1a. Unadjusted series 1 200

1 000 Series: unadjusted Sample 1 1,925 Observations 1,925

800

Mean 2.89E-07 Median -0.155450 Maximum 5.746657 Minimum -5.052641 Std. Dev. 0.985866 Skewness 1.158130 Kurtosis 11.82035

600 400 200

Series: adjusted Sample 1 1,925 Observations 1,925

1 000

Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis

800 600 400 200

Jarque-Bera 6,670.421 Probability 0.000000 0 -4

-2

0

2

4

6

-1.08E-11 -0.158435 5.746721 -5.745700 0.985867 3.444171 22.34144

Jarque-Bera33,811.06 Probability 0.000000

0 -6

-4

-2

0

2

4

6

According to the histograms, the departure of the distributions of price changes from the normal distribution is obvious. The Jarque-Bera test confirmed (at 1 percent significance level) the inexistence of the normal distribution. The distributions of price changes have a positive skew and a high excess kurtosis. 19


Non-normality of the distributions is explained, at least in part by: (i) frequent adjustment of administered prices; (ii) changes in tax regime, (iii) high weight in the consumption basket of foodstuffs, with high dependency on weather conditions. For identification of the variables with relevance on inflation determination, Granger causality was tested and bivariate VAR models were constructed, using inflation rate series (measured by CPI, CORE111, CORE212 and CORE313) and seasonally adjusted macroeconomic data series between Jun. 1997 – Dec. 2001. Using the results of bivariate VARs, multivariate VARs were constructed for every inflation rate measure (CPI, CORE1, CORE2 and CORE3). Series that were used in the econometric analysis: –

Consumer price index (CPI);

–

Core inflation (CORE1, CORE2 and CORE3);

–

Administered price index;

–

Industrial output index;

–

Gross average wage;

–

Net average wage;

–

Unit labor cost (ULC) (1994 = 100);

–

Unemployment rate;

–

Industrial production price index (IPPI);

–

Broad money (M2);

–

Narrow money (M1);

–

Net domestic assets in the banking system;

–

Real interest rate on deposits with the NBR;

–

Treasury bills real yield;

–

Real 1M BUBOR;

–

Nominal exchange rate (USD/ROL);

–

Government budget balance as a percentage of the budgetary income.

All series are in logs (except the real interest rates and the government budget balance) and seasonally adjusted (except T-bill yield and exchange rate).

11

CORE1 = CPI – administered prices CORE2 = CORE1 – prices of seasonal goods 13 CORE3 = 23 percent trimmed mean 12

20


4.1. Determinants of the CPI inflation rate Bivariate Granger causality tests (marginal significance level of the F-test) between CPI inflation rate and selected macroeconomic variables for 1, 2, …, 8 lags are presented in Table 2: Table 2 1

2

3

4

5

6

7

8

0.009 0.441 0.213 0.239 0.217 0.635 0.014 0.273 0.223 0.662

0.015 0.403 0.434 0.163 0.521 0.524 0.014 0.553 0.562 0.814

0.039 0.524 0.560 0.116 0.644 0.017 0.037 0.524 0.677 0.371

0.077 0.775 0.544 0.146 0.506 0.023 0.035 0.084 0.471 0.293

0.092 0.709 0.477 0.131 0.636 0.008 0.000 0.009 0.022 0.218

0.074 0.716 0.697 0.258 0.808 0.022 0.000 0.004 0.035 0.123

0.064 0.801 0.802 0.367 0.877 0.020 0.000 0.009 0.024 0.147

0.001 0.884 0.804 0.411 0.507 0.037 0.001 0.007 0.005 0.154

0.548 0.442 0.348 0.348 0.015

0.636 0.730 0.697 0.389 0.017

0.397 0.841 0.821 0.111 0.043

0.622 0.709 0.370 0.140 0.081

0.670 0.666 0.463 0.185 0.149

0.437 0.671 0.484 0.058 0.192

0.575 0.848 0.474 0.018 0.159

0.478 0.930 0.620 0.039 0.077

Lags Variables administered prices industrial output gross wage net wage ULC unemployment rate IPPI M2 M1 net domestic assets real interest rate on deposits with the NBR real yield on treasury bills real 1M BUBOR nominal exchange rate government budget balance

According to the marginal significance level of the F-test, at 5 percent significance level, CPI inflation rate is Granger caused by: –

administered prices;

–

unemployment rate;

–

industrial production price index;

–

narrow money (M1);

–

broad money (M2);

–

nominal exchange rate (USD/ROL);

–

government budget balance.

In constructing bivariate VAR models, for the determination of the number of lags, the marginal significance level of the F-test from the Granger causality test, the results of the Johansen cointegration test and the Akaike and Schwarz information criteria were taken into account. The impulse-response functions constructed from the bivariate VAR models between CPI inflation rate and macroeconomic variables are presented in Annex 3a.

21


According to impulse-response functions, inflation is influenced by the following macroeconomic variables: –

administered price index;

–

unemployment rate;

–

broad money (M2);

–

narrow money (M1);

–

net domestic assets in the banking system;

–

nominal exchange rate (USD/ROL).

In order to test the possibility of inflation forecasting using multivariate models, a multivariate VAR model was constructed, with 4 lags14 with the following macroeconomic variables: inflation rate (measured by CPI), broad money (M2), real interest rate on deposits with the NBR, nominal exchange rate and balance of the government budget. The diagnostic tests for the inflation regression equation confirmed its stability (CUSUM and CUSUM of Squares), the Jarque-Bera test rejected the null hypothesis of a normal distribution of the residuals (at 5 percent significance level) and the correlogram and correlogram of squared residuals did not show any autocorrelation or ARCH in the residuals. The multivariate model excluded two macroeconomic variables, although the bivariate VAR models showed a statistically significant relation - the unemployment rate and the administered price index – because the impulse/response functions in Graph 2 showed that the dynamic relations between inflation rate and these variables were counterintuitive (a rise in unemployment rate would cause a rise in inflation rate and a rise in administered prices would cause a drop in inflation rate). Moreover, because of the positive correlation between M1, M2 and the net domestic assets in the banking system, only the broad money was included in the multivariate model. Four multivariate VAR models were constructed, by introducing, apart from broad money and nominal exchange rate, other variables, although the link between these and the inflation rate was not confirmed by the bivariate models (unemployment rate, administered prices, gross and net wage, real interest rate for deposits with the NBR, government budget balance), the most robust results being obtained for the model that included broad money, nominal exchange rate, real interest rate for deposits with the NBR and government budget balance. Even in this case, the statistical relations were weak (according with the results of the impulse/response functions15 and variance decomposition – Graphs 2 and 3).

14

The number of lags was chosen based on the existence of a cointegration relation between the variables, the results of the bivariate Granger causality tests, the number of lags in the bivariate VAR models and the informational criteria. 15 The following ordering was used: nominal exchange rate, broad money, real interest rate on deposits with the NBR, government budget balance, consumer price index.

22


Graph 2 Response of inflation rate to one SD innovation +/- 2 SE of CPI

of M2

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1

0.0

0.0

-0.1

-0.1

-0.2

-0.2

-0.3

-0.3 1

2

3

4

5

6

7

8

9

10

1

2

3

of real NBR deposit interest rate

4

5

6

7

8

9

10

8

9

10

of nominal exchange rate

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1

0.0

0.0

-0.1

-0.1

-0.2

-0.2

-0.3

-0.3 1

2

3

4

5

6

7

8

9

10

9

10

1

2

3

4

5

6

7

of government budget balance 0.4 0.3 0.2 0.1 0.0 -0.1 -0.2 -0.3 1

2

3

4

5

6

7

8

23


Graph 3 Variance decomposition – percent of CPI variance due to CPI

broad money

80

80

60

60

40

40

20

20

0

0 1

2

3

4

5

6

7

8

9

10

1

2

3

real NBR deposit interest rate

5

6

7

8

9

10

8

9

10

nominal exchange rate

80

80

60

60

40

40

20

20

0

0 1

2

3

4

5

6

7

8

9

10

8

9

10

government budget balance 80

60

40

20

0 1

24

4

2

3

4

5

6

7

1

2

3

4

5

6

7


4.2. Determinants of the CORE1 inflation rate Bivariate Granger causality tests (marginal significance level of the F-test) between CORE1 inflation rate and macroeconomic variables for 1, 2, …, 8 lags are presented in Table 3: Table 3 1

2

3

4

5

6

7

8

0.017 0.367 0.232 0.250 0.195 0.699 0.063 0.299 0.245 0.728

0.007 0.584 0.257 0.126 0.166 0.403 0.109 0.186 0.246 0.706

0.263 0.725 0.874 0.481 0.903 0.105 0.114 0.961 0.817 0.905

0.381 0.988 0.274 0.249 0.807 0.108 0.260 0.019 0.090 0.856

0.207 0.980 0.447 0.420 0.836 0.123 0.255 0.052 0.067 0.752

0.052 0.945 0.397 0.414 0.797 0.088 0.151 0.061 0.077 0.575

0.145 0.828 0.492 0.491 0.785 0.094 0.189 0.296 0.365 0.051

0.094 0.837 0.643 0.656 0.935 0.272 0.342 0.241 0.056 0.078

0.542 0.370 0.095 0.413 0.194

0.558 0.324 0.167 0.372 0.012

0.606 0.770 0.079 0.017 0.008

0.556 0.599 0.131 0.068 0.032

0.446 0.399 0.020 0.123 0.044

0.188 0.525 0.040 0.059 0.051

0.217 0.294 0.025 0.020 0.053

0.151 0.491 0.026 0.046 0.069

Lags Variables administered prices industrial output gross wage net wage ULC unemployment rate IPPI M2 M1 net domestic assets real interest rate on deposits with the NBR real yield on treasury bills real 1M BUBOR nominal exchange rate government budget balance

According to marginal significance level of the F-test, at 5 percent significance level, CORE1 inflation rate is Granger caused by: –

administered prices,

–

broad money,

–

real 1M BUBOR,

–

nominal exchange rate (USD/ROL),

–


In constructing bivariate VAR models, for determination of the number of lags, the marginal significance level of the F-test from the Granger causality test, the results of the Johansen cointegration test and the Akaike and Schwarz information criteria were taken into account. The impulse-response functions constructed from the bivariate VAR models between CORE1 inflation rate and macroeconomic variables are presented in Annex 3b.

25


According to impulse-response function, the CORE1 inflation rate is influenced by: –

administered price index,

–

unemployment rate,

–

broad money,

–

narrow money,

–

net domestic assets,

–

real 1M BUBOR,

–

nominal exchange rate (USD/ROL),

–


Based on the results of the bivariate models, a 4 lags16 multivariate VAR model was constructed with the following variables17: CORE1 inflation rate, broad money, real interest rate for deposits with the NBR, nominal exchange rate and government budget balance. The diagnostic tests for the inflation regression equation confirmed its stability (CUSUM and CUSUM of Squares), the Jarque-Bera test did not reject the null hypothesis of a normal distribution of the residuals (at 5 percent significance level) and the correlogram and correlogram of squared residuals did not show any autocorrelation or ARCH in the residuals. This model confirms the existence of statistically significant links between the CORE1 inflation rate and the selected macroeconomic variables. The relations between CORE1 inflation rate and selected variables are more powerful than those resulting from the model in which CPI was used as a measure of inflation rate. Using the ordering: nominal exchange rate, broad money, real interest rate for deposits with the NBR, government budget balance, CORE1 inflation rate, the impulse-response functions and variance decomposition are presented in Graphs 4 and 5.

16

The number of lags was set by taking into account the existence of a cointegration relation between the selected variables, the results of bivariate Granger causality tests, the number of lags for bivariate VAR models and informational criteria. 17 Because of the reasons mentioned above, unemployment rate and administered prices index were excluded.

26


Graph 4 Response of CORE1 to one SD innovation +/- 2 SE of CORE1

of broad money

0.3

0.3

0.2

0.2

0.1

0.1

0.0

0.0

-0.1

-0.1

-0.2

-0.2 1

2

3

4

5

6

7

8

9

10

1

2

3


4

5

6

7

8

9

10

8

9

10


0.3

0.3

0.2

0.2

0.1

0.1

0.0

0.0

-0.1

-0.1

-0.2

-0.2 1

2

3

4

5

6

7

8

9

10

9

10

1

2

3

4

5

6

7

of government budget balance 0.3

0.2

0.1

0.0

-0.1

-0.2 1

2

3

4

5

6

7

8

27


Graph 5

Variance decomposition – percent of CORE1 variance due to broad money

CORE1 60

60

50

50

40

40

30

30

20

20

10

10 0

0 1

2

3

4

5

6

7

8

9

10

1

2

3

real NBR deposit interest rate

5

6

7

8

9

10

8

9

10

nominal exchange rate

60

60

50

50

40

40

30

30

20

20

10

10

0

0 1

2

3

4

5

6

7

8

9

10

9

10

government budget balance 60 50 40 30 20 10 0 1

28

4

2

3

4

5

6

7

8

1

2

3

4

5

6

7


4.3. Determinants of the CORE2 and CORE3 inflation rate Replacing CORE1 with CORE2 and CORE3 and keeping the other macroeconomic variables, the number of lags and the ordering (for comparability reasons), two multivariate VAR models were constructed. According to Johansen cointegration test, for the two VAR models, the variables are cointegrated; the diagnostic tests for residuals of the CORE2 and CORE3 regression equation and the CUSUM and CUSUM of squares tests confirmed the stability of those two models. However, the relations between CORE2 and CORE3 inflation rates and the selected macroeconomic variables are weaker than those reflected by the CORE1 multivariate VAR model. The impulse/response functions18 for the two multivariate VAR models are shown in the Graphs 6 and 7. Graph 6 Response of CORE2 to one SD innovation +/- 2 SE of CORE2

of broad money

0.3

0.3

0.2

0.2

0.1

0.1

0.0

0.0

-0.1

-0.1

-0.2

-0.2 1

2

3

4

5

6

7

8

9

10

1

2


3

4

5

6

7

8

9

10

9

10


0.3

0.3

0.2

0.2

0.1

0.1

0.0

0.0

-0.1

-0.1

-0.2

-0.2 1

2

3

4

5

6

7

8

9

10

9

10

1

2

3

4

5

6

7

8

of government budget balance 0.3

0.2

0.1

0.0

-0.1

-0.2 1

18

2

3

4

5

6

7

8

The same ordering as in the multivariate CPI VAR model was used.

29


Graph 7 Response of CORE3 to one SD innovation +/- 2 SE of CORE3

of broad money

0.20

0.20

0.15

0.15

0.10

0.10

0.05

0.05

0.00

0.00

-0.05

-0.05

-0.10

-0.10

-0.15

-0.15 1

2

3

4

5

6

7

8

9

10

1

2

3


5

6

7

8

9

10

8

9

10


0.20

0.20

0.15

0.15

0.10

0.10

0.05

0.05

0.00

0.00

-0.05

-0.05

-0.10

-0.10

-0.15

-0.15 1

2

3

4

5

6

7

8

9

10

9

10

of government budget balance 0.20 0.15 0.10 0.05 0.00 -0.05 -0.10 -0.15 1

30

4

2

3

4

5

6

7

8

1

2

3

4

5

6

7


5. CONCLUSIONS The NBR’s current monetary strategy includes elements such as the disinflationary stance, flexibility of the exchange rate and the important role of monetary aggregates, which are also present in the direct inflation targeting regime. However, at present, Romania fulfils only partially the preconditions required for the implementation of this type of regime (Table 4), and a 2-3 year period is necessary in order to prepare an appropriate framework. Given the current economic development stage and the forecast for the main macroeconomic indicators in the Preaccession Economic Programme, the year 2004 can be considered as an adequate moment for adopting direct inflation targeting. In the period prior to the change of the monetary policy regime, the central bank should promote a proactive attitude by gradually introducing some elements that define inflation targeting: –

consistent pursuance of the inflation target and complete freeing of monetary policy from supporting other macroeconomic objectives;

–

consolidation of central bank credibility by avoiding the adjustment of the initially announced inflation target during the year;

–

strengthening the central bank’s accountability for the assumed objective by publishing an inflation report that should include an inflation forecast.

Concurrently with incorporating several elements of inflation targeting in the monetary strategy, the NBR should initiate the legislative procedures for the formal adoption of the new regime. A similar approach was taken in New Zealand (the first country that introduced inflation targeting), i.e. the strategy was adopted de facto in 1988 and de jure two years later, in March 1990, when the new statute of the central bank came into force. A decision is also necessary to be made regarding the price index based on which the inflation target will be set, other central banks that pursue inflation targeting opting either for the total CPI or for excluding from the CPI some items which differ from country to country. In the case of Romania, the choice is between the consumer price index (which has the advantage of accessibility and transparency) and the core inflation indices (which, by excluding the influence of administered prices and/or seasonal factors, allow the insulation of the shocks exogenous to the monetary policy). If one of the core inflation measures is preferred, a campaign for preparing the general public is necessary to be initiated so that, when the new monetary strategy is implemented, the announced inflation target should not be wrongly perceived, thus jeopardising the anchoring of inflationary expectations. Within the analysed monetary strategy, a crucial role is given to the inflation forecast, which is generally done by employing econometric models. In the case of Romania, the performed tests prove that such an approach is premature, given the fragility of the relationships between the leading indicators and inflation (measured by CPI, CORE1, CORE2). Once inflation rate stabilises at one-digit levels and the macroeconomic equilibria become stronger, these links are expected to improve and the use of econometric tools for making reliable medium-term forecasts will be possible. 31


Table 4. Preconditions for adopting direct inflation targeting and their fulfilment in Romania Prerequisites

Rationale

Are these requirements met?

Institutional prerequisites Instrumental independence of the central bank

The central bank needs discretion to attain its inflation objective.

Yes

Effective monetary policy instruments

The central bank needs at least one effective instrument to inform the general public on the monetary policy stance and to anchor inflationary expectations.

No

Accountability of the central bank

The central bank's responsibilities for attaining the inflation target should be clearly specified.

No

Transparency of monetary policy decisions

The central bank should communicate clearly and regularly its monetary policy decisions and their rationale to the government, financial markets and the general public.

Partially

Flexible exchange rate regime

The flexibility required for attaining the inflation target is not compatible with a fixed exchange rate regime.

Yes

Joint responsibility of government and central The government should commit itself to the inflation bank target, i.e. should promote economic policies that are consistent with the target.

Yes

Harmonisation of monetary and fiscal policies

The government should support, through its fiscal policy, the attainment of the set target.

No

Public support

The public should trust the strategy, as influencing expectations is crucial for the well functioning of this strategy.

No

A well-developed financial system

Financial markets should be developed and liquid enough to put pressures towards price stability and to provide information on inflationary expectations.

No

Choice of an appropriate price index

The inflation target should be set in terms of an index that is well understood by the general public and representative of developments in the purchasing power of money.

Undecided

Setting a fluctuation band for the inflation target and the time horizon

The degree of flexibility and the time schedule for attaining the inflation target should be pre-set, so that responsibility for attaining/missing the target be assumed.

No

Adequate knowledge of the transmission mechanism of monetary policy

Monetary policy makers should have timely and complete information on the effects of their actions on macroeconomic variables.

No

Forecasting inflation

The inflation targeting strategy should ensure the anchoring of inflationary expectations to the announced target, meaning that the central bank should rely on inflation forecasts in making monetary policy decisions and it should be proactive.

No

Technical prerequisites

32


BIBLIOGRAPHY Ball Laurence, Mankiw Gregory

“Relative Price Changes as Aggregate Supply Shocks”, The Quarterly Journal of Economics, vol. CX, nr. 1, 1995

Bernanke Ben, Laubach Thomas, Mishkin Frederic, Posen Adam,

“Inflation Targeting – Lessons From the International Experience”, Princeton University Press, Chicester West Sussex, 1999

Christoffersen Peter, Wescott Robert

“Is Poland Ready for Inflation Targeting?”, IMF Working Paper WP/99/41, 1999

Cukiermann Alex

“Central Bank Strategy, Credibility, and Independence”, Cambridge, Mass.: The MIT Press, 1992

Debelle Guy

“Inflation Targeting in Practice”, IMF Research Department Paper No. 35 (Washington: International Monetary Fund), 1997

Debelle Guy, Masson Paul, Savastano Miguel, Sharma Sunil

“Inflation Targeting as a Framework for Monetary Policy”, IMF Staff Papers-Economic Issues No. 15 (Washington, International Monetary Fund), 1998

Fry Maxwell, Julius DeAnne, Mahadeva Lavan, Roger Sandra, Sterne Gabriel

”Key issues in the choice of monetary policy framework” in Mahadeva Lavan and Sterne Gabriel ”Monetary Policy Frameworks in a Global Context”, 2000, Bank of England

Masson Paul, Savastano Miguel, Sharma Sunil

“Can Inflation Targeting Be a Framework for Monetary Policy in Developing Countries?”, article published in Finance & Development, based on IMF Working Paper 97/130, “The Scope for Inflation Targeting in Developing Countries”, 1998

Mishkin Frederic

“Inflation Targeting in Emerging Market Countries”, NBER Working Paper No. 7618 (Cambridge, Massachusetts: National Bureau of Economic Research), 2000

Niepelt Dirk

“The Fiscal Myth of the Price Level”, mimeo, Institute for International Economic Studies, Stockholm University, 2001

33


Peeters Marga

“Achieving Price Stability in the Euro Area: Monetary or Inflation Targeting?”, De Nederlandsche Bank, Economic Research and Special Studies Department, 2000

Popa Cristian

„Tintele alternative în orientarea politicii monetare”, BNR – Caiete de studii nr. 9, 2000

Pujol Thierry, Griffiths Mark

“Moderate Inflation in Poland: A Real Story”, IMF Working Paper WP/96/57, 1996

Svensson Lars

”Inflation Targeting as a Monetary Policy Rule”, Sveriges Riksbank, WP No. 62, 1998

34

ANNEXES

Annex 1

Characteristics of inflation targeting regime in different countries New Zealand

Canada

Chile

Israel

United Kingdom

Sweden

Czech Republic

Poland

Hungary

Date of introduction

1988*

1991

1991

1992

1992

1993

1998

1999

2001

Initial target

0-2%

2-4%

15-20%

14-15%

1-4%

2%

5.5-6.5%

2003

5 years (election cycle)

unlimited

2002-2005

annually

annually

CPI

net inflation (1998-Mar.2001) CPI (since Apr. 2001)

CPI

CPI

Country

Targeted index

core inflation

CPI

CPI

CPI

RPIX (CPI – mortgage rates)

Target announcement

agreement between MoF and CB

agreement between MoF and CB

…

Government

Treasury Chancelor

central bank

agreement between government and CB

central bank

central bank

quarterly

semiannually

semiannually semiannually

quarterly

quarterly

quarterly

quarterly

quarterly

yes

no

yes

no

yes

no

yes

Inflation report Forecast publication

yes

yes

Source: Landerretche Oscar, Morande Felipe, Schmidt-Hebbel Klaus – Inflation targets and stabilisation in Chile in Mahadeva Lavan and Gabriel Sterne – Monetary Policy Framework in a Global Context, Bank of England, 2000; Debelle Guy – Inflation targeting in practice, IMF Working Paper WP-97/35, 1997; Reports of central banks. Legend: * – the direct inflation targeting regime was adopted de facto in 1988 and de jure in 1990. ** – period from target adoption to attaining a stationary inflation level. x – a stationary inflation level has not yet been reached. … – missing data

1*

Annex 2

Macroeconomic context in several countries when direct inflation targeting (IT) was adopted

Country

Year prior to IT implementation

Inflation rate (annual average)

Current account balance

Government budget balance

(%)

(percent in GDP)

(percent in GDP)

New Zealand

1989

5.7

-3

…

Chile

1990

26.6

-1.7

0.8

Canada

1990

4.7

-3.9

-3.1

Israel

1991

19.7

-0.6

-4.5

United Kingdom

1992

3.7

-1.7

0.0

Finland

1992

2.6

-4.6

0.0

Spain

1994

5.3

-1.4

-6.4

Czech Republic

1997

8.5

-6.1

-0.9

Poland

1998

11.8

-4.4

-1.0

Brazil

1998

3.2

-4.3

-8.0

Hungary

2000

10.8

-4.3

-3.7

Source: IFS, Reports of central banks. … – missing data

2*

Annex 3a Impulse-response functions constructed from the bivariate VAR models between CPI inflation rate and selected macroeconomic variables

CPI response to an administered prices shock

CPI response to an industrial output shock 0.2

0.2

0.1

0.1

0.0 0.0 -0.1 -0.1

-0.2

-0.3

-0.2 1

2

3

4

5

6

7

8

9

10

1

2

3

CPI response to a gross wage shock

4

5

6

7

8

9

10

CPI response to a net wage shock

0.20

0.2

0.15 0.1

0.10 0.05

0.0 0.00 -0.05

-0.1

-0.10 -0.15 1

2

3

4

5

6

7

8

9

10

-0.2 1

CPI response to a ULC shock

2

3

4

5

6

7

8

9

10

CPI response to an unemployment rate shock

0.15

0.3

0.10

0.2

0.05 0.1

0.00 0.0

-0.05 -0.1

-0.10 -0.15 1

2

3

4

5

6

7

8

9

10

-0.2 1

CPI response to an IPP shock

2

3

4

5

6

7

8

9

10

CPI response to a broad money shock 0.3

0.3

0.2

0.2 0.1

0.1 0.0

0.0

-0.1

-0.2

-0.1 1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

3*

Annex 3a

CPI response to a narrow money shock

CPI response to a net domestic assets shock

0.3

0.2

0.2

0.1

0.1

0.0

0.0

-0.1

-0.1

-0.2

-0.2

-0.3 1

2

3

4

5

6

7

8

9

10

1

CPI response to a real NBR deposit interest rate shock

2

3

4

5

6

7

8

9

10

CPI response to a real treasury bills yield shock

0.20

0.20

0.15

0.15

0.10

0.10

0.05

0.05

0.00

0.00

-0.05

-0.05 -0.10

-0.10

-0.15

-0.15 1

2

3

4

5

6

7

8

9

1

10

CPI response to a real 1M BUBOR shock

2

3

4

5

6

7

8

9

10

CPI response to a nominal exchange rate shock

0.15

0.2

0.10 0.1 0.05 0.00

0.0

-0.05 -0.1 -0.10 -0.15 1

2

3

4

5

6

7

8

9

10

-0.2 1

2

3

4

5

6

7

8

9

10

CPI response to a government budget balance shock 0.2 0.1 0.0 -0.1 -0.2 -0.3 1

2

3

4

5

6

7

8

9

10

4*

Annex 3b Impulse-response functions constructed from the bivariate VAR models between CORE1 inflation rate and selected macroeconomic variables CORE1 response to an industrial output shock

CORE1 response to an administered prices shock 0.20

0.2

0.15 0.1

0.10 0.0

0.05 0.00

-0.1

-0.05 -0.2

-0.10 -0.3

-0.15 1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

CORE1 response to a net wage shock

CORE1 response to a gross wage shock 0.15

0.20

0.10

0.15

0.05

0.10

0.00

0.05

-0.05

0.00

-0.10

-0.05 -0.10

-0.15 1

2

3

4

5

6

7

8

9

1

10

2

3

4

5

6

7

8

9

10

CORE1 response to an unemployment rate shock

CORE1 response to a ULC shock 0.20

0.15

0.15 0.10

0.10 0.05

0.05

0.00

0.00

-0.05 -0.05

-0.10 -0.15

-0.10 1

2

3

4

5

6

7

8

9

1

10

2

3

4

5

6

7

8

9

10

CORE1 response to a broad money shock

CORE1 response to an IPP shock 0.3

0.25 0.20

0.2 0.15

0.1

0.10 0.05

0.0 0.00

-0.1

-0.05 1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

5*

Annex 3b

CORE1 response to a narrow money shock

CORE1 response to a net domestic assets shock

0.2

0.2

0.1

0.1

0.0

0.0

-0.1

-0.1

-0.2

-0.2 1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

CORE1 response to a real treasury bills yield shock

CORE1 response to a real NBR deposit interest rate shock 0.2

0.2

0.1

0.1

0.0

0.0

-0.1

-0.1

-0.2

-0.2

1

1

2

3

4

5

6

7

8

9

2

3

4

5

6

7

8

9

10

10

CORE1 response to a nominal exchange rate shock

CORE1 response to a real 1M BUBOR shock 0.2

0.20 0.15

0.1 0.10

0.0

0.05 0.00

-0.1

-0.05

-0.2

-0.10 -0.15

-0.3 1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

CORE1 response to a government budget balance shock 0.3

0.2

0.1

0.0

-0.1

-0.2 1

2

3

4

5

6

7

8

9

10

6*