You are on page 1of 56

Seminar on Economic Policy in Europe Prof. Dr.

Dirk Wentzel SS 2012

The Future of Monetary Integration in Europe after the Crisis

Topic 4: Central Bank Independence & Macroeconomic Performance

Onieit, Mathias Bachelor International Business ID#: 303469 onimat@hs-pforzheim.de

Espinoza Chavarri, Diego Bachelor International Business ID#: 303785 espdie@hs-pforzheim.de

Submission date: 05/07/2012

Table of Contents
List of Abbreviations .............................................................................................. II List of Illustrations ................................................................................................. III List of Tables ......................................................................................................... IV

1 2 3

Abstract .............................................................................................................. 6 Introduction........................................................................................................ 6 Historical Evolution of Central Bank independence ......................................... 7 3.1 Changes in Central Bank Independence ...................................................... 7 3.2 Reasons ...................................................................................................... 12

Central Bank independence and Inflation........................................................ 13 4.1 Theoretical predictions ............................................................................. 13

4.1.1 Impact of Inflation ................................................................................ 13 4.1.2 The Credibility Problem ....................................................................... 14 4.1.3 The Solution: Conservatism and Independence ................................... 16 4.2 Empirical findings .................................................................................... 17 Measuring Central Bank independence ............................................. 17

4.2.1

4.2.1.1 Legal Central Bank independence ................................................... 17 4.2.1.1.1 Criteria for legal Central Bank independence ............................. 17 4.2.1.1.2 Variable aggregation ................................................................... 19 4.2.1.2 Governor turnover rate (TOR) ......................................................... 20 4.2.1.3 Questionnaire on Central Bank independence ................................. 20 4.2.2 4.2.3 4.2.4 5 Modeling the relationship .................................................................. 21 Results and interpretation ................................................................... 21 Central Bank independence and Inflation variability......................... 24

Central Bank independence and real macroeconomic variables ..................... 25 5.1 5.2 5.3 Central Bank independence, GDP growth and investment ...................... 25 Central Bank independence and interest rates ......................................... 26 Central Bank independence, wages and unemployment .......................... 28

Criticisms of Central Bank independence ....................................................... 29 6.1 Methodological issues .............................................................................. 30

3 6.2 6.3 Subjectivity .............................................................................................. 32 Robustness................................................................................................ 35 Control Variables ............................................................................... 35 Observation Periods ........................................................................... 38

6.3.1 6.3.2 6.4 6.5 6.6 6.7

Causality .................................................................................................. 39 Disinflationary credibility ........................................................................ 44 Wage demands ......................................................................................... 47 Alternatives .............................................................................................. 48 Forms of exchange rate fixation......................................................... 48 Inflation contracts............................................................................... 49 Inflation targets .................................................................................. 50

6.7.1 6.7.2 6.7.3 7

Conclusion ....................................................................................................... 50

References.............................................................................................................. 52

List of Abbreviations
GDP EEC ECB CBI LCBI TOR GMT A/PD Gross Domestic Product European Economic Community European Central Bank Central Bank independence Legal central bank independence index Turnover rate of central bank governor Legal central bank index by Grilli, Masciandaro, Tabellini (1991) Legal central bank index by Alesina (1988) based on Parkin and Bade (1985) CWN QCBI IQ FOI Legal central bank index by Cukierman, Webb and Neyapti (1992) Central bank independence index based on a questionnaire Institutional Quality Financial opposition to inflation

List of Illustrations
Figure 1: The evolution of central bank independence and inflation ....................... 9 Figure 2: Average Aggregate Index of Legal Independence, in Chile and 9 Latin American Economies .............................................................................. 10 Figure 3: Legal Central Bank Independence in Selected Former Socialist Economies ......................................................................................................... 11 Figure 4: Partial relation of Inflation to the Legal Independence and to the Turnover Rate of Governors of Central Banks in Industrial and Developing Countries ................................................................................................. 22 Figure 5: Relation between legal CBI and inflation in two different periods ....... 23 Figure 6: Scatterplot: Central Bank independence and Average Real GNP Growth ................................................................................................................ 26 Figure 7: Scatterplot: Central Bank independence and Average Real Interest Rate ................................................................................................................ 27 Figure 8: Scatterplot: Central Bank independence and Variance of Real Interest Rate ......................................................................................................... 28 Figure 9: Scatterplot: Central Bank independence and Average Unemployment . 29 Figure 10: Varying independence rank of the United Kingdom ........................... 34 Figure 11: Correlation between CBI and FOI and CBI and Average Inflation ..... 40 Figure 12: Inflation aversion for 10 core EU countries and average inflation rates ................................................................................................................ 43

List of Tables
Table 1: Institutional determinants of CBI in the literature .................................. 32 Table 2: Correlation coefficients between different measures of CBI ................. 33

6 1. Abstract This paper critically analyzes the historical evolution of central bank independence, its theoretical foundations and the empirical studies on its impact on inflation and real macroeconomic variables. The results show that to date, no reliable confirmation of the favorable effect of central bank independence on price stability has been provided. Moreover, several aspects of the underlying theory remain questionable. The impact of central bank independence appears to be contingent upon other factors that have yet to be explored in detail.

2. Introduction

The recent events surrounding the financial crisis in Greece and its bailout by the European Central Bank in violation of article III-183 of the Maastricht Treaty which states that The union shall not be liable for or assume the commitments of central governments raises the question whether central bank is really independent from political influence. This, in turn, raises also the question whether central bank independence is an important aspect of the monetary system.

Throughout the last decade, central bank independence, defined by The New Palgrave Dictionary of Economics as freedom of monetary policymakers from direct political or governmental influence in the conduct of policy has gained increasing popularity among economists and politicians around the globe.

The aim of this paper is to present the theoretical foundations and empirical findings which have led to the widespread acceptance of independence as a fundamental element of central bank design. This work is organized as follows:

Chapter 3 outlines the institutional changes that have occurred over the last twenty years and discusses the reasons for the shift towards higher degrees of central bank independence.

7 Chapter 4 illustrates the theoretical foundations and predicted benefits of central bank independence and presents empirical evidence for the negative relation between central bank independence and the average rate of inflation. In Chapter 5 the relation between central bank independence and real macroeconomic variables such as GDP growth and unemployment are analyzed. Chapter 6 provides a survey of major points of criticism in literature and thorough evaluation of them. Finally, Chapter 7 summarizes the preceding chapter and draws a conclusion.

3. Historical Evolution of Central Bank Independence 3.1 Changes in Central Bank Independence

Today, central banks legal and actual independence is considerably higher than 20 years ago. Before the 1990s, central banks autonomy did not seem to be that needful for maintaining price stability and economic welfare. At that time, most of central banks did not even have legal autonomy or instrument independence, limiting their scope of action by being just a department of ministries of finance.

Nevertheless, it cannot be merely affirmed that legal and actual independence did not exist at all. In fact, in some developed countries such as Japan, UK, USA and West Germany, central banks possessed a certain degree of legal autonomy, with price stability being one of their main objectives. However, unlike in our days, in the majority of countries price stability did not receive special attention from the central bank, belonging thus just to a group of several objectives such as steady growth, higher levels of employment or the financing of the governments public expenditures. It must be also noted that in another group of countries central bank did not even possess instrument independence so that price stability was maintained merely by the ministry of finance. (Cukierman, 2006)

8 Although some central banks possessed a certain degree of legal autonomy, the level of actual independence was basically lower than the level indicated in the law, particularly in developing countries. In other words, actual independence showed wide deviations from legal independence (Cukierman, 2006).

But things are different today. Throughout the last 20 years, most central banks have substantially increased their levels of statutory independence. Nowadays, central banks main objective is to ensure price and financial stability. There is much evidence confirming the dramatic increase of legal central bank independence in the 1990s worldwide.

This statement is based on a number of legal indices developed by different authors such as Cukierman (1992) and Cukierman, Webb, and Neyapti (1992), on updates of the Grilli, Masciandaro, and Tabellini (1991) index, and the index of Cukierman, Miller, and Neyapti (2002) regarding twenty-six former socialist economies in the 1990s. Additionally, it has been also taken into account the work by Jacome and Vazquez (2005), which provides an index for twenty-four Latin American and Caribbean countries (Cukierman, 2006).

Figure 1 displays the evolution of central bank independence from the 1980s to 2003 for 69 countries, separately for OECD and non-OECD countries. As the graphic shows, the majority of central banks have become more independent. As an example, OECD countries have on average experienced a reduction in average inflation by 15 percentages while CBI has on average increased by 0.30. (Markwardt and Hielscher, 2011)

Source: Markwardt and Hielscher, (2011). Indicators of central bank independence are taken from Cukierman et al. (1992) and Crowe and Meade (2007). Countries with inflation rates above 50% are excluded from the figure.

In addition average inflation rates have decreased in almost all countries, both OECD and non-OECD countries alike. Furthermore, the Maastricht Treaty signed by the European Economic Community (EEC), has enabled the European Central Bank (ECB) to pursue monetary policy independent of national governments. For this reason, all the countries that joined the European Monetary Union have upgraded their level of legal independence to that of the Bundesbank since the 1980s. In the same manner, the United Kingdom granted the Bank of England full autonomy in 1997. (Cukierman, 2006)

Figure 2 shows the evolution of average legal independence in Chile (dark bar) and nine Latin American countries (light bar) during the last fifty years based on the legal Central Bank Independence Index (CWN Index) by Cukierman et al. (1992).

10

Source: Cukierman, (2006). Calculations based on data from Cukierman, Webb and
1

Neyapti

(1992)

and

Jacome

and

Vazquez

(2005).

The graphic shows that on average, legal independence has increased over the years. After around forty years of relative immobility in central bank legislation, the level of legal independence took a quantum leap in the 1990s, not only in developed countries but also in developing countries (Cukierman, 2004).

Chile shows a similar trend to that of the other countries, but the level attained over the 1990s is higher than the average level of the other Latin American countries. Although Jacome and Vazquez (2005) did not take into account every country for their study, it illustrates that in general, CBI increased steadily over the years.

Notes: The figure combines data on the aggregate index of legal independence, CWN, from table A1 in Cukierman, Webb and Neyapti (1992) for the forty years between 1950 and 1989 with data on the same index for the 1990s from Appendix II in Jacome and Vazquez (2005) for Chile and for nine other Latin American countries, namely Argentina, Bolivia, Colombia, Honduras, Mexico, Nicaragua, Peru, Uruguay and Venezuela.

11

Figure 3 presents the average value of the legal CBI index (CWN) for six Former Socialist Economies, namely Croatia, Hungary, Macedonia, Poland, Romania and Slovenia. The legal reforms they instituted, mainly in the early 1990s, represent about a tripling of the CWN index developed in Cukierman, Webb, and Neyapti (1992). (Cukierman, 2006)

Source: Adapted from table 1 in Cukierman, Miller and Neyapti (2002).2

The bar for period 1 reflects the average level of legal independence in those countries prior to the central bank reforms of the nineties. The bar for period 2 reflects the average level of legal independence in those countries after the last central bank reform between the early nineties and 1997. (Cukierman, Miller and Neyapti, 2002).

The former socialist economies had two central bank reforms in the 1990s at intervals of about five years, in order to incorporate a uniformly higher level of in2

Notes: The figure shows the average value of the CWN index (see remark to figure 1 for details) for six Former Socialist Economies which were not part of the Soviet empire and had, consequently, central bank laws prior to the downfall of the Soviet Union.

12 dependence. This fact supports the view that the trend toward legal independence intensified in the 1990s (Cukierman, Miller and Neyapti, 2002). In summary, the sources cited above appear to support the conclusion that legal independence levels experienced a sustained increase in the 1990s worldwide.

3.2 Reasons

The indicators showing the tendency towards increased central bank independence (CBI) in the last 20 years were presented in the previous section. But why did central bank independence level actually rise so much since the 1990s? According to Cukierman (2006), this quantum leap of legal independence was the result of a combination between global and regional factors. The following reasons were put forward by Cukierman (2006) and try to explain the reasons for this increase.

The stagflation of the 1970s and the poor economic performance of high-inflation countries in Latin America in comparison to the good performance of lowinflation countries such as Germany or Japan increased a worldwide quest for price stability.

In addition, as mentioned above the low inflation level achieved by the highly independent Bundesbank and the Central Bank of Japan until the 1980s seemed to give evidence that central bank independence could be an effective tool to achieve price stability. Complementarily, new theories such as Rogoffs work The Optimal Degree of Commitment to an intermediate Monetary Target (1985) suggested a relation between central bank independence and the level of inflation.

Furthermore, regional treaties like the Maastricht Treaty signed by the European Economic Community (EEC) in 1992, helped to raise the central bank independence since a precondition for membership in the European Monetary Union (EMU) was the upgrade of their central banks independence.

Finally, globalization increased foreign investment by dismantling controls on capital flows. The incentive for investors to enter new foreign markets raised the

13 necessity of governments to maintain price stability, in order to capture the biggest quantity of private investment.

According to Cukierman (2006) these events finally led to the general acceptance that inflation deteriorates the economy and price stability should be placed as the most important goal for central banks. To achieve fiscal stability and consequently a superior economic performance, central bank independence could be the accurate tool.

These factors explain at least in part the significant development of central bank independence outlined in chapter 3.1.

4. CBI and Inflation 4.1 Theoretical predictions 4.1.1. Impact of Inflation Experience has shown over the years that high and volatile inflation can have dramatic economic and social consequences. High inflation if it is not controlled on time- can lead to a breakdown of the economy, being therefore inflation and output growth negatively correlated in high-inflation countries (King, 2004).

It is relevant to mention that the impact of inflation depends partly on whether inflation is anticipated or unanticipated. As a principle feature of the anticipated inflation, it can be pointed out that economic agents begin their protection by making predictions about future inflation rates. For instance, in case of a steady inflation increase, labour unions may negotiate with employers to increase the wages level in order to protect their real wages. They would basically try to maintain their real living standards. As a result, companies costs would rise, forcing employers to settle a higher selling price so that no changes in their profits are produced. Consequently, this would create further pressure on wages. This process is known as wage-price spiral which results in further inflation or even unemployment. (Fischer, 1982)

14 On the other side, unanticipated inflation refers to the difficulty of economic actors to predict correctly the rate of inflation in the near future, especially if inflation is volatile every year. This can result in errors in their inflation forecasts, losing, as a consequence, part of their real income. This is the so called money illusion. Here, the economic actors confuse nominal and real values and make bad decisions regarding their financial plans. (Riley, 2006)

Additionally, from an investors point of view, budgeting becomes more difficult as a result of the constant rise of prices and costs. Price instability could therefore demotivate them to invest in some countries. It should not be forgotten that lower investment has a negative effect on the economys long run growth potential. Moreover, at least until exchange rate mechanisms adjust relative prices, a loss in international competitiveness is caused whenever a country experiences a higher rate of inflation than another, resulting in a worse trade performance. For the country in question this causes a higher unemployment. (Cikowicz and Rzoca, 2010)

Furthermore, the consequences of inflation on society are considerable. After an increase in inflation a redistribution of income normally takes place. There is a negative impact of inflation on savers. As a result of a rise in the price level, money loses its value; thus savers will lose money if nominal interest rates are lower than inflation rate since it leads to negative real interest rate. (Mankiw, 2008)

Based on the reasons presented above, it seems to be of extreme importance to have inflation continuously under control. The task for the government is to find a monetary institutional structure that stabilizes prices. Nowadays, governments are more aware of this matter, paying special attention to the causes, consequences and solutions for inflation. Theories and facts appear to show a negative relation between inflation and central bank independence, suggesting as a remedy for high inflation rates the increase of central bank independence.

4.1.2. The Credibility Problem

15 This section analyses the theoretical foundations the effect of central bank independence on inflation. Barron and Gordons theory (1983) is maybe the strongest influence for all later central bank independence developments. They based their studies on the work of Kydland and Prescott (1977) and the time-inconsistent behavior explained in it.

Barron and Gordon (1983) consider a social welfare maximizer who has complete control over the rate of inflation and whose goals are determined over employment and inflation. They assert that any deviation from prefixed employment and inflation target levels causes increasing marginal benefits that outweigh marginal losses from increased inflation. For instance, inflation reduces real wages due to the fact that nominal wage contracts are fixed over a period of time. Inflation would however create more output since it would be cheaper for employers to contract additional workers and increase output. Furthermore, monetary stimulation of the economy through lower interest rates and consequently higher investment thanks to lower capital costs would be more effective, as labor costs remain unchanged until the expiration of the wage contracts. This would result in higher output and lower rates of unemployment, goals favored by the government especially in times of election.

For that reason, although the central bank announces an inflation target, there is always an incentive for the welfare maximizing government to exercise influence on the central bank and its monetary policy so that it lets the inflation rate rise above the announced rate in order to attain higher output. This incentive increases especially once wages are set (Hayo and Hefeker, 2007).

Because private agents are assumed to be forward-looking, they recognize the welfare maximizers incentive to increase the money supply beyond the announced target, and incorporate this expectation in their wage demands. (Barron & Gordon, 1983). Consequently, there is an increase in the inflation rate as well as in the wage level. In other words, no output and no employment gain will be produced, but just a positive rate of inflation which lies above the natural rate. This is the so-called inflationary bias (Hayo and Hefeker, 2007).

16

Therefore, a mechanism should be devised which makes policy announcements credible again and eliminates the inflationary bias.

4.1.3 The Solution: Conservatism and Independence In order to solve the credibility problem, Rogoff (1985) suggested the appointment of a conservative central banker and the granting of independence to the central bank.

Society can make itself better off by selecting an agent to head the independent central bank who is known to place greater weight on inflation stabilization (relative to unemployment stabilization) than is embodied in the social loss function. (Rogoff, 1985). According to him, Central bank independence can be conducted firstly by nominating a conservative central banker and giving him the authority to apply arbitrarily any monetary policy. In other words, to nominate a central banker who presents a stronger ant-inflationary attitude than the rest of the society and granting him independence from government influence.

The mechanism suggests appointing someone whose preferences are known to diverge from those of the welfare maximizing authority. If someone who puts more relative weight on avoiding inflation than unemployment were to set monetary policy, the rate of inflation would be lower (Hayo and Hefeker, 2007).

The reason is that wage setters would no longer need to anticipate a higher inflation rate above the announced inflation targets since government would no longer exercise influence on the monetary policy in order to increase output and employment. Knowing the inflation target -fixed and carried out by the conservative central banker- wage setters do not have the necessity to demand higher wages than needed since they do not have to fear that the conservative central banker will renege on his announced target rate (Hayo and Hefeker, 2007)

17 Thus, appointing a conservative central banker can help to reduce the inflation bias, although the possible decisions taken by the Central bank are not always the most accurate solutions from a social point of view (Rogoff, 1985).

4.2 Empirical findings

In pursuit of a confirmation of the theoretical assumptions, since the beginning of the 1990s several economists conducted empirical studies regarding the real effect of CBI on average inflation, the most prominent of them being Parkin and Bade (1988), Alesina (1988), Grilli, Masciandaro and Tabellini (1991) and Cukierman, Webb and Neyapti (1992), in the order of their publications. Before a relation between CBI and inflation can be established, a method to measure CBI has to be devised first. The following section will explain the approaches of the aforementioned authors. For the sake of conciseness and to avoid redundancy, we limit our illustration of the methodology to the work of Cukierman, Webb and Neyapti (1992). We believe that this is justified for two reasons. First, the 4 studies used very similar approaches and reached the same conclusions. Second, a wider sample of countries and a new measure for central bank independence, the turnover rate of the central bank governor is used. Third, the study of Cukierman et al. (1992) is the most recent of the named studies, so it should take into account possible methodological issues occurred in previous publications. Subsequently, the results will be analyzed.

4.2.1 Measuring Central Bank Independence

4.2.1.1 Legal Central Bank independence

4.2.1.1.1 Criteria for legal Central Bank Independence

Defining a measure for central bank indepedence is not a self-evident task. As Cukierman, Webb and Neyapti (1992) put it: Unavoidably, there were subjective or arbitrary decisions in coding, classifying, and weighing. Other authors made

18 similar remarks3, each showing the difficulty of finding an appropriate measure that reflects CBI. The closest proxy the authors find for representing a banks independence are the statutory rules encoded in the law, which would be the firstbest solution if they were followed exactly in reality. The CBI indicator based on legal provisions is often referred to as the legal indicator. To compose their legal indicator, Cukierman, Webb and Neyapti (1992) created 4 different groups of legal characteristics. The first cluster regards the person of the central bank governor. Each countrys laws were scrutinized to find out how the highest executive authority can be influenced by the government, namely through his appointment, dismissal and term of office. If the government has great influence on deciding about the person of the governor, it is more likely to see a governor appointed that acts in line with the governments expectations. If the theory holds that the central bank is more inflation-averse than the government, the power of the government to select the governor could compromise any objective of the bank to hold prices stable. Similarily, a relatively short term of office could induce the governor to fulfill the governments monetary desires, in the hope that he will not be replaced. Long terms of office, especially relative to the politicians terms of office, can in turn increase independent behaviour by the central bank governor, as he can define and apply his own monetary policies without the risk of being penalyzed by dismissal. Moreover, the possibility for the central bank governor to hold other offices in government can compromise his independence, so this aspect too is included in this group of characteristics.

The second group of characteristics is concerned with monetary policy formulation. It should be obvious that a high degree of participation of the government in this process can lead to higher inflation if the government prefers an expansionary policy while the central bank is committed to a more restrictive intervention in the money market.

Alesina (1988): These tests require first a classification of the degree of of dif-

ferent Central Banks, which is not easy to compute.

19 Beside policy authority, the second category also embraces an index for the resolution of conflicts that may arise between government and the central bank during the policy determination process. Who will have the final say in case no compromise can be reached? If it is the central bank then the central bank can be assumed to be more independent. On the other side of the scale, if the last word is with the government, de jure policy setting authority will not influence de facto independence positively. Between the two, various shades of grey exist. A third characteristic regarding the role of the central bank in the governments budgeting process complements the policy index group.

Not less important than the previous two groups is the third, which covers the banks objectives. Ceteris paribus, if a central banks statutes state price stability as the only or most important objective, Cukierman, Webb and Neyapti (1992) consider the central bank as more independent, compared to, for instance, a provision where economic growth or full employment is the primary goal and price stability is subordinated, at best. It might be surprising that the objective enters the CBI indicator, despite the fact that from a logical point of view, the objective itself cannot be an indicator of CBI, as the definition of CBI is the extent to which the objective (namely of low inflation) can be attained. However, if the CBI is interpreted in a broader sense, as the independence of the central bank in the pursuit of low inflation, the objective of low inflation can be seen as a prerequisite and thus be included in the array of indices that make up CBI. The fourth and final group of legal independence indices addresses lending practices. Limiting the ability of the government to borrow directly from the government, be it through absolute prohibition or clearly defined lending conditions such as interest rate and maturity, can enhance CBI. If, on the other hand, the government had carte blanche as far as borrowing from the central bank is concerned, a tendency towards high inflation could be expected, provided that the government resorted to this source of debt funding in an abusive manner.

4.2.1.1.2 Variable aggregation

20 Once they have attributed a value between 0 and 1 to each particular index of the groups of characteristics, with 1 being the value denoting the highest conduciveness to independence, the average of each group is created, before the groups are weighted according to their presumed significance for independence, and summed up to result in the legal CBI index.

4.2.1.2 Governor turnover rate (TOR)

In addition and in contrast to previous studies, Cukierman, Webb and Neyapti (1992) not only relied on the legal CBI indicator, but also on two informal CBI indicators, one of them being the turnover rate of the central bank governor. The authors correctly recognized4 that the legal CBI can only approximate real CBI, either because legal provisions are incomplete and the void is filled by traditional conduct and the bank officials personal discretion, or because the laws are not followed in practice. To overcome the gap between legal CBI and real CBI, the authors use the average actual term of office of the central bank governor as a proxy. The rationale behind this measure is that with a high replacement frequency, the government has more opportunities to place a person at the top of the central bank that endorses the governments monetary policies, provided that their influence on governor selection is sufficiently high. Furthermore, if the term of office of the central bank governor does not extend beyond the term of office of the elected government, this may deter the former from implementing long-term policies focused on price stability. However, Cukierman et al. (1992) warn that caution should be applied when interpreting a very low TOR, since it could imply the presence of either a very independent governor, or of a very subservient one.

4.2.1.3 Questionnaire on Central Bank independence

As an example, they offer the example of Argentina, where the governor resigns whenever the government changes, despite the legal provision of a 4-year term of office (pp. 362-363)

21 To complement the other two measures, Cukierman, Webb and Neyapti (1992) created a third variable based on a questionnaire sent to a number of representative specialists on monetary policy in the central banks of various countries. Respondants to this questionnaire had to answer questions on criteria similar to those used to compose the legal CBI indicator, but which focus entirely on the actual practice, thus providing another proxy for bridging the gap between statutes and reality.

4.2.2 Modelling the relationship

In order to explore a possible relationship between the different measures of CBI and average inflation, Cukierman, Webb and Neyapti (1992) used multivariate regression analysis with various combinations of the CBI indicators on data from 72 countries and 4 periods from 1950 until 1989. To avoid bias due to outlying countries with very high inflation, the authors employed the annual real depreciation rate D as relative, normalized measure:

(1) D = / +1 , resulting in a value between 0 and 1.

The transformed inflation rate D is then regressed on the various indices: (2) D= *LCBI + TOR + *QCBI + i + C

where LCBI is legal CBI, TOR is the turnover rate of the governor, QCBI is the CBI indicator based on the questionnaire, i is the intercept and C a number of control variables for one-time period effects.

4.2.3 Results and Interpretation

The regression on the legal CBI variable and the TOR for the entire sample yielded the following equation:

22 (3) D= -0.02 LCBI + 0.28 TOR + 0.09 + C (, and denote significance at the 10%, 5%, and 1% level, respectively).

Contrary to the expectations formulated in their hypothesis, and despite being negatively related to the transformed inflation rate, the coefficient of the legal CBI index is not significant. The coefficient of the TOR has the expected sign and is highly significant at the 1% level. Cukierman, Webb and Neyapti (1992) explain this result with the fact that the relatively high inflation rates of the developing nations raise the average inflation rate across the entire sample and thus render the comparatively small variations in the rates of the industrialized countries insignificant. A similar domination effect could explain the highly significant TOR. Figure 4 demonstrates this graphically.

(Source: Cukierman et al. 1992)

In each case, the upper line dominates the significance of the variable measured, with a more horizontal slope reflecting an insignificant measure.

To overcome this masking issue, the authors split the sample into subsamples of 21 developed and 51 developing countries, respectively. Regression with the developed countries only yielded the following result:

23 (4) D= -0.06 LCBI - 0.08 TOR + 0.09 + C

With a significance level of 5% and a negative sign, the results seem to confirm the hypothesis as far as the legal CBI index is concerned. The TOR index, however, shows the opposite of the expected sign, while being significant. Attributing this to the anomalous case of Iceland (very high inflation rate, extremely low TOR), Cukierman, Webb and Neyapti (1992) dropped this country to find that without it, TOR becomes insignificant and its coefficient positive. The following scatter plot (Figure 5) illustrates the negative relation between LCBI and inflation for two different periods:

(Graphic adapted from Walsh 1997)

As for the subsample of developing countries, the results show that TOR is is highly significant and positively related to D, whereas the legal CBI index is insignificant: (5) D= 0.01 LCBI + 0.28 TOR + 0.11 + C

The authors also tested the effect of the introduction of the questionnaire index compiled for a set of 22 countries, leading to the following regression equation:

24

(6) D= 0.16*LCBI + 0.57TOR + 0.46*QCBI + 0.27 + C

They conclude that with a 1% significance level and a negative sign, central bank independence approximated by the questionnaire index results in lower inflation, and the very significant TOR index shows that high governor turnover has an adverse effect on inflation.

In summary, Cukierman et al. (1992) found that industrialized countries tend to have a lower average rate of inflation when the corresponding central bank is relatively independent, measured according to a range of legal determinants. The turnover rate of the central bank governor seems to yield mixed results for that group of countries. If the central bank governor is replaced more often, the inflation rate tends to be lower, unless extremely anomalous nations are removed from the sample, in which case the turnover rate becomes insignificant for the industrialized countries. Regarding the developing countries examined in their study, they found that a higher turnover rate results in higher average inflation, and that central bank independence according to legal statutes is not an explanatory variable in that group of countries. The questionnaire index, although limited to a rather small subset of 22 countries, equally shows that, if used as a proxy,central bank independence reduces average inflation.

4.2.4 CBI and Inflation Variability

Because not only the level of inflation, but also its variation over time causes economic costs (Cukierman 1984), Cukierman et al.(1992) deemed it worthwhile shedding light on the question whether CBI also has an effect on inflation variability. With a slightly modified model using the standard deviation S as dependent variable in the model: (7) S = *LCBI + TOR + *QCBI + i + C

25 The results are similar to those obtained for average inflation. The legal CBI index is completely insignificant in the entire country set, whereas the TOR shows significance at the 1% level. As far as the subsample of industrial countries is concerned, the probability that the coefficient of the legal CBI index deviated randomly from zero is 10% but again shrinks to 5% when the outlier Iceland is removed from the list. In the case of the developing countries, the TOR has a highly significant (1%) influence on the variability of inflation, whereas the legal indicator does not appear to explain variations.

5. Central Bank independence and real macroeconomic variables

This section explains briefly the relation between central bank independence and economic performance in the areas of inflation, growth and investment, real interest rates and the accommodation of wage and employment. 5.1 Central Bank independence, GDP growth and investment

Grilli, Masciandaro, and Tabellini (1991) find that real growth and central bank independence do not present any relation in developed economies. Two years later it was also corroborated by Alesina and Summers (1993) and Cukierman, Kalaitzidakis, Summers, and Webb (1993).

Figures 6 reflects the relationship between central bank independence and the level of economic growth. Figure 6 shows that countries like Switzerland, which has an extremely independent central bank show slower growth while countries such as Sweden and United Kingdom shows the same level of economic performance. Another example is a comparison between Denmark and Japan both of which possesses the same degree of central bank independence but present an entirely different level of average real growth. In conclusion, as the figure shows, no relation between CBI and average real growth emerges.5

Analogous results are obtained if one uses growth of GNP per capita (Alesina and Summer,1993)

26

Source: Alesina and Summers (1993)

With respect to developing economies, there appears to be no link between legal independence and growth rate per capita income. However, it should be mentioned that independence of a central bank - taking into account political vulnerability of the central bank and related measures of turnover -has effectively an impact on the growth rate. As supporting evidence, using data from the 1960s to the 1980s, Cukierman et al. (1993) found that high political vulnerability of the central bank governor and related measures of turnover are effectively negatively associated with per capita growth.

Additionally, Cukierman, Kalaitzidakis, Summers, and Webb (1993) present a similar negative impact on the share of investment in GDP in some developing countries. These results support the hypothesis that private investment is lower under weak central bankers, reducing the long-run growth rate (Cukierman, 2006).

5.2 Central Bank independence and interest rates

27 Alesina and Summers (1993) find that neither nominal nor real rates of interest are negatively correlated with legal independence in developed economies. Figure 7 presents the relationship between central bank independence and real interest rates.

Source: Alesina and Summers (1993)

They also show that, on the one hand, the average real return to depositors was higher in developed economies when they presented higher levels of legal independence. On the other hand, in developing countries, the variability of both nominal and real deposit interest rates is positively associated with the turnover of central bank governors. (Alesina and Summers, 1993)

It should also be underlined that expansionary monetary policy may influence real rates in the short run; however a systematically expansionary monetary policy under a dependent central bank - operates to reduce average real rates over a longer period. In addition, as shown in Figure 8, there is a clear negative relationship between central bank independence and the variability of ex post real interest rates. (Alesina and Summers, 1993)

28

Source: Alesina and Summers (1993)

Based on these findings, the conclusion presented is that the variability of both reald and nominak real interest rate is lower, and the average real return to depositors is higher, in countries with higher levels of independence (Cukierman, 2006). 5.3 Central Bank independence, wages and employment As can be seen in Cukierman, Rodriguez, and Webb (1998) central banks of industrial economies, which possess higher levels of legal independence, accommodate nominal wage increases to a lower degree than in economies with lower central bank independence.

This finding in line with Rogoffs (1985) theory, which states that a more conservative and independent central bank accommodate wage increases to a lower degree than any flexible central bank.

With respect to unemployment, it also does not seem to be related to the level of central bank independence. This is related to the fact noted by Alesina and Summers (1993) who argue that the correlation between unemployment performance and real GNP growth performance is also low.

29 Figures 9 shows the analysis of unemployment in relation to central bank independence. Here, it can be observed that average unemployment is not closely related to the measures of central bank independence.

Source: Alesina and Summers (1993)

6. Criticisms of Central Bank Independence

Given the apparent unanimity of theory and early empirical studies regarding the favorable effect of CBI on inflation and its impact on institutional change in the past two decades, various critics raised concerns regarding the unconditional case for more independence. These criticisms range from doubts about the theoretical foundation of the CBI influence on inflation to objections regarding the reliability of the empirical findings. The following section will present a selection of these criticisms, along with a critical evaluation of their relevance to the importance of CBI for achieving price stability.

30 6.1 Methodological Issues

Numerous empirical studies seem to confirm the hypothesis that CBI is negatively related to price stability.6 However, it is commonly known in the scientific world that the confirmation or rejection of such a hypothesis highly depends on a number of (methodological) criteria and a failure to translate the theoretical hypothesis into a mathematical one is an immense source of bias. The results significantly depend on the selection of the data set, its transformation and aggregation, and on the mathematical model employed.

Forder (1999) critically analyzed whether this kind of methodological errors occurred in the most prominent studies. In the work of Cukierman, Webb and Neyapti (1992), for example, he noticed that when their analysis yielded a result that did not confirm the initial hypothesis, they split the country set into two subsets for developed and developing countries, respectively, which then resulted in the desired confirmation that (at least) in industrial countries, legal CBI is significant. Forder critizes this a posteriori data selection as being contrary to scientific research principles, because such a modification could be conducted infinitely until the desired outcome is reached. More importantly, he argues that Cukierman et al. (1992) mistakenly draw the conclusion that just because legal CBI is not a significant determinant of inflation in the industrial country group, and TOR not in the developing country group, these measures are not determinants of independence in their respective countries. Instead, so his argument goes, they have merely shown that these measures are not related to inflation in their respective countries, which does not imply that they are automatically bad measures for CBI.

However, there are two important points that invalidate Forders argument. First, there should be no restriction on using new pathes if during a study previous assumptions are proven wrong (as in this case the assumption that the effect of legal CBI is the same in both, developed and developing countries). Second, the authors
6

See Chapter 4.2

31 provided plausible reasons for why the initial model did not work7, and for the theoretical underpinning of deviations in the importance of legal CBI in both country groups.

Nevertheless, there are decisions in the analysis of Cukierman et al. (1992), which cannot be neglected as easily. Because of having Iceland in the list, the coefficient of the TOR is, contrary to the theory, negative and statistically significant. This by no means can be a reason to remove that country from the list, just to obtain the desired sign. Unlike the case of splitting the countries into two groups, Cukierman et al. (1992) do not provide a reasonable explanation except for the fact that Iceland is an anomalous case it can be assumed that if Iceland had been a positive outlier, deviating the overall data towards confirmation of the hypothesis, the authors would not have removed it from the list. Dropping that country from the data set must therefore be interpreted as discretionary modification without justification and must indeed lead one to question the validity of the results.

A similar point of criticism is brought up against Alesina (1988), who discards financial independence as the right measure for CBI, only because Parkin and Bade (1988) found no association whatsoever between financial independence and inflation. Once again, Forder stresses the importance of distinguishing between a measure that is not related to inflation because it is a bad measure which does not capture the empirical essence of CBI, and a measure that is not related to inflation simply because the there is no relation between independence and inflation. According to him, it was a flagrant methodological error to exclude a posteriori the financial independence as indicator for CBI, because it failed to show correlation with inflation. Researchers should verify a priori whether their measure captures the empirical notion of independence.

See Chapter 4.2.3

32 6.2 Subjectivity

Closely related to the methodological issues presented in the previous section is the issue of subjectivity. This type of bias comes in 3 major forms (criteria selection, criteria valuation and criteria weighing) [Eijffinger and Schaling 1993; Forder 1999; Maslowska 2007] and can significantly affect the outcome of a statistical analysis.

The first occasion for different measures and potentially different study outcomes is the selection of indicators for CBI. The following Table 1 from Maslowska (2007) perfectly illustrates how different the composition of the legal CBI index for the most prominent indices is.

(Source: Maslowska, 2007)

There are various reasons that might cause different value attribution. Forder (1999) proposes the closeness of the author to the Central Bank of a particular country, which could provide him with insights other authors do not have access to. He also suggests that the wish to confirm a hypothesis subconciously influences the author in his valuation.

33 Finally, it might be that different weighing is caused by different cultural backgrounds and resulting perceptions of what factors are important and which are not.

Given these numerous sources of subjectivity bias, it is surprising that the studies reach the same conclusion8: central bank independence is negatively related to average inflation and this relationship is significant. Forder (1999) warns that this apparent mutual confirmation is deceiving. He convincingly argues that although the different CBI indices are significantly related to inflation, these results might be scientifically useless, since the measures employed possibly do not reflect CBI. To support his thesis, he shows that once Germany and Switzerland, both countries whose central banks are widely accepted as being independent (Forder calls it conventional wisdom), are removed from the indices, the correlation coefficients shrink to astonishingly low values. Table 2 illustrates the impact of removing Germany and Switzerland on the correlation coefficient.

(A/PB denotes the index of Alesina (1988) based on the work of Parkin and Bade (1985), GMT the index of Grilli, Masciandaro and Tabellini (1991) and CWN the index of Cukierman, Webb and Neyapti (1992))

Given that the low correlation coefficients simply reflect the fact that the countries, if ranked according to their independence, show very different orders, Forder concludes that each of these rankings (measures) are little better than random (p. 32). The case of England illustrates the problem. Figure 10 shows how the rank (vertical scale) varies heavily depending on the legal index used (horizontal scale), onscillating between 1 and 10 of 12 countries.

This affirmation does not embrace all existing studies, instead, the most prominent and most often cited ones. Furthermore, taking the study of Cukierman et al. (1992) into account, this is limited to industrial countries.

34

(Source: Mangano, 1998)

Apparently, there appears to be no real consensus as to what countries are independent, so Forder concludes that the measures do not measure central bank independence, regardless of the fact that each of the measures is negatively related to inflation. The causal chain factual independence measured independence inflation could not be established.

The criticisms brought up by Forder are judged as being too extreme by Berger, de Haan, and Eijffinger (2002). They state that CBI is an unobservable concept, and therefore, any proxies created to represent CBI are necessarily noisy indicators. Therefore, they argue, researchers should not focus on just one index, as it is done by many of them, but instead rely on a mix of various CBI indices to test their hypotheses. However, Mangano (1998, p. 482) warns that any combination of unreliable measures, no matter how elaborate, is still an unreliable measure in itself. Thus, the subjectivity bias cannot effectively be overcome by combining subjective and inconsistent measures.

Moreover, Berger et al. (2002) question the reasonability of Forders (1999) approach of excluding Germany and Switzerland from the data to obtain more effective rank correlations between the different measures. They state that there is no reason to exclude these countries instead of other, possibly low-independence countries like Greece. However, they miss the point that unlike Greece, Germany and Switzerland are considered as highly independent across all the studies inves-

35 tigated. If they were to exclude Greece, they would have to exclude Portugal and New Zealand as well.

To overcome the issue of subjectivity, Forder suggests the development of a methodology of statute-interpretation, which would enable researchers to reach consistent approaches to measurement. Moreover, he proposes that the central banks be ranked anonymously, without letting the person who conducts the ranking know which bank belongs to which country.

These suggestions look quite appealing at the first glance, however, they could proof to be hard to turn into reality. Language is so complex that developing a framework for consistently interpreting statutes can be just as difficult as the valuation of raw statutes. As far as the anonymous ranking is concerned, it should be kept in mind that this would require a translation of the statutes into a language that can be understood by the person who is to rank the central banks. In this case, however, subjectivity could occur already during translation, rendering this approach ineffective. Forder therefore correctly concedes that it might be possible that the subjectivity issues cannot be satisfactorily overcome.

6.3 Robustness

6.3.1 Control Variables

The fact that the most prominent studies showing a negative relation between CBI and inflation, due to a lack of control variables, are also the most fragile ones in terms of statistical robustness, soon raised the question whether they were subject to omitted-variable bias. This is why subsequent studies on CBI tended to include other variables in their models, variables that might have such a big impact on inflation differentials that the CBI indices no longer would have a significant influence.

36 The results of these studies do not generally invalidate the impact of CBI, however, they are not unanimous either. Walsh (1997), for example finds that the CBI variable remains significant even when various control variables are included. The additional variables he includes are openness, the natural rate of unemployment, government budget deficit, and a dummy variable representing the degree of conservativeness of the government. The openness of a countries economy, measured by the share of imports in GDP, is included because Romer (1993) asserts that inflation-aversion increases when the importance of international economic exchange rises. This assertion is supported empirically. The natural rate of unemployment is one of Walshs control variables, because a higher rate could induce the government to stimulate economic growth through monetary expansion, which in turn leads to a higher inflation rate (Walsh 1997). Public deficit in terms of a share of GDP are included because Walsh expects the government of highly indebted countries to seek a reduction of these debts through devaluation. If this is the case, a positive relation should be expected. Finally, conservatism of the government is included because such a government is expected to support low-inflation policies, which reduces the importance of a conservative central banker. On the other hand, a similar study conducted by Sturm and de Haan (2001) found that the TOR index often is not a significant determinant of average inflation, and when it is, only when high inflation countries are included in the sample. The control variables used are openness, political instability, the log of GDP per capita, exchange rate regime, and debt-to-GDP ratio. The different results could be due to the use of different control variables and the use of an updated TOR dataset which contains a larger number of observations.

Two further studies that found no significant relation are those by Fuhrer (1997) and Campillo and Miron (1997), both dismissed on the grounds that they employ the legal CBI index on samples of developing countries, despite the fact that previous studies have shown that legal provisions do not have any significant impact on inflation in those countries, a result that was corroborated with convincing theoretical foundations (Sturm and de Haan 2001, Berger et al. 2002).

37

In another study, Klomp and de Haan (2010) investigated whether the contradictory findings of studies including control variables are due to heterogeneity among the countries examined. Using a random coefficient model with the HildrethHouck estimator, they found that CBI is only relevant in about 20% of the countries, leading them to the conclusion that specific conditions must be present in order for CBI to have a significant effect on inflation.

Continuing this suggested path, in a recent paper Markwardt and Hielscher (2011) investigated whether such an influencial precondition for the effectiveness of CBI is the quality of political institutions. The authors claim that the gain in monetary policy reputation, which, according to the model of Rogoff (1985), would reduce the inflationary bias, can only be attained if the institutional environment, represented by political stability, rule of law, and democratic accountability, conveys enough trust in the effectiveness of an independent central bank. Without high institutional quality, CBI would not have any beneficial effects on credibility, and consequently would not improve price stability.

To test their hypothesis, Markwardt and Hielscher (2011) use a regression model of the form (8):

where represents the inflation differential, 80/89,i the initial level of average inflation between 1980 and 1989, CBI the change in CBI, 0 a constant, an error term, and IQ the institutional quality. At the center of interest lies the effect of the interaction term IQ*CBI. An increase in CBI will only have a significant effect, if the factor IQ is sufficiently large.

They find that in case of developed countries, when control variables are included, not only CBI and CBI^2 are significant variables, but also the interaction term IQ*CBI. If political stability is used as a proxy to measure institutional quality, the term is significant at the 10%-level, if democratic accountability is used as proxy, the significance of IQ*CBI even reaches the 1%-level.

38 As for the developing countries, the results show that neither CBI alone, nor the interaction term is significant. Markwardt and Hielscher (2011) conclude that the beneficial effect of CBI on inflation is subject to two conditions: the change in CBI must be sufficiently large and the level of IQ must be high enough.

Markwardt and Hielscher (2011) provide an interesting explanation as to why CBI matters in some countries, and does not in others. Moreover, it provides an enhancement to the explanatory power of the legal CBI based on statutes and the term IQ*CBI could possibly be much closer to actual central bank independence than CBI alone. Nevertheless, a critical analysis of the study reveals some flaws. On the methodological side, it appears odd that after finding that the interaction term is not significant if control variables are omitted, and significant if they are included, the authors decide to rely on the model with control variables during the rest of their investigation, on the basis that it is the more robust model. However, a robustness test is a means of verifying that the variables analyzed remain significant even if other supposedly significant variables are included in the model. This implies that the analyzed variables should be significant without control variables, which is not the case in the model of Markwardt and Hielscher (2011).

On the theoretical side, it should be expected that aspects of institutional quality, such as rule of law, would improve the explanatory power of the legal CBI index in case of developing countries. The reason why legal CBI was never found to be a significant determinant of inflation was attributed to the fact that rule of law in not deeply embedded in those countries. Therefore, if rule of law is taken into account, legal CBI can be expected to be relevant in developing countries, an assumption that is not confirmed by the study of Markwardt and Hielscher.

6.3.2 Observation periods

Critics also voiced doubts regarding the consistency of the CBI inflation relation over different time periods. Berger et al. (2002) argue that it is perfectly normal to

39 find a weaker impact of CBI under a fixed exchange rate system than under a floating exchange rate system, a reasoning that is supported empirically by Walsh (1997). Nevertheless, they point out that there is some noticeable effect even in case of a fixed rate regime.

6.4 Causality

Already in the year 1995, Posen came forward with the idea that central bank independence might not be an exogenous variable. Instead of being the cause for a certain degree of inflation, CBI would be the result of a third variable, the latter also being the explanatory variable for inflation itself.

As a common source of both, CBI and inflation, Posen (1995) proposes an inflation-averse society, more precisely inflation-averse interest groups with the ability to effectively exert influence over central bank policy making and the political process to bring about CBI. Among several potential lobby groups like pensioners, importers and mortgage holders, while each of them presenting a strong interest in price stability, Posen singled out the financial sector as the most influential source of inflation opposition within a society. Not only does this lobby group feature the properties necessary for effective collective action, namely relatively low costs in relation to the potential gains, and easy monitoring of the members participation (Olsen, 1965), furthermore the central bank and the financial sector are quite naturally interwoven, e.g. through the sharing of a partly common labor pool and mutual supervision (Posen, 1995).

To test his hypothesis, Posen (1995) constructed a composite index for effective financial opposition to inflation (FOI) which draws on four different measures: the degree of universal banking in a country, regulatory influence of the central bank on the financial sector, fragmentation of a countrys political party system, and the degree of federalism.

40 A first hint towards such a relationship provides Figure 11, which shows the correlation between FOI and CBI, and FOI and average inflation, respectively, in OECD countries over the period 1950-1989.

Subsequently, regression analyses were conducted to determine the influence of FOI on legal CBI, FOI on inflation, and FOI and legal CBI on inflation, respectively.

The empirical findings are intriguing and appear to confirm Posens hypothesis that high FOI is the actual cause of low inflation. In detail the data reveals that FOI is a highly positive (at the 1 %-level) determinant of inflation, whereas legal CBI is not significant if used in the same regression.

41 Posen (1995) therefore concludes that both, a society committed to price stability and an independent central bank as vehicle to realize this preference are necessary to bring about low inflation. Neither of them alone would suffice.

Berger, de Haan and Eijffinger (2002) criticize Posens findings on the basis that the legal CBI indicator was employed for developing countries, despite the fact that earlier studies found it to be insignificant in that country group. However, Berger et al. (2002) mistakenly neglect the fact the country sample used by Posen only included low and medium inflation countries, thus excluding hyperinflationary countries for the reason that in case of steady hyperinflation, financial institutions might adapt their business practices to be consistent with high inflation, therefore eliminating their interest in low inflation and the need for an independent central bank to implement corresponding policies. Under this assumption, which Posen (1995) illustrates with the example of the Brazilian financial sector, which exhibits business practices highly focused on reaping profits in a high-inflation environment, there is no point to limit the validity of the legal CBI-argument to developed countries; instead, it provides a useful element to explain the lack of correlation between CBI and inflation in countries with high rates of inflation. Moreover, Berger et al. (2002) fail to recognize that legal CBI is not only insignificant in the overall country sample if FOI is included as explanatory variable, but also in the OECD-country sample.

Another study which apparently contradicts Posens (1995) claims is the one published by de Haan and Kooi in 2000. In this study, the authors find that CBI remains a significant explanatory variable, even if FOI is included as regressor. However, instead of using the legal CBI, the TOR index is used, and the influence of TOR only remains significant if high-inflation countries are included. This finding, although appearing to be contrary to it, is in line with Posens (1995) findings that CBI becomes insignificant in low and medium inflation countries. Moreover, the argument could be put forward that in hyper-inflation countries, TOR might be an endogenous variable explained by high levels of inflation and the resulting lack of trust in and dismissal of the central bank governor in office.

42 Hayo (1998) proposes a similar theory, under which the inflation-aversion of the society as a whole, instead of influential interest groups, determine the level of CBI and consequently the level of average inflation. In his historical-feedback approach, he argues that the (hyper-)inflationary history of a country makes its population more inflation-averse, and that positive experiences with CBI and its disinflationary effects increases the societys propensity to make the central bank even more independent. According to Hayo, this could also be a reason for why legal CBI indicators have not proven to be significant explanatory variables for price level increases in developing countries. Exogenous shocks in these nations might, due to a more rigid economic environment, result in higher real costs to a larger group of the population, which would lead to a lack of consensus regarding the need of monetary stabilisation policies. Legal CBI would therefore not necessarily lead to more actual independence.

Using data from public opinion polls in nine member states of the European Community between 1976 and 1993, he finds a strong correlation between public inflation-aversion and the average inflation-rate (see Figure xxx for an illustration of the correlation of a more recent data set which embraces 12 countries). Due to the lack of a sufficient amount of data, Hayo (1998) refrains from conducting a multivariate regression analysis using both, CBI and the indicator for inflationaversion.

43

(Figure 12 adapted from Hayo and Hefeker 2007)

Posen (1995) and Hayo (1998) present compelling arguments for the case that inflation-aversion influences the level of CBI and the level of average inflation. The idea that a nation must have an incentive to contain inflation is quite obvious, as Posen (1995, p. 254) puts it: If there are distributive consequences, there is no reason to assume that the adoption of CBI is self-enforcing. The preferences for price stability embodied by CBI require political support. If CBI does not embody such preferences, it will not affect inflation over the long run; if such prefer- ences were universally supported, independence would be unnecessary. Nevertheless, it remains unclear how much of the inflation-aversion is channeled through CBI, and how much of it affects inflation levels in other ways. The fact that Posen (1995) found legal CBI to be insignificant when FOI is included as control variable, while being significant when examined alone, allows the assumption that non-neglectable parts of the prime cause for low inflation are transmitted through other channels. These channels or preconditions could be found in other

44 institutional arrangements, as the study by Markwardt and Hielscher (2011) seems to confirm.9 Of the two theories, the one of Posen is the more convincing one, since the one of Hayo at least indirectly contradicts the theoretical assumptions postulated by Rogoff in 1985. According to Rogoff, CBI can reduce the inflation bias, if a conservative central banker is appointed, placing more emphasis on price stability than the rest of the society. In this case, however, there is no reason for the society to become supportive of monetary stabilization before legal CBI can have a significant impact on price stability. The theories become compatible again if the appointment of the conservative central banker is driven by a certain level of public desire for a stable prices. In any case, there is reason to believe that CBI is an exogenous, as well as an endogenous variable, and CBI is most likely neither a necessary, nor sufficient condition for price stability.

6.5 Disinflationary Credibility

The argument for central bank independence is mainly based on Rogoffs (1985) prediction that CBI would lead to more credibility as far as monetary policy announcements are concerned, which in turn would approach the wage setters expected rate of inflation to the real inflation rate, ultimately leading to lower economic costs in case of lower than expected money growth. According to Posen (1995b), this change in credibility and the resulting change in behavior of private agents constitutes an essential pillar of the theoretical construct of CBI and its supposed benefits. He therefore investigated whether the change in CBI really brings about the predicted credibility bonus for disinflationary monetary policy. Without the existence of this link, so he argues, the CBI argument would be deprived of its theoretical foundations.

see Chapter 6.3.1

45 Because disinflationary credibility is even more difficult to measure than CBI, Posen (1995b) examines four empirically measurable effects that are to be expected as the result of higher credibility. First, costs of disinflation should be lower in case of CBI, due to the fact that these costs are caused by the deviation of actual from expected inflation. If increased credibility due to CBI narrows the gap between actual and expected inflation, these costs must be lower. Second, nominal wage rigidity should be greater in case of CBI. Posen (1995b) argues that with higher credibility and consequently lower inflation and less uncertainty regarding inflation, wage setters would be less inclined to renegotiate wage contracts very often, as they incur costs whenever they have to renegotiate. The third effect is very similar but concerns the product market-price side. With higher CBI and more credibility, suppliers should tend to stick to a certain level of prices for a longer period of time. Posen (1995b) admits that too many other factors affect the prices of goods to be able to effectively measure the impact of credibility on it. However, using the New Keynesian approach from Ball, Mankiw and Romer (1988), which states that good prices are more rigid under low inflation (menu printing argument), Posen asserts more credibility comes along with higher disinflation costs. The final prediction regards the time span of the disinflation period. With credibility-inducing CBI, the central bank would make a bigger effort to correctly time the implementation of the previously announced target rate. At the same time, wage setters would renegotiate contracts to be in effect at the time the credibly announced target rate would set in. Thus, the duration of disinflation would be zero in case of a totally credible monetary regime10 (Posen [1995b] based on Fischer [1985]). To test each of the predictions, Posen mainly relied on OLS regression analysis using the data of 17 OECD countries for the decades between 1950 and 1989 and the CWN indicator.

10

Posen concedes that the duration would not be zero in reality, due to information imperfection. Nevertheless, a clear effect should still be noticeable. (p. 9)

46 To his surprise, Posen (1995b) finds that all four predictions are largely rejected, with only the fourth one regarding the disinflation time span showing mixed results, depending on which method for measuring the length of period is used. CBI does not seem to reduce inflationary costs (instead, it appears to be increasing them), reduce the length of the disinflation period, or increase nominal price rigidities. Even though a confirmation of two of the four predictions are subject to additional conditions, the fact these predictions do not hold lead to the conclusion that an increase in disflationary credibility due to CBI is not the mechanism that reduces inflation. Posen (1995b) therefore argues that either the link between CBI and inflation is missing, or the measures CBI indicator utilized does not capture actual CBI correctly. The former hypothesis could in part find corroboration in Posens (1995a) analysis of endogeneity. Credibility might not be the link assumed by the theoretical literature (Rogoff 1985), but instead both CBI and low inflation are the result of the financial sectors inflation-aversion11. The latter hypothesis is an equally suitable explanation for the failed confirmation of the CBI-credibility-inflation causal chain, given the existing criticisms in literature regarding CBI indicators12.

Posens (1995b) findings constitute a serious attack on the case for more CBI and its beneficial effects. Void of its theoretical foundations, the argument for making central banks more independent becomes weak, and even more so if the criticisms regarding the proper measurement of CBI is taken into account. With both, the theoretical and the empirical connection between CBI and inflation contested, there is little ammunition left for the proponents of more independence. Nevertheless, as is the case with other points of criticism, Posens (1995b) findings cannot be regarded as an absolute and unconditional refutation of established theories.

The author himself admits that at least two of his four predicted credibility effects depend on the validity of other theories. On the one hand, higher product-price

11 12

see Chapter 6.4 see Chapter 6.1 & 6.2

47 rigidity is based on the additional assumption that producers are less often inclined to adjust prices under low inflation than they would under higher inflation (Ball, Mankiw and Romer, 1988). On the other hand, the disinflation period will only be shorter as suggested by Fischer (1985), if monetary policy implementation and wage setting can be properly timed, which in reality is hard to assume if implementation lags and imperfect information are taken into account. Moreover, the higher disinflation costs due to nominal wage rigidity expected under higher credibility would only occur if a low inflation environment would really induce less frequent wage renegotiation, an assumption which is mainly based on negotiation costs and which neglects laws and traditional practices.

In addition, none of the predicted results are sufficient conditions, they are necessary conditions at best, provided that the additional assumptions hold. And again, the question of the correct measurement arises: throughout his paper, Posen (1995b) uses legal CBI measures, which themselves are already a major object of criticism. On top of this, the measures for disinflation periods and nominal wage rigidities are possible source of bias13.

Therefore, the theory of the credibility enhancing effect of CBI has neither been confirmed, nor refuted, even if Posens stresses the fact that all of his predictions remain unconfirmed after his analysis. Instead, another crack in the foundation of the CBI theory has been revealed.

6.6 Wage demands

According to the theory of CBI outlined in chapter 4.1, the inflationary bias arises because wage setters adjust their wage demands to expected levels of inflation. This implies, however, that wage setters are inflation-averse an assumption that is not self-evident (Hayo and Hefeker 2007).

13

Even though Posen (1995b) used two different measures of wage rigidity with comparable results, it cannot be said that either of these non-correlating measures are reliable indicators.

48 Berger, Hefeker and Schb (2004) claim that labor unions are only inflationaverse if the outside option (such as unemployment benefits) is measured in nominal terms. Another point of criticism regarding the behavior of wage setters is put forward by Guzzo and Velasco (1999), who argue that the appointment of a highly liberal central banker, as opposed to the conservative one suggested by the CBI theory, would discipline the labor unions in their wage depends, in fear of high inflation14.

6.7 Alternatives

The literature discussed so far in this paper focused on central bank independence as a means of achieving the goal of low inflation. However, there is no reason to believe that this is the only viable tool for attaining price stability without compromising economic growth. This section therefore briefly presents several alternative ways to reach stable price levels.

6.7.1 Forms of exchange rate fixation

Hayo and Hefeker (2007) suggest that some form of currency peg could be used instead of or in addition to CBI to obtain the desired level of price stability, arguing that with free capital mobility in connection with the value of the own currency fixed to that of a country with notoriety of price stability, domestic monetary policy would become impossible. However, Hayo and Hefeker claim that the potential benefits of such an arrangement for inflation can suffer from a lack of credibility, since currency fixes can and have been abandoned overnight, due to changes in policy or capital attacks on the currency in question. A credible fix should therefore come in form of a currency board or currency union.

14

This theory is valid only for monopolistic labor unions; for a refined analysis regarding different levels of labor union concentration, see Cukierman and Lippi (1999) and Lippi (2002, 2003)

49 Even though experience has shown that pegs are indeed seldom long-lasting arrangements (Obstfeld and Rogoff, 1995), there is no reason to conclude that this induces low credibility on monetary expansionary expectations in the way a dependent central bank would. In case of a dependent central bank, the inflationary bias is expected to occur because the monetary authority, through influence by government, has the objective and the tools to renege on previously announced inflation promises. In contrast, under a fixed exchange rate with free capital movements, there is practically no room for discretionary monetary policy, as the exchange rate mechanisms causes the price level to equilibrate. A very rigid fix, however, limits the domestic ability to respond to exogenous shocks to fiscal policy measures. In order to retain the possibility of monetary shock accommodation, a currency band is the appropriate monetary design. Forms of fixed exchange rate can therefore be seen as a viable tool for maintaining price stability and is used in a wide range of countries today, either with or without complementary CBI.

6.7.2 Inflation contracts

Another alternative suggested by Hayo and Hefeker (2007) is a form of contract that modifies the incentives of the central banker in the way proposed by Walsh (1995). Based on the principal-agent theory, Walsh argues that the suboptimal degree of shock-accommodation which arises if a conservative central banker is appointed, could be overcome if the central banker is provided with incentives such as monetary punishments that make him apply stabilization policies without engaging in systematic monetary expansion. Thus, an inflation contract could reduce the inflationary bias the way a conservative and independent central banker would. However, as Obstfeld and Rogoff (1996) point out, this solution has two major disadvantages. First, in order to provide the right incentives, the preferences function of the central banker would have to be found out, which could prove to be extremely difficult. Second, a major difficulty would be the definition and recognition of shocks that are to be mitigated via monetary expansion.

50

6.7.3 Inflation targets

Instead of granting goal independency to central banks, Hayo and Hefeker (2007) suggest that the central bank and the government should negotiate an inflation target, and that the central bank be held accountable in case the targets are not met. Despite the simplicity of such an arrangement, Hayo and Hefeker (2007) admit that the effectiveness depends on the correct interpretation of such an inflation target contract, and on the enforcement of the punitive measures in case of unmet targets. Moreover, in their paper, Hayo and Hefeker do not mention the necessity of having a conservative central banker in place, which is an important condition, since with a liberal central banker, inflation target negotiations would lead to higher rates.

7. Conclusion

During the past two decades, central banks around the globe experienced a significant shift towards more independency from political intervention, almost immediately following the publication of Rogoffs (1985) theoretical predictions and the empirical confirmation that central bank independence reduces average rates of inflation without compromising other goals such as growth and low rates of unemployment. Whether the surprisingly quick adaption of the CBI case by politicians can be attributed to the convictional power of Rogoffs theory, to the opportunity of delegating legal monetary policy authority in order to avoid blame from the electorate in case of high inflation, or even to the possible usefulness of CBI to signal price stability to creditors in order to reduce lending costs, the facts illustrated in this paper show that it is by no means established that CBI systematically reduces inflation rates.

51 On the empirical side, there appears to be no real consensus as to which countries central banks are actually independent. Consequently, studies proving the negative correlation between CBI and inflation are subject to subjectivity bias and have to be considered with the utmost care. On the theoretical side, criticisms outlined in this paper regarding causality raises doubts whether CBI really influences inflation, or whether CBI and inflation have a common source, with little or none direct interaction between them. And even if there is direct interaction between CBI and inflation, its magnitude seems to depend on additional factors, as recent research has shown. Even worse, not even the inflationary bias-reducing positive effect of CBI on disinflationary credibility could be confirmed. Given these and other issues, it is difficult to make a justified recommendation for more independent central banks. And even if there were no doubt about the theoretical impact of CBI on inflation, the way the independence of the highly independent ECB was compromised during the Greek bailout in 2010 makes it questionable whether actual independence could ever be achieved. Maybe the case for CBI has to be relativized to independent unless in case of emergency. But then, the problem of defining these cases of emergency arises. If alternatives to central bank independence are considered, such as the performance contract which is difficult to implement, or the exchange rate peg which limits the ability for monetary policy, it might be reasonable to implement CBI anyway, in the hope that the theoretical assumptions hold. After all, as CBI is relatively easy to implement and appears to have no negative impact on the economy, this might be a free lunch no economy will want to forgo.

52 References

Alesina, A., 1988. Macroeconomics and Politics. NBER Macroeconomics Annual. MIT Press, Cambridge, MA, pp. 13-61.

Alesina, A., and L. Summers (1993) Central Bank Independence and Macroeconomic Performance: Some Comparative Evidence. Journal of Money, Credit, and Banking 25(2): 15162.

Ball, L., Mankiw, N.G., Romer, D., 1988, The New Keynensian Economics and the Output-Inflation Tradeoff, Brookings Papers on Economic Activity, Spring

Barro, R. J., and D. B. Gordon (1983) A Positive Theory of Monetary Policy in a Natural Rate Model. Journal of Political Economy 91(2): 589610.

Berger H., Haan, J. de, Eijffinger S.C.W., 2002. Central Bank Independence: An update of theory and evidence, Journal of economic surveys Vol. 15, No. 1

Berger, H., Hefeker, C., Schb, R., 2004, Optimal Central Bank Conservativeness and Monopoly Labor Unions, IMF Staff papers 51, 585-605

Campillo, M. and Miron, J.A., 1997, Why does inflation differ across countries? In C. D. Romer and D.H. Romer, (eds), Reducing Inflation: Motivation and Strategy. Chicago: University of Chicago Press.

Cikowicz, P.,and A. Rzoca (2010) Inflation and corporate investment in selected OECD countries in the years 1960-2005 - an empirical analysis. Warsaw School of Economics

Cukierman, A., 1984. Inflation, Stagflation, Relative Prices, and Imperfect Information. Cambridge: Cambridge University Press.

Cukierman, A. (1992) Central Bank Strategy, Credibility, and Independence, Cambridge (Mass.): MIT Press.

53

Cukierman A, S. B. Webb, and B. Neyapti. (1992). Measuring the independence of central banks and its effect on policy outcomes. World Bank Economic Review, Vol. 6 No. 3, 353-398

Cukierman, A., (2008) Central bank independence and monetary policymaking institutions- past, present and future. European Journal of Political Economy 24, 722-736.

Cukierman, A., G. Miller, and B. Neyapti (2002) Central Bank Reform, Liberalization and Inflation in Transition Economies: An International Perspective, Journal of Monetary Economics 49, 237-264.

Cukierman, A., P. Kalaitzidakis, L. Summers, and S. Webb. (1993) Central Bank Independence, Growth, Investment, and Real Rates. Carnegie-Rochester Conference Series on Public Policy 39 (Autumn): 95145.

Cukierman, A., P. Rodriguez, and S. Webb (1998) Central Bank Autonomy and Exchange Rate Regimes: Their Effects on Monetary Accommodation and Activism. In Positive Political Economy: Theory and Evidence, edited by S. Eijffinger and H. Huizinga. Cambridge University Press.

Eij nger Sylvester C. W., and Erik Schaling. (1993), Central Bank Independence in Twelve Industrial Countries, Banca Nationale del Lavoro Quarterly Review, Vol. 184, pp. 49-89.

Fischer, S. and J. Huizinga (1982) Inflation, Unemployment and Public Opinion Polls, Journal of Money, Credit and Banking 14, 39-51.

Fischer, S., 1985, Contracts, Credibility, and Disinflation, in V. Argy and J. Nevile, (eds), Inflation and Unemployment: Theory, Experience, and Policymaking, London: George Allen & Unwin

54 Forder, James. (1999), Central Bank Independence: Reassessing the Measurements, Journal of Economic Issues, Vol. XXXIII, No. 1., pp. 23 - 40.

Fuhrer, J. C., 1997, Central bank independence and inflation targeting: monetary policy paradigms for the next millennium?, New England Economic Review, 1 2, 19-36

Grilli, V., Masciandaro, D., Tabellini, G., 1991. Political and monetary institutions and public financial policies in the industrial countries. Economic Policy 13, 342-392.

Guzzo, V., Velasco, A., 1999, The case for a populist central banker, European Economic Review 43, 1317-1344

Haan, J. de, Kooi, W., 2000, Does central bank independence really matter? New evidence for developing countries using a new indicator, Journal of Banking and Finance 24, 643-664

Hayo, B., 1998, Inflation Culture, Central Bank Independence and Price Stability, European Journal of Political Economy 14, 241-263

Hayo, B., Hefeker, C., 2007, Does central bank independence cause low inflation: A skeptical view

Hielscher, K., Markwardt, G., 2011, The Role of Political Institutions for the Effectiveness of Central Bank Independence, CESIFO Working paper no. 3396

Jacome, L. I., and F. Vazquez (2005) Any Link between Legal Central Bank Independence and Inflation? Evidence from Latin America and the Caribbean. Working paper 05/75. Washington: International Monetary Fund.

Klomp, J., de Haan, J., 2010, Central bank independence and inflation revisited, Public Choice (2010), Springer Verlag, 445-457

55 King, Mervyn. (2004). The Institutions Of Monetary Policy. American Economic Review, v94(2,May), 1-13.

King, Mervyn (2004). The Inflation-Targeting Debate. University of Chicago Pres. (p. 11 - 16)

Kydland, F. W. and E. C. Prescott (1977) Rules Rather than Discretion: The Inconsistency of the Optimal Plans, Journal of Political Economy 85, 473-491.

Mankiw, G. (2008) Principles of Economics. Mason, Ohio. 676- 685.

Markwardt, G. and Hielscher, K. (2011) The Role of Political Institutions for the Effectiveness of Central Bank Independence. Cesifo Working paper No. 3396

Maslowska, A., 2007, Discussion on the Inconsistency of Central Bank Independence Measures, Discussion Paper No. 21, Turun Kauppakorkeakoulu, Turun Yliopisto

Obstfeld, M. and K. Rogoff (1995) The Mirage of Fixed Exchange Rates, Journal of Economic Perspectives 9, 73-96

Olson, M. (1965). The logic of collective action. Cambridge, MA: Harvard University Press.

Parkin, M., Bade, R., 1988. Central bank laws and monetary policies. Unpublished manuscript. University of Western Ontario, London, Canada.

Persson, T. y Tabellini, G. (1993): Monetary and Fiscal Policy. MIT Press, Cambridge, Massachusetts.

Posen, A. 1995, Declarations are not enough: financial sector sources of central bank independence. In B. Bernanke and J. Rotemberg, (eds), NBER Macroeconomics Annual 1995, Cambridge MA: MIT Press.

56 Posen, A. 1995b, Central bank independence and disinflationary credibility: a missing link?, Working paper - Staff Reports, Federal Reserve Bank of New York

Riley, G. (2006). Consequences of Inflation. Eton College. Department of Economics and Politics. Eton, England.

Rogoff, K. (1985) The Optimal Degree of Commitment to an Intermediate Monetary Target, Quarterly Journal of Economics 100, 1169-1190.

Romer, David, Openness and Inflation: Theory and Evidence, Quarterly Journal of Economics, 108 (4), 1993, pp. 869903.

Sturm, J.-E., Haan, J. de, 2001, Inflation in developing countries: does central bank independence matter? New evidence based on a new data set, CCSO working papers: working paper series of the CCSO Center for Economic Research ; 3

Walsh, C. E. (1995) Optimal Contracts for Central Bankers, American Economic Review, 85 150-167

Walsh, C.E., 1997, Inflation and Central Bank Independence: Is Japan Really an Outlier?. Monetary and Economic Studies, May, 89-117

You might also like