# A critical assessment of measures of central bank independence.

I. INTRODUCTIONThe institutional relationship between central banks and government has been a subject of considerable interest ever since David Ricardo showed in the early nineteenth century how the close relationship between the Bank of England and government led to inflation and depreciation of the pound. As long as central banks adhered to the gold standard and eligibility rules for discounting, price stability would be achieved; however, as monetary systems shifted from commodity to fiduciary systems, the relationship between price stability and central bank institutional design became a more important policy issue especially during the past three decades. Financial liberalization, greater acceptance of long-run inflation as a monetary phenomena, and rejection of the long-run tradeoff between inflation and unemployment elevated the discussion of central bank institutional design and price stability. As a result establishing or extending central bank independence occurred in a wide number of countries: New Zealand (1989); Colombia, Italy, and Portugal (1992); Korea, Japan, and the United Kingdom (1997); and Sweden (1998). The 1992 Maastricht Treaty adapted formal independence as the foundation for the European Central Bank. The International Monetary Fund adapted independence as a policy goal and used the 1997 Asian Financial Crisis, especially in Korea, to encourage greater central bank independence for countries needing assistance.

Support for more independent central banks has in significant part been based on correlations between inflation and measures of central bank independence over time and across countries, frequently based on single-variable regression models. Measuring central bank independence and using the measures to predict monetary policy outcomes was first proposed by Bade and Parkin (1978, 1982, 1988); (1) extended by Cukierman (1992) as well as Cukierman, Webb, and Neyapti (1992); and has evolved into a substantial body of literature claiming their exists a statistically significant inverse relationship between inflation and measures of central bank independence. Klomp and de Haan (2010) provide perhaps the most extensive investigation to date. They combined 59 studies in a meta-regression analysis and concluded there is a negative and significant relationship between inflation and central bank independence for Organization for Economic Cooperation and Development countries.

Two studies typify the literature--the frequently cited study by Alesina and Summers (1993) and a study by Carlstrom and Fuerst (2009) updating Alesina and Summers (AS) using a comprehensive and recent set of measures of central bank independence developed by Fry et al. (2000). AS constructed measures of central bank independence for 16 industrial countries over the period from 1955 to 1988, estimated correlations between the average inflation rate and measures of independence over the period from 1955 to 1988, and concluded their results verified what others found: "a near perfect negative correlation between inflation and central bank independence" (1993, 154). Carlstrom and Fuerst (2009) extended the AS study: first, they estimated the relationship between the average inflation rate and Fry et al. (FJMRS) measures of independence for 26 industrial countries over the period from 1988 to 2000 and second, they estimated the relationship for the pooled AS and FJMRS measures of independence for a total of 42 industrial countries. Carlstrom and Fuerst concluded the relationship has remained stable over the entire period from 1955 to 2000 and 63% of the decline in the average inflation rate from the period from 1955 to 1988 period to the 1988 to 2000 period can be attributed to increased central bank independence. (2)

The measurement literature has been influential. The literature has supported efforts to enhance central bank independence and to quantify the conventional wisdom that more independent central banks generate better price stability outcomes. The acceptance of the measures of central bank independence and statistical results is manifested by their recent inclusion in textbooks (Hubbard 2008, 449; Mishkin 2006, 354; Wright and Quadrini 2009, 167) and their prominent place in Blinder's (1999, 57-58) well-known and frequently cited essay on central bank policy.

The measurement literature, however, has been questioned at two levels. First, a few researchers question the entire framework of measuring independence and statistical results and conclude they provide misinformation about the fundamental relationship between the central bank and government. They emphasize the difficulty of measuring central bank independence in the case of Japan and the United States as well as the predictive power of the estimated relationship for these two countries (Cargill 1989, 1995a); the lack of statistical robustness of the relationship between inflation and independence (Cargill 1995b; Campillo and Miron 1997; Fujiki 1996; Oatley 1999); and the direction of causation (Posen 1998). Second, others such as Pollard (1993) are less critical but nonetheless advise caution in interpreting the results because much of the literature focuses on de jure rather than de facto measures of independence; there are differences in the independence ranking of a few central banks, especially the Bank of Japan; and, like Posen (1998), there is a question as to the direction of causation implied by the correlations between central bank independence and inflation.

The concerns, however, have proved unconvincing for a variety of reasons: publication in nonmainstream outlets; focus on case studies regarded as outliers; large number of studies reporting stable relationships relative to the few that question the stability of the statistical relationships; and, the conventional wisdom central bank independence contributes to lower inflation rates. Recent studies like Carlstron and Fuerst (2009) and Klomp and de Haan (2010) indicate confidence in the measures and statistical results remains firm.

The measurement literature has made at least two important contributions: first, the measurement matrix used to establish a measure of independence for a specific country identifies important elements defining the institutional relationship between the central bank and government that influence monetary policy outcomes especially provided by Cukierman (1992) and second, the measurements when viewed from the perspective of ordinal rather than cardinal measures of independence highlight important institutional redesigns of central banks during the past several decades. However, the literature's focus on using the measures as cardinal variables suitable for statistical analysis and for drawing broad policy implications is problematic.

This article presents evidence that the information content of two well-known measures is not as great as claimed; the statistical robustness of the correlations claimed by Carlstrom and Fuerst (2009) is questionable; and most important, the measures incorporate important classification problems inherent in the measurement methodology that have not been fully appreciated in the literature in the case of Japan and, especially, the United States as recently revealed by Meltzer's detailed history of the Federal Reserve (2003 and 2009). The article argues the measures of independence are problematic and inappropriate for statistical analysis; rather, the measures can more usefully be considered as ordinal measures of independence to indicate the direction of major institutional redesigns of central banks.

II. THE INFORMATION CONTENT OF THE MEASURES OF INDEPENDENCE

Table 1 presents the AS and FJMRS measures of independence ordered from the most to the least independent central banks. Table 2 presents estimates of the standard model using the AS, FJMRS, and pooled data for average inflation with two representations of central bank independence:

(1) Inflation = [[beta].sub.0] + [[beta].sub.1]CBI.

Central bank independence is first represented by the measures in Table 1 and second by a dummy variable (Dummy 1) set to 1.0 for central banks with a measure equal to or greater than 90 bifurcating each sample into groups of more independent and less independent central banks. The dummy variable is set to 1.0 for three of the 16 central banks for the AS measures and for 10 of the 26 central banks for the FJMRS measures. In the pooled regression following Carlstrom and Fuerst (2009), Equation (1) includes dummy variables to account for shifts in [[beta].sub.0] (Dummy2) and [[beta].sub.1] (Dummy3).

The results in Table 2 using the explicit measures of central bank independence generate the standard result but should be considered in the context of regressions using the much simpler dummy variable representation of independence. Regressions with a dummy variable bifurcating the sample into more and less independent central banks generates the same results with about the same explanatory power measured by [R.sup.2]. As a group, the more independent central banks are associated with lower average inflation rates using the AS, FJMRS, and pooled data. The dummy variable results reaffirm the standard result; however, they also suggest the information content of the specific measures of independence is not great and can be approximated by a dummy variable.

However, the more important issue is that a review of the construction methodology, the AS and FJMRS measures of independence, suggest that estimation of the standard model for the FJMRS measures or the AS-FJMRS pooled measures is inappropriate. The AS measures of independence are a composite of the original Bade and Parkin (1982) measures for 12 countries extended to 16 countries and measures constructed by Grilli, Masciandaro, and Tabellini (1991). The AS measures are based on the method proposed by Bade and Parkin in that the measure for a given country is based on the chartering legislation and/or other legal documents that establishes the relationship between the central bank and the government. The AS measure for each country is constant over the period from 1955 to 1988 presumably because AS, Bade and Parkin, and Grilli, Masciandaro, and Tabellini concluded no significant legislative change in the institutional relationship between the central bank and the government occurred during the entire period. Assuming this is correct, it is then appropriate to regress the average inflation rate over the period from 1955 to 1988 on the average level of independence over the period from 1955 to 1988.

In contrast, the FJMRS measures of independence cannot be correlated with an average inflation rate over the period from 1988 to 2000 because they apply only to the last 3 years of the sample period by construction. Likewise, it is inappropriate to pool the measures with the AS measures of independence. The FJMRS measures are based on the central banks' assessment of their independence expressed in responses to a Bank of England survey of 114 central banks starting September 1998 and lasting over a year (Fry et al. 2000, Appendix 2 and 3). The FJMRS measures of independence rely on the central bank's interpretation of its independence for a specific period of time in contrast to a noncentral bank researcher's interpretation of the bank's charter over a long period of time. FJMRS based their measures of independence for 92 central banks (Fry et al. 2000, Table A.1) on the survey results, of which 26 industrial countries were selected by Carlstrom and Fuerst (2009). The survey questions regarding independence and accountability (Fry et al. 2000, 174-75) apply to the time of the survey; hence, the FJMRS measures of independence are representative for 1998, 1999, and 2000 but not necessarily representative for the entire period from 1988 to 2000.

The inherent problem in using the FJMRS measures without determining whether they are representative for the entire estimation period is revealed by a review of the recent history of the Bank of Japan and the Bank of Korea. The FJMRS independence measure of independence for these two banks significantly overstates their independence before 1998. The problem is not with the FJMRS measures but their use as a regressor explaining the average inflation rate over the period from 1988 to 2000.

The high index value of 93 for the Bank of Japan reflects the unprecedented redesign of the Bank of Japan in June 1997 that became effective on April l, 1998. Cargill, Ito, and Hutchison (2000) document the significant institutional redesign of the Bank of Japan in June 1997 in response to the Big Bang Announcement of the Hashimoto government in November 1996. The Big Bang Announcement set forth a policy to modernize Japan's financial system. The institutional redesign of the Bank of Japan meaningfully increased the Bank of Japan's formal independence and according to calculations made by Cargill, Ito, and Hutchison (2000, 109), the level of independence was meaningfully increased. Before 1998, the Bank of Japan was widely regarded as one of the most dependent central banks in the world.

The Bank of Korea experienced institutional similar redesign at the same time (Cargill 2001, 2010). Like the Bank of Japan, the Bank of Korea was a highly dependent central bank before 1998. A Presidential Commission for Financial Reform established in January 1997 released reports in April, June, and November of 1997 recommending a major redesign of Korea's financial system including a more independent and transparent Bank of Korea. The Commission's recommendation became the basis for major institutional redesign of Korea's financial and central banking institutions in response to the near collapse of the Korean economy in late 1997; the 1997 Asian Financial Crisis; and pressure from the International Monetary Fund to enhance the independence of the Bank of Korea as well as other reforms as a condition for securing a $58 billion line of credit. The Bank of Korea Act was revised in late 1997 and became effective from April 1, 1998. Hence, the value of 73 for Korea is appropriate only for 1998, 1999, and 2000 but significantly overstates independence of the Bank of Korea for the period from 1988 to 1997.

III. ROBUSTNESS OF THE STATISTICAL RESULTS

A few researchers have questioned the claimed statistical robustness of the standard results for measures of central bank independence constructed for the postwar period up to the late 1980s. This section investigates whether the statistical results reported by Carlstrom and Fuerst for the more recent FJMRS measures are as stable as claimed setting aside the issues raised above. The [[beta].sub.1] coefficient is negative and significant at the 1% level for all 26 countries (Table 2); however, this result is sensitive to the presence of the most independent central banks in the sample.

Equation (1) is estimated dropping the most independent central bank (25 observations), then estimated dropping the next most independent central bank (24 observations), then estimated dropping the next most independent central bank (23 observations), and so on until the ninth most independent central bank is dropped (16 observations). The independence coefficient is negative in all 10 regressions (Table 3); however, the index coefficient is insignificant at the 5% confidence level after dropping just three of the most independent central banks in the fourth regression. That is, the lack of a relationship between inflation and central bank independence cannot be rejected at the standard 5% confidence level after only three of the most independent central banks are removed from the estimation sample. The exercise was repeated using the dummy variable approach to representing central bank independence (Table 4) with no change in results.

IV. MISCLASSIFICATION OF CENTRAL BANKS

Most researchers admit the measures are not exact; however, the measurement problems are more serious than lack of precision and should question the prominence given to the measures of independence and their correlations with monetary policy outcomes. The problems fall into three categories: first, classifications of the Bank of Japan that cannot be supported by review of the Bank of Japan's charter and history; second, misclassification because of an overreliance on the central bank's charter or de jure independence rather than de facto independence; and third, varying classifications for a given central bank that reflect differing subjective interpretation of the central bank's charter.

A. Misclassification of the Bank of Japan Based on the De Jure Approach

The measures of independence place major importance on the central bank's legal operating framework, but in the case of Japan several well-known studies have misclassified the independence of the Bank of Japan.

The AS measure for the Bank of Japan is 62.5 placing the Bank of Japan as the fourth most independent central bank among the 16 country sample with only Germany (100), Switzerland (100), and the United States (87.5) ranked as more independent. The AS Japan measure is a combination of the Bade and Parkin measure which regards the Bank of Japan and the Federal Reserve as equally independent and the Grilli, Masciandaro, and Tabellini measure (3) that regards the Bank of Japan as possessing 50% of the independence of the Federal Reserve. The AS and Bade and Parkin measures of independence for Japan are difficult to justify considering the history of the Bank of Japan. Pollard (1993) and others have noted the differences in ranking the Bank of Japan's level of independence and correctly conclude a lower ranking is appropriate based on the Bank of Japan Law.

The 1942 Bank of Japan Law remained the operating legal framework for the Bank of Japan until the Bank of Japan Law was revised in June 1997. The originating charter of the Bank of Japan in 1882 and the two renewals of the charter in 1912 and 1942 rendered the Bank of Japan one of the most formally dependent central banks in the world. Article 42 of the 1942 Bank of Japan Law placed the Bank of Japan under the supervision of the Ministry of Finance while Article 43 gave the Ministry of Finance power to order the Bank to undertake any necessary business or to alter the By Laws as well as other necessary action. The Ministry of Finance could dismiss high-level officials at will and changes in the Bank's policy instruments involved meaningful Ministry of Finance involvement. The dependence of the Bank of Japan on the government is documented in Cargill (1989), Cargill and Royama (1988), and Cargill, Ito, and Hutchison (1997, 2000).

B. De Jure versus De Facto Independence

There is great potential for misclassifying a central bank's independence by placing too much reliance on de jure rather than de facto independence even when the de jure ranking is consistent with the central bank's legal framework. The distinction between de jure and de facto independence has long been recognized in the measurement literature (Bade and Parkin 1982; Cukierman 1992; Cukierman, Webb, and Neyapti 1992; Pollard 1993; and, most recently, Alpanda and Honig 2010); however, Mayer (1976) should be credited with being one of the first to point out that formal rules outlined in the central bank's charter at best can only approximate the relationship between the central bank and government and at worst, can be misleading. It is difficult to quantify de facto independence and available measures of de facto independence are not suitable for the standard regression models used in the measurement literature.

Cukierman (1992) and Cukierman, Webb, and Neyapti (1992) were the first to deal with the problem in two ways. First, the turnover rate of central bank leadership was utilized as a proxy for de facto independence; however, this is not a suitable proxy because low turnover rates may imply acquiescence to the government and high turnover rates may be because of unrelated factors such as scandals like those that occurred at the Bank of Japan in 1998. Second, de facto independence is measured by a survey of monetary policy specialists in 23 countries; however, the survey responses are subjective, the survey results are subject to small sample size problems and the resulting measures of independence represent limited time periods. (4)

The FJMRS measures provide more comprehensive and consistent de facto measures of independence than the Cukierman et al. de facto measures; however, while both measures provide a more accurate reflection of independence, neither is appropriate for regressions estimating the relationship between average inflation over long periods of time and de facto measures of independence because of inherent mismatch between representative time periods as discussed in Section II.

Two case studies illustrate the classification problems that arise from focusing on de jure independence. First, every published Bank of Japan measure of independence before 1997 is constant for the postwar period; however, a constant measure is inconsistent with the evolution of central banking in Japan. The de jure perspective assigns a constant measure to Japan because the 1942 Bank of Japan Law with some revisions made in 1949 remained the operating document of the Bank until the Bank of Japan Law was revised in 1997; however, the Bank of Japan actually achieved a meaningful but difficult to quantify increase in de facto independence after 1973. Cargill, Ito, and Hutchison (1997, 2000) explain how the Bank of Japan after the "wild inflation" period of the early 1970s was able to operate more independently from the Ministry of Finance despite the language in the 1942 Bank of Japan Law. This political independence was manifested by the decision of the Bank of Japan in May 1989 to commence increasing the discount rate from a low of 2.5% to ultimately 6.0% over about a 2-year period. The Ministry of Finance opposed this decision in private, and when the Bank of Japan raised the discount rate in May 1989 despite the Ministry's concerns the issue became public. It was unusual for disagreements between the Bank of Japan and the Ministry to become public. The Bank of Japan was acting as a far more de facto independent central bank while from a de jure perspective the Bank of Japan remained one of the world's most dependent central banks.

After the Bank of Japan achieved enhanced de jure independence in 1998 it shifted to de facto dependence in the first few years of the new century as it came under political pressure to adapt more aggressive monetary ease to reverse the deflation process. In anticipation of the end of Governor Masaru Hayami's term in March 2003, Prime Minister Koizumi made it clear the new Governor would have to be a "deflation" fighter (Cargill and Sakamoto 2008). The government made it further known privately and publicly that if the Bank of Japan did not adapt more aggressive policy its newfound independence would be compromised by a formal inflation target. One might argue that an inflation target is normally considered a point in favor of independence; however, in this case the Bank of Japan viewed such a rule as a weakening of its new-found formal independence. Formal independence did not insulate the Bank from political pressure and despite the increase in de jure independence, the Bank of Japan experienced a decline in de facto independence under the Koizumi and post-Koizumi administrations.

Second, the constant and high Federal Reserve measure of independence over much of the postwar period in the absence of major legislative changes (5) is an even more dubious assumption based on the recent history of the Federal Reserve provided by Meltzer (2003, 2009) and reviewed by Cargill (2011). Meltzer's history presents convincing evidence the Federal Reserve on several occasions abdicated its de facto independence to accommodate the government. The de jure independence of the Federal Reserve grossly overstates its de facto independence and its de facto independence has varied considerably. Meltzer presents evidence the newly independent Federal Reserve after the Accord of 1951 generated a stable macroeconomic environment in the 1950s because small government deficits under the Eisenhower and Kennedy administrations provided a favorable environment to pursue price stability. Under the more activist Johnson, Nixon, and Carter administrations, the Federal Reserve abdicated its de facto independence to work within government and coordinated monetary policy with fiscal policy that directly led to the Great Inflation from 1965 to 1985. Abrams (2006) likewise documents the influence on Federal Reserve policy by the Nixon administration. The Federal Reserve regained its de facto independence under Paul Volcker in 1979 and continued to function more or less as a de facto independent central bank through the end of Greenspan's tenure as Chair in 2006.

The role of the Federal Reserve in the Great Inflation in contrast to the postwar inflation record of the Bank of Japan also manifests the poor predictive power of the widely accepted regressions between inflation and measures of central bank independence. The comparative inflation records of Japan and the United States through the 1980s contradict the conventional wisdom that independent central banks generate lower inflation rates. Despite being de jure independent the Federal Reserve was responsible for the Great Inflation in part because of de facto dependence on government; yet, during this period every measure of independence ranked the Federal Reserve as one of the more independent central banks in the world during the Great Inflation period. In contrast, the Bank of Japan as one of the world's most de jure dependent central banks achieved an impressive price stability record from 1950 with the start of reindustrialization through the late 1980s (Cargill, Ito, and Hutchison 1997). The successful price stabilization policies of the Bank of Japan became widely recognized in the 1980s with some researchers referring to the Bank of Japan as a "model" or "credible" central bank in terms of pursing price stability (Hutchison 1987). The Federal Reserve has seldom been referred to in such terms. It is difficult to rationalize the widespread acceptance of the correlations between inflation and measured independence with their lack of predictive power for two of the more important central banks in the world.

Hence, efforts like Cukierman (1992); Cukierman, Webb, and Neyapti (1992); and Fry et al. (2000) to construct more meaningful measures of independence are commendable; however, the more accurate measures are not suitable for the standard regression framework widely used in the literature. Too much reliance has been based on de jure measures of independence in the measurement literature.

C. Lack of Measurement Consistency

One would expect some differences between the various measures of independence, but some of the differences appear large. The different rankings of the Bank of Japan have already been addressed but the same point can be made for other central banks. Table 5 compares the AS and the Cukierman, Webb, and Neyapti measures over several time periods. The simple correlation between the two measures is 0.81; however, this masks considerable variation in the measures of independence for almost a third of the 16 central banks. Adapting a 20% point difference between the two measures as "significant," there are 5 out of 16 countries for which there is a significant difference over the entire period or a significant difference for some of the subperiods: Belgium, Japan, Norway, Spain, and Switzerland. The approach to measuring independence and the subjective decisions made in determining the final measure suggest meaningful potential for misclassifying a central bank's independence. (6) In addition, note the AS measure is time invariant while Cukierman, Webb, and Neyapti vary the measure of independence even though both measures are based on codification of the central bank's de jure relationship with the government.

V. REPRESENTATIONS OF INDEPENDENCES ARE BETTER CONSIDERED AS BOARD ORDINAL MEASURES

The measures of central bank independence can be viewed either as ordinal or cardinal measures of independence. The above discussion suggests the measures as cardinal measures used in econometric modeling are problematic. De facto measures like Fry et al. (2000) are more accurate measures of independence but by construction are not suitable for establishing a relationship between average inflation over long periods of time and independence. The problems are not easily dismissed by the frequent claim they are only approximations to establish an overall average relationship between monetary policy outcomes and central bank independence. As ordinal measures, however, they provide interesting information about differences between central banks across countries and for a given central bank over time when the differences between central bank institutional design is large or the institutional change for a given central bank is significant. Three examples illustrate the useful application of the measures as ordinal indicators.

First, the measures of independence in Table 1 verify the magnitude of institutional redesign of central banking institutions that commenced in the 1980s and continued through the 1990s. The average central bank independence measure increased and the variance decreased for the 16 and 26 countries. The increase in independence and decline in variance is more prominent for the 16 countries. Major institutional redesign of central banks has occurred over the past three decades and as a result, central bank independence is now higher and differences between the independence of central banks have narrowed. The measures verify this result.

Second, Takahashi (1998) and Cargill, Ito, and Hutchison (2000, 109) estimated a measure of independence for the Bank of Japan using the Cukierman, Webb, and Neyapti matrix framework after the revised Bank of Japan Law became effective April 1, 1998. Both studies calculated the new measure at about 0.38 based on a scale from 0.0 to 1.0 compared to the 1989 Japan measure of 0.18 computed by Cukierman, Webb, and Neyapti (1992, 389). The change suggests a meaningful increase in de jure independence.

Third, in a similar fashion, Cargill (200l, 2010) constructed an independence measure for the Bank of Korea after the Bank of Korea Act was revised December 1997. The new measure of independence incorporating some additional changes in the Bank of Korea Law in 2003 is 0.36 compared to the 1989 measure of 0.27 assigned to the Bank of Korea by Cukierman, Webb, and Neyapti (1992, 389). The increase in de jure independence, however, is larger than suggested by comparison with Cukierman, Webb, and Neyapti. Their measures of Bank of Korea independence over the period from 1950 to 1989 covered in their study overstate the independence of the Bank of Korea compared to the Bank of Japan. The Bank of Japan measure of 0.18 is constant over the period from 1950 to 1989 while the Bank of Korea measure varies: 0.30 from 1950 to 1971, 0.32 from 1972 to 1979, and 0.27 from 1980 to 1989. The Bank of Korea measure should be no higher than the Bank of Japan and probably lower based on the postwar history of the Bank of Korea and the degree of administrative control over the Korean financial system by the government compared to Japan. (7) In any event, the revised measure of independence as well as the measure developed by Fry et al. (2000) verifies that the revised Bank of Korea Act meaningfully increased Bank of Korea independence from government in 1998.

VI. CONCLUDING COMMENT

Attempts to measure central bank independence have made important contributions to understanding the elements of de jure independence and, from an ordinal perspective, the measures provide insights into changes in the institutional design of central banks over time and across countries as illustrated for central banks in general, the Bank of Korea and the Bank of Japan. At the same time, the measures of independence are not as informative as claimed, the statistical results are not as robust as claimed, and more importantly, the potential for classification problems is large. This article focused on Korea, Japan, and the United States; however, researchers familiar with the monetary policy records and institutional design of other central banks might find similar misclassification problems. Overall, the contributions of the measurement literature are meaningful but the measurement literature has overemphasized the reliability of the measures, the statistical results based on the measures, and the policy implications drawn from the statistical results.

ABBREVIATIONS

AS: Alesina and Summers

FJMRS: Fry et al.

doi: 10.1111/j.1465-7295.2011.00427.x

REFERENCES

Abrams, B. A. "How Richard Nixon Pressured Arthur Burns: Evidence from the Nixon Tapes." Journal of Economic Perspectives, 20, 2006, 177-88.

Alesina, A., and L. H. Summers. "Central Bank Independence and Macroeconomic Performance: Some Comparative Evidence." Journal of Money, Credit and Banking, 25, 1993, 151-62.

Alpanda, S., and A. Honig. "Political Monetary Cycles and a De Facto Ranking of Central Bank Independence." Journal of International Money and Finance, 29, 2010, 1003-23.

Bade, R., and M. Parkin. "Central Bank Laws and Monetary Policy: A Preliminary Investigation," in The Australian Monetary System in the 1970s, edited by M. G. Porter. Victoria, Australia: Monash University, 1978.

--. "Central Bank Laws and Monetary Policy." Unpublished Manuscript, University of Western Ontario, 1982.

--. "Central Bank Laws and Monetary Policy." Unpublished Manuscript, University of Western Ontario, 1988. Accessed December 4, 2011. http://economics.uwo.ca/faculty/parkin/CentralBank Laws.pdf.

Blinder A. S. Central Banking in Theory and Practice. Cambridge, MA: The MIT Press, 1999.

Brumm, H. J. "Inflation and Central Bank Independence: Conventional Wisdom Redux." Journal of Money, Credit and Banking, 32(Pt 1), 2000, 807-19.

Burdekin, R., C. Wihlborg, and T. Willet. "A Monetary Constitution Case for an Independent European Central Bank." The World Economy, 15, 1992, 231-49.

Campillo, M., and J.A. Miron. "Why Does Inflation Diverge Across Countries'?" in Reducing Inflation: Motivation and Strategy, edited by C. D. Romer and D. H. Romer. Chicago: University of Chicago Press, 1997.

Cargill, T. F. Central Bank Independence and Regulatory Responsibilities: The Bank of Japan and the Federal Reserve. New York: Salomon Brothers Center for the Study of Financial Institutions, New York University, 1989.

--. "The Bank of Japan and Federal Reserve: An Essay on Central Bank Independence," in Monetarism and the Methodology of Economics: Essays in Honour of Thomas Mayer, edited by K. D. Hoover and S. M. Sheffrin. England: Edward Elgar Publishing Co., 1995a.

--. "The Statistical Association between Central Bank Independence and Inflation," Banca Nazionale del Lavoro Quarterly Review, 193, 1995b, 159-72.

--. "Central Bank Independence in Korea." Journal of the Korean Economy, 2, 2001, 1-33.

--. "The Bank of Korea in Historical and Comparative Perspective," in Academic Paper Series on Korea, Vol. 3, Korea Economic Institute and the Korea Institute of International Economic Policy, 2010.

--. "Meltzer's History of the Federal Reserve: A Review Essay." International Finance, 14, 2011, 183-207.

Cargill, T. F., T. Ito, and M. M. Hutchison. The Political Economy of Japanese Monetary Policy. Cambridge, MA: MIT Press, 1997.

--. Financial Policy and Central Banking in Japan. Cambridge, MA: MIT Press, 2000.

Cargill, T. F., and S. Royama. The Transition of Finance in Japan and the United States. Stanford, CA: Hoover Institution Press, 1988.

Cargill, T. F., and T. Sakamoto. Japan Since 1980. New York: Cambridge University Press, 2008.

Carlstrom, C. T., and T. S. Fuerst. "Central Bank Independence and Inflation: A Note." Economic Inquiry, 47, 2009, 182-86.

Cukierman, A. S. Central Bank Strategy, Credibility, and Independence: Theory and Evidence. Cambridge, MA: MIT Press, 1992.

Cukierman, A., S. Webb, and B. Neyapti. "Measuring the Independence of Central Banks and Its Effect on Policy Outcomes." World Bank Economic Review, 6(3), 1992, 353-98.

Eijffinger, S., and E. Schaling. "Central Bank Independence in Twelve Industrial Countries." Banca Nazionale del Lavoro Quarterly Review, 184, 1993, 1041.

Fry, M., D. Julius, L. Mahadeva, S. Roger, and G. Sterne. "Key Issues in the Choice of Monetary Policy Framework," in Monetary Policy Frameworks in a Global Context, edited by L. Mahadeva and G. Sterne. London: Routledge, 2000.

Fujiki, H. "Central Bank Independence Indexes in Economic Analysis: A Reappraisal." Monetary and Economic Studies, 14, 1996, 80-101.

Grilli, V., D. Masciandaro, and G. Tabellini. "Political and Monetary Institutions and Public Financial Policies in the Industrial Countries." Economic Policy, 13, 1991, 341-92.

Hubbard, G. R. Money, the Financial System, and the Economy. Boston: Pearson/Addison Wesley, 2008.

Hutchison, M. "A Model Central Bank?" FRBSF Weekly Letter, Federal Reserve Bank of San Francisco, 1987.

Klomp, J. G., and J. de Haan. "Inflation and Central Bank Independence: A Meta Regression Analysis." Journal of Economic Surveys, 24, 2010, 593-621.

Mayer, T. "Structure and Operations of the Federal Reserve System: Some Needed Reforms," Compendium of Papers Prepared for the Financial Institutions and the Nation's Economy Study. Committee on Banking, Currency and Housing, 94th Congress, Second Session, GPO, Washington, June 1976.

Meltzer, A. H. A History of the Federal Reserve, Volume 1:1913-1951, Chicago: University of Chicago Press, 2003.

--. A History of the Federal Reserve, Volume 2: 1951 1986, Chicago: University of Chicago Press, 2009.

Mishkin, F. S. The Economics of Money, Banking and Financial Markets. Boston: Pearson/Addison Wesley, 2006.

Oatley, T. "Central Bank Independence and Inflation: Corporatism, Partisanship, and Alternative Indices of Central Bank Independence." Public Choice, 98, 1999, 399-413.

Pollard, P. S. "Central Bank Independence and Economic Performance." Review. Federal Reserve Bank of St. Louis, 1993, 21-36.

Posen, A. "Do Better Institutions Make Better Policy'?" International Finance, 1, 1998, 173-205.

Seliger, B. "Review of Myung-Chang, Chung (ed.), The Bank of Korea: A History of Fifty Years, Bank of Korea, 2000." Korean Studies Review, (1), 2002. Accessed December 4, 2011. http:// koreaweb.ws/ks/ksr/ksr02-01.htm.

Takahashi, T. Kaisei Nichigin-ho no igi to konngo no kasdai (Significance and Challenges in the Future of the Revised Bank of Japan Law). Nissay Research Institute, Report, 1998 [In Japanese].

Wright, R.E., and V. Quadrini. "Money and Banking." 2011. Flatworld Knowledge. Accessed December 4, 2011. http://catalog.flatworldknowledge.com/ catalog/editions/619.

THOMAS F. CARGILL *

* T.F.C. is professor of economics, University of Nevada, Reno. He expresses appreciation to Hiroshi Fujiki, Federico Guerrero, Thomas Mayer, Allen Metzler, Michael Parkin, and Shigenori Shiratsuka for comments on an earlier version presented at the Bank of Japan, Institute for Monetary and Economic Studies, May 12, 2009. He also expresses appreciation to several referees of this journal. Any errors are the responsibility of the author.

Cargill: Department of Economics, University of Nevada, Reno, Reno, NV 89557-0206. Phone 775-849-1588, Fax 775-849-1586, E-mail tcargill1588@charter.net

(1.) Bade and Parkin (1982) developed a 1-4 scale independence ranking of 12 central banks based on their respective charters with independence-influencing attributes divided into Policy Type and Financial Type summarized in matrix form. Despite being frequently citied, the 1982 paper is not published and generally unavailable. A 1988 version of the 1982 paper is available at http://economics.uwo.ca/ faculty/parkin/.

(2.) The average inflation rate for 16 industrial countries declined from 5.58% to 2.73% over the 1955-1988 and 1988-2000 periods. The average inflation rate for 26 industrial countries over the period 1988-2000 was 3.30%. The variance of the inflation rate also declined. At the same time, the average index of central bank independence increased from 58.98 to 90.25 for the 16 countries on a scale of 0-100 while the index for the 26 countries averaged 83.35 over the period 1988-2000.

(3.) The Japan misclassification is also present in index developed by Eijffinger and Schaling (1993); however, like Grilli, Masciandaro, and Tabellini (1991), the Bank of Japan is assigned a low independence value by Cukierman, Webb, and Neyapti (1992) and Burdekin, Wihlborg, and Willett (1992).

(4.) Cukierman et al. report survey-based measures of independence for 23 developed and developing countries. Japan and the United States are not included. The de facto measures of independence reported by Cukierman, Webb, and Neyapti (1992, 368) represent 6 of the 16 countries in the AS measures and pertain to the 1980-1989 period.

(5.) There were legislative events impacting the Federal Reserve's relationship to government. Concurrent Resolution 133 in 1975 and the Full Employment and Growth Act of 1978 imposed monetary aggregate targets and other changes on the Federal Reserve; however, it is debatable whether these significantly impacted the de jure independence of the Federal Reserve.

(6.) Pollard (1993, 29) also notes the differences in rankings for a few countries, especially Japan.

(7.) During the period from 1950 to the mid-1980s the Bank of Korea policy was constrained by authoritarian governments who used central bank policy as part of a government-directed industrial policy. Policy conflicts on occasion resulted in dismissal of Bank of Korea governors (Seliger 2001). Price stability and prudential responsibilities took second place to government industrial objectives. In the 1980s government industrial policy began to unravel, Korea adapted financial liberalization and completed a transition from military rule to democratic government. Nonetheless, the Bank of Korea continued to be constrained by government policy objectives because of inertia from the past, continued to operate under administrative dominance by the Ministry of Finance and Economy, and suffered under the willingness of a growing class of elected politicians to use central bank policy to support client industries and sectors. Most would regard the Korean financial regime for much of the postwar period more rigid and administratively controlled than the Japanese financial regime.

TABLE 1 AS Measures of Independence and Average Inflation Rate over the Period from 1955 to 1988 and FJMRS Measures of Independence and Average Inflation Rate over the Period from 1988 to 2000 with Central Banks Ranked from Highest to Lowest Level of Independence AS Measures: 1955-1988 Country Measure Average Inflation Rate Germany 100 3 Switzerland 100 3.2 United States 87.5 4.1 Canada 62.5 4.5 Denmark 62.5 6.5 Japan 62.5 4.9 The Netherlands 62.5 4.2 Australia 50 6.4 Belgium 50 4.1 France 50 6.1 Norway 50 6.1 Sweden 50 6.1 United Kingdom 50 6.7 Italy 43.75 7.3 Spain 37.5 8.5 New Zealand 25 7.6 FJMRS Measures: 1988-2000 Country Measure Average Inflation Rate Sweden 97 1.65 Germany 96 2.41 Japan 93 1.1 United States 92 3.25 Canada 91 2.54 The Netherlands 91 2.41 Finland 91 2.7 France 90 2.01 Switzerland 90 2.27 Singapore 90 1.98 New Zealand 89 2.68 Denmark 88 2.44 Italy 88 4.14 Ireland 87 2.79 Greece 86 5.29 Portugal 85 6.36 Taiwan 85 2.73 Spain 80 4.35 Belgium 77 2.22 United Kingdom 77 3.98 Hong Kong 74 6.1 Australia 73 3.33 Korea 73 5.51 Austria 68 2.43 Iceland 59 6.17 Norway 57 2.85 Source: Carlstrom and Fuerst (2009, 183). TABLE 2 Regression Results for Two Representations of Central Bank Independence for the Periods: 1955-1988, 1988-2000, and 1955-2000 Estimation Coefficient Significance Period Variable Coefficient t-Statistic of Coefficient 1955-1988 Constant 9.44 * 13.64 0.00 Index -0.07 * -5.90 0.00 Constant 6.08 * 17.34 0.00 Dummy 1 -2.64 * -3.27 0.01 1988-2000 Constant 8.81 * 4.20 0.00 Index -0.07 * -2.65 0.01 Constant 3.96 * 12.84 0.00 Dummy 1 -1.73 * -3.47 0.00 1955-2000 Constant 9.44 * 10.30 0.00 Index -0.07 * -4.45 0.00 Dummy 2 -0.63 -0.30 0.77 Dummy 3 0.00 -0.03 0.98 Constant 5.94 * 18.03 0.00 Dummy 1 -1.92 * -3.69 0.00 Dummy 2 -1.31 -0.55 0.58 Dummy 3 -0.01 -0.25 0.80 Estimation [R.sup.2] Period Variable Adjusted Number Obs 1955-1988 Constant 0.69 16 Index Constant 0.39 16 Dummy 1 1988-2000 Constant 0.19 26 Index Constant 0.31 26 Dummy 1 1955-2000 Constant 0.60 42 Index Dummy 2 Dummy 3 Constant 0.55 42 Dummy 1 Dummy 2 Dummy 3 * Coefficient significant at .05 or lower. TABLE 3 Regression Results Based on the FJMRS Measures of Independence for 26 Industrial Countries, 1988-2000 Regression Countries in Coefficient Number Regression Variable Coefficient t-Statistic 1 1-26 Constant 8.81 * 4.20 Index -0.07 * -2.65 2 2-26 Constant 8.51 * 3.89 Index -0.06 * -2.37 3 3-26 Constant 8.46 * 3.67 Index -0.06 * -2.21 4 4-26 Constant 7.88 * 3.40 Index -0.05 -1.91 5 5-26 Constant 7.99 * 3.31 Index -0.06 -1.87 6 6-26 Constant 7.84 * 3.12 Index -0.05 -1.72 7 7-26 Constant 7.61 * 2.91 Index -0.05 -1.54 8 8-26 Constant 7.45 * 2.71 Index -0.05 -1.39 9 9-26 Constant 6.96 * 2.45 Index -0.04 -1.14 10 10-26 Constant 6.46 * 2.18 Index -0.03 -0.89 Regression Countries in Significance [R.sup.2] Number Regression Variable of Coefficient Adjusted 1 1-26 Constant 0.00 0.19 Index 0.02 2 2-26 Constant 0.00 0.16 Index 0.03 3 3-26 Constant 0.00 0.14 Index 0.04 4 4-26 Constant 0.00 0.11 Index 0.07 5 5-26 Constant 0.00 0.11 Index 0.08 6 6-26 Constant 0.01 0.09 Index 0.10 7 7-26 Constant 0.01 0.07 Index 0.14 8 8-26 Constant 0.02 0.05 Index 0.18 9 9-26 Constant 0.03 0.02 Index 0.27 10 10-26 Constant 0.05 -0.01 Index 0.39 Regression Countries in Number Regression Variable Number Obs 1 1-26 Constant 26 Index 2 2-26 Constant 25 Index 3 3-26 Constant 24 Index 4 4-26 Constant 23 Index 5 5-26 Constant 22 Index 6 6-26 Constant 21 Index 7 7-26 Constant 20 Index 8 8-26 Constant 19 Index 9 9-26 Constant 18 Index 10 10-26 Constant 17 Index * Coefficient significant at .05 or lower. TABLE 4 Regression Results for Inflation as a Function of a Dummy Defined by Different Levels of Independence Based on the FJMRS Measures of Independence Regression Dummy Set to 1.0 Coefficient Number for Countries Variable Coefficient t-Statistic 1 1-10 Constant 3.96 * 12.83 Dummy 1 -1.73 * -3.47 2 1-18 Constant 4.07 * 8.15 Dummy 1 -1.12 -1.87 3 1-23 Constant 3.82 * 4.41 Dummy 1 -0.59 -0.64 Regression Dummy Set to 1.0 Significance [R.sup.2] Number for Countries Variable of Coefficient Adjusted 1 1-10 Constant 0.00 0.31 Dummy 1 0.00 2 1-18 Constant 0.00 0.09 Dummy 1 0.07 3 1-23 Constant 0.00 -0.02 Dummy 1 0.53 Regression Dummy Set to 1.0 Number for Countries Variable Number Obs 1 1-10 Constant 26 Dummy 1 2 1-18 Constant 26 Dummy 1 3 1-23 Constant 26 Dummy 1 * Coefficient significant at .05 or lower. TABLE 5 Comparison of the AS and and Cukierman, Webb, and Neyapti Measures of Independence for 16 Industrial Countries, 1955-1988 Percentage Cukierman, Webb, Point Country Time Period and Neyapti AS Difference Australia 1955-1959 NA 50.00 NA 1960-1971 52.17 50.00 2.17 1972-1979 52.17 50.00 2.17 1980-1988 52.17 50.00 2.17 Belgium 1955-1959 21.74 50.00 -28.26 1960-1971 21.74 50.00 -28.26 1972-1979 24.64 50.00 -25.36 1980-1988 24.64 50.00 -25.36 Canada 1955-1959 65.22 62.50 2.72 1960-1971 65.22 62.50 2.72 1972-1979 65.22 62.50 2.72 1980-1988 65.22 62.50 2.72 Denmark 1955-1959 72.46 62.50 9.96 1960-1971 72.46 62.50 9.96 1972-1979 72.46 62.50 9.96 1980-1988 72.46 62.50 9.96 France 1955-1959 40.58 50.00 -9.42 1960-1971 52.17 50.00 2.17 1972-1979 34.78 50.00 -15.22 1980-1988 34.78 50.00 -15.22 Germany 1955-1959 100.00 100.00 0.00 1960-1971 100.00 100.00 0.00 1972-1979 100.00 100.00 0.00 1980-1988 100.00 100.00 0.00 Italy 1955-1959 36.23 43.75 -7.52 1960-1971 36.23 43.75 -7.52 1972-1979 36.23 43.75 -7.52 1980-1988 36.23 43.75 -7.52 Japan 1955-1959 26.09 62.50 -36.41 1960-1971 26.09 62.50 -36.41 1972-1979 26.09 62.50 -36.41 1980-1988 26.09 62.50 -36.41 Netherlands 1955-1959 60.87 62.50 -1.63 1960-1971 60.87 62.50 -1.63 1972-1979 60.87 62.50 -1.63 1980-1988 60.87 62.50 -1.63 New Zealand 1955-1959 26.09 25.00 1.09 1960-1971 34.78 25.00 9.78 1972-1979 34.78 25.00 9.78 1980-1988 34.78 25.00 9.78 Norway 1955-1959 28.99 50.00 -21.01 1960-1971 21.74 50.00 -28.26 1972-1979 24.64 50.00 -25.36 1980-1988 24.64 50.00 -25.36 Spain 1955-1959 18.84 37.50 -18.66 1960-1971 13.04 37.50 -24.46 1972-1979 13.04 37.50 -24.46 1980-1988 33.33 37.50 -4.17 Sweden 1955-1959 42.03 50.00 -7.97 1960-1971 42.03 50.00 -7.97 1972-1979 42.03 50.00 -7.97 1980-1988 42.03 50.00 -7.97 Switzerland 1955-1959 76.81 100.00 -23.19 1960-1971 76.81 100.00 -23.19 1972-1979 76.81 100.00 -23.19 1980-1988 92.75 100.00 -7.25 United Kingdom 1955-1959 30.43 50.00 -19.57 1960-1971 62.32 50.00 12.32 1972-1979 39.13 50.00 -10.87 1980-1988 39.13 50.00 -10.87 United States 1955-1959 69.57 87.50 -17.93 1960-1971 69.57 87.50 -17.93 1972-1979 69.57 87.50 -17.93 1980-1988 69.57 87.50 -17.93

Printer friendly Cite/link Email Feedback | |

Author: | Cargill, Thomas F. |
---|---|

Publication: | Economic Inquiry |

Article Type: | Report |

Geographic Code: | 9JAPA |

Date: | Jan 1, 2013 |

Words: | 8067 |

Previous Article: | Financial restraints and private investment: evidence from a nonstationary panel. |

Next Article: | Inflation dynamics and labor market specifications: a Bayesian dynamic stochastic general equilibrium approach for Japan's economy. |

Topics: |