The Resiliency of Macroeconomic Policy Regimes

Transcription

1 The Resiliency of Macroeconomic Policy Regimes A Comparative Analysis of Central Bank Reforms 1 Draft Paper prepared for the 30 th ECPR Joint Sessions of Workshops, Turin, March 2002, Workshop 3: How Economic Integration Matters: Europeanisation, Discourse and Policy Regimes Michelle Beyeler Institute of Political Science University of Bern Lerchenweg 36 CH 3000 Bern 9 michelle.beyeler@ipw.unibe.ch 1 The paper is written in the context of the research project Critical Junctures. A Comparative Analysis of Institutional Reforms generously supported by the Swiss National Science Foundation.

2 2 I Introduction To what extend has European economic integration forced domestic macroeconomic policy regimes to change? This is the central question addressed in this paper. On one hand it is often argued, that European integration especially the process of European Monetary Union has forced formerly interventionist regimes to change in a paradigmatic way towards monetarism. A sign for the institutionalization of the new paradigm is seen in the wave of central bank reforms in industrialized democracies making them more independent from government interference and establishing price stability as their primary goal (Surel 2000). On the other hand, the Europeanization literature points to path-dependent development and the importance of mediating factors, which lead to differential responses by the member states to Europeanization pressures (Börzel and Risse 2000; Cowles, Caporaso and Risse 2001; Héritier and Knill 2000; Radaelli 2000; Schmidt 2001). When it comes to macroeconomic policy regimes scholars defend the limited and contingent impact view by referring to fiscal policies, where the main responsibility still remains at the national level. But, what about monetary institutions? Have the pressures stemming from European economic integration in this domain been so powerful, that national trajectories and institutional particularities unlike in other policy fields have been simply overridden? I argue that the mechanisms the Europeanization literature has detected for other policy regimes also hold for the case of monetary institutions: Domestic path and constellations are important for explaining the reform-pattern. Certainly, Monetary Union has provided a powerful incentive for legal adjustment of the central-bankgovernment relations in those European countries, which used to have legally dependent central banks. However, a closer look at monetary policy outcomes shows that the legal reforms have been mostly codifications of changes in behavioral norms and practices. The real forces behind these changes are learning processes resulting from the failure of old policy concepts and the enhanced necessity for a credible and transparent economic policy in integrated economies. These processes were supported by an increasing centrality of the independence question in the national and international discourse on monetary institutions. The idea that central banks have to be independent has become the dominant view among a large coalition of politicians, economists, investors, and central bankers. This argument is based on a comparative study of central bank reforms. Prevailing comparative studies on this topic mostly focus only on a few cases (Surel 2000, Elgie, 1998 #1987). Additionally, the focus has been mostly on European central banks. This does not allow separating the pressures stemming from Europeanization from more

3 3 international pressures, which impact on all systems. I adopt a research strategy, which is although still case-oriented based on a wider set of cases including non- European states. The procedure draws on the comparative method described by Ragin (1987, 2000). The paper proceeds in the following manner. After a review of the positions on the effects of European economic integration on domestic macroeconomic institutions and policy regimes, I present a configurational model of institutional change. This model is applied in the empirical section for analyzing the pattern of central bank reforms and monetary polices in 22 advanced industrialized countries. II Conceptualizing the Impact of European Economic Integration There is no doubt, that the process of economic integration in the European Union has an impact on its member states macroeconomic policymaking. With the realization of the single market and the introduction of a single currency more and more economic policy decisions, such as exchange rate policies or monetary policies are now taken at the European level. This certainly has an effect on the administration and decision making procedures. Less clear is, to what extent this process has also transformed traditional paradigms and rules of action. Or, if we put the question in a new institutionalist vocabulary: Has Europeanization caused pathsifting institutional reforms? Surel (2000) concludes based on case studies of France, Great Britain and Italy, that EMU has forced member states to abandon their established principles and frames of monetary policy by making their central banks independent according to the requirements of the Maastricht treaty. The former institutional framework in the three countries has given the government the legal power to intervene in their economies monetary policies. This power, after the reform has been delegated to the central banks, coupled with a binding mandate for price stability policies. Surels analyses show that the degree of central bank independence after the reform has been considerably higher in the two EMUcountries France and Italy, than in Great Britain. He argues, that EMU has left national institutions with no possibilities to resist (Surel 2000: 158). This argument is compatible with the claim, that in case the European Union prescribes an institutional model, to which the member states must comply, they have only little discretion (Knill and Lehmkuhl 1999). In such a case the most important determinant of whether there will be an impact or not is goodness of fit between the prescribed European model and the domestic one. If this fit is good no important changes at the national level, will have to take place. Only when there is a misfit, reforms are needed.

4 4 On the other hand, there is an implicit assumption behind the argument, which challenges a major claim of the Europeanization literature: It is assumed, that the mechanisms facilitating or inhibiting adaptation to Europeanization pressures, usually emphasized by the literature, are irrelevant in the case of monetary institutions. Börzel and Risse (2000: 7) state: Policy or institutional misfit, however, is only the necessary condition for domestic change. Whether misfits produce a substantial effect at the domestic level, depends on the presence of some factors facilitating adaptation as the sufficient condition of domestic change. These factors, located at the domestic level, are according to the two authors the existence or absence of multiple veto points, change agents and formal or informal institutions facilitating change. Also Héritier and Knill (2000: 2) hold, that a mismatch between national and European policies does not guarantee a reform: Rather to bring a policy change about reform capacity is needed. They argue, that this capacity is determined by formal and factual vetoplayers. The direction of change is also dependent on the member-states prevailing belief system. The responses to Europeanization will thus be differential. Similarly, Schimdt (2001) identifies five mediating factors that explain divergence in memberstate adjustment to Europeanization pressures: economic vulnerability, political institutional capacity, policy legacies, policy preferences and discourse. Are the mechanisms and mediating factors identified by the Europeanization literature really not relevant in the case of the institution of central bank independence? Have the pressures been so strong, that differential responses were unlikely? The answer I provide to these questions is no. Also in this domain domestic reform capabilities are important to explain the pattern of reform. In the remainder of this article, this argument is developed theoretically and empirically. III A Configurational Model of Institutional Reforms The frame of analyses is a configurational model of institutional reforms. This model draws from general concepts of new institutionalists and those of the Europeanization literature. It is configurational insofar, as it considers the claim, that in order to observe change, a conjunction of several factors is needed (Cortell and Peterson 1999) and that more than one combination of causes can lead to the same outcome. The model includes three main aspects: reform triggers, reform capacity, and reform response (cf. figure 1).

5 5 Figure 1: Configurational Model of Institutional Reforms Reform triggers Ideational, political-institutional, and economical pressures for an institutional model, which is not compatible with the current institution misfit. + Reform capacity A) Politics + Discourse B) Institutional capacity C) International Position = Reform response 1) Inertia 2) Adoption 3) Transformation Reform Triggers The points of departure are reform-triggers. Institutions, once in place, tend to be stable (Pierson 2000) 2. The tendency to stability implies that institutions will not be changed, unless there is an incentive to do so. Such incentives are ideational, political-institutional or economical. They can be single events or gradual developments stemming form internal or external sources (cf. Cortell and Peterson 1999; Wilson 2000). It is however important to see that the presence of such an event or development can only then be conceived as a reform-trigger, if the current institutional design can not respond to it. This means, that we cannot just look at changed circumstances to detect reform triggers. We also have to consider domestic legacies, institutional path, and vulnerabilities in order to assess, whether pressures for institutional change emerge form these events or developments. The concept of goodness of fit is used prominently in the Europeanization literature (Börzel and Risse 2000; Cowles, Caporaso and Risse 2001; Schmidt 2001). It is argued there, that the degree of adaptational pressure is defined by the level of compatibility between European and domestic policies and institutions. If the European model already resembles to a large degree the domestic one, compliance is possible without substantial reforms. A misfit is thus seen as a necessary condition to observe domestic change at all. When doing empirical studies on Europeanization, the first point should be to assess, to what degree domestic policies or institutions differ from the policies or institutions required by Europe. The goodness of fit concept can also be translated to more general analyses of change: If it is possible to assess the 2 Several properties ascribed to institutions are responsible for this stability: First, institutions are susceptible to increasing return processes: after high initial setup costs, once formed, institutions generate learning effects, coordination effects and adaptive expectations which increasingly lower the costs of social transactions (North 1990: 94). Second, political institutions are webbed into complex, interdependent and not very transparent systems. Third, existing structures shape the power-relations between social groups. The group that benefits form the current structure will use its influence to inhibit change (Pierson 2000). Fourth, institutions once internalized by actors, help them to identify their roles, to act in an appropriate way and to fill their social actions with meaning (March and Olsen 1989).

6 6 degree of misfit between current institutions and structures that would be necessary to cope with emerging problems, we can translate the concept to other reform triggers than just Europeanization. Reform Response In order to conceptualize institutional reform-responses it is important to make clear the underlying definition of institution. I adopt a new institutionalist conception, where institutions are defined as humanly devised constraints that shape social action (North 1990; Pierson 2000). Institutions are altered by two types of developments changes in formal rules at one hand, and changes in informal rules at the other. These two processes not necessarily have to occur at the same time. In the contrary, it can well be, that only the behavioral norms are changed, while the formal institutions still remain the same. And, as well is it possible, that a considerable reform of laws has relatively little effect, because the behavioral norms of the old law have remained the same. Even, if a legislative reform marks a large departure from the previous legislation, policy outcomes not necessarily have to change that much. Rather, as North already convincingly argued: change is overwhelmingly incremental (1999: 89): Although a wholesale change in the formal rules may take place, at the same time there will be many informal constraints that have a great survival tenacity [ ] The result over time tends to be [ ] far less revolutionary. (North 1990: 91). If one takes into account the three types of reform-responses take place: (1) inertia, (2) adoption, and (3) transformation. Inertia occurs, when neither the formal rules, nor the behavioral norms are changed. Adaptation takes place, if only one of the two is changed. And finally, transformation describes a reform process where both, formal and informal rules are changed. Reform capacity The presence of triggers for reform a priori gives no answer, about what type of reform response can be expected. The crucial intervening factor is the reform capacity of the political system. This reform capacity is determined by the political reform capacity, institutional reform capacity and the international position of the political system. The political reform capacity relates to the reform-preferences of the government and its ability to legitimate these preferences vis à vis the public (see also section on policy ideas and discourse). The institutional reform capacity relates to the presence or absence of institutional veto-players/veto-points in a system (Haverland 2000). Finally, the international position stands for the dependence of a system on other nations in terms of financial, but also political help.

7 7 Figure 2: Reform Expectations at Different Domestic Configurations Reform-Trigger Misfit in formal and/or informal dimensions of the institution Reform Capacity Reform Response Politics and Institutional International (Expectation) Discourse Reform Capacity Position For Reform High (Does not matter) Transformation For Reform Low High dependence Transformation For Reform Low Low dependence Adaptation Against Reform (Does not matter) Low dependence Inertia Against Reform (Does not matter) High dependence Adaptation In figure 2 the three elements are linked together. The point of departure is, that reform-pressures have created a misfit in both, informal, as well as formal dimensions of the institution. The figure lists the expected reform responses given different reform-capacity configurations. Policy Discourse: The Analysis of Ideational Pressures and Ideational Reform-Facilitators Special attention in this frame to analyze reforms is given to the role of ides and discourse. Policy discourse serves two purposes: First, it aims at finding shared beliefs related to the appropriateness of policy programs among policy-related elites. Second, it characterizes a process of persuasion dedicated at convincing the public, that these policy programs are in their interest. In this case the discourse aims at legitimating the policy ideas that are considered to be appropriate by the elites (Schmidt 2000). In alienation to Schmidt the term coordinative discourse is used to characterize the first and communicative discourse for the second. She argues, that the type of discourse lead in a country depends on the countries institutional framework. In multi-actor systems, because of the need for co-ordination and negotiation between elites, coordinative discourse prevails, while communicative discourse is dominant in single-actor systems. Discourse is conceptualized as a factor mediating domestic adaptation to external pressures or in the classifications of the model of institutional change presented here as an aspect of political reform capacity. But, discourse does also play a role at another stage of the model: it is important for tracing the development of ideational reform pressures. Ideational reform pressures are powerful in the moment a specific institutional model is promoted to be the solution. The type of discourse that takes place at this stage of institutional change is coordinative in type and takes places at different levels international, European, and domestic.

8 8 VI Analyzing the Pattern of Central Bank Reforms The majority of work on reforms of political institutions is based on single or comparative case studies. This also applies for studies on shifts in central bank independence 3. The reason for this dominance of case study research is that the assumptions about causal conditions in reform processes do not correspond to the assumptions that have to be met in quantitative studies. We do not expect the relationship between reform triggers and the reform outcomes to be linear, nor can reform triggers and reform capacity be conceived to operate in an additive manner. Reforms are conceptualized to be the outcome of the joint presence of certain factors at a critical point in time (critical junctures). Additionally, very diverse combinations of factors can produce the same reform-outcome (different path). Case studies are useful tools to observe sequences and casual connections that are practically impossible to model with quantitative techniques (Ragin 2000: 29). However, with small N studies, it is difficult to gain confidence that the findings also apply to a broader range of cases. On the other hand, if broad databases are included, trust in generalizations is higher, at the expense of reducing causation to simple patterns of co-variation between variables 4. One possibility to overcome this problem is a comparative strategy, which uses set theory to analyze configurations of conditions in a medium-sized population (Ragin 1987, 2000). This research strategy, which remains case-oriented, is applied here to study central bank reforms in the universe of industrialized democracies 5. The aim behind this undertaking is to test the configurational model of institutional change, which so far has only been used as a frame for analysis in case studies, for a larger set of cases. This empirical section starts with an outline of the possible reform triggers for monetary institutions. Then legal and behavioral reform patterns in the 22 included cases are described. Finally reform triggers, reform capacities, and reform responses are linked together. Triggers for Reforms of Monetary Institutions Ideational Reform Pressures Ideational pressures stem form the increasing dominance of the idea, that governments can only credibly commit themselves to a low inflation policy, if central

9 9 banks are independent. These ideas have been promoted by a group of economists, domestic finance ministries and central bankers, through European institutions, and also through international economic institutions. The scientific discourse of the question of central bank independence is related to the main macroeconomic debate of the 20 th century the dispute between Keynsianists and Monetarists. Whereas the former argued that discretionary monetary policy could be used to generate overall benefits for society (e.g. smoothing business cycles, preventing recessions and stimulating employment and growth), the latter replied that attempts to steer the economy would only produce inflation 6. Monetarists believed that the best solution for monetary policy is to adopt a fixed rule targeting the money stock (Friedman 1982 (1962): 54). The modern version of the rules versus discretion debate is called the time-inconsistency problem of monetary policy (Barro and Gordon 1983; Kydland and Prescott 1977; Rogoff 1985) 7. According to the timeinconsistency literature, discretionary monetary regimes will always produce an inflationary bias, even if policymakers are inflation-averse. Governments, thus, have to credibly commit themselves to a non-inflationary rule. Such a credible commitment is only possible, if monetary policy is delegated to an independent central bank, managed by a conservative central banker. Scholars defending the time-inconsistency argument for delegation of monetary policy to an independent agency have developed sophisticated indicators for measuring central bank independence and have provided a large empirical work showing a negative correlation between the degree of independence and inflation (see Eijefinger and Haan 1996 for an overview). After having found empirical support, that central bank independence reduces inflation without hindering growth Grilli, Masciandaro and 3 See for instance Elgie and Thompson 1998 (France and England, from the 17th century to 1997); Goodman 1992 (on Germany, Italy and France from the 50s to the 80s); Marcussen 1999b (Great Britain, France and Sweden in the 90s); Stockdale 1999 (US and England 20th century); Surel 2000 (France, Italy and Great Britain in the 90s). 4 See Ragin (2000) for a general discussion of the difference between variable-oriented and caseoriented research. 5 The 15 EU member stats, Norway, Switzerland, Australia, Canada, Japan, New Zealand, USA. 6 According to monetarists, expansive monetary policy will, instead of affecting growth, just crowd out through an increase in price levels. For a nice and entertaining overview of this economic debate see (Krugman 1994: Chapter 1). 7 The time inconsistency problem can be described as follows: The policymaker prefers a monetary policy which aims at price stability but still allows for a flexible response in case of an output shock (by increasing inflation to affect the real wage, employment and output). At the point in time, when wage-setters negotiate their nominal wage, they do not know, whether there will be an output shock. But they know, that if yes, it is accompanied by a policy response. In order to be on the safe side, they write a inflationary markup into their wage contracts. Because high wage claims in the case of stable prices would lead to an output-contraction, the policymaker has to accommodate by adjusting the inflation-rate. As a result, the inflation-level will be sub optimal and the policymaker looses its ability to react to real output shocks (cf. Lohmann 1992).

10 10 Tabellini (1991: 375) conclude in an influential article that having an independent bank is like a free lunch 8. The free-lunch thesis has been put into question by other comparativists, which claim, that the crucial variable is centralization of wage bargaining (Franzese and Hall 2000: 192; Hall and Franzese 1998). Corporatism is also seen as a good institutional alternative to independent central banks, since governments have the possibility to stimulate employment through monetary polices without having an inflation problem, because they can effectively negotiate wage/price-restraint with employers associations and trade unions (Iversen 1998a; Iversen 1998b; Scharpf 1988). While these studies still accept the assumptions made by the timeinconsistency literature but do not accept the free-lunch hypothesis, others rejected the whole theory (Kaltenthaler and Anderson 2000; Posen 1993). But, despite of this scientific debate, pro-independence ideas clearly dominate the discourse. Domestic Reform Pressures: Problems with High Inflation In the 70s and early 80s inflation has become a top priority problem in many western democracies. Apart form the oil-crises also other factors account for this. During the 70s Keynesian type of macroeconomic intervention was very fashionable and used by many countries. Fine-tuning of economies is however difficult. If not carried out with caution, a long time horizon, and awareness of the long and variable lags in its effects (Blinder 2000), the result will be a disturbance of prices (Notermans 1998). Such disturbances in prices can be both, deflation as a result of to restrictive monetary policies and inflation as a result of to expansionary policies. In both cases labor markets and capital markets will adjust to the deflationary or inflationary expectations in a way, which enhances the problem. Because of these self-enforcing processes, both types of price disturbances, if not stopped early enough, endanger growth and full employment 9. The problem of deflation is easier to cure. A turn to 8 This article has even been cited in a debate on central bank independence in the British Parliament (cf. Marcussen 1998). 9 Notermans (1998: 12) describes the problem as follows: In the case of a reduction of the price level demand and supply for investment is reduced. Deflation increases real debts, which makes borrowing for industrial investment less attractive. Lending money is unattractive because the risk for debt default is enhanced and holding money is more profitable. Less investment means less growth and more unemployment. If the labor market reacts with cuts in nominal wages, the price level will go down further and the vicious cycle is complete. In case of inflation initially the demand and supply of investment is increased. Accordingly, also employment and growth are stimulated. Tight labor markets react with higher nominal wages and as long as monetary policies are committed to full employment, this will lead to higher prices again. In the longer term, the result of inflation will be a flight out of money either into foreign assets, debts or real estate (since speculation becomes more attractive than productive investment).

11 11 more expansive policies is needed for correction. Solving the problem of inflation by turning to more restrictive policies is politically risky. The benefits of such a correction will only appear after a period of painful adjustment. In the short term the growth and unemployment problem is aggravated. In the 70s, the problem of the inflation spiral was enhanced additionally through the introduction of flexible exchange rates in the Bretton Woods system: Under flexible exchange rates inflation produces a run on the currency. The resulting exchange rate appreciation raises the price for imported goods, what increases the domestic price level even further. One attempted solution to address these problems, was to introduce a fixed exchange rate system again (Notermans 1998). Several countries started to peg their currency to a stable one. Another approach, tried in Europe was to establish exchange rate co-operation again. The failure of the European currency Snake, which operated only for a few years in the 70s, shows, that exchange rate cooperation would not work, until the inflation spirals had been broken (cf. McNamara 1998). The main task of countries bedeviled with to high inflation has thus remained to introduce price stability policies through monetary restriction by going through painful economic adjustment. Allotting the central bankers more power to decide on monetary policies is one strategy. Europeanization and Globalization At the European level both, monetary cooperation culminating in European Monetary Union and the completion of the single market are processes, which promote central bank reforms. In the case of Monetary Union, central bank independence is the model required by EU. Art. 108 of the Maastricht Treaty (effective in 1993) states that Each member state shall ensure [ ] that its national legislation including the statutes of its national central bank is compatible with this Treaty and the Statute of the ESCB2 10. In order to be compatible both was required, making price stability the primary goal of monetary policy and an independent central bank (EU 1997). Due to the treaty, the decision of a country to join the final stage of EMU included a commitment of this country for central bank independence. The second process at the European level, which may create reform pressures, is the single market. By allowing free movement of capital, services, goods, and labor, the European internal market, which was completed in the late 80s represents an extreme regional form of the economic integration process at the international level globalization. Economic integration is an indirect pressure for central bank reforms. On one hand, it increases the costs of using monetary policy and reduces its effectiveness (Garrett 1998; Milner and Keohane 1996; Scharpf 2000). On the other 10 European System of Central Banks

12 12 hand, because of increasing competition between national capital markets, the need for the reputation for being stable increases. Identifying Reform Triggers Overall the pressures for reform aim at an institutional model witch is based on the monetary policy paradigm of preservation of price stability and independent central banks. These reform-pressures constitute reform-triggers for those countries, where the goodness of fit between the prevailing institutional design and this model is poor. The assessment of the goodness of fit poses two practical problems: First, at what point in time should we measure goodness of fit and, second, how should we measure it. The first problem relates to the problem of handling dynamics and sequences in general. If a factor has a causal impact, the dependent variable should change after this factor appeared. However, most pressures for change mentioned above, like Europeanization or globalization are processes and not single events. Even in the case of the Maastricht Treaty, where a clear point in time can be specified for the legal effectiveness, the negotiation process itself could also have an impact. Moreover, if we look at the misfit right before the treaty became effective, we do not know, whether the misfit has been considerably larger some time before and the agreement specified in the treaty has only been a formalization of a trend taking place for other reasons. Related to this last point is the question of how to measure a misfit. What measurement criteria should we chose? In this study the focus is on the institution regulating the relationship between central bankers and governments. One way to assess the degree of independence is to focus on central bank statutes. Different scholars have applied this strategy in the late 80s and early 90s. Several such indicators for the degree of central bank independence have been constructed (see for instance Cukierman, Webb and Neyapti 1992). In the late 90s the same criteria have been applied to reassess the degree of independence after statutory reforms (Elgie 1998; Elgie and Thompson 1998; Sousa 2001; Surel 2000). The statutory reading strategy however, only allows assessing misfit in the formal dimension of the institution. But, changes in the formal structure are only one side of the coin, the other are informal rules and practices. James Forder even holds:...the true power of an institution is determined by much more than the surface appearance of the law and one needs to know the actual practice in policy-making rather than the formal rules (2001: 203). Stockdale (1999) argues that a change in behavioral norms and practices laid the fundament for a subsequent legal reform. This implies that legislative changes can be merely the formalization of current behavioral independence of the central bank, and thus do not affect the subsequent behavioral independence of the central bank. And,

13 13 even more interesting is the observation, that developments of formal and informal independence not necessarily have to parallel each other (Stockdale 1999: 19). It is thus possible, that central banks are behaviorally independent for a long period, without a statutory reform taking place. Having these points the analysis of goodness of fit should considers both misfit in legal and behavioral dimensions of central bank independence. In order to catch shifts in central bank independence before Maastricht, the period under study has to go back at least to the early 80s. The Reform Pattern in 22 Countries Reforms of Central Bank Legislation Table 1 reports the degrees of legal central bank independence in the 80s and today. In the 80s a majority of central banks in industrialized democracies were legally dependent from their governments. This changed during the 90s. By the turn of the century twelve of these countries belong to the Euro-zone and have a common central bank, which is constructed to be one of the most independent banks in the world (Elgie 1998). Outside of EMU, Sweden and the United Kingdom and outside of Europe, New Zealand (1989) and Japan (1997) adopted new laws that considerably changed the relationship between governments and central banks (cf. table 1). The mentioned reforms all followed a specific pattern: legally a clear cut restriction of government inference on central bank action was introduced and price stability was defined as the most important goal of monetary policy (cf. appendix table 1).

14 14 Table 1: Degree of Central Bank Independence in 22 Advanced Industrialized Democracies Indepen Legal status of central banks in the 80s a Legal status of central banks today b -dence EMU-country Other EMU-country Other (1) High Austria, Germany Switzerland Austria, Belgium, Germany, Finland, France, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain Japan, New Zealand, Sweden, Switzerland (2) Netherlands USA UK, USA, (3) Canada, Denmark Canada, Denmark (4) Ireland, Luxembourg Australia, Japan Australia (5) Low Belgium, Finland, France, Greece, Italy, Portugal, Spain New Zealand, Norway, Sweden, UK Norway Notes: Countries experiencing large reforms are: Belgium, Finland, France, Greece, Ireland, Italy, Japan, Luxembourg, New Zealand, Portugal, Spain, Sweden, and UK (criteria at least two categories away from 80s score). a Countries ranked according to several measures of legal independence as proposed by Freitag (1999: 108). He constructed an index using different measures for central bank indicators. Greece, Luxembourg, Portugal and Spain, which were not included in Freitags analysis, were added using Cukiermans (1992) assessment of legal independence. b Ranking of the countries according to the reforms stated in appendix table 1. If no significant reform took place, the value of the 80s is reported. The reform pattern of legal central bank independence shown in table 1, already allows to draw conclusions about reform-triggers: First, all reforms that occurred in advanced industrialized democracies are going in the direction of more independence. There is no case where formal central bank dependence has been enhanced. Central bank independence seems to become the dominant institutional model in modern states 11. Second, and related to this point is the observation, that only those countries that had a dependent central bank in the 80s experienced large reforms. Reforms are considered to be large, if the value of central bank independence is at least two points away from the 1980s score. Since all cases, where large legal reforms have taken place are countries having a dependent central bank in the 80s, but, not all countries having a dependent central bank have reformed them, we can conclude, that a legally dependent central bank in the 80s is a necessary, but not a sufficient condition for a large reform. Third, all EMU countries have adjusted to the treaty requirements. The joint presence of two conditions: namely having a dependent central bank and membership in EMU can thus be interpreted as a sufficient condition for a large reform of the central bank legislation. Fourth, belonging to the single market is not a sufficient condition for a reform. The Maastricht Treaty as a consequence of strong opposition to the project from Great Britain permits opting-outs. This option was chosen by Great Britain and Denmark. 11 The trend towards central bank independence can also be observed in developing countries (cf. Cotarelli 1997).

15 15 The opt-out contains, that the two countries can still join the EMU, when the time is considered appropriate. Sweden because it joined the European Union after the negotiation of the treaty does not have this special status. But, the Swedish government has explicitly issued a government bill stating that Sweden is not joining EMU and that this decision can only be overturned by a general referendum. The Swedes, however, chose to interpret the treaty in a differential way: while they didn t consider membership in the currency union as compulsory, they still argued that they are obliged by EU law to implement the statutory requirements for central bank independence as stated in the Maastricht treaty (cf. Marcussen 1999). They reformed their central bank together with many EMU-states in The reform was enough pronounced to make Sweden ready for EMU-membership with respect to the institutional requirements considering the central bank. Great Britain also reformed its central bank however the degree of central bank independence after the reform is still a bit smaller than the one of full EMU-states (cf. Surel 2000). Denmark has not remarkably changed its legislation. Norway, participating through the EEA in the single market without being a member of the Union has not changed its central bank legislation. Norway even remains alone in the category of countries having a largely dependent central bank this is a strong case for the rejection of the hypothesis that participation in the single market is sufficient for a formal reform. Shifts in Behavioral Central Bank Independence Informal constraints pose a practical problem for empirical researcher. How can we measure norms, conventions or codes of conduct, which are not written down and do not manifest themselves in a material manner? Although, most empirical political economy studies use measures for central bank independence based on formal criteria, some attempts to catch behavioral independence have been made (cf. Eijefinger and Haan 1996: 26-28). For instance, behavioral independence has been measured through the turnover rate of central bank governors or questionnaires sent to central bankers (Cukierman, Webb and Neyapti 1992, Cukierman 1992). Another research strategy is to measure the behavior indirectly through the outcomes of a regime. Iversen (1998b, 2000) relies on an indicator of the relative differential growth in exchange rates across countries to operationalize the monetary regime (Iversen 2000: 212). This measure is based on the assumption, that relatively appreciating currencies reflect more restrictive, non-accommodating monetary polices, while relatively depreciating currencies stand for accommodating regimes. The rationale behind this assumption is, that an increase in (money) supply ceteris paribus will lead to a decrease in prices (exchange rates can be interpreted as the price of a currency expressed in terms of other currencies). Based on similar arguments, also the inflation rate or the interest rate could be taken as indicators for central bank

16 16 behavior. Such indicators have two major weaknesses. First, central bank behavior is only one among many factors influencing them. Especially in open economies with flexible exchange rates, much of their variation is determined by developments, which are stemming from abroad and are out of the reach of national central banks. Second, intensified interconnectedness and increased exchange rate co-operation between countries complicates the interpretation of the indicators considerably. However, if these quantitative indicators are used together with more qualitative information on economic policy development and government central bank relations, they are very useful tools of analysis. Table 2: Reduction of Inflation Rates from the 80s to the 90s Country Average inflation 80s Average inflation 90s Difference 90s and 80s Portugal Greece New Zealand Italy Ireland Australia Spain Sweden Norway Finland France UK Canada Denmark USA Belgium Luxembourg Switzerland Japan Austria Netherlands Germany Mean Sources: (OECD 1999; OECD 2000), Own calculations. A look at macro-indicators for monetary policy, for instance inflation rates, indicates that achieving disinflation and price stability has become a task in all countries. Case studies show, that the turn to low inflation policies also had an impact on behavioral central bank independence. Looking at the relation between the French core executive and the Bank of France, Elgie and Thompson (1998) argue, that the transformation of the old regime which had manifested itself in a highly dependent central bank started already in the early 1980s. In 1982, after severe exchange rate crises, Mitterand had to reverse his Keynesian policies in favor of austerity. By the mid-1980s the Socialists contributed to the formation of a new

17 17 national consensus that recognized the need for lower inflation and external equilibrium (Goodman 1992: 141). Subsequently, also proposals for central bank reforms came on the table (Elgie and Thompson 1998: 121). Although, during the 1980s legal reform never became a top-priority issue, the bank was given de facto responsibility for managing liquidity levels (Elgie and Thompson 1998: 124). A similar story can be told for the British case. After England had left the European Exchange Mechanism in 1992, the Bank began to play a more important role in the development and implementation of monetary policy. This new role manifested itself in a new Inflation Report, through which the Bank was made responsible for monitoring the government s progress in pursuing the inflation target (Elgie and Thomson 1998: 77). If we consider shifts in legal central bank independence and reduction of inflation rates at the same time, it becomes clear, that indeed behavioral and legal responses to reform-triggers not necessarily coincide. Also those countries, without large statutory reforms achieved considerable reductions of the inflation rate (cf. table 3). Table 3: Reduction of inflation rates and shifts in legal central bank independence Reduction of average inflation rates between the 80s and the 90s b more than 4 percentage points less than 4 percentage points Large reform of the legal Yes FIN, FRA, GRE, IRE, ITA, NZL, POR, SPA, SWE, UK BEL, JAP, LUX structure a No AUS, CAN, DEN, NOR AUT, GER, NET, SWI, USA Notes: a For the criteria to decide on cases which experienced large reforms in the legal structure refer to table 1. b cf. table 2 According to table 3 we can divide the 22 countries into 4 groups. The group, encompassing the five countries, which had neither considerably reformed the central bank statutes, nor substantially reduced the inflation rate, are Austria, Germany, Switzerland, Netherlands and the USA. All five neither had a dependent central bank in the 80s, nor a considerable inflation problem. These countries are not interesting for the analysis of central bank reforms, simply because there was no misfit. Three cases where not confronted with a virulent inflation spiral, but still considerably reformed their central bank statutes: Belgium, Japan, and Luxembourg. In these countries the legal reform was not caused by big changes in monetary policies, since price stability has been established for a long time. The explanation for the reform can thus not lie in the domestic economic situation. Rather, the reforms have been adaptations to ideational and institutional reform-triggers.

18 18 Since 1973 the Japanese monetary authorities adhere to an unambiguously nonaccommodating policy (Iversen and Pontusson 2000: Footnote 10). The reform was a very late formal institutionalization of a policy that has been conducted for many years before. But why would Japan institutionalize central bank independence in 1977? One explication lies in the decreasing competitiveness of the Japanese economy as a consequence of the continuing appreciation of the US Dollar against the Yen in the run-up to the Asia-crisis. The institutionalization of central bank independence can be interpreted as a symbolic measure to signalize the markets, that Japan even when facing economic problems will pursue a credible and stable monetary policy. Belgium has participated in all European exchange rate arrangements and except for a short period of crises in 1981/82 been a country with a stable currency and low inflation. The central incentive for reform of the central bank statutes was the countries membership in EMU. The reform has only been an adaptation to the institutional requirements of the treaty and not a fundamental transformation of the monetary policy regime. This also holds for Luxembourg, which created the Central Bank of Luxembourg in order to meat the treaty requirements. The remaining 14 cases have all faced considerable changes in their monetary policy practices, while being different on other categories. Ten cases had a large reform of the central bank statutes. Four did not. Among the 14 we have European as well as non-european countries, EMU members and opt-outers, countries with only low dependence on international markets, and countries which are highly integrated in the international economy. The commonality they have is, that all have suffered under high levels of inflation. Indeed the strongest trigger for change in macroeconomic practices seems to be the failure of old policy concepts, which lead to a process of learning and to the adaptation of macroeconomic policies that guarantees price stability. The path to arrive to these policies has however been differential: In some countries, this process has been abrupt as for instance in the case of Mitterands turnaround 1982 in France. In other countries, for instance in Denmark or Norway, the process of change had been slower and subtler. Some countries, have needed pronounced reforms of the formal government-central-bank relationship in order to achieve price stability (e.g. New Zealand, Italy, Portugal, Spain and Greece), some have not (Australia, Canada, Denmark, Norway, but also Finland, France, Ireland, Sweden, and UK, which managed to stabilize inflation already under the previous regime cf. appendix 2). The explication for these differential responses lies in the reform capacities of these countries.

19 19 Explaining Differences in the Reform Patterns of Formal Central Bank Independence the Link to Reform Capacity and Policy Ideas Only in five of the 22 countries substantial changes in legal and behavioral dimensions of the macroeconomic policy regime coincide: New Zealand, and the four southern European cases. These are the cases, where a clear institutional transformation has taken place. When New Zealand reformed its central bank legislation, in 1989, it had a political system with virtually no veto-players and a strong, centralized government. Lijphart (1999) classifies before 1989 New Zealand as a model case of a Westminster Democracy. Institutional capacity for a large reform has thus clearly been given. The introduction of central bank independence was only one measure of a whole package of economic reforms in the country introduced by the labor government (Silverstone, Bollard and Lattimore 1996). It is difficult to say, why this government started the economic reforms. Wallis (1997) argues, that the transition was enabled by a conspiracy of technocrats. The institutional design introduced by New Zealand supports this view. Unlike, the central bank reforms triggered by EMU, which have been clearly stimulated by the German model (Dyson and Featherstone 1999: 794), New Zealand adopted a design, which was inspired by the time-inconsistency and principal-agent theory 12. The southern European countries, unlike the rest of the EMU group, have had considerable problems in achieving price stability. Divided governments and high institutional constraints for reforms, like in Italy are one explanation, the difficult transitions of the former dictatorships another. In these cases, certainly, the prospect to borrow credibility and reputation by the means of a European currency union had been a powerful motivator for transforming previous macroeconomic regimes. (For the case of Italy cf. Sbragia 2001). In the other EMU countries legislative reforms of central bank statutes had only been codifications of practices that have already been established for long time. Maastricht has thus been the final tip in a process of change and did not lead to a major overhaul of domestic policy styles (Sandholtz 1993). The discourse on the institutional frame for monetary polices, driven by a small elite consisting of economists, central bankers, and finance ministers (Dyson 2000), had however an impact on the type of institutional solution, that was chosen to formally institutionalize the changed practices. It would be wrong to conclude that Europeanization had no impact at all. But Europeanization forces laid mostly in the capacity of European actors to frame an institutional model, to be the solution for the inflation problem. But also here, we have to look at the domestic level to understand, why some countries responded to 12 After the government commits to an inflation target, the central bank can independently decide over monetary policy instruments to achieve this target. In case the target is clearly missed, mechanisms to punish central bankers have been introduced.