The Quality of Monetary Policy and Inflation Performance: Globalization and its Aftermath

THE QUALITY OF MONETARY POLICY AND INFLATIONPERFORMANCE: GLOBALIZATION AND ITS AFTERMATH*

manc_2195 617..645

byMARTIN T. BOHL

Westfälische-Wilhelms UniversitätDAVID G. MAYES

University of Aucklandand

PIERRE L. SIKLOS†Wilfrid Laurier University and Freie Universität

With few exceptions the last decades have seen reductions in inflationaround the world. This experience is the result of many factors. In thispaper, we seek to isolate one of the factors, namely, improvements in thequality of monetary policy. We estimate a gravity—like model andpropose an exhaustive analysis of the potential role for a large number ofinstitutional factors. Briefly, we find that institutional factors helpexplain inflation relative to the US experience, which is used as thebenchmark. Nevertheless, a role for greater central bank autonomy is afeature of the 1980s and early 1990s only.

. . . a policy mistake by a major central bank would give rise to risks of globalinflation. [CHINA MONETARY POLICY REPORT, QUARTER ONE, 2009,PEOPLE’S BANK OF CHINA]

1 Introduction

With a few unfortunate exceptions the last three decades has seen reductions ininflation around the world to the point that many would argue that furtherimprovements in price stability would offer only limited welfare gains. Thisexperience is the result of many factors, some of which are country-specific(e.g. see IMF, 2006), others reflect the spread of both trade and capital. In thispaper, we seek to isolate one of the common factors, namely, improvements inthe quality of monetary policy. As with any such study we have to be able to

* Manuscript received 19.8.09; final version received 19.8.09.† Keynote presentation for the University of Manchester’s Centre for Growth and Business

Cycle Research fifth conference on Growth and Business Cycle in Theory and Practice(25–26 June 2009). The research for this paper was partially conducted when the first andthird authors were visiting scholars at the Bank of Finland in 2007, while the third authorwas also Bundesbank Professor at the Freie Universität, Berlin, and the second author wasAdvisor to the Board at the Bank of Finland. Siklos is also Director of the ViessmannEuropean Research Centre, Senior Fellow at the Rimini Centre for Economic Analysis andthe Centre for International Governance Innovation (CIGI). A separate appendix is avail-able on request.

JEL Classification Codes: E42, E58, C33

The Manchester School Vol 79 No. S1 617–645 June 2011doi: 10.1111/j.1467-9957.2011.02195.x

© 2011 The AuthorsThe Manchester School © 2011 Blackwell Publishing Ltd and The University of ManchesterPublished by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK, and 350 Main Street, Malden, MA 02148, USA.

617

take account of the main range of other factors that could explain cross-country differences in inflation performance. We do this with a panel study of20 Organization for Economic Cooperation and Development (OECD) coun-tries covering the last 30 years of data so that we can draw some conclusionsabout the role of monetary policy.

There is considerable consensus at present over what constitutes ‘good’monetary policy (e.g. see Woodford, 2003). Nevertheless, not only is therecontinuing debate about the role that monetary policy has played in theglobal disinflation effort but there are disagreements about whether somemonetary policy frameworks have been more effective than others (e.g. seeFreytag and Schneider, 2007). Over much of the period the Swiss andGerman performance has been best, with monetary policy based on thecontrol of monetary aggregates. However, since the Volcker years, the USAhas also performed well relying on a somewhat more pragmatic, somewould say, flexible, policy. Still others have outperformed the UnitedStates, particularly with the rise of inflation targeting, and Switzerland andGermany no longer stand out as exceptional. Of course the comparison iscomplicated by the creation of the euro area, where a single monetarypolicy affects the inflation rates of 16 countries, 10 of which are in oursample.

Several papers attempt to document the extent to which inflation ratesaround the world have been driven by global factors (e.g. a partial listincludes Tootell, 1998; Gamber and Hung, 2001; Tytell and Wei, 2004;Bean, 2006; Pain et al., 2006; Borio and Filardo, 2007). In particular, therehas been keen interest in the role that China’s inflation rate has played indriving down inflation rates in various parts of the world (e.g. see Côté andDe Resende 2008). A related literature has documented how the MaastrichtTreaty, and the requirement of inflation convergence among EuropeanUnion member countries, may also have indirectly contributed to inflationdevelopments around the globe (e.g. see Siklos, 2008b, and referencestherein). There continues to be an ongoing debate about the role that thechoice of a monetary strategy plays in generating a particular inflationoutcome. For example, the jury is still out about whether the adoption ofexplicit inflation targets independently contributes to anchoring inflationaryexpectations (e.g. Mishkin and Schmidt-Hebbel, 2007). Most would notobject to interpreting an inflation targeting strategy as one that has suc-ceeded in reducing inflation. The controversy is whether this kind of mon-etary policy works better than others that have been tried in the past, andwere found wanting, such as a rigid exchange rate peg. Finally, there is alingering view that certain institutional characteristics that describe boththe relationship between the central bank and government, and the mannerin which the central bank sets the course of monetary policy, should also bea significant factor in overall inflation performance (e.g. Acemoglu et al.,2008). After all, the drop in inflation worldwide also took place when

The Manchester School618

© 2011 The AuthorsThe Manchester School © 2011 Blackwell Publishing Ltd and The University of Manchester

central banks both became more autonomous, accountable and transparent(e.g. Dincer and Eichengreen, 2007; van der Cruijsen and Eijffinger, 2008).Needless to say, there are sceptics of the notion that there are significantlinks between central bank institutional characteristics and inflation perfor-mance (e.g. Hayo and Hefeker, 2008).

Figure 1 plots inflation in the 1990s, as well as inflation since the 1960sfor different regions of the world. The apparent convergence of inflationrates seems clear from the data for the 1990s. However, a longer run per-spective suggests that similar cross-country inflation rates are not exactly anew phenomenon. In other words, we have seen inflation rates across theworld that were both low and stable before, although both the institutionaland policy environments of the 1960s versus the 1990s are, of course, mark-edly different.

This paper is firmly in the camp of studies relying on a fairly largecross-section of countries over the past decade and a half that asks: to whatextent are institutional or global factors, whether of the economic or politicalvarieties (e.g. see Eichengreen and Leblang, 2006), responsible for the fall inglobal inflation? However, unlike other studies (e.g. Borio and Filardo, 2007)we eschew reliance on a global measure of the slack in output, in part becausethis proxy has proved somewhat controversial (e.g. see Ihrig et al., 2007).Instead, we consider a measure of economic distance, originally proposed byAlesina and Grilli (1992) which, given the difficulties inherent in proxyingglobal factors in the first place, seems both an intuitively plausible andtractable way of capturing international influences on inflation in a cross-sectional model of the kind used here.

The rest of the paper is organized as follows. First, we briefly survey someof the approaches that have been taken to address the impact of globalizationon inflation performance. Next, we outline the methodology used to estimateour version of the determinants, institutional or otherwise, of inflation in across-sectional framework. Then, we turn to a description of the data and adiscussion of some stylized facts, prior to presenting the empirical results. Thepaper concludes with a summary and explores potential avenues for futureresearch.

Briefly, we find that institutional factors play a role in explaining infla-tion relative to the US experience, which is used as the benchmark. Never-theless, any reduction in inflation stemming from greater central bankautonomy is a feature of the 1980s and early 1990s. Thereafter, central banksin the OECD look very much alike. For this reason, the ongoing financialcrisis will provide in future not only a means of testing the robustness of ourresults but may well permit researchers to sort out the question whether somemonetary policy strategies are better than others. In particular, we may wellbe able to answer more definitively whether, as Mishkin (2009) argues, ‘. . .increased commitment to stabilizing inflation [is] the right thing to do whenwe are in the throes of a financial crisis . . . ’

Quality of Monetary Policy 619


(a)

–2

0

2

4

6

8

10

1990 1992 1996 1998 2000 2002 2004 2006

Industrial ChinaEuro area USA

Per

cent

Periods of 'high' inflationin China (1993–95)excluded

(b)

0

10

20

30

40

50

60

70

80

1960 1970 1980 1990 2000

DevelopingEuro areaIndustialUSA

Per

cent

Fig. 1 (a) Inflation in Four Regions: 1990–2006. (b) Inflation in Four Regions: 1960–2006Note: Inflation is defined here as 100 times the first log difference in the consumer price index.

Data are annual from the International Monetary Fund’s International Financial StatisticsCD-ROM.



2 Background Literature

There exist a variety of approaches taken to explore the connection betweeninflation and its global determinants. Some estimate a Phillips curve, or someversion of it, and ask whether the trade-off has weakened in recent years.Various reasons are given such as general improvements in the conduct ofmonetary policy or perhaps the so-called Great Moderation (Bernanke, 2004)which began during the mid 1980s, interpreted as a reduction in both thefrequency and size of economic shocks (e.g. Ball and Sheridan, 2005; Borioand Filardo, 2007).

Alternatively, global factors, especially during the second half of the1990s and early in the 21st century, may have contributed significantlyto explaining cross-country inflation performance. Nevertheless, the evi-dence concerning the role of globalization is not yet conclusive. Disagree-ments about the relative importance of global factors stem partly fromthe manner in which the aggregate supply curve is empirically specified,namely the extent to which backward versus forward-looking elements inthe trade-off are allowed to determine current inflation. Perhaps moreimportantly, there have been concerns expressed over the measurement andreliability of global measures of output slack. For example, Borio andFilardo (2007) estimate a weighted average of international output gaps,where the latter are approximated via Hodrick-Prescott filtering (otherfilters are also considered) and find that these help us understand the deter-mination of inflation. Ihrig et al. (2007) report that how the global outputgap measure is estimated matters and that Borio and Filardo’s results arenot robust.

A different approach has been to determine empirically the extent towhich inflation from China, whose economic importance has grown consid-erably in recent years, has been imported. This development has taken placeat the same time as views about the extent to which pass-through effects fromexchange rate changes to inflation have moderated considerably in recentyears as inflation expectations have become better anchored (e.g. Balliu andFuji, 2004; Campa and Gonzalez-Minguez, 2007). The role of pass-througheffects is not entirely divorced from an earlier literature, starting with Romer(1993), who reports a negative correlation between inflation and the degree ofeconomic openness.1 Accordingly, as globalization has grown in influence sohave economies become more open with the implication that they havebecome less inflation prone. Also contributing to this outcome has been therelaxation in capital controls (e.g. see Edwards, 2007).

1The sum of exports and imports to gross domestic product. While openness and other phenom-ena associated with what is commonly referred to as globalization may have exerted animpact on inflationary developments, Woodford (2007) finds it difficult to square this resultwith the notion of an increased desire on the part of central banks to set an independentcourse for monetary policy.



Another literature examines whether the international environment hasbecome more competitive, again in no small part because of the rise of Chinaand the remaining BRIC countries (viz., Brazil, Russia and India), therebydriving down inflation on a global scale (e.g. see Guerrieri et al., 2008).

Finally, an altogether different approach tries to find some commonfeatures in the data. These may emerge as a form of convergence in inflationacross various parts of the world, which can be approximated according towhether inflation differentials between pairs of countries are stationary. Themore cointegration that is found between inflation rates the fewer commontrends there are. Siklos and Wohar (1997) show that, in a sample of Ncountries, convergence requires not only that there must be N-1 cointegratingvectors but that they must be of the (1,1) form. Alternatively, as in Busettiet al. (2007), the simple inflation differential must be stationary. Clearly, ifthere is a global connection that drives inflation rates to be attracted to eachother, this can imply that inflation rates are cointegrated. Unfortunately, thisapproach is not directly informative about the source of the cointegration,namely whether it stems from institutional factors, the monetary policy strat-egy that is followed, or other factors.2

Indeed, this kind of strategy has been indirectly criticized because theappearance convergence in inflation is obtained by a regression to the mean.Ball and Sheridan (2005) contend that the superior inflation performance ofan inflation targeting regime, a policy choice that involves considerable insti-tutional transformation (e.g. see Bernanke et al., 1999), is merely a statisticalartifact (Ball, 2006). They estimate a regression of the form:

π π β β π εit it it it− = + +− −1 0 1 1 (1)

where π it is average inflation in country i over period t, or regime, while π it−1

represents average inflation in an earlier ‘period’, or regime. The regime inquestion is, of course, inflation targeting and whatever other regime precededit. An estimate of the parameter b1 is an empirical indication of the size andstatistical significance of a ‘regime shift’.

While Ball and Sheridan present evidence suggesting that inflation tar-geting did not represent an improvement (also see Dueker and Fischer, 2006),there have also been several critics of the approach implicit in equation (1),including Vega and Winkelried (2005), and Ragan (2005), who point out,among other faults with the Ball and Sheridan approach, that the results canbe sensitive to the choice of the sample period chosen to compare one regimewith an earlier one, and that the specification relies exclusively on realizedinflation. Inflation targeting is perhaps most distinctive in its attempt todirectly anchor inflationary expectations.3 There is often insufficient scope

2Stock and Watson (2003) ascribe to ‘luck’ a significant portion of macroeconomic performanceduring the Great Moderation.

3Other monetary policy frameworks also try to anchor expectations. A currency board e.g. triesto relate them to inflation in country of the backing currency.



given to the subtle but important differences that exist concerning the idio-syncratic features of different inflation targeting regimes. In addition, there istoo little recognition of the fact that inflation targeting has outlasted theobvious alternative ‘global’ monetary policy strategy, namely Bretton Woods(e.g. see Rose, 2007; Siklos, 2009).

The Ball and Sheridan test does not make allowances for the notion thatthe choice of the regime is possibly endogenous. Hence, countries may havechosen inflation targeting because of a poor earlier historical record withinflation while some (e.g. Lin and Ye, 2007) contend that the adoption ofinflation targeting amounts to a form of ‘window dressing’. Therefore, thiskind of regime cannot independently explain the successful reduction ininflation. However, the transparency normally entailed by inflation targetingmakes it rather difficult for the policy to be simply window dressing—one ofthe sources of its success is that the authorities have to make a visiblecommitment for the policy to be credible. Moreover, the test of accountabil-ity for an inflation targeting regime is more demanding than for competingmonetary policy strategies. Notably, the central bank is more autonomousbut, in the bargain, must publicly explain itself if the agreed to inflationcontrol ranges are violated.

Finally, the Ball and Sheridan approach essentially amounts to treatinginflation targeting as a ‘straw man’ thereby setting up the test for failure of theregime while ignoring other forces that may impinge on inflation perfor-mance, such as the role of ‘globalization’. This is the view taken by, amongothers, Rogoff (2003) who argues that global factors, difficult as they may beto measure, lie behind a significant portion of the reduction in inflationworldwide in recent years. Performance under inflation targeting somehowneeds to be compared with how performance might have been under analternative, and perceived to be, good monetary policy.

Another global factor, emphasized in the literature that addresses theconsequences of institutional arrangements more explicitly, concerns theimpact of central bank independence. To say that the evidence in favour ofcentral bank independence reducing inflation is mixed is an understatement.Supporters of a connection between central bank independence and levels ofinflation abound (e.g. Grilli et al., 1991; Cukierman, 1992; Eijffinger and deHaan 1996; Siklos, 2002) but a number of valid concerns and objectionsabout how de jure forms of central bank independence are constructed havealso been levelled at this literature (e.g. Banaian et al., 1998; Mangano,1998).4

Just as criticisms of the inflation strategy have emerged, there have beensimilar doubts raised about the resort to monetary aggregates to control

4Concerns about the subjectivity of de jure measures of central bank independence led someauthors to resort to proxies for de facto forms of central bank independence, such asturnover rates of central bank governors.



inflation (Bernanke and Mishkin, 1992), in spite of the prominent role of thiskind of indicator as one of the pillars in the monetary policy strategy of theEuropean Central Bank.

Finally, exchange rate regime choice considerations have often loomedlarge in the literature that considers the driving forces behind inflation per-formance (e.g. see Reinhart and Rogoff, 2004). In a related fashion, crises ofthe financial variety may also, perhaps indirectly, contribute to inflationperformance. Attempting to take account of this possibility also raises anumber of thorny econometric issues, some of which are highlighted in thefollowing section.5

Neglected in all the foregoing arguments is that the mere declaration of adesire for lower inflation, or inheriting a lower inflation from abroad, does nottake place in a vacuum. Institutions, and how they adapt to changing domesticand international circumstances, may also have had a role in changes ininflation performance worldwide. It is with this in mind that the methodologi-cal approach followed in this paper, while not entirely dissimilar to the Ball andSheridan approach, is substantively different in emphasis, as is shown below.

3 Methodological Considerations

3.1 Estimation Strategy

Quarterly data from 20 OECD countries, for the most part since 1990, areused in a panel setting.6 There are two novel aspects to the study. First, weestimate what essentially amounts to a gravity-like model. Second, wepropose generally a more exhaustive analysis of the potential role of a largenumber of institutional factors than has heretofore been done.

Hence, if we treat a particular country as having followed the mostdesirable monetary policies, on average, then the tests below emphasize theattraction, or pull, of that particular benchmark country in influencing theothers’ inflation. We have used the USA as our main focus, primarily becausethe US record on inflation arguably sets the standard for most countries, at

5In the sample of countries considered below, relatively few of them suffered either a banking ora currency crisis. The only exceptions are: Finland (1991–3), Japan (1997), Norway (1991)and Sweden (1991, 1993). In contrast, hardly a year has gone by during the 1990–2006period examined when one or several countries did not experience a banking or a currencycrisis. It is therefore unlikely that directly accounting for crises will be fruitful. For acompendium of crisis events, see Laeven and Valencia (2008).

6It is natural to want to consider as large as possible a panel to investigate the issues under study.However, if one takes seriously the results and Acemoglu et al. (2008)—and the commentson this study were rather critical—then institutional factors matter a great deal more whenthere are significant constraints on the political principals. There are, in addition, as willbecome clearer in the discussion below, data availability constraints. All the countries inour data set, save Portugal and Spain, are ones with high constraints on the executive. Seethe appendix in Acemoglu et al. (2008).



least over the period covered in this study.7 This transforms the problem intoone that asks: what determines the differential in country pair’s inflationrates. The basic specification can thus be written as follows:

π α α α ε*ijt t ij ijt= + + + ′ +0 b Zijt (2)

where π*ijt is the inflation differential between country i and j,8 and a arethree constants, the first common to all country pairs and years, the secondspecific to year t, and the third specific to country pairs but constantthrough time (e.g. these would include country and region dummies). Zijt isa vector of ‘gravity’ variables. Following the usual convention in the litera-ture bijt = b′ (though these are testable hypothesis, of course). Cheng andWall (2005) give an overview of specifications such as equation (2) andsuggest that a model which includes country fixed pairs effects performsbest in statistical terms. Note that one test of convergence, namely the unitroot or cointegration version of the test for convergence, would set at = aij

= b�Zijt = 0 implying that inflation differentials are stationary. One of manytests for panel unit roots can also be used to confirm or reject this hypoth-esis, as will be shown.

Critical to equation (2) is the definition of Zijt. Relying on the argu-ments made above these consist of measures of central bank autonomy,transparency and accountability, qualitative measures of governance, thetype of exchange rate regime or monetary policy strategy in place, ameasure of economic distance suggested by Alesina and Grilli (1992), and ameasure of relative ‘stress’ in monetary policy as a proxy for ‘distance’ inthe stance of monetary policy. Qualitative indicators of transparencyfrom Siklos (2002, Chapter 6, 2008b) were also employed. An alternativemeasure of transparency from Dincer and Eichengreen (2007) was used.Governance variables are constructed from Siklos (2002, 2008b). Theycapture the size and type of committee structure, the frequency of meetings,whether committee votes are publicly made, whether committee decisionmaking is sanctioned by legislation or not. There is a growing body ofevidence that associates the emphasis on greater transparency and account-ability in central banking with better inflation performance (e.g. see Eijffin-ger and Geraats, 2006). Exchange rate regimes and monetary policystrategy classifications are from Levy-Yeyati and Sturzenegger (2002) andReinhart and Rogoff (2004).9

7An alternative benchmark could be the euro area. Indeed, the tests below were also repeated forthis case and the conclusions are largely the same. One difficulty with this alternativebenchmark is that the euro area has only formally existed since 1999, although pseudo-datacan be generated as far back as the 1980s. While there may be other benchmarks one couldconsider it is difficult to conceive of better ones than either the USA or the euro area.

8Following Cukierman (1992), we define inflation as ln(1 + %DP/100), where %DP is the per centchange in consumer prices.

9Both of these indicators of exchange rate regimes have since been updated.



While measures of central bank autonomy, such as the seminal onecreated by Cukierman (1992), and since used by many authors (see Siklos,2008a, and references therein), have been criticized for valid reasons (e.g. seeBanaian et al., 1998; Acemoglu et al., 2008), they contain a grain of truthabout inflation outcomes (also see Polillo and Guillén, 2005). Finally, recentimprovements are refinements in the measurement of political and economicindependence of central banks (Arnone et al., 2007), building upon the origi-nal work by Grilli et al. (1991), also seem to be associated with changes ininflation performance, at least for several regions of the world.

Economic distance refers to an indicator of divergence in output perfor-mance defined as the ratio of standard deviations adjusted for the correlationin output growth between two countries based on a four-year movingaverage. Since Alesina and Grilli (1992) introduce this measure as a short-hand way of evaluating the likely costs of monetary union this is a naturalvariable to use in the present context to represent some of the economic forcescontributing to inflation convergence. Presumably, the smaller are the differ-ences in output performance between pairs of countries the smaller the costsof inflation convergence. In other words, economic distance can be thoughtof as a proxy for how globalization has influenced differences in economicperformance across countries, i.e. a global real factor that could be reflectedin the inflation differential vis-à-vis the USA. More precisely, the Alesina andGrilli (1992) statistic is evaluated as

σ σ ρi j ij( ) + +( )⎡⎣ ⎤⎦2 2 1 2

1 (3)

where s is the standard deviation of output growth in countries i and j, andrij is the simple correlation in output growth between countries i and j.Equation (3) makes it clear that economic distance is a function of thevariation in output growth (or inflation) in one country relative to somebenchmark, as well as how closely income (or inflation) are correlatedbetween these same pairs of countries. As output growth volatility in countryi rises over time relative to that of the USA, this will also likely reduce thecorrelation between the growth rate of the i,j country pair. Consequently,economic distance will rise and this reflects divergences in economic perfor-mance between pairs of countries. The data reveal in (not shown) US thatthere has been considerable variation in economic distance over the decade ofthe 1990s and, while cross-country differences persist, differences have nar-rowed since the early 1990s.

Contrary to the belief held by some that economic growth in the USAwas becoming decoupled from the rest of the world the evidence suggests, atleast for many of the OECD economies in our sample, that growth correla-tions have actually risen over time. Nevertheless, it is interesting to note thatthere is a discrepancy between those economies that chose an inflation tar-geting strategy, who have generally seen their growth correlations with the



USA falling, and other non-inflation targeting countries.10 Interested readerscan examine the relevant plots in an appendix (not shown).11

The measure of ‘distance’ in the stance of monetary policy is simply thedifference between country pairs ij in a measure of ‘stress’ introduced byClarida et al. (1998) that is meant to ‘. . . gauge how different interest ratesmight be but for a binding commitment to an exchange rate arrangement,monetary union, or currency board.’ The proposed stress indicator for eachcountry i is evaluated as the difference between the central bank’s interest rateinstrument and the rate predicted by an interest rate rule.

STRESS i zi i t jt n t jt= − ( ) − ( )+, β π γE EΩ Ω (4)

where i is the central bank nominal interest rate instrument, p is inflation, andz are other variables in the central bank’s reaction function (usually theoutput gap, but in open economies the real exchange rate too). STRESSij isthen the differential in stress measures between pairs of countries i and j. Thelarger is the value of the STRESS variable, the greater the ‘distance’ in themonetary policy stance between countries i and j. The weights b and g reflectthe central bank’s preferences for inflation versus output gap control. Forreasons explained below, instead of estimating these coefficients we imposevalues that are thought to capture how the ‘typical’ central bank trades-offinflation and output gap objectives in OECD economies.

3.2 Data and Stylized Facts

Annual data since at least 1990 are used to estimate a version of equation (2).The data are essentially either quantitative (e.g. inflation, output), or quali-tative (e.g. central bank autonomy, transparency) represented by categoricalvariables. More detailed sources are provided in a separate appendix avail-able on request.

3.2.1 Inflation. Figure 2 plots the inflation differentials vis-à-vis the USAduring the 1990s while the same plot showing the differential during thedecade of the 1980s is shown in Fig. 3. By the 1990s US inflation is not muchdifferent than inflation in almost all of the OECD countries considered in thisstudy. The results are not substantially different when the benchmark is the

10To be fair, the decoupling view was primarily associated either with the evolution of economicgrowth between emerging markets and the USA, and not necessarily between Organizationfor Economic Cooperation and Development economies and the USA. Siklos (2008b)shows that the decoupling hypothesis does not hold even when emerging market economiesare considered. However, there has been a reduction in growth correlations between econo-mies that have adopted inflation targeting and the USA that is noticeably stronger thanamong economies that do not target inflation and the US economic growth. These samecorrelations were not generated for the euro area and the economies in our sample since halfthe countries in this study eventually adopted the euro.

11It is also noteworthy that correlations can be interrupted by crisis (e.g. the 1992 Exchange RateMechanism crisis).



–3

–2

–101 1

990

1995

2000

2005

Au

str

ia

–3

–2

–1012 1

990

1995

2000

2005

Au

str

alia

–2.0

–1.5

–1.0

–0.5

0.0

0.5

1990

1995

2000

2005

Be

lgiu

m

–3

–2

–1012 1

990

1995

2000

2005

Ca

na

da

–3

–2

–101 1

990

1995

2000

2005

De

nm

ark

–1.2

–0.8

–0.4

0.0

0.4

0.8

1990

1995

2000

2005

Eu

ro a

rea

–3

–2

–101 1

990

1995

2000

2005

Fin

lan

d

–2.0

–1.6

–1.2

–0.8

–0.4

0.0

0.4

1990

1995

2000

2005

Fra

nce

–3

–2

–10123 1

990

1995

2000

2005

Ge

rma

ny

–4

–2024 1

990

1995

2000

2005

Ire

lan

d

–2

–10123 1

990

1995

2000

2005

Ita

ly

–5

–4

–3

–2

–101 1

990

1995

2000

2005

Ja

pa

n

–3

–2

–1012 1

990

1995

2000

2005

Ne

the

rla

nd

s

–3

–2

–101 1

990

1995

2000

2005

No

rwa

y

–3

–2

–1012 1

990

1995

2000

2005

Ne

w Z

ea

lan

d

–202468 1

990

1995

2000

2005

Po

rtu

ga

l

–4

–20246 1

990

1995

2000

2005

Sw

ed

en

–10123 1

990

1995

2000

2005

Sp

ain

–3

–2

–1012 1

990

1995

2000

2005

Sw

itze

rla

nd

–2

–101234 1

990

1995

2000

2005

UK

Inflation, Country i less US inflation

Fig

.2In

flati

onD

iffe

rent

ials

vis-

à-vi

sU

Sin

flati

on:1

990–

2006

Not

e:T

hedi

ffer

enti

alis

defin

edas

the

infla

tion

rate

inco

untr

yil

ess

infla

tion

inth

eU

SA.I

nflat

ion

ishe

rede

fined

asin

Fig

.1.



–8

–6

–4

–202 1

980

1985

Au

str

ia

–4

–202468 1

980

1985

Au

str

alia

–8

–4048 1

980

1985

Be

lgiu

m

–4

–20246 1

980

1985

Canada

–2

–101234 1

980

1985

De

nm

ark

–4

–2024 1

980

1985

Eu

ro a

rea

–20246 1

980

1985

Fin

lan

d

–20246 1

980

1985

Fra

nce

–6

–4

–20 1

980

1985

Germ

any

–4048

12 1

980

1985

Ire

lan

d

02468

10 1

980

1985

Ita

ly

–6

–4

–20 1

980

1985

Ja

pa

n

–6

–4

–20 1

980

1985

Ne

the

rla

nd

s

–4

–20246 1

980

1985

No

rwa

y

048

12 1

980

1985

Ne

w Z

ea

lan

d

05

10

15

20 1

980

1985

Po

rtu

ga

l

0246 1980

1985

Sw

ed

en

02468 1980

1985

Spain

–8

–6

–4

–20 1

980

1985

Sw

itze

rla

nd

–101234 1

980

1985

UK

Inflation, country i less US inflation

Fig

.3In

flati

onD

iffe

rent

ials

vis-

à-vi

sU

Sin

flati

on:1

980–

9N

ote:

The

diff

eren

tial

isde

fined

asth

ein

flati

onra

tein

coun

try

iles

sin

flati

onin

the

USA

.Infl

atio

nis

here

defin

edas

inF

ig.1

.



euro area, again for data since the 1990s. Note that some countries (e.g.Portugal, Sweden, even the UK) have experienced larger disinflations relativeto the USA than many other countries. Indeed, by the end of the 1990s,inflation differentials become small, often below 1 per cent. In contrast, the1980s (Fig. 3) sees US inflation as substantially lower than it is elsewhere inthe OECD. Moreover, the order of magnitude of the inflation differentialsduring the 1980s is many times higher than in the 1990s, and thereafter. Thechange in cross-country inflation performance is nothing short of dramatic.One of the only exceptions is Japan which returns consistently lower inflationthan in the USA.

Is the comparison between the 1980s and 1990s and beyond the correctone? To the extent that the last two decades or so have seen the introductionof major changes in the relationship between central bank and their govern-ments, on the one hand, and the public, on the other, as well as the absenceof major aggregate supply shocks, the answer is yes. Table 1 presents decanaldata on mean inflation differentials between the USA and the remainingOECD economies in our sample since 1960. Whereas relatively few countrieshad inflation rates that were, on average, lower than in the USA during the1960s (only one country), or the 1970s (four countries), the situation changesrapidly thereafter with six countries during the 1980s and 15 countries dis-playing inflation performance superior to that in the USA.12 Moreover, astylized fact often ignored, is the substantial drop in inflation volatility, withthe most notable drop occurring beginning during the 1990s. Volatility playsa role in the diminution of economic distance, as we shall see below.13

3.2.2 A Measure of the Perceived Success Monetary Policy Framework:Inflation Expectations. An obvious means of assessing credibility is toexamine a proxy for inflation expectations. Unfortunately, comparable to datagoing back decades is difficult to collect. Figures 4 and 5 plot forecast errorsfrom the International Monetary Fund’s World Economic Outlook since 1980and, depending upon availability, forecast errors based on the poll ofForecasters in The Economist or the Consensus Forecasts (http://www.consensuseconomics.com). A band around these forecasts of 10.5 per cent isalso shown. Admittedly, the size of the band is somewhat arbitrary but such aband translates, after 10 years, into a cumulative error of approximately 5 percent. This is judged to be sufficiently large so that failure to remain within this

12Not all the differentials are, of course, statistically significant. Blanchard and Simon (2001)drew attention to the drop in volatility of inflation and real economic growth beginning inthe mid 1980s that has since come to be called the Great Moderation.

13There are other ways to bring in a role for volatility but these will not be considered here anyfurther. The components of economic distance were previously described. A plot of eco-nomic distance is relegated to an appendix. It is interesting to note that while economicdistance has diminished between the USA and the other Organization for EconomicCooperation and Development economies in our data, the same is not necessarily true ofemerging markets (e.g. China; results not shown).



Ta

ble

1Su

mm

ar

ySt

at

ist

ics

Sam

ple

AU

AU

DB

EC

AD

NK

FI

FR

GE

IRIT

JPN

EN

ON

ZP

TS

ES

PS

WU

KE

A

1960

–9M

ean

1.06

-0.0

90.

500.

243.

242.

701.

380.

091.

911.

403.

011.

891.

381.

031.

681.

273.

650.

881.

35St

d.D

ev.

2.04

1.53

1.50

0.56

2.33

3.16

1.55

2.06

1.68

3.01

2.20

1.60

2.09

1.17

0.66

2.27

4.04

1.92

1.23

Obs

erva

tion

s9

99

99

99

99

99

99

99

99

99

1970

–9M

ean

-0.9

32.

480.

010.

262.

022.

981.

67-1

.96

5.09

4.68

1.75

-0.0

31.

193.

998.

711.

386.

50-2

.02

4.95

1.66

Std.

Dev

.2.

863.

062.

991.

601.

774.

021.

322.

803.

493.

654.

613.

353.

033.

035.

882.

494.

583.

674.

142.

09O

bser

vati

ons

1010

1010

1010

1010

1010

1010

1010

1010

1010

109

1980

–9M

ean

-1.6

02.

71-0

.60

0.91

1.29

1.68

1.70

-2.6

03.

415.

13-2

.87

-2.5

32.

635.

9310

.82.

254.

34-2

.15

1.75

0.48

Std.

Dev

.2.

203.

183.

081.

951.

501.

752.

581.

904.

033.

351.

441.

912.

453.

896.

231.

562.

782.

531.

091.

82O

bser

vati

ons

1010

1010

1010

1010

1010

1010

1010

1010

1010

1010

1990

–200

6M

ean

-0.6

5-0

.08

-0.7

5-0

.64

-0.7

8-0

.99

-1.0

1-0

.87

0.05

0.48

-2.3

1-0

.42

-0.6

2-0

.67

1.74

-0.4

30.

90-1

.11

0.33

-0.4

4St

d.D

ev.

0.83

1.23

0.48

0.86

0.89

0.95

0.58

1.22

1.39

1.14

1.00

1.05

0.81

1.00

2.51

2.28

0.87

1.14

1.23

0.54

Obs

erva

tion

s17

1717

1717

1717

1717

1717

1717

1717

1717

1717

17

Not

e:A

U=

Aus

tria

,AU

D=

Aus

tral

ia,B

E=

Bel

gium

,CA

=C

anad

a,D

NK

=D

enm

ark,

FI

=F

inla

nd,F

R=

Fra

nce,

GE

=G

erm

any,

IR=

Irel

and,

IT=

Ital

y,JP

=Ja

pan,

NE

=N

ethe

rlan

ds,N

O=

Nor

way

,N

Z=

New

Zea

land

,PT

=P

ortu

gal,

SE=

Swed

en,S

P=

Spai

n,SW

=Sw

itze

rlan

d,U

K=

Uni

ted

Kin

gdom

,EA

=eu

roar

ea.A

lso,

see

text

and

data

appe

ndix

for

addi

tion

alde

tails

abou

tso

urce

s.D

ata

are

annu

al.



–0.4

–0.20.0

0.2

0.4

0.6

0.8

80

85

90

95

00

05

Austr

ia

–0.8

–0.6

–0.4

–0.20.0

0.2

80

85

90

95

00

05

Austr

alia

–0.6

–0.4

–0.20.0

0.2

0.4

80

85

90

95

00

05

Belg

ium

–0.8

–0.6

–0.4

–0.20.0

0.2

80

85

90

95

00

05

Canada

–0.8

–0.6

–0.4

–0.20.0

0.2

80

85

90

95

00

05

Denm

ark

80

85

90

95

00

05

Fin

land

–0.8

–0.6

–0.4

–0.20.0

0.2

80

85

90

95

00

05

Fra

nce

–3

–2

–101

80

85

90

95

00

05

Germ

any

–2

–1012

80

85

90

95

00

05

Irela

nd

–4

–3

–2

–101

80

85

90

95

00

05

Italy

–0.4

–0.20.0

0.2

0.4

80

85

90

95

00

05

Japan

–10.0

–00.5

00.0

00.5

10.0

80

85

90

95

00

05

Neth

erlands

80

85

90

95

00

05

Norw

ay

–2

–1012

80

85

90

95

00

05

New

Zeala

nd

–505

10

80

85

90

95

00

05

Port

ugal

–5

–4

–3

–2

–101

80

85

90

95

00

05

Sw

eden

80

85

90

95

00

05

Spain

–0.2

5

–0.2

0

–0.1

5

–0.1

0

–0.0

5

0.0

0

0.0

5

0.1

0

80

85

90

95

00

05

Sw

itzerland

–2

–10123

80

85

90

95

00

05

UK

80

85

90

95

00

05

US

A

Actual iess inflation forecast

–1.0

–0.50.0

0.5

1.0

–1.0

–0.8

–0.6

–0.4

–0.20.0

0.2

–1.5

–1.0

–0.50.0

0.5

1.0

–1.0

–0.8

–0.6

–0.4

–0.20.0

0.2

Fig

.4In

flati

onF

orec

ast

Err

ors

Bas

edon

Wor

ldE

cono

mic

Out

look

(WE

O)

Est

imat

es:1

980–

2006

Not

e:A

ctua

linfl

atio

n(a

sde

fined

inF

ig.1

)le

ssth

ein

flati

onfo

reca

st.S

ourc

esof

data

are

give

nin

Fig

.1an

dth

ete

xt.



–0.8

–0.40.0

0.4

0.8

1990

1995

2000

2005

Au

str

ia

–3

–2

–10123

1990

1995

2000

2005

Au

str

alia

–0.8

–0.40.0

0.4

0.8

1990

1995

2000

20

05

Be

lgiu

m

–2

–1012

19

90

19

95

20

00

20

05

Ca

na

da

19

90

19

95

20

00

20

05

De

nm

ark

–0.8

–0.40.0

0.4

0.8

1990

1995

2000

2005

Fra

nce

–3

–2

–1012

1990

1995

2000

2005

Ge

rma

ny

–0.8

–0.40.0

0.4

0.8

1990

1995

2000

20

05

Ita

ly

19

90

19

95

20

00

20

05

Ja

pa

n

19

90

19

95

20

00

20

05

Ne

the

rla

nd

s

1990

1995

2000

2005

No

rwa

y

–3

–2

–10123

1990

1995

2000

2005

Ne

w Z

ea

lan

d

–2

–101234

1990

1995

2000

20

05

Sw

ed

en

–0

.8

–0

.40.0

0.4

0.8

19

90

19

95

20

00

20

05

Sp

ain

–2

–1012

19

90

19

95

20

00

20

05

Sw

itze

rla

nd

–3

–2

–10123

1990

1995

2000

2005

UK

1990

1995

2000

2005

US

A

Actual less inflation forecast

–1

.5

–1

.0

–0

.50.0

0.5

1.0

–0

.8

–0

.40.0

0.4

0.8

1.2

–1.6

–1

.5

–1

.0

–0

.50.0

0.5

1.0

–1.2

–0.8

–0.40.0

0.4

0.8

1.2

–1.0

–0.50.0

0.5

1.0 Fig

.5In

flati

onF

orec

ast

Err

ors:

1990

–200

6B

ased

onC

onse

nsus

/Eco

nom

ist

For

ecas

tsN

ote:

Act

uali

nflat

ion

(as

defin

edin

Fig

.1)

less

the

infla

tion

fore

cast

.Sou

rces

ofda

taar

egi

ven

inF

ig.1

and

inth

ete

xt.



band over an extended period of time is likely to reduce the credibility of anyregime. Notwithstanding the fact that the methodology used to generate theseforecasts can differ by source, we find that whereas forecast errors were largeduring the 1980s the tolerance band is clearly much more visible when datasince the 1990s only are examined. Again, this is suggestive of a shift in thebehaviour and performance of inflation between the two decades.

3.2.3 Measuring the Stance of Monetary Policy. How should one proxythe stance of monetary policy in such a cross-sectional setting for over twodecades worth of data? Two alternatives are considered. While it is standardpractice to assume nowadays that an interest rate instrument is used to depicthow central banks set the stance of monetary policy this was not always so.Indeed, when the conduct of monetary policy over two decades or more ofdata is considered, it is unlikely that an interest rate instrument will berelevant throughout. Some central banks (viz., the European Central Bankand the Bank of Canada; see Armour et al., 1996; Roffia and Zaghini, 2007)have, for some time, shown that the money gap is a useful predictor ofinflation. Essentially, this amounts to finding the difference between a narrowmonetary aggregate (e.g. M1) and the level of money predicted by a long-runcointegrating relationship that defines money demand. This means estimatingan expression of the form

m y i Tt t t t= + + + +α α α α ε0 1 2 3 (5)

where m is the logarithm of the chosen money supply measure, y is thelogarithm of real gross domestic product, i is a short-term interest rate and Tis a time trend which may or may not be necessary as a proxy for the effect offinancial innovations, an unobserved variable (at least until perhaps well afterthe fact). If the estimates reveal that there is some cointegration present in thedata then a linear combination of some of the variables is stationary. Quiteoften, for a variety of countries, received empirical evidence reveals a singlecointegrating vector. In that case, et is stationary. The money gap then issimply the error correction term found by lagging the error term in equation(5) one period. The appendix to the paper (not shown) plots estimates of themoney gap for each country in the data set.

If we now consider the measurement of the stance or stress of monetarypolicy (see equation (4)) based on a Taylor rule approximation we also facea few difficulties. One approach might be to estimate some forward-lookingrule that appears to best fit each country’s interest rate behaviour. Of course,during the period under study it is likely that the exchange rate played agreater or lesser role over time in setting interest rates (e.g. see Taylor, 2007,2008). Moreover, estimation of such a rule requires a fair amount of data andone also has to be mindful of the choice of instruments when estimating theserules as it can have an important influence on the results (e.g. Clarida et al.,1998; Siklos and Bohl, 2008). Given that a version of Taylor’s original



formulation seems to capture reasonably well the rule a central bank wouldhave followed, on average, we follow Poole’s (2006) approach and rely on thefollowing definition of the instrument rule, i.e.

ˆ *, ,i yit i i i t iy i t= + −( ) +−+

−ρ γ π π γπ 1 1100 � � (6)

where pt-1 is last period’s inflation rate, p+ is a target for inflation, �y is anestimate of the output gap, while gp, γ �y , respectively, represent the weights acentral bank attaches to inflation rates that depart from the target and theoutput gap, and r is the sum of the target rate of inflation and an estimate ofthe ‘equilibrium’ or ‘natural’ real interest rate. All coefficients are estimatedfor country i. We follow Poole and assume that p = 2%, the equilibrium realinterest rate is 1.5 per cent, so that r = 3.5%, gp = 1.5 to satisfy the Taylorprinciple, while two alternative values for the output gap weight are consid-ered, namely γ �y = 0 5 0 8. , . .14 Consequently, an indication that the stance ofmonetary policy is tight implies that the difference between the actual policyrate according to the Taylor rule and the predicted interest rate (iit ) is positivewhile loose policies would result in a negative value in the expression

STRESS i iit it it= − ˆ (7)

Finally, it is also plausible that some of the determinants of inflation con-sidered in this study will interact with each other. However, to prevent usfrom estimating an over-parameterized specification, only the interactionbetween the stance of monetary policy and proxies for central bank inde-pendence is considered. Since a great deal of the controversy over theimpact of policies and institutions on inflation has revolved around theinfluence of central bank autonomy, and how the monetary authority actu-ally implements policy, the implied interaction seems to us a natural one toconsider.

3.2.4 An Enduring Mystery?. While equation (2) admits a wide variety ofcovariates to determine inflation differentials vis-à-vis the USA we havedeliberately omitted an additional variable that may also have an importantinfluence on inflation. Increasingly, it is being said, that policy makersignored developments in asset prices and this may explain in part the currenteconomic predicament. One difficulty is that it is not clear how policy makersshould respond to asset prices, nor whether the response ought to be tailoredto some asset prices and not others. It is equally unclear whether one cancome up with a Taylor type principle for asset prices. Whereas the focus oflate has been on housing prices, at other times attention was paid to equityprices. Indeed, it is likely that certain asset prices, such as equity prices, are

14Poole (2006) considers two alternatives for the weight on the output gap because the smallerweight fits the data better, but results in interest rates that are too smooth while the correctamount of volatility in US interest rates is obtained by increasing γ �y .



likely to have a greater economic impact in some countries than in otherswhere housing prices matter more. Central banks, of course, have longargued that implementation of a monetary policy that yields low and stableinflation will take care of the rest and that, in any event, concerns over assetprice developments is a feature of policy communication. For all thesereasons, the results presented below eschew a direct link between asset priceand goods price inflation though, indirectly, the impact of one on the othermay show up, e.g. through interest rate developments. Figure 6 illustrates, fora selection of the countries considered in this study, the bivariate relationshipbetween the Bank for International Settlements’s aggregate index of asset

–3

–2

–1

0

1

2

Infla

tio

n d

iffe

ren

tia

l

–4 –3 –2 –1 0 1 2 3

Asset prices

–2.5

–2.0

–1.5

–1.0

–0.5

0.0

0.5

1.0

1.5

Infla

tio

n d

iffe

ren

tia

l–8 –4 0 4 8 12

Asset prices

–1.2

–0.8

–0.4

0.0

0.4

0.8

Infla

tio

n d

iffe

ren

tia

l

–10 –5 0 5 10 15

Asset prices

–2

–1

0

1

2

3

4

Infla

tio

n d

iffe

ren

tia

l

–10 –5 0 5 10

Asset prices

–4.5

–4.0

–3.5

–3.0

–2.5

–2.0

–1.5

–1.0

–0.5

Infla

tio

n d

iffe

ren

tia

l

–6 –4 –2 0 2 4 6

Asset prices

–2.8

–2.4

–2.0

–1.6

–1.2

–0.8

–0.4

0.0

0.4

0.8

1.2

Infla

tio

n d

iffe

ren

tia

l–12 –8 –4 0 4 8

Asset prices

–3

–2

–1

0

1

2

3

4

5

Infla

tio

n d

iffe

ren

tia

l

–40 –20 0 20 40 60

Asset prices

–2.4

–2.0

–1.6

–1.2

–0.8

–0.4

0.0

0.4

0.8

1.2

1.6

Infla

tio

n d

iffe

ren

tia

l

–8 –4 0 4 8

Asset prices

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

Infla

tio

n

–15 –10 –5 0 5 10 15 20

Asset prices

AUSTRALIA CANADA EURO AREA

JAPAN NEW ZEALAND SWEDEN

SWITZERLAND UK USA

Fig. 6 Asset Price Inflation and the Inflation Differential: 1990–2008Note: The vertical axis plots the inflation differential vis-à-vis the USA (except for the USA

which shows the inflation rate). The horizontal axis plots the rate of change (100 times the logfourth difference) in the Bank for International Settlements’s aggregate price index. The leastsquares regression line is also plotted while the distribution of the time series is also shown on

each axis.



prices and the dependent variable in (2) for the 1990–2008 period.15 There areno obvious stylized facts that emerge from this Figure. Indeed, if one were toreproduce the plot for a sub-sample before inflation control policies becamewidely used (i.e. 1970–89) an entirely different picture would emerge in mostof the cases considered (not shown). The mystery deepens still when correla-tions between equity, housing and aggregate asset price measures and eitherthe inflation differential against the USA or each country’s output gap isexamined (results not shown). For example, in four of the nine countriesconsidered there is significant correlation between equity and housing prices,but it is positive in some countries and negative in others. Moreover, when weaggregate asset prices, or the output gap is examined, the correlation virtuallydisappears (not shown). Clearly, including a role for asset prices is best left aspart of the agenda for future research.

4 Empirical Results

Table 2 provides a series of panel unit root test results. These consistentlyshow that the null of a unit root in the inflation differential is rejectedregardless of the sample or the test employed. Hence, it is unlikely that theresults to be discussed below are an artefact of the non-stationarity of thedependent variable in question. The remaining variables are stationary byconstruction or were found to be so using similar testing procedures (notshown). While the tests employed are fairly general there have been a numberof recent developments in panel unit root testing that could conceivablyoverturn the results shown in Table 2. However, tests for a subset of the panelused in Table 2 (not shown but also see Siklos, 2008b) suggest that thefindings of the stationarity in the inflation differential in the chosen panel of

15I am grateful to Claudio Borio of the Bank for International Settlements for making availablethe asset price data.

Table 2Panel Unit Root Test: Differential VIS-À-VIS US Inflation

Panels 1980–2006 1980–98 1990–2006 1999–2006

All countriesLLC -4.29 (0.00) -5.37 (0.00) -5.64 (0.00) NAIPS -6.10 (0.00) -5.35 (0.00) -5.70 (0.00) NA

EA & Non-EA countriesLLC NA NA -4.42 (0.00) -7.01 (0.00)IPS NA NA -4.25 (0.00) -2.62 (0.00)

Note: All countries are the panel of countries listed in Table 1. EA & Non-EA countries are the euro area andthe remaining countries in the sample that are not euro area members. Data are annual. NA means notapplicable. LLC is the Levin, Lin and Chu panel unit root test, IPS is the Im, Pesaran and Shin panel unit roottest. The LLC assumes a common unit root, while the IPS test is based on individual unit root behaviour. Inboth cases then null is the unit root. The panels are not balanced.



countries is robust to other assumptions about the degree of heterogeneityand other cross-country variations that could affect the panel unit root teststatistics.

Next, we turn to the benchmark Ordinary Least Squares fixed effectspanel. The results are provided in Table 3. Both country-specific and time-specific fixed effects cannot be rejected.16 It is immediately clear that estimatesare highly sensitive to sample choice. In particular, estimation over the full1980–2006 period suggests that only economic distance has a somewhat weakimpact on the inflation differential vis-à-vis the USA with a rise in the formerleading to an increase in the latter, as previously hypothesized. Nevertheless,on closer inspection, one sees that the full sample masks a notable change ininflation dynamics between the periods 1980–98 and 1999–2006.17 The earlierperiod reveals that a combination of the stance of monetary policy and theimpact of accumulated forecast errors significantly influenced cross-countryinflation differentials against the USA. Presumably, worse than expecteddomestic inflation forecast performance over time, a proxy for past centralbank performance, led to an improvement in actual relative inflation perfor-mance. Of course, we are unable to identify the extent to which this outcomecan be explained by the reaction of the monetary authorities to this indicatorof their performance but it does appear that the statistical insignificance ofthe other control variables, notably the stance of monetary policy and centralbank autonomy, cannot explain inflation differentials in the 1980–98 period.However, matters change when we examine the 1990–2006 sample. Both thestance of monetary policy and the interaction of the stance of monetarypolicy and central bank autonomy contribute to explaining inflation relativeto the USA. More precisely, when nominal interest rates are higher thannecessary, according to the hypothesized Taylor rule (columns (4)), the pre-dictable relative tightening of policy reduces the inflation differential. In thecase where the stance of policy is proxied by a money gap, a rise increases thedifferential. This is to be expected because a positive money gap indicates aloosening of monetary policy and, hence, is expected to be relatively infla-tionary. Interestingly, the signs on the interaction coefficient can be inter-preted as suggesting that the combination of a tighter (looser) monetarypolicy stance and greater central bank autonomy permits a slightly higher(lower) inflation differential. This could be considered as a little bit of evi-dence that there is a credibility bonus from greater central bank independenceor transparency.

16Only the joint test of country and time-specific fixed effects is shown. However, separate testingdoes not change any of the conclusions. Both types of fixed effects, where relevant, werefound to be necessary.

17Since it is not immediately clear when to date the effective start of European Monetary Unionwe also consider a separate sub-sample beginning in 1995 (see Mayes and Virén, 2005). Ourconclusions are largely unchanged. An appendix provides the results for this sub-sample(not shown).



Ta

ble

3P

an

elO

rdin

ary

Lea

stSq

uare

Est

ima

tes

Var

iabl

es

1980

–200

619

80–1

998

1990

–200

619

99–2

006

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Con

stan

t1.

12(0

.93)

1.31

(0.5

7)**

1.62

(0.5

1)*

0.51

(0.7

9)0.

17(0

.62)

-0.9

2(1

.26)

-1.3

4(0

.79)

***

-1.0

9(1

.26)

For

ecas

ter

rors

-0.1

7(0

.16)

-0.2

1(0

.07)

*-0

.31

(0.0

7)*

-0.1

2(0

.09)

-0.1

7(0

.08)

**0.

12(0

.08)

0.13

(0.1

2)0.

12(0

.09)

Mon

etar

yP

olic

ySt

ance

-0.0

2(0

.14)

0.02

(0.0

3)-0

.02

(0.0

8)-0

.31

(0.1

7)**

*0.

09(0

.04)

*0.

07(0

.19)

0.02

(0.0

3)0.

04(0

.12)

Cen

tral

bank

inde

pend

ence

-0.0

9(1

.93)

-1.0

9(1

.16)

0.58

(0.9

8)0.

01(0

.08)

0.90

(1.3

4)-0

.002

(0.1

0)0.

49(1

.08)

0.21

(1.3

6)

Inte

ract

ion

-0.4

0(0

.28)

-0.0

7(0

.08)

-0.3

8(0

.17)

**0.

02(0

.01)

***

-0.1

7(0

.07)

*-0

.01

(0.0

1)-0

.05

(0.0

6)-0

.09

(0.1

5)D

ista

nce

0.17

(0.1

0)**

*0.

22(0

.09)

*0.

19(0

.89)

**-0

.01

(0.0

8)0.

02(0

.06)

-0.3

4(0

.08)

*-0

.31

(0.1

1)-0

.34

(0.1

0)*

Fix

edef

fect

s9.

62(0

.00)

10.7

2(0

.00)

8.78

(0.0

0)4.

42(0

.00)

4.16

(0.0

0)13

.69

(0.0

0)14

.23

(0.0

0)13

.10

(0.0

0)O

bser

vati

ons

431

360

360

242

235

7876

78R

20.

680.

680.

680.

570.

560.

820.

820.

82

Not

e:T

heF

ixed

effe

cts

test

isfo

rth

eN

ullt

hatb

oth

cros

s-se

ctio

nan

dti

me-

spec

ific

fixed

effe

cts

are

join

tly

insi

gnifi

cant

(sep

arat

ete

stin

gof

cros

s-se

ctio

nve

rsus

tim

e-sp

ecifi

cfix

edef

fect

sdi

dno

tal

ter

any

ofth

eco

nclu

sion

ssh

own)

.The

F-s

tati

stic

isgi

ven

wit

hp-

valu

esin

pare

nthe

sis.

The

expl

anat

ory

vari

able

sar

ede

fined

inth

ete

xt.T

he‘I

nter

acti

on’t

erm

refe

rsto

the

prod

uct

ofth

e‘M

PSt

ance

’va

riab

lean

dth

epr

oxy

for

cent

ral

bank

inde

pend

ence

.P

roxi

esus

edfo

rce

ntra

lba

nkin

depe

nden

cein

the

resu

lts

repo

rted

abov

ear

e(c

olum

nsnu

mbe

r):(

1),(

2),(

3),(

8)C

ukie

rman

-Sik

los;

(4),

(6)

Din

cer-

Eic

heng

reen

;(5)

,(7)

Gri

lli-M

asci

anda

ro-T

abel

lini;

(6).

For

ecas

ter

rors

for

the

1980

–98

and

1980

–200

6ar

eba

sed

onW

orld

Eco

nom

icO

utlo

okfo

reca

sts

for

infla

tion

;for

the

1990

–200

6,19

99–2

006

base

don

Eco

nom

ist

fore

cast

s.M

PSt

ance

ispr

oxie

dac

cord

ing

toeq

uati

on(7

)in

all

colu

mns

exce

ptco

lum

n(5

)w

here

the

mon

eyga

pis

used

.The

pane

lis

unba

lanc

ed.

*Si

gnifi

esst

atis

tica

llysi

gnifi

cant

atth

e1

per

cent

,**

5pe

rce

nt,*

**10

per

cent

leve

l.H

eter

oske

dast

icro

bust

stan

dard

erro

rsar

ein

pare

nthe

sis.



Finally, when we examine the most recent sample (1999–2006) onlyeconomic distance has a statistically significant impact on the inflation dif-ferential. It is conceivable that the bonus provided by institutional factors aswell as attainment of low and stable inflation in the panel of countriesconsidered removed a significant role for the other factors considered.

As previously discussed, it is likely that the relationship between theinflation differential and the various determinants considered is an endog-enous one. Consequently, Table 4 presents estimates of a panel based onGeneralized Method of Moments estimation. These include, in addition tothe usual selection of lags as variables as instruments, a variety of gover-nance indicators created by the World Bank as well as exchange rate andinflation targeting dummy indicators (see notes to Table 4). Not surpris-ingly, the change of estimation strategy influences the results. There arethree notable differences between the results in Tables 3 and 4. First, eco-nomic distance vis-à-vis the USA no longer reliably explains the inflationdifferential, regardless of the sample period considered. Second, the cumu-lative impact of inflation forecast errors have a much larger impact in the1980–98 sample than when Ordinary Least Square is used. Third, variousproxies for central bank autonomy and transparency are more robustlyrelated to the inflation differential. Perhaps most interesting of all is thatwhile the central bank institutional indicator has a large negative impact onthe US inflation differential in the early sample (1980–98) the sign for thisvariable turns positive in the post 1990 period. The interaction with thestance of monetary policy offsets this reaction somewhat but does noteliminate it entirely. While the results are somewhat sensitive to the choiceof different central bank indicators another interpretation is simply that,after an initial beneficial impact on inflation, the link between the inflationdifferential and central bank autonomy is a transitory one. Finally, it isworth briefly commenting on the country-specific intercept terms. It is inter-esting that while these are almost always statistically significant for the fullsample, suggesting a positive inflation differential exists against US infla-tion, these same intercepts become largely insignificant, or becoming nega-tive and significant, especially in the post 1999 period. This outcome simplycaptures, statistically, the stylized fact reported earlier, namely the dramaticconvergence in relative inflation performance across the OECD countriesexamined since the late 1990s.

5 Conclusions

This paper has considered whether various determinants of the quality ofmonetary policy in 20 OECD countries since the 1980s have significantlyaffected inflation vis-à-vis the USA. The USA is chosen as the benchmarkbecause its monetary policy over period considered (1980–2006) is often



Ta

ble

4P

an

elG

ener

aliz

edM

etho

dsof

Mom

ents

Est

ima

tes

Var

iabl

es

1980

–200

619

80–9

819

90–2

006

1999

–200

6

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

For

ecas

ter

rors

-0.3

2(0

.10)

***

-0.7

0(0

.23)

*-0

.73

(0.2

4)*

-0.1

7(0

.10)

***

-0.1

0(0

.08)

0.04

(0.0

7)-0

.05

(0.1

2)-0

.12

(0.1

3)M

onet

ary

Pol

icy

Stan

ce-0

.08

(0.3

6)-0

.32

(0.1

2)*

-0.1

7(0

.31)

-0.1

5(0

.21)

0.24

(0.0

4)*

0.20

(1.6

0)0.

08(0

.04)

**0.

08(0

.16)

Cen

tral

bank

inde

pend

ence

5.08

(2.2

7)**

-6.7

4(3

.38)

**-1

.24

(3.0

0)-0

.08

(0.1

3)3.

08(1

.08)

*-0

.06

(0.2

3)1.

99(0

.76)

*-4

.52

(5.1

1)

Inte

ract

ion

-0.3

4(0

.30)

1.11

(0.3

8)*

0.38

(0.7

5)0.

02(0

2)-0

.45

(0.1

1)*

0.00

1(0

.10)

-0.1

3(0

.12)

**-0

.46

(0.2

2)**

Dis

tanc

e-0

.07

(0.1

6)0.

22(0

.19)

0.10

(0.1

7)-0

.11

(0.0

6)**

*-0

.05

(0.0

4)-0

.04

(0.1

)-0

.13

(0.0

6)-0

.23

(0.1

4)**

*O

bser

vati

ons

410

333

339

242

235

106

7678

Inte

rcep

tsC

oeff

.(s

.e.)

Coe

ff.

s.e.

Coe

ff.

s.e.

Coe

ff.

s.e.

Coe

ff.

s.e.

Coe

ff.

s.e.

Coe

ff.

s.e.

Coe

ff.

s.e.

Aus

tral

ia4.

19(1

.21)

*4.

371.

63*

3.76

1.48

*1.

451.

43-1

.08

.78

-0.1

23.

460.

252.

174.

484.

00A

ustr

ia3.

48(1

.56)

**2.

071.

25*

1.22

1.86

0.82

0.87

-1.6

10.

78**

NA

NA

NA

Bel

gium

1.02

(0.8

8)2.

601.

37**

0.72

0.92

0.25

0.70

-2.4

00.

53*

NA

NA

NA

Can

ada

3.47

(1.2

7)*

2.27

1.46

2.14

1.61

0.94

1.65

-1.7

70.

73*

-0.3

03.

20-0

.98

1.13

3.04

3.21

Den

mar

k3.

08(1

.16)

*3.

071.

26**

2.19

1.29

***

0.44

0.82

-1.8

90.

57*

-0.0

71.

43-1

.43

0.41

*2.

442.

56F

inla

nd3.

97(1

.30)

*2.

971.

07*

2.92

1.31

**0.

841.

610.

290.

78N

AN

AN

AF

ranc

e3.

22(0

.97)

*3.

721.

31*

2.71

1.08

*0.

280.

96-2

.16

0.57

*N

AN

AN

AG

erm

any

4.21

(2.1

9)**

*3.

961.

92**

2.84

2.52

1.78

1.78

-1.4

21.

22N

AN

AN

AIr

elan

d6.

21(1

.19)

*5.

642.

22*

6.37

1.48

*1.

471.

46-1

.90

0.91

**N

AN

AN

AIt

aly

7.54

(1.2

7)*

10.2

52.

25*

9.83

2.04

*3.

741.

86**

1.36

0.84

NA

NA

NA

Japa

n0.

11(1

.20)

-0.3

61.

04-1

.49

1.47

-1.0

21.

43-3

.70

0.60

*-2

.25

2.31

-3.4

30.

97*

0.33

3.15

Net

herl

ands

1.65

(1.5

1)0.

771.

23-0

.10

1.74

1.38

1.15

-1.7

50.

65*

NA

NA

NA

New

Zea

land

7.75

(2.0

6)*

9.58

2.56

*8.

272.

31*

2.19

2.26

-1.0

21.

02-0

.26

4.01

-0.6

81.

855.

185.

22N

orw

ay3.

47(1

.00)

*4.

831.

75*

3.68

1.30

*0.

971.

27-1

.61

0.62

*-1

.50

2.84

-0.4

11.

082.

271.

83P

ortu

gal

15.4

9(2

.88)

*21

.99

6.58

*24

.10

5.97

*8.

313.

87**

3.70

2.88

NA

NA

NA

Spai

n6.

87(1

.45)

*8.

421.

78*

6.92

1.59

*3.

471.

78**

0.53

0.77

NA

NA

NA

Swed

en6.

36(1

.45)

*5.

862.

65**

7.08

1.61

*2.

552.

84-1

.21

1.38

-1.3

14.

54-1

.62

2.16

4.58

5.67

Swit

zerl

and

2.06

(1.4

7)1.

331.

880.

291.

480.

111.

50-2

.61

.89*

-1.6

22.

64-2

.45

1.38

***

2.98

4.39

UK

7.65

(1.4

1)*

9.65

2.64

*8.

652.

07*

4.12

2.95

0.36

1.66

0.04

3.69

-0.7

01.

643.

813.

84E

uro

area

NA

NA

NA

NA

NA

0.13

2.59

-1.6

60.

53*

3.60

4.03

Not

e:Se

eno

tes

toT

able

3.G

MM

(Are

llano

-Bon

dtw

o-st

epes

tim

ator

)use

sla

gsof

the

regr

esso

rsas

wel

las

the

voic

ean

dac

coun

tabi

lity,

polit

ical

stab

ility

,and

qual

ity

ofgo

vern

men

tind

icat

ors

from

the

Wor

ldB

ank

gove

rnan

cein

dica

tors

data

set

(Kau

fman

net

al.,

1999

,20

07;

Wor

ldB

ank,

2006

),th

eex

chan

gera

tein

dica

tors

from

Lev

y-Y

eyat

ian

dSt

urze

negg

er,

and

adu

mm

yva

riab

lefo

rth

enu

mbe

rof

year

sa

coun

try

has

adop

ted

anex

plic

itin

flati

onta

rget

,w

here

appl

icab

le.

Afix

edba

ndw

ith

(New

ey-W

est)

,B

artl

ett

kern

el,

hete

rosk

edas

tici

tyad

just

edst

anda

rder

rors

.N

Am

eans

not

appl

icab

le.

The

pane

lis

unba

lanc

ed.I

nco

lum

ns(6

)an

d(7

)an

othe

rin

tera

ctio

nte

rm(s

tanc

eti

mes

cent

ralb

ank

inde

pend

ence

)is

also

adde

dbu

tal

lthe

coef

ficie

nts

wer

est

atis

tica

llyin

sign

ifica

nt,e

ven

atth

e10

per

cent

leve

land

are

omit

ted.



viewed as having set the standard for other central banks.18 In spite of theproblems that arise when panels are estimated with fixed effects (e.g. see Roseet al., 2005) this study finds that institutional factors play a role in influencinginflation performance relative to the US experience. Nevertheless, the impactis a more nuanced one than previously imagined with the bonus generatedby the introduction of institutional reforms such as greater central bankautonomy or transparency a feature of the 1980s and early 1990s. By the late1990s, it becomes much more difficult to discriminate between the inflationperformance of the USA and its OECD counterparts.

Owing to the selection of countries, samples, estimation techniquesemployed, not to mention the various institutional and non-institutionalproxies used to measure the quality of monetary policy over time, ours is notthe last word on the subject. For example, considering the non-linearity in thedynamics of inflation differentials over time, a characteristic of interest ratedifferentials over the same period, may also prove to be a useful avenue tofollow in future research. Finally, notwithstanding the considerable limita-tions on data availability, expanding the panel of countries to include emerg-ing market economies, among others, may also yield important insights intothe evolution of inflation and its connection with the implementation ofmonetary policy. Needless to say, these all represent fruitful avenues forfuture research.

References

Acemoglu, D., Johnson, S., Querubin, P. and Robinson, J. (2008). ‘When Does PolicyReform Work? The Case of Central Bank Independence’, Brookings Papers onEconomic Activity, Vol. 2, Spring, pp. 351–417.

Alesina, A. and Grilli, V. (1992). ‘The European Central Bank: Reshaping MonetaryPolitics in Europe’, in M. B. Canzoneri, V. Grilli and P. R. Masson (eds),Establishing a Central Bank: Issues in Europe and Lessons from the U.S., Cam-bridge, Cambridge University Press.

Armour, J., Attah-Mensah, J., Engert, W. and Hendry, S. (1996). ‘A Distant-Early-Warning Model of Inflation Based on M1 Disequilibria’, Bank of CanadaWorking Paper 96-5.

Arnone, M., Laurens, B. J., Segolato, J.-F. and Sommer, M. (2007). ‘Central BankAutonomy and Inflation: Lessons from Global Trends’, IMF Working PaperWP/07/88, April.

Ball, L. (2006). ‘Has Globalization Changed Inflation?’, NBER Working Paper 12687,November.

Ball, L. and Sheridan, N. (2005). ‘Does Inflation Targeting Matter? ’, in B. S. Ber-nanke and M. Woodford (eds), The Inflation-Targeting Debate, Chicago andLondon, University of Chicago press, pp. 249–276.

18US policy can be regarded as a standard for correcting the problem of high inflation and thenkeeping it under control. German (or Swiss) policy on the other hand could be regarded asa standard for inflation control over the period as a whole. The creation of the euro areamakes Germany a more difficult standard to follow over our data period. However, aspointed out previously, subject to sample limitations, the results generally hold when theEuropean Central Bank is used as the benchmark.



Balliu, J. and Fuji, E. (2004). ‘Exchange Rate Pass-Through and the Inflation Envi-ronment in Industrialized Countries: an Empirical Investigation’, Bank ofCanada Working Paper 2004-21.

Banaian, K., Burdekin, R. and Willett, T. (1998). ‘Reconsidering the Principal Com-ponents of Central Bank Independence: the More the Merrier? ’, Public Choice,Vol. 97, October, pp. 1–12.

Bean, C. (2006). ‘Globalization and Inflation’, Bank of England Quarterly Bulletin,Vol. 46, No. 4, pp. 468–75.

Bernanke, B. (2004). ‘The Great Moderation’, Remarks by Governor Ben S. Ber-nanke At the meetings of the Eastern Economic Association, Washington, DCFebruary 20, 2004. http://www.federalreserve.gov

Bernanke, B. and Mishkin, F. (1992). ‘Central Bank Behavior and the Strategy ofMonetary Policy: Observations from Six Industrial Countries’, in O. Blanchardand S. Fischer (eds), NBER Macroeconomics Annual, Cambridge, Mass., Cam-bridge University Press, pp. 183–228.

Bernanke, B., Laubach, T., Mishkin, F. and Posen, A. (1999). Inflation Targeting,Princeton, NJ, Princeton University Press.

Blanchard, O. and Simon, J. (2001). ‘The Long and Large Decline in U.S.Output Volatility’, Brookings Papers on Economic Activity, Vol. 1, pp. 135–164.

Borio, C. and Filardo, A. (2007). ‘Globalization and Inflation: New Cross-Country Evidence on the Global Determinants of Domestic Inflation’, BIS 227,May.

Busetti, F., Forni, L., Harvey, A. and Venditti, F. (2007). ‘Inflation Convergence andDivergence within the European Monetary Union’, International Journal ofCentral Banking, Vol. 3, June, pp. 95–121.

Campa, L. G. and Gonzalez-Minguez, J. (2007). ‘Exchange-Rate Pass-Through toImport Prices in the Euro Area’, in F. di Mauro and R. Anderton (eds), TheExternal Dimension of the Euro Area: Assessing the Linkages, Cambridge, Cam-bridge University Press, pp. 63–94.

Cheng, I.-H. and Wall, H. J. (2005). ‘Controlling for Heterogeneity in Gravity Modelsof Trade and Integration’, Review of the Federal Reserve of St. Louis, Vol. 87,January–February, pp. 49–63.

Clarida, R., Gali, J. and Gertler, M. (1998). ‘Monetary Policy Rules in Practice: SomeInternational Evidence’, European Economic Review, Vol. 42, June, pp. 1033–1067.

Côté, D. and De Resende, C. (2008). ‘Globalization and Inflation: the Role of China’,Bank of Canada Working Paper 2008-35.

van der Cruijsen, C. and Eijffinger, S. (2008). ‘The Economic Impact of CentralBankTransparency’, in P. L. Siklos, M. T. Bohl and M. E. Wohar (eds), Chal-lenges in Central Banking, Cambridge, Cambridge University Press, forthcoming.

Cukierman, A. (1992). Central Bank Strategy, Credibility, and Independence, Cam-bridge, Mass., Cambridge University Press.

Dincer, N. and Eichengreen, B. (2007). ‘Central Bank Transparency: Where, Why,and with What Effects? ’, NBER Working Paper 13003.

Dueker, M. and Fischer, A. (2006). ‘Do Inflation Targeters Outperform Non-Targeters? ’, Review Federal Reserve Bank of St. Louis, Vol. 88, September–October, pp. 431–450.

Edwards, S. (2007). ‘Capital Controls, Capital Flow Contractions, and Macroeco-nomic Vulnerability’, Journal of International Money and Finance, Vol. 26, pp.814–840.

Eichengreen, B. and Leblang, D. (2006). ‘Globalization and Democracy’, BISWorking Paper 219, December.



Eijffinger, S. and De Haan, J. (1996). ‘The Political Economy of Central Bank Inde-pendence’, Special Papers in International Economics 19. Princeton: PrincetonUniversity, Department of Economics, International Finance Section.

Eijffinger, S. C. W. and Geraats, P. (2006). ‘How Transparent Are Central Banks? ’,European Journal of Political Economy, Vol. 22, March, pp. 1–21.

Freytag, A. and Schneider, F. (2007). ‘Monetary Commitment, Institutional Con-straints and Inflation: Empirical Evidence for OECD Countries Since the 1970s’,CESifo Working Paper 2001.

Gamber, E. and Hung, J. (2001). ‘Has the Rise in Globalization Reduced U.S.Inflation in the 1990s? ’, Economic Inquiry, Vol. 39, No. January, pp. 647–66.

Grilli, V., Masciandaro, D. and Tabellini, G. (1991). ‘Political and Monetary Insti-tutions and Public Financial Policies in the Industrial Countries’, EconomicPolicy, Vol. 13, October, pp. 341–392.

Guerrieri, L., Gust, C. and Lopez-Salido, D. (2008). ‘International Competition andInflation: a New Keynesian Perspective’, Board of Governors of the FederalReserve System (U.S.), International Finance Discussion Papers 918.

Hayo, B. and Hefeker, C. (2008). ‘The Complex Relationship between Central BankIndependence and Inflation’, in P. L. Siklos, M. T. Bohl and M. E. Wohar (eds),Challenges in Central Banking, Cambridge, Cambridge University Press, forth-coming.

Ihrig, J., Kamin, S., Lindner, D. and Marquez, J. (2007). ‘Some Simple Tests of theGlobalization Hypothesis’, International Finance Discussion Paper 891, April.

International Monetary Fund (IMF) (2006). World Economic Outlook [WEO],‘Chapter III: How Has Globalization Affected Inflation?’ April.

Kaufmann, D., Kraay, A. and Zoido-Lobatón, P. (1999). ‘Governance Matters’,Policy Research Paper 2152, October.

Kaufmann, D., Kraay, A. and Matruzzi, M. (2007). ‘Governance Matters VI: Gov-ernance Indicators for 1996–2006’, World Bank Policy Research Paper 4280, July.

Laeven, L. and Valencia, F. (2008). ‘Systemic Banking Crises: a New Database’, IMFWorking Paper 08/224, September.

Levy-Yeyati, E. and Sturzenegger, F. (2002). ‘Exchange Rate Regimes and EconomicPerformance’, IMF Staff Papers, Vol. 47, Special Issue, pp. 62–98.

Lin, S. and Ye, H. (2007). ‘Does Inflation Targeting Really Make a Difference?Evaluating the Treatment Effect of Inflation Targeting in Seven Industrial Coun-tries’, Journal Of Monetary Economics, Vol. 54, November, pp. 2521–2533.

Mangano, G. (1998). ‘Measuring Central Bank Independence: a Tale of Subjectivityand of Its Consequences’, Oxford Economic Papers, Vol. 50, July 1998, pp.468–492.

Mayes, D. and Virén, M. (2005). ‘Monetary Policy Problems for Currency Unions:Asymmetry and the Problem of Aggregation in the Euro Area’, Economic Mod-elling, Vol. 22, pp. 219–251.

Mishkin, F. (2009). ‘In Praise of an Explicit Number for Inflation’, Financial Times, 11January.

Mishkin, F. and Schmidt-Hebbel, K. (2007). ‘Does Inflation Targeting Make a Dif-ference? ’, NBER Working Paper 12876.

Pain, N., Kospke, I. and Sollie, M. (2006). ‘Globalization and Inflation in the OECDEconomies’, OECD Economics Department Working Paper 524, November.

Polillo, S. and Guillén, M. F. (2005). ‘Globalization Pressures and the State: theGlobal Spread of Central Bank Independence’, American Journal of Sociology,Vol. 110, May, pp. 1764–1802.

Poole, W. (2006). ‘The Fed’s Monetary Policy Rule’, Review of the Federal ReserveBank of St. Louis, Vol. 88, January–February, pp. 1–11.



Ragan, C. (2005). ‘The Exchange Rate and Canadian Inflation Targeting’, Bank ofCanada Working Paper 05-34.

Reinhart, C. and Rogoff, K. (2004). ‘The Modern History of Exchange Rate Arrange-ments: a Reinterpretation’, Quarterly Journal of Economics, Vol. 119, February,pp. 1–48.

Roffia, B. and Zaghini, A. (2007). ‘Excess Money Growth and Inflation Dynamics’,ECB Working Paper 749, May.

Rogoff, K. (2003). ‘Globalization and the Global Disinflation,’, paper presented forthe Federal Reserve Bank of Kansas, 2003 Jackson Hole Conference.

Romer, D. (1993). ‘Openness and Inflation: Theory and Evidence’, Quarterly Journalof Economics, Vol. 108, No. 4, pp. 869–903.

Rose, A. K. (2007). ‘A Stable International Monetary System Emerges: InflationTargeting Is Bretton Woods, Reversed’, Journal of International Money andFinance, Vol. 26, September, pp. 663–681.

Rose, A. K., Fatas, A. and Mihov, I. (2005). ‘Quantitative Goals for Monetary Policy’,Journal of Money, Credit and Banking, Vol. 39, August, pp. 1163–1176.

Siklos, P. L. (2002). The Changing Face of Central Banking, Cambridge, CambridgeUniversity Press.

Siklos, P. L. (2008a). ‘No Single Definition of Central Bank Independence Is Right forAll Countries’, European Journal of Political Economy, Vol. 24, December, pp.802–816.

Siklos, P. L. (2008b). ‘Inflation Targeting Around the World’, Emerging MarketsTrade and Finance, Vol. 44, November–December, pp. 17–37.

Siklos, P. L. (2009). ‘As Good As It Gets? The International Dimension to Canada’sMonetary Policy Strategy Choices’, C.D. Howe Commetary (July), http://www.cdhowe.org.

Siklos, P. L. and Bohl, M. (2008). ‘Policy Words and Policy Deeds: the ECB and theEuro’, International Journal of Finance and Economics, Vol. 13, July, pp. 247–265.

Siklos, P. L. and Wohar, M. (1997). ‘Convergence in Interest Rates and InflationRates across Countries and over Time’, Review of International Economics, Vol.5, February, pp. 129–141.

Stock, J. and Watson, M. (2003). ‘Forecasting Output and Inflation: the Role of AssetPrices’, Journal of Economic Literature, Vol. 41, September, pp. 788–829.

Taylor, J. (2007). ‘The Explanatory Power of Monetary Policy Rules’, Business Eco-nomics, Vol. 42, October, pp. 8–15.

Taylor, J. (2008). ‘Globalization and Monetary Policy: Missions Impossible’, in J.Gali and M. J. Gertler (eds), International Dimensions of Monetary Policy,Chicago, University of Chicago Press, forthcoming.

Tootell, G. B. (1998). ‘Globalization and U.S. Inflation’, Federal Reserve Bank ofBoston, New England Economic Review, July/August.

Tytell, I. and Wei, S. (2004). ‘Does Financial Globalization Induce Better Macroeco-nomic Policies’, IMF Working Paper 04/84, (Washington: IMF).

Vega, C. and Winkelried, D. (2005). ‘Inflation Targeting and Inflation Behavior: aSuccessful Story? ’, International Journal of Central Banking, Vol. 1, December,pp. 153–175.

Woodford, M. (2003). Interest and Prices, Princeton, N.J, Princeton University Press.Woodford, M. (2007). ‘Globalization and Monetary Control’, Working Paper,

Columbia University, July.World Bank (2006). ‘Worldwide Governance Indicators: 1996–2006’, http://

go.worldbank.org/ATJXPHZMH0.



The Quality of Monetary Policy and Inflation Performance: Globalization and its Aftermath

Documents

Transcript of The Quality of Monetary Policy and Inflation Performance: Globalization and its Aftermath