Monetary Policy Transparency in the Inflation Targeting Countries: the Czech Republic, Hungary and Poland 1

Transcription

1 Monetary Poliy Transpareny in the Inflation Targeting Countries: the Czeh Republi, Hungary and Poland 1 Mariusz Jarmuzek *, Luan T. Orlowski **, Artur Radziwill *** Abstrat This paper evaluates transpareny of monetary poliy in the three New Member States of the EU that are pursuing diret inflation targeting strategy. Two dimensions of transpareny are quantified: (1) the institutional transpareny that indiates the extent to whih the entral bank disloses information related to its poliymaking proess and (2) the behavioural transpareny that reflets the larity of indiation given to the finanial market partiipants by entral bank with respet to its poliy stane. The paper shows an ambiguous assoiation between these two measures of transpareny, whih may be attributed to the ative exhange rate management poliy that undermines the atual transpareny as proxied by the behavioural measure. Keywords: monetary poliy, institutional transpareny, behavioural transpareny, diret inflation targeting, New Member States, monetary onvergene, euro JEL Classifiations: E52, E58, P52 1 This researh was supported by a grant from the CERGE-EI Foundation under a program of the Global Development Network. All opinions expressed are those of the authors and not neessarily of CERGE-EI or the GDN. The paper has benefited from omments by Wii Arulampalam, Stephen Broadberry, Marek Dabrowski, Martin Ellison, Jan Filaček, Jens Hölsher, Dennis Leeh, Marus Miller, Neil Rankin and Andrze Slawinski as well as the partiipants of the Eonomi Poliy Seminar at the University of Warwik. The authors are grateful to Rafal Beneki of the Bank Millennium and Radomir Ja of the Commerzbank Seurities for helpful omments on the data and finanial markets developments in NMS, and also to Halina B. Orlowski for extensive omments as well as oneptual and editorial assistane. The standard dislaimer applies. * Center for Soial and Eonomi Researh (CASE) and University of Warwik ** Sared Heart University and Center for Soial and Eonomi Researh (CASE) *** Center for Soial and Eonomi Researh (CASE) and Birkbek College, University of London. Corresponding author: Luan T. Orlowski, Sared Heart University, Department of Eonomis and Finane, 5151 Park Ave., Fairfield, CT Tel orlowskil@saredheart.edu

2 1. Introdution The main obetive of this study is to evaluate monetary poliy transpareny in the three New Member States (NMS) of the European Union (EU) that are pursuing monetary poliy based on diret inflation targeting (DIT) framework i.e. the Czeh Republi, Hungary and Poland. The ommon hallenge for these NMS is to meet the Maastriht riteria and to harmonise their monetary poliies with that of the European Central Bank (ECB) in preparation for adopting the euro. The main presumption of this paper is that the andidates to the euro need to take steps towards greater monetary poliy transpareny, omparable to that of the ECB. To this end enhaned transpareny an be viewed as an important aspet underlying a suessful monetary onvergene to the euro. The ECB s transpareny, therefore, serves as a benhmark for evaluating our results. Although the ECB does not follow the DIT strategy, its poliy is onsidered seemingly transparent as it identifies a well-presribed inflation target. For the purpose of our analysis, we define transpareny as the entral bank s far-reahing ommuniation of monetary poliy goals, strategies and operational tatis that is refleted in the larity of finanial market partiipants. We thus evaluate transpareny from the institutional and the behavioural standpoint. The institutional transpareny is understood as the extent to whih entral bank disloses information that is related to the poliymaking proess (Eifinger and Geraats, 2002) and the behavioural transpareny reflets the larity of indiation given to the finanial market partiipants by the entral bank with respet to its poliy stane. We further analyse the link between the institutional and the behavioural transpareny and put forth some hypothesis explaining a seemingly weak assoiation between the two. We measure the institutional transpareny in NMS by employing the methodology of Eiffinger and Geraats (2002). By adopting their survey questions we obtain the transpareny index for the three entral banks in NMS. This index omprises politial, eonomi, proedural, poliy and operational aspets of transpareny. We subsequently extend our evaluation of institutional transpareny by devising a supplementary questionnaire that aptures the speifi features of the inflation targeting framework. Our behavioural measure is based on the eonometri model originally proposed by Haldane and Reed (2000), but modified in our study to aount for speifi onditions of the three NMS. We believe that our dual approah to measuring monetary poliy transpareny ould bring a new dimension to the literature addressing various aspets of monetary onvergene to the euro. The paper is organised as follows. Setion 2 elaborates on the importane of monetary poliy transpareny from the institutional and the behavioural standpoint. Setion 3 presents the empirial investigation of the institutional transpareny by assessing information dislosure patterns in the Czeh Republi, Hungary and Poland. Setion 4 measures the behavioural transpareny in the three NMS and Setion 5 examines the link between both measures. Setion 6 enapsulates our findings. 2. The Importane of Monetary Poliy Transpareny In general terms, transpareny of monetary poliy an be asertained as an effort to mitigate information asymmetry between the monetary authority and the publi at large. Speifially, a transparent entral bank is prone to dislose full information about its poliy goals, strategies and tatis (Bernanke, et. al., 1999) as well as about its modelling assumptions, methods, and offiial foreasts. To this end, transpareny leads to more aurate expetations of the publi at large

3 and thus enhanes the redibility of the entral bank as finanial market partiipants an more easily udge its intentions. Chortareas et al. (2003) show that transpareny tends to lead to a lower average inflation and a lower variability in inflation and output. Thus in essene, transpareny is onsidered to be an important element in the design of a modern entral banking. This importane is undersored by the IMF (1999) in its Code of Good Praties on Transpareny in Monetary and Finanial Poliies. The signifiane of transpareny is also affirmed in pratie. Aording to the survey results of 91 entral banks onduted by Fry et al. (2000), 74% of them onsider transpareny to be an important feature of entral banking as important or very important. In this survey transpareny omes third in the rank of importane after institutional independene and the ability to ontain inflation expetations. Suh pereption is understandable given potential benefits from diminishing the information asymmetry between the monetary authority and the eonomi agents. First, the entral bank gains an opportunity to math its ations to its publi statements, whih is an effetive venue for strengthening redibility (IMF, 1999). Seond, transpareny helps align private setor expetations with the offiial foreasts of inflation and other monetary variables, whih redues their volatility and makes a ountry s monetary onditions less risky (Favero et al., 2000). Third, a higher degree of transpareny reinfores the expetations hannel of monetary poliy transmission relative to alternative hannels suh as the aggregate demand, the exhange rate, et. (Svensson, 1999). In the ontext of our researh, this gain is partiularly important, sine the expetations hannel plays a pivotal role in the DIT poliy framework, thus transpareny an be viewed as an important prerequisite for a viable introdution of the DIT (Mishkin, 2000), partiularly by monetary authorities in transition eonomies (Orlowski, 2000, 2001). We distinguish two basi approahes to measuring monetary poliy transpareny: (1) the institutional approah evaluates the extent of information dislosed by entral banks through available douments (Eiffinger and Geraats, 2002) and (2) the behavioural approah takes into onsideration the larity of indiation given to the finanial market partiipants by entral bank with respet to its poliy stane and an be assessed by measuring the impat of poliy announements on market interest rates (Haldane and Read, 2000). We ondut our analysis along both approahes. This allows us to determine the extent to whih entral banks an improve behavioural transpareny and thus the effetiveness of monetary poliy by simply hanging patterns of information dislosure Institutional Transpareny: The Eiffinger-Geraats Approah We find the Eiffinger and Geraats (2002) approah partiularly useful in our assessment of monetary poliy transpareny in NMS as it allows for a meaningful omparison of ommon riteria aross ountries as well as for an independent evaluation 2. They pereive transpareny as a multidimensional onept that omprises politial, eonomi, proedural, poliy and operational aspets of entral banking. These dimensions an be illustrated in relation to the simple model of the eonomy: 2 A similar approah is presented by De Haan et.al. (2001), although it is less omprehensive and based on arbitrary weighting of responses. Bernanke et al. (1999) evaluate transpareny of seleted ountries and Blinder et al. (2001) offer informal disussion of transpareny of maor entral banks. These studies do not provide ommon riteria for ross-ountry omparison. Fry et al. (2000) onstrut an index of poliy explanations for 91 entral banks on the basis of their survey that may lead to biased results. Other studies measure transpareny along one seleted dimension, for example, Chortareas et al. (2001) fous exlusively on the degree of detail of published foreasts in their survey of 87 ountries.

4 2 *) + A( y k ) 2 E ( i r ) + d L = ( π π y (1) y = y a π (2) E π = π + b( y y) + s (3) E i = f ( π, π, y) + e (4) Equation (1) is the entral bank loss funtion, π * is the inflation target, k y is the output target, y is the potential level of output, k>1 represents the politial pressure at high levels of output 3, and A is the relative weight attahed to output stabilisation that determines the willingness of a entral bank to allow deviations from inflationary targets (Cukierman, 2001b). Aordingly, a transparent entral bank will explain a politial deision about its preferene assigned to the output gap (a high parameter A) or prie stability. Equation (2) haraterises the aggregate demand where i is nominal interest rate, r is long-term interest rate and d is aggregate demand shok. In essene, it reflets eonomi growth onditions thus it pertains to eonomi transpareny. Equation (3) is a new-keynesian Phillips urve with the slope b and supply shok s. It is a basis for formulating a entral bank s reation funtion, whih underlines its poliy deisions and their eonomi reperussions. Equation (4) represents an instrument rule of a entral bank that generally should be onsistent with its loss funtion and the struture of the eonomy, while e denotes finanial shoks leading to the imperfet ontrol of a entral bank over i. Through its publi dislosure, a entral bank provides information about proedural aspets of interest rate deisions. Given this set-up, politial transpareny implies openness about poliy obetives, inluding an expliit prioritisation of potentially onfliting goals and quantitative targets or in other words openness about parameters of equation (1). This ould be benefiial as formal obetives and quantitative targets might diminish the unertainty about poliymakers preferenes and therefore redue inflationary bias, as shown by Nolan and Shaling (1996). Also IMF (1999) underlines the importane of mathing poliymakers ations to publi statements in order to enhane their reputation. Last but not least, institutional arrangements and the entral bank independene in partiular play a ruial role in the evaluation of politial transpareny as they might be ritial for preventing deviations from stated obetives. 4 Eonomi transpareny relates to the release of maroeonomi fundamentals underlying monetary poliy deisions inluding the available eonomi data, employed poliy models and the internal foreasts that the entral bank relies on. In other words it involves the symmetri information about both eonomi variables and strutural parameters in equations (2) and (3). Aording to Tarkka and Mayes (1999) monetary poliy is more preditable due to the release of the entral bank foreast. Furthermore, Geraats (2000) shows that the publiation of omprehensive entral bank foreasts or its eonomi model failitates reputation building and allows for greater flexibility to stabilise eonomi shoks. Geraats (2001) argues that only perfet transpareny about eonomi information ompletely eliminates the inflation bias. Chortareas et al. (2002) find that the greater degree of detail in the entral banks published foreasts is assoiated with lower average inflation. In a different vein, Gersbah (1998) and Cukierman (2001a) find that greater eonomi transpareny has a negative impat on stabilisation effetiveness in the absene of preferene unertainty. Geraats (2002b) also argues that eonomi transpareny might lead to greater politial pressure on a entral bank when its independene is not well established. 3 This an be attributed to the politial yles driven by eletions (Blanhard and Fisher, 1989) that an reate pressure on the government to support targeting output above the natural rate (Alesina, 1988). 4 The evidene of the importane of independene for redibility of monetary poliy and prie stability is doumented, among others, by Cukierman (1996).

5 Proedural transpareny means openness about the deision making proess and it might entail poliy rule or strategy and also release of minutes and voting reords that show the onsiderations underpinning the deisions. In other words, it involves openness about the funtional form of equation (4). Buiter (1999) and Gersbah and Hahn (2001a, 2001b) seem to suggest the positive onsequenes of dislosure of non-attributed voting reords and minutes, while there is still ontroversy about the release of attributed douments. The release of voting reords has an important informational value for finanial markets, as indiated by Clare and Courtenay (2000). In a similar vein, Buiter (1999) laims that finanial market partiipants an extrapolate the assessment of market risk based on the disagreement among the voting members of the monetary poliy ounil. 5 Poliy transpareny involves prompt announement and adequate explanation of the deision as well as poliy inlination to signal possible future ations, i.e. availability of information about i in equation (2). Tabellini (1987), Dotsey (1987), Rudin (1988), Cosimano and Van Huyk (1993) show that higher serey about short-term monetary targets impedes ahievement of monetary obetives partiularly if the knowledge about the long-term target is low and the information exhange between a entral bank and eonomi agents is imperfet. However, neither the importane of poliy explanation nor indiation has been unambiguously evaluated in the literature. Operational transpareny relates to ontrol errors in implementation of monetary poliy and disturbanes that affet the transmission of monetary poliy or the symmetry in knowledge about shoks s, d and e in equations (2), (3) and (4), respetively. Faust and Svensson (2002) argue that operational transpareny tends to redue the inflation bias and improve soial welfare; however, they also show (Faust and Svensson, 2001) that operational transpareny is likely to be quite limited in pratie. Jensen (2001) argues that although high operational transpareny might alleviate the low redibility problem, it may also impair the entral bank s ability to stabilise output. All five dimensions of the institutional transpareny desribed above are assessed for the entral banks of NMS in Setion Behavioural Transpareny: the Haldane-Read Model The behavioural approah to transpareny measures market pereptions that an be assessed diretly in a survey of finanial market partiipants or through analyzing impat of poliy announements on market interest rate movements. Haldane and Read (2000) propose a simple theoretial framework underlining the behavioural measure of transpareny. Their approah is based on the model of monetary poliy surprises that aptures interations between the private setor and the monetary authorities. It ontains three behavioural equations 6 : x t+ k = xt+ k + β tit+ m + ε t+ k t t m α 1 (5) * ( x x ) i + = δ (6) t t ( t+ it+ m+ ) t+ it+ m+ / 1 E (7) t 1 = t 5 It is worthy noting that a number of influential entral banks (inluding U.S. Federal Reserve, Bank of Japan and Bank of England) publish voting reords with some time lag following their meetings. 6 Equation (5) an be interpreted as a redued form of equations (2) and (3); equation (6) restates (4).

6 Equation (5) desribes the monetary poliy transmission mehanism, in whih x t stands for a vetor of the variables embedded in the entral bank s reation funtion 7, k is the average transmission poliy lag, β is a oeffiient explaining the monetary transmission mehanism 8 and t i t + m is the entral bank s offiial interest rate at time t with maturity m. Some information asymmetry is assumed, as ε t+1 (i 0) is known only to the monetary authorities at the time preeding deision-making on interest rates. Equation (6) an be pereived as a poliy feedbak rule that speifies oeffiient δ refleting disrepanies between the vetor of feedbak variables and the vetor of poliy targets. 9 Future poliy targets x * t+i (i 0) that are not known to the private setor agents reate another soure of information asymmetry. The private setor determines the yield urve refleting a predited path of future offiial interest rates based on the expetations theory of the term struture that is inluded in equation (7). The eonomi rationale behind this model is that a entral bank uses equation (5) embedding monetary transmission mehanism to infer x t when a shok ε t ours at the beginning of eah period (t-1). Subsequently, a entral bank sets future poliy targets x * t. Sine this part of the proess annot be observed by the private setor, preditions regarding a path of offiial * interest rates are based on its own udgements of the urrent and expeted values of x t and x t inluded in equations (5) and (6). These expetations an be extrated from the yield urve by analysing the forward rates. If a shift in the forward rates ours following the deision announement about offiial interest rates at time t relative to t-1 in whih the expetations were formed, the sets of feedbak variables x t and poliy targets x * t for a entral bank and the private setor were different. This an be interpreted as the lak of transpareny that an impair redibility of monetary authorities beause they were not able to provide suffiient information about their reation funtion and to onvine the market partiipants about their ommitment to ahieving the poliy goals. Based on this argument, the eonometri analysis of market interest rates responses to hanges in offiial interest rates may provide a useful behavioural measure of transpareny. We ondut this analysis in the empirial setion of our paper. 3. Institutional Transpareny in New Member States 3.1. Eiffinger and Geraats Index for NMS In this setion we employ questionnaire developed by Eiffinger and Geraats (2002) in order to ompile the transpareny index for the entral banks in the three NMS ountries that have adopted the DIT strategy. The questionnaire ontains five questions related to the five transpareny types. Sine eah question is based on three riteria and the sore for eah answer ranges from zero to one (all answers are equally weighted), the minimum sore of the evaluated entral bank an be zero and the maximum an be fifteen. The questionnaire is synthesized in Table 1, whih also shows the results of our evaluation. We derive answers to eah question from the information that was in the publi domain in English in May Subsequently, we obtain the set of indies haraterising eah transpareny dimension: politial, eonomi, proedural, poliy and operational. They are ompared to the ECB benhmark sore repliated from the Eiffinger and Geraats (2002).. Insert Table 1 around here. 7 These variables an be inflation, output, or employment (Svensson, 1997). 8 Haldane and Read (2000) assume that β is known for both the entral bank and private setor. 9 It is assumed that both the entral bank and the private setor know the δ parameter.

7 The assessment of politial transpareny involves inquiry about the existene of formal obetives, a quantitative target and independene. All three entral banks in our sample have formal obetives of monetary poliy, all expliitly speify a quantitative target for the prie stability and all have guaranteed institutional independene from the government, mostly in the form of expliit instrument independene 10. In the ase of Poland this independene is guaranteed by the Constitution. A similar legislation is to be introdued in the Czeh Republi in an effort to omply with the EU law. In sum, in terms of politial transpareny the three entral banks are in line with the ECB benhmark sore. In evaluating eonomi transpareny, we fous on three riteria: the dislosure of eonomi data, the poliy models and the offiial foreasts. To assess the timely release of eonomi data, Eiffinger and Geraats look at the dislosure pratie of money supply, inflation, GDP, unemployment rate and apaity utilisation data 11. These data, with the exeption of apaity utilisation, are published in all three ountries on at least quarterly basis, either by the entral bank or the national statistial offie. Estimation of apaity utilisation still poses a methodologial hallenge for ountries undergoing maor strutural hanges and is available only for Hungary, whih reeives a full sore for this question, while the two other NMS reeive a redued sore. The seond query is based on the assumption that in order to interpret the entral bank s poliy ations it is important to know its underlying poliy models. Among the three NMS, only the National Bank of Poland (NBP) is yet to publily dislose its model. Furthermore, NBP does not release numerial internal foreasts for inflation and/or output. Sine the Czeh National Bank (CNB) and the National Bank of Hungary (NBH) publish quarterly foreasts, they reeive a full sore. In sum, only Poland seems to lag behind the ECB in eonomi transpareny at the time of our analysis. To measure proedural transpareny, we look for the entral banks desription of monetary poliy strategy, the publiation of minutes and voting reords. All entral banks in our sample pursue an expliit monetary poliy strategy, namely the DIT framework. With respet to the minutes, only the CNB releases a omprehensive aount of poliy deliberations within a reasonable amount of time, however, not detailed and non-attributed. Neither the NBH nor the NBP publishes the minutes, although both banks release poliy announements following monthly meetings of their governing ounils. The third riterion is based on the assumption that monetary authority should also release voting reords that would dislose how eah deision on its main operating instrument or target was reahed. The NBP is the only entral bank that publishes attributed voting reords. The CNB disloses non-attributed voting reords and therefore sores a half point, while the NBH does not dislose any voting reords and sores a zero. In summary, in the ategory of proedural transpareny the NBH sore is at par with the benhmark sore of the ECB that does not provide minutes, or voting reords, while both the CNB and the NBP outperform it. Poliy transpareny entails timely dislosure of poliy deisions, its explanation and the future bias. In terms of a prompt announement of monetary poliy deisions, all entral banks reeive a full sore. Two of them publish a poliy explanation, although in different forms. The CNB inludes the statement of poliy deision in its minutes, while the NBP provides 10 Formal guarantees of entral bank independene in NMS are in fat a refletion of the historially strong politial pressures. 11 Following the disussion of maor fators affeting monetary poliy ation outlined in the entral bank douments one might argue that some indiators of the fisal poliy stane along with the external imbalane and exhange rate misalignment ould be also onsidered under this ategory. In fat, the general government balane and the urrent aount data are available in eah of the examined NMS, at least on a quarterly basis. In addition, the inlusion of other fators suh as labour produtivity, real wages and unit labour ost ould further enhane transpareny.

8 explanation in the press release following eah meeting of the Monetary Poliy Counil. Only the NBH does not provide poliy explanation and therefore reeives a zero sore. It is not a ommon pratise among the examined entral banks to publish the poliy inlination, with the exeption of the NBP that announes a poliy bias, and thus sores a full point. To sum up, in terms of poliy transpareny the NBP emerges as a leader with a maximum sore, followed by the CNB. Both outperform the ECB, whih does not provide any poliy inlinations and publishes solely a vague poliy explanation. Hungary lags behind the two other NMS and the ECB. Operational transpareny enompasses ontrol errors, transmission disturbanes and evaluation of poliy outome. Control errors pertain to the degree of ahieving the poliy operating targets. All three entral banks provide suh evaluation on at least annual basis and all get a full sore. The seond riterion inquires whether a entral bank regularly provides information on unantiipated maroeonomi disturbanes that affet the poliy transmission proess. In other words, it tells us whether or not the entral bank inludes a disussion of past foreast errors. Only the NBH inludes suh disussion, while the CNB and the NBP provide merely a general assessment of developments in monetary poliy transmission mehanism, thus both sore half point. The next issue is whether a entral bank provides regularly an evaluation of the poliy outome relative to its maroeonomi obetives. In this respet all three entral banks obtain a full sore, although their analyses do not appear equally thorough. In summary, all examined entral banks seem to be more transparent in evaluating their poliy performane than the ECB. This underpins their redibility-building efforts, even though their analyses and poliy explanations are at times vague - the fator that is not aptured by the Eiffinger and Geraats index. As shown in Table 1, the overall sore does not fully reflet the pronouned differenes in the speifi transpareny riteria for the individual NMS. Moreover, the results imply that the institutional transpareny ahieved by the three NMS is impressively high and not inferior to that of the ECB. In the following setions we disuss why suh a favourable assessment might be somewhat misleading, as the high institutional transpareny does not translate into high behavioural transpareny The Inflation Targeting Dimension In this setion we dwell further on the qualitative aspets of institutional transpareny in the three DIT ountries. We want to address the highly diversified DIT onditions in the individual NMS that are examined in the literature (Orlowski, 2003). We, therefore, evaluate the monetary regimes in NMS against the DIT blueprint that ontains a unique, forward-looking inflation target, speified as a linear traetory with a ontinuous target monitoring and adustment on a rolling basis, and well-presribed reation rules on missing the target (Svensson, 1999; Orlowski, 2001; Jonas and Mishkin, 2003). Consistently, we supplement the analysis based on the original Eiffinger and Geraats questionnaire by devising a set of DIT-speifi questions about a number of poliy goals, exhange rate poliy, forward-looking harater, target speifiation, traetory, monitoring, adustment rules and overshooting proedures. Our additional questionnaire (Appendix 1) is applied to the three NMS as well as to the ECB sine we need to derive a benhmark sore. The results are reported in Table 2.. Insert Table 2 around here. We first ask a fundamental question whether the inflation target onstitutes the unique goal of monetary poliy. The CNB and the NBP reeive a full sore, as they do not speify other expliit goals beyond the one stated in the DIT strategy, although the CNB is also pursuing a

9 managed float but without a target referene rate. On the ontrary, the NBH is targeting the exhange rate band around a predetermined referene rate in addition to the inflation target thus it sores zero. The ECB reeives a full sore sine it follows a learly predetermined inflation target of maximum two perent based on the harmonised CPI. We then look at interventions in the foreign exhange market sine they may redue transpareny in the implementation of DIT strategy. The NBP is the only examined bank that does not intervene in the foreign exhange market and thus reeives a full sore. The CNB that has engaged only in small-sale interventions reeives a half-point, while the NBH sores no point for its prevalent heavy interventions. Sine the ECB intervenes infrequently and on a relatively small sale, it sores a half point. Next, we evaluate the onformity of poliy rule with DIT strategy. Aording to Rudebush and Svensson (1999), inflation foreast targeting requires the poliy instrument to be adusted in suh a way that a onditional inflation foreast (the intermediate target variable) hits the inflation target at an appropriate level. The ECB emphasises the forward-looking harater of onduted poliies, but only the NBH and the CNB formulate their monetary poliy based on suh a forward-looking rule, whih also implies publishing the inflation foreasts. Both entral banks publish a disussion about the outlook for the main determinants of inflation, taking into onsideration the upside and downside risks of missing the target. Consequently, they reeive a full sore. The NBP does not release foreasts, and thus it does not onform to the forwardlooking poliy rule. The next question deals with the hoie of the inflation target. In priniple, ore inflation has the advantage of greater stability and smaller vulnerability to shoks. Using ore inflation redues the danger of missing the target due to prie volatility that is outside the sope of monetary poliy. However, there are also substantial drawbaks of adopting ore inflation as a target, as it does not enompass adustments in administratively regulated pries. Also, ore inflation indexes have not proved to be more stable than the headline index in transition ountries (Jonas and Mishkin, 2003). The main argument in favouring CPI over ore inflation as a poliy target is that the private setor ustomarily uses it as a basis for forming inflation expetations. The ECB and the three analysed entral banks speify their poliy target on the basis of the headline CPI 12. In addition, all three employ various measures of ore inflation to the prie hanges that they an influene. The CNB used to employ net inflation as its main analytial and ommuniative indiator of inflation; however, it has swithed to targeting headline inflation sine April Thus all entral banks in our analysis reeive a full sore for the hoie of the inflation target. The subsequent question is based on the assumption that the inflation targeting entral bank should announe target traetory. Only CNB and NBP release information about pursuing inflation target traetory for the medium-term. They apply the target traetory for headline inflation by means of a ontinuous band. We assign a half point to NBH that provides operating year-end targets for the next two years, as well as the ultimate target of 3 perent to be met by 2005 instead of a ontinuous band. The next question asks whether a entral bank employs a rolling basis for quarterly assessment of target fulfilment. The two entral banks that meet this standard are the CNB and the NBP, as indiated in their Inflation Reports. The NBH reeives no point, sine it follows the year-end targets only, in spite of publishing a target traetory for the medium- term. We note that the pratie of the ECB is to provide suh assessment monthly. We assign a half point to the CNB and NBH for an ex post target hange of the inflation target. Neither one of them expliitly exludes a possibility of a target hange; however, suh ation has not taken plae sine the introdution of DIT. The same assessment applies to the 12 More preisely, the ECB inflation target is based on the Harmonised Index of Consumer Pries for the euro area.

10 ECB. The NBP has an inlination to hange the inflation target when meeting it beomes problemati (as in 1999 and in 2002) and therefore reeives no point. Ahieving the inflation target is the main fator determining the deision-making proess with respet to hanges in monetary poliy instrument at any stage. Nonetheless, a number of possible exogenous shoks or unforeseen events, the auses of whih remain fully or predominantly outside of monetary poliy ontrol should be taken into aount. Attempts to orret suh shoks in the short run would likely be ostly for monetary poliy. Thus, it is extremely important for monetary authorities pursuing DIT to work out target overshooting proedures. Only the CNB obtains a half point for releasing preisely defined esape lauses. Suh lause is also missing in the ase of the ECB. In ontrast to our previous evaluation based on the Eiffinger and Geraats index, this extended analysis shows that there are marked differenes in institutional transpareny among the three entral banks. The CNB poliy appears to be the most transparent as it is the only entral bank in our analysis that outsores the ECB. The overall sores of NBH and NBP lag behind. It shall be noted, however, that the NBP sores relatively better on questions related to the exlusivity of inflationary goal, while the NBH on the forward-looking harater of the poliy. 4. Behavioural Measures of Transpareny: Empirial Results In this setion we test whether the monetary poliy in the three NMS that are pursuing DIT is transparent based on the observed dynamis of finanial markets. Under a transparent poliy, a entral bank provides finanial market partiipants with a learly defined indiation of its poliy stane. Consequently, market interest rates losely follow the offiial rates. We ask whether the introdution of DIT strategy atually inreased suh defined behavioural transpareny. In order to address this issue we estimate the reation of short-term money market interest rates to hanges in offiial rates. This methodology follows Haldane and Read (2000) who argue that in a transparent monetary poliy regime, short-term interest rates antiipate to some extent hanges in offiial rates. Their argument relies on information asymmetry and the existene of a stationary stohasti equilibrium with full knowledge of the authorities reation funtion 13. The original eonometri model proposed by Haldane and Reed (2000) is as follows: ( L) t+ it+ m+ + γ t it+ m + δ D t it+ m et+ m t+ i t+ m+ = + β + + α (8) for =1,3,6,24,60,120,240, where stands for maturity of the forward rate expressed in months and t is a time index. It assumes that m=1, as the main instrument usually affets monthly money market rate; β(l) is a vetor of polynomial lag operator (L). The lagged dependent variables are used to remove autoorrelation, whereas D is a regime-shift dummy variable aimed at apturing the introdution of DIT. Both variables are speified in first differenes in order to examine ontemporaneous movements in interest rates. We modify the original speifiation beause of the underdeveloped forward interest rates markets in the examined NMS. Thus the speifiation 13 The original idea of Haldane and Read (2000) is to extrat measures of interest rate surprises along the yield urve. Therefore, the authors also examine whether the monetary poliy is redible by looking at the long-end of the yield urve. This approah is however not feasible in the ase of analysed NMS, as they had to bring inflation down (whih has shifted their yield urves signifiantly over the last deade) and as they lak developed forward rates market. These fators ould lead to misleading results in evaluation of redibility, but are not essential in the evaluation of transpareny that is the fous of this researh.

11 that is employed in the model exludes forward interest rates and embeds short-term money market interest rates. The modified equation an be expressed as follows: it = + β, ( L) it + γ it + δ D it ε t, α + (9) The graphial presentation of the data series for the three NMS and the ECB is shown in Figures Insert Figure 1 around here.. Insert Figure 2 around here.. Insert Figure 3 around here.. Insert Figure 4 around here. The eonometri estimation of the Equation (9) at the short-end of the yield urve 15 provides the evidene of a substantial improvement in transpareny of the CNB and the NBP (Table 3). The oeffiients for the variables measuring reation to hanges in the offiial repo rate are statistially signifiant and their estimated values are lower during the DIT period relative to the pre-dit. This improvement onfirms expetations about positive results of enhaned ommuniation strategy as well as the quality of analytial publiations sine the introdution of the DIT that seemingly allowed for more larity about entral banks reation funtions among finanial market partiipants. Based on lower values of estimated oeffiients γ and δ for Poland, one an draw a onlusion that the NBP is more transparent than the CNB. This is somewhat puzzling given the higher CNB sores on institutional transpareny. Nevertheless, they both lag behind the ECB, whose γ oeffiient is slightly lower than that of the NBP. The most intriguing are the results for Hungary. The empirial evidene suggests that the NBH s transpareny has delined following the introdution of DIT, leaving it behind both the NBP and the CNB, while under the exhange rate peg NBH seemed to be ahead of both of them. The noted deterioration in NBH behavioural transpareny has taken plae in spite of the improvement in information dislosure over the last three years. One may therefore argue that the ative exhange rate management of NBH and also the CNB tends to deteriorate the behavioural transpareny in ontrast to the NBP s pure float.. Insert Table 3 around here. 14 Our eonometri analysis is based on the offiial interest rates data available on the CNB, the NBH, and the NBP web sites as well as the data on money market interest rates obtained from the Reuters Database and the EoWin Database. The time series that are used in the estimations are sampled daily and over the period of January 1, 1996 to July 21, 2003 for the Czeh Republi, Deember 25, 1996 to May 23, 2003 for Hungary, and Marh 2, 1998 to May 23, 2003 for Poland. We gratefully aknowledge helpful omments on the data and finanial markets developments in NMS reeived from Rafal Beneki of the Bank Millennium and Radomir Ja of the Commerzbank Seurities. 15 Detailed results are provided in the annex.

12 5. The Link between Institutional and Behavioural Transpareny The behavioural measure of transpareny suggests that Poland has the most and Hungary the least transparent monetary poliy. These results are not fully onsistent with the institutional measure based on the Eiffinger and Geraats index, by whih the CNB appears to be the most transparent. The CNB s leadership seems to be onfirmed by the assessment based on our extended index of transpareny that inorporates the speifi features of the DIT strategy. Based on this DIT-speifi analysis, the CNB again emerges as the most transparent entral bank, although this result is at odds with the behavioural measure. Looking at individual ategories of institutional transpareny, the sores are more differentiated. For instane, the sore for the eonomi transpareny is inonsistent with the behavioural measure; it therefore seems ambiguous whether the publiation of the model, apaity utilisation and even internal foreasts has a signifiant impat on the behavioural transpareny. Moreover, while we strongly believe that the politial aspet of transpareny, in partiular the entral bank independene, is essential, there is not enough variation in sores among the examined NMS to draw a definite onlusion about its link to the behavioural transpareny. On the other hand, the poliy transpareny orresponds very well to the ranking of the behavioural measure. This suggests that the poliy explanation along with the statement of poliy inlination may have been strongly orrelated with the behavioural transpareny. Taking into onsideration our DIT-speifi index, the institutional transpareny seems to be orrelated with the behavioural transpareny in terms of a single poliy goal and the absene of ative exhange rate poliy. Poland is leading on both ounts. The two other NMS may be suffering from the fear of floating that poses a onstraint on their poliy ondut within the DIT framework. This might be explained by their high degree of openness relative to Poland. As a result, their frequent resorting to foreign exhange market interventions misleads the finanial markets and thus inhibits the behavioural transpareny. Although the NBP lags behind on some important aspets of DIT in partiular by relying on the bakward- rather than forward-looking poliy rule and by not publishing the inflation foreasts - these shortomings are more than offset by its firm ommitment to the inflation goal and by refraining from foreign exhange market interventions 16. In our opinion, it is therefore exatly the uniqueness of goal and absene of ative exhange rate poliy that beomes ruial in determining the behavioural transpareny. However, the analysis of Hungary also suggests that, when the exhange rate based poliy is not de fato abandoned, targeting the exhange rate translates into a more transparent monetary poliy than targeting inflation. This finding is potentially important for preparations for the euro aession and the preeding partiipation in the ERM II, as the exhange rate target needs to be given more onsideration. This, in turn, may potentially hinder the monetary poliy transpareny in NMS to some degree. 6. Conluding Remarks Our general finding is that the monetary authorities of Poland, Hungary and the Czeh Republi have advaned their monetary poliy transpareny to the level roughly omparable to that of the ECB. Sine improved transpareny an be viewed as an effetive venue of enhaning monetary poliy redibility, we feel entitled to argue that the three examined entral banks are ripe to oin the European System of Central Banks as its fully redible members. The demonstrated degree 16 This is onsistent with the assessment of the suess in DIT implementation by Jonas and Mishkin (2003).

13 of alignment of monetary poliy transpareny between the three entral banks and the ECB an be also interpreted as a progress on the road to the euro. Nevertheless, our examination of the three monetary authorities by the individual riteria of institutional transpareny identifies a number of defiienies that they still need to overome, in spite of their overall suess. Speifially, the NBP may improve transpareny by inreasing reliane of its poliy deisions on inflation foreasts and other forward-looking variables, as well as by publily dislosing its underlying poliy model and the offiial inflation foreast. Both the CNB and the NBP ould expand the range of eonomi data made available for publi dissemination. At minimum, the NBH needs to improve its proedural transpareny with respet to a publi dislosure of minutes and voting reords of the governing ounil meetings. Perhaps more importantly, the NBH ould beome more transparent if it provided a lear prioritisation of the inflation target over the exhange rate stability goal. In sum, by the institutional transpareny definition, the CNB appears to be the most transparent among the examined entral banks, even outperforming the ECB. On the behavioural transpareny grounds, the NBP emerges as a lear leader. This is a noteworthy aomplishment sine it implies an important venue for strengthening redibility and promoting the role of expetation hannel in the ondut of monetary poliy. While the behavioural transpareny has improved in the Czeh Republi and Poland following the adoption of DIT, it has delined in Hungary, as implied by our empirial exerise. However, more definite onlusions an only be drawn on the basis of further testing that would apture the dynami evolution of entral banking environment in the three NMS. In the near future the three NMS will be obliged to oin the ERM II on their way to adopting the euro. In doing so, they will have to supplement their predominant inflation target by the exhange rate stability obetive. It remains to be seen whether suh a strategy will dilute their ommitment to prie stability and, therefore, hinder their monetary poliy transpareny. In response to suh fears, new means of fostering transpareny will have to be explored and devised for suh a omplex poliy framework stemming from the neessity to fulfil the Maastriht onvergene riteria.

14 Referenes Alesina, A. (1988), Maroeonomis and Politis, in: Fisher, S. (ed.) NBER Maroeonomis Annual, Cambridge, MA: The MIT Press. Bernanke, B. S., T. Laubah, F. S. Mishkin, and A. S. Posen, (1999), Inflation Targeting: Lessons from the International Experiene, Prineton, New Jersey: Prineton University Press. Blanhard, O. and S. Fisher, (1989), Letures on Maroeonomis, Cambridge, MA: The MIT Press Blinder, A., C. Goodhart, P. Hildebrand, D. Lipton, and C. Wyplosz, (2001), How Do Central Banks Talk?, Geneva Report on the World Eonomy 3, ICMB, May. Buiter, W. H., (1999), Alie in Euroland, Journal of Common Market Studies, June, 37 (2), Chortareas, G., D. Stasavage and G. Sterne (2002), Does it pay to be transparent? International evidene from entral bank foreasts, Federal Bank of St. Louis Review, Vol. 84, No. 4, July-August, pp Chortareas, G., D. Stasavage and G. Sterne (2003), Does Monetary Poliy Transpareny Redue Disinflation Costs, The Manhester Shool, vol.71, no.5, pp Clare, A. and R. Courtenay (2000), Assessing the Impat of Maroeonomi News Announements on Seurity Pries over Different Monetary Poliy Regimes, Paper presented at the Bundesbank/CFS Conferene on Transpareny in Monetary Poliy on 16/17 Otober 2000 in Frankfurt am Main. Cosimano, T. F. and J. B. Van Huyk, (1993), Central Bank Serey, Interest Rates, and Monetary Control, Eonomi Inquiry, July, 31 (3), Cukierman, A., (1996), The Eonomis of Central Banking in: Wolf Holger (ed.), Maroeonomi Poliy and Finanial Systems, Proeedings of the Deember 1995 Tunis Congress of the International Eonomi Assoiation, The Mamillan Press. Cukierman, A. (2001a), Aountability, Credibility, Transpareny and Stabilization Poliy in the Eurosystem, in: Charles Wyplosz (ed) The Impat on Europe and the Developing Contries, Oxford: Oxford University Press, Cukierman, A. (2001b), Are Contemporary Central Banks Transparent About Eonomi Models and Obetives and What Differene Does It Make? Bundesbank Disussion Paper 05/01, January. de Haan, J., F. Amtenbrink, and S. C. Eiffinger (1999), Aountability of entral banks: aspets and quantifiation, Bana Nazionale del Lavoro Quarterly Review, vol. 209, pp Dotsey, M., (1987), Monetary Poliy, Serey, and Federal Funds Rate Behavior, Journal of Monetary Eonomis, Deember, 20 (3), Eiffinger, S. C.W. and P. M. Geraats (2002), How Transparent Are Central Banks?, CEPR Disussion Paper No. 3188, February. Faust, J. and L. E.O. Svensson (2001), Transpareny and Credibility: Monetary Poliy with Unobservable Goals, International Eonomi Review, May, 42 (2), Faust, J. and L. E.O. Svensson, (2002), The Equilibrium Degree of Transpareny and Control in Monetary Poliy, Journal of Money, Credit and Banking, Vol. 34, No. 2, pp Favero, C., X. Freixas, T. Persson and C. Wyplosz, (2000), One Money, Many Countries, Monitoring the European Central Bank No. 2, London: CEPR. Fry, M., J. DeAnne, L. Mahadeva, S. Roger, and G. Sterne, (2000), Key Issues in the Choie of Monetary Poliy Framework, in Lavan Mahadeva and Gabriel Sterne, eds., Monetary Poliy Frameworks in a Global Context, London: Routledge, pp

15 Geraats, P. M., (2000), Why Adopt Transpareny? The Publiation of Central Bank Foreasts, CEPR Disussion Paper 2582, Otober. Geraats, P. M., (2001), Preommitment, Transpareny and Monetary Poliy, Bundesbank Disussion Paper 12/01 September. Geraats, P. M. (2002a), Transpareny of Monetary Poliy: Does the Institutional Framework Matter?, mimeo, University of Cambridge, May. Geraats, P. M. (2002b), Central Bank Transpareny, Eonomi Journal Vol. 112, Issue 483, pp Gersbah, H. (1998), On the Negative Soial Value of Central Banks Knowledge Transpareny, Eonomis of Governane Vol. 4, No. 2, pp Gersbah, H. and V. Hahn (2001a), Should Individual Voting Reords of Central Bankers Be Published?, Bundesbank Disussion Paper 02/01 January. Gersbah, H. and V. Hahn (2001b), Voting Transpareny and Confliting Interests in Central Bank Counils, Bundesbank Disussion Paper 03/01 January Haldane, A. and V. Read (2000), Monetary poliy surprises and the yield urve, Bank of England Working Paper no IMF (1999), Code of Good Praties on Transpareny in Monetary and Finanial Poliies : Washington DC, Issing, O. (2001), The Euro Area and the Single Monetary Poliy, Osterreihishe Nationalbank, Working Paper 44. Jensen, H. (2002), Optimal Degrees of Transpareny in Monetary Poliymaking,.The Sandinavian Journal of Eonomis, Vol. 104, Issue 3, pp Jonas, J. and F. S. Mishkin (2003), Inflation targeting in transition ountries: Experiene and prospets, NBER Working Paper No Matousek, R. and A. Tai (2003), Diret Inflation Targeting and Nominal Convergene: The Czeh Case, Open Eonomies Review, Vol. 14, No. 3, pp Mishkin, F. S. (2000), Inflation Targeting in Emerging-Market Countries, Amerian Eonomi Review: AEA Papers and Proeedings 90(2), pp Newey W., K. West, (1987), A simple positive semi-definite, heterosedastiity and autoorrelation onstintent ovariane matrix, Eonometria, 55. Nolan, C. and E. Shaling, (1996), Monetary Poliy Unertainty and Central Bank Aountability, Bank of England Working Paper 54 Otober. Orlowski L. T. (2000), Diret Inflation Targeting in Central Europe, Post-Soviet Geography and Eonomis 41(2), pp Orlowski L. T. (2001), From Inflation Targeting to the Euro-peg: a Model of Monetary Convergene for Transition Eonomies, Eonomi Systems 25(3), pp Orlowski, L. T. (2003), Monetary Convergene and Risk Premiums in the EU Aession Countries, Open Eonomies Review, Vol. 14, No. 3, pp Rudin, J. R. (1988), Central Bank Serey, Fed Wathing, and the Preditability of Interest Rates, Journal of Monetary Eonomis, September, 22 (2), Rudebush, G. and L. E. O. Svensson (1999), Poliy Rules for Inflation Targeting in: John B. Taylor, ed. Monetary Poliy Rules, University of Chiago Press, pp Svensson, L. E. O. (2000), Open-eonomy Inflation Targeting, Journal of International Eonomis, 50, pp Svensson, L. E. O. (1999), Inflation Targeting as a Monetary Poliy Rule, Journal of Monetary Eonomis, 43(3), pp Tabellini, G. (1987), Serey of Monetary Poliy and the Variability of Interest Rates, Journal of Money, Credit and Banking, November, 19 (4), Tarkka, J. and D. Mayes (1999), The Value of Publishing Offiial Central Bank Foreasts, Bank of Finland Disussion Paper 22/99 Deember.