The New Political Economy of Central Banking

Comment

Carluccio Bianchi

PAVIA UNIVERSITY, ITALY
 

There seems to be a widespread agreement today, both among academic economists and central bankers, that achieving and maintaining long run price stability should be the primary, if not the unique, goal of monetary policy. This consensus has been produced by theoretical developments and empirical experience together. At the first level, starting from the 1960s it has been recognized that there is no long run trade off between inflation and unemployment, at least at high level of price increases; furthermore popular dislike about inflation has been progressively accompanied by the professional recognition that the phenomenon has sizable economic and social costs; finally the actual powers of monetary policy to influence or even determine altogether the inflation rate have been acknowledged. From the empirical standpoint the experience of the so-called Great Inflation of the 1970s has led to an increasing awareness that high and persistent inflation cannot allow a better performance of the real economy: if a country ends up with no gains in output and only higher prices, then it seems reasonable to look for institutional devices capable of reducing structural inflation. Among such devices the most important seems to be the delegation of monetary policy to an independent Central Bank (CB) in the presumption that an effective causal link exists between monetary policy autonomy and low inflation.

But what are the ultimate causes and the alleged consequences of CB independence? This is the subject of Prof. Eijffinger’s paper, whose title may be slightly misleading if the reader expects to find a discussion of contemporary changes either in the theoretical literature or in the actual practice of Central Bankers with respect to monetary policy frameworks, strategies or behavior; rather Prof. Eijffinger’s contribution has the nature of a comprehensive critical survey of the most recent theoretical and empirical literature about the determinants and the degree of CB independence.

The paper is divided into two parts: the first one deals with the ultimate explanations, and the likely consequences at the same time, of CB independence; the second one is concerned with the possible determinants of the different degrees of monetary autonomy that one finds in the real world, together with a discussion of the empirical evidence purporting to assess the robustness of each hypothesis.

Obviously the first and most fundamental explanation for the experienced tendency towards CB independence is given by the already mentioned widespread belief that monetary delegation leads to lower long run basic inflation; other, partially related, adduced motivations are lower inflation variability, higher and perhaps more stable economic growth.

With regards to the fundamental argument concerning the relationship between CB independence and the rate of inflation, Prof. Eijffinger mentions three possible reasons backing the general presumption about the alleged causal link, namely:

i) the public choice view;
ii) the Sargent Wallace argument;
iii) the time inconsistency approach.

According to the public choice literature, Central Bankers are exposed to strong political pressures to adopt accommodating and lax monetary policies, because of budgetary and seigniorage considerations; thus a more independent CB will be less subject to political influences, promoting higher price stability (this line of argument calls for personal and policy independence).

The Sargent Wallace argument claims that if fiscal policy is dominant, monetary authorities will sooner or later be compelled to monetize the budget deficit, ultimately generating inflation; on the contrary a strong and independent CB can force the Government to take measures to reduce the deficit and/or the volume of outstanding debt (this argument calls instead for financial independence).

The most prominent justification for CB autonomy is however that one based on the time inconsistency literature. The substance of the argument is very well known so that just a brief sketch is needed. Policymakers seek to minimize a loss function whose arguments are inflation and output (or unemployment). Actual output is positively correlated to unanticipated inflation, since wages and prices move sluggishly over time owing to nominal rigidities. Thus policymakers have a structural temptation to surprise agents in order to reap the benefits of unexpected inflation. Rational agents however recognize the Government incentive to opportunistic behavior and incorporate it into their price expectations; the final outcome is then a suboptimal equilibrium where unemployment is always at its natural level but inflation is higher (thus setting the case for policy independence). The literature has outlined a number of possible solutions to the time inconsistency problem; among them the delegation of monetary policy to an independent authority with a longer time horizon and a greater preference for price stability will reduce the inflationary bias and improve society’s welfare. This is the substance of Rogoff’s (1985) well known contribution: the appointment of a conservative Central Banker will lead to a preferable outcome where both the level and the variance of inflation are lower. When supply shocks are allowed, there is a trade off between credibility and flexibility, since output variance is an increasing function of the conservativeness of monetary authorities; this result, however, only means that the optimal value of inflation aversion on the part of the CB is not infinite. As Prof. Eijffinger notes, in Rogoff’s model the CB cannot be overridden ex post when extreme productivity shocks occur. Lohmann (1992) however has extended the basic model by introducing a finite cost of overriding the CB, finding that an equilibrium exists where it is still optimal to delegate monetary policy to an independent institution assigning a high, but less than infinite, weight to inflation in its loss function.

The time inconsistency approach seems to provide the most attractive and persuasive rationale for CB independence. It must be recognized, however, that this model has been recently submitted to extensive criticism. Two years ago, for instance, McCallum (1995) argued that in a repeated game context Central Bankers come to realize the futility of opportunistic behavior, so that forward looking monetary authorities do not behave in that way. Last year, at the Jackson Hole Symposium, Merving King and John Taylor (1996) claimed that Central Bankers do not wish to push actual output beyond potential, thus depriving the inflationary bias of its theoretical foundation. Finally, and more recently, Alan Blinder (1997) put forward a similar argument, strengthened by his practical experience as a member of the Federal Open Market Committee (FOMC): Central Bankers know that they can push unemployment below the NAIRU level; but they also know that they can do it at the expense of higher inflation; and they are simply not ready to do it on the basis of a "responsible behavior" that can be met in many other situations in the real world. From an empirical standpoint, moreover, the time inconsistency approach seems to be weakened by the non-existence of the implied positive correlation between the NAIRU level and the size of the inflationary bias.

Prof. Eijffinger however is not willing to accept this mounting criticism. He shares many other economists’ opinion that the dynamic inconsistency argument captures the substance of the positive attitude towards inflation shown by politicians, Governments and the general public. Similarly he is eager to dismiss the reputational solutions to the time inconsistency dilemma on the ground that in the presence of asymmetric information there is no presumption that reputation may work for the CB (similarly the presence of uncertainty, together with the observation that credibility is not lost or gained in a single stroke, as in the Barro-Gordon model, work against the reputational models). Prof. Eijffinger also rejects the optimal contracting approach on the basis of the argument that social planners do not exist and that in any case a contract does not overcome the motivation for dynamic inconsistency, but merely relocates it. Anyway, Svensson (1995) has recently shown that the same results of an optimal contract can be achieved through delegation of authority to a Central Banker whose desired level of inflation is lower than society’s, even though it is admitted that it may be difficult or impossible to identify ex ante the potential CB’s preferences.

Monetary independence, besides leading to lower absolute inflation, should also produce less inflation variability. This result can be achieved thanks to a more stable money supply growth, triggered by a reduced dependence on electoral pressures (this is the substance of Alesina’s (1988) argument about the existence of a political business cycle). Friedman’s contribution (1977) adds another reason why Central Banks may reduce price variability: if there is a correlation between high and variable inflation, a commitment to price stability would also produce a lower inflation variance.

While the motivations for CB independence based upon the presumption of the existence of a link with inflation seem to be sufficiently agreed upon, more controversy exists around the contention that greater monetary autonomy also leads to an improved growth performance. Even from a theoretical viewpoint no definite, clear-cut conclusions have been reached: on the one hand we have the idea that low inflation, via the Mundell-Tobin effect, is associated with higher real interest rates that hinder growth; on the other hand, an environment of price stability should decrease the risk premium connected to high and variable inflation, thus stimulating economic development, which could also benefit from a reduction in uncertainty and related allocative costs. In a paper presented at last year’s Jackson Hole conference, Fischer (1996) argued that the inflation-growth link should be negative at high inflation rates, but positive at very low rates because of the existence of asymmetric price adjustments. Summers (1996) endorsed this view by adding the argument that the zero floor on nominal interest rates would produce longer recessions and exacerbate recovery problems. Empirical evidence on the inflation-growth relationship is inconclusive too: while there seems to be sufficient support for the thesis that very high inflation (such as 40% per year) hinders growth, no significant correlations emerge at normal and low levels of price increase.

The second part of Prof. Eijffinger’s paper is dedicated to the determinants of CB independence. After remarking that the degree of monetary autonomy varies widely among OECD countries, he seeks for the possible explanations of such different institutional arrangements and surveys the empirical evidence purporting to test the validity of the proposed explanations. Altogether seven possible arguments seem relevant to the author, namely:

1) the natural rate of unemployment;
2) the size of Government debt;
3) political instability;
4) supervision of financial institutions;
5) financial opposition to inflation (FOI);
6) public opposition to inflation (POI);
7) other residual determinants.

The first motive is linked to the time inconsistency approach. For any given output target, the higher the equilibrium or natural unemployment rate, the wider the inflation bias and therefore the stronger the incentive for an independent CB. Prof. Eijffinger recognizes, however, that empirical tests provide no support for this view. As mentioned beforehands, moreover, the most recent European experience seems at variance with this explanation, since high and persistent unemployment rates have been associated with a low and falling inflation.

A large Government debt implies high interest payments whose burden can be reduced by lower nominal interest rates associated with reduced inflationary expectations promoted by an independent CB. This argument, put forward by Cukierman (1994), has not found any empirical support too.

The link between CB independence and political instability is even not so clear at the theoretical level. On the one hand if politicians have a high chance of not being reelected, they have interest in delegation, but, on the other hand, incumbent parties have exactly opposite concerns. Cukierman (1992, 1994) tries to reach more definite conclusions either referring to the degree of national consensus or to the extent of political polarization. Empirical tests of this hypothesis are however inconclusive, since the various studies refer to different group of countries, use alternative measures of CB independence and adopt various proxies for the phenomenon of political instability.

With regards to the issue of financial supervision, Prof. Eijffinger reports the results of a study by Goodhart and Schoenmaker (1993) according to which, considering a sample of 26 mainly industrialized countries, in about half of them the function of supervising agency is combined with the responsibility for monetary policy, so that this motivation seems to have little impact on the subject of CB independence. Even from a theoretical viewpoint, anyway, arguments in favor of a separation, such as a possible conflict of interests between activities or the bad publicity associated with failures or rescue operations, coexist with opposed arguments such as the goal of a smooth operation of the payments system and the role of lender of last resort of the CB. As with the previously dismissed motivation for CB independence, empirical studies yield no uniform conclusion on this subject too.

A more promising line of research seems to characterize the alleged connection between CB independence and FOI. Prof. Eijffinger reports the results of two interesting studies by Posen (1993), according to which the supposed causal link between monetary autonomy and low inflation is illusory, since there is no noticeable cross countries correlation between measures of CB independence and inflation performance. Indeed Central Bankers will be prepared to take a strong anti-inflationary stance only if supported by the financial sector, so that it may be argued that high FOI generates both CB independence and low inflation as a joint product, without implying a causal link between these two variables. Empirical evidence, again, provides mixed results, even though there seems to be a well defined relationship between the size of financial markets and the degree of CB independence, which in a way would give some support to Posen’s theses. In a different context, Prof. Blinder’s recent discussion of his experience at the FOMC seems to endorse this line of thought, mainly when he declares that Central Bankers too often tend to pay too much attention to the mood or desires of financial markets and are continuously tempted to deliver the policy that the same markets expect or demand (1997, p. 15), despite the well recognized fact that financial markets’ time horizons are often short-sighted and their opinions highly volatile.

FOI is thus very important; this concept is of course narrower, but probably more influential, at least in the short run, than the more general phenomenon of POI, which is likely to be a structural determinant of CB independence strongly associated with a nation’s past experience about inflation itself. But while this long run association is indisputable, and empirically verified, in the medium run Eijffinger and Schaling (1995) find evidence of a correlation between society’s preferences for unemployment (versus inflation) and the degree of CB independence, thus yielding indirect support to the dynamic inconsistency approach.

Finally other possible determinants of monetary autonomy are briefly recalled in the terminal part of the paper: they may refer to the political characteristics of a given country, such as the existence of a system of checks and balances, or again to economic parameters and outcomes such as the amplitude of supply shocks or the slope of the Phillips curve. By using the time inconsistency approach again, Eijffinger and Schaling (1995) argue that a higher degree of CB independence should be supported by a lower variance of shocks and a steeper short run trade off between inflation and unemployment.

As a result of his critical survey, Prof. Eijffinger feels allowed to conclude that the most recent tendency towards CB independence is to be regarded positively, since this institution will lead to lower absolute inflation with no costs in term of real growth. He mentions however the existence of two possible caveats to this conclusion that make it less sound: the first one is the absence of any well defined correlation between CB independence and economic growth, together with the possible existence of high disinflation costs at low levels of inflation; the second one is the presence of a conflict, at least in the short run, between CB autonomy and accountability.

Prof. Eijffinger’s comprehensive survey raises a number of interesting issues worth to be discussed. Some points have already been pointed up in the previous discussion of the paper. Two arguments however seem to deserve a more accurate study and reflection. The first one has to do with the asserted link between CB independence and low inflation. While Prof. Eijffinger explicitly declares that independence is not a sufficient and/or a necessary condition for price stability, he is still eager to accept the generally admitted causal link. But, if one looks at the historical record and especially at the European experience, the paramount importance of institutional factors emerges quite clearly. As early as 1981 the Governor of the Bank of Italy precisely stated the necessary conditions for the so-called monetary constitution called for the task of achieving price stability. In the concluding remarks of his annual Report for that year he said, "The return to a stable currency requires a real change in the monetary constitution, involving the functions of the CB and the procedures for determining public expenditure and the distribution of income. The first condition is that the power to create money should be completely independent from the agents that determine expenditure ... The second condition is ... the existence of rules of procedure that subordinate major expenditure decisions to the need of monetary stability ... Public expenditure decisions must be made subject to rules that compel substantial respect for the need to match new expenditure by additional revenue .. In view of its economic and monetary effects, the immensity of interests involved and the power of representation and delegation exercised through it, collective bargaining at national and company level is a true economic force comparable to the mechanism by which public sector expenditure is determined; it is in equal need of criteria that prevent the drift towards monetary instability... [Thus] Central bank autonomy, reinforcement of budgetary procedures and a code for collective bargaining are prerequisites for the return to monetary stability" (1981, pp. 181-183).

More recently similar arguments have been repeated by the Governor of the Bank of France when explaining the reasons underlying the nation’s success in bringing down inflation, "...Monetary policy alone, of course, could not suffice. We also embarked on an appropriate fiscal policy over that time ... our tightening of fiscal policy mutually reinforced our monetary policy. In our case, and I think it is important to be reminded of this, we also attached great importance to revenue and wage policies. We had some sort of a priori vision of the wage developments, which has proven extremely useful in our disinflation process. I think it would be a mistake not to mention that as fully part of the process. On top of that we implemented a number of structural reforms of major importance, including price liberalization, exchange control liberalization, credit liberalization, privatization, and so forth. I think these structural reforms were the fourth policy instrument, which was undoubtedly key in that process" (1996, p. 332).

A second comment on Prof. Eijffinger’s paper bears on the crucial distinction between long run vs. short run strategies and behavior. While the long term commitment to price stability may be endorsed, there is still much disagreement, and research needs to be done, about the appropriate behavior of Central Banks in the most important short term perspective. In the short run a whole host of shocks may hit the economy: the theoretical literature prefers to concentrate on productivity shocks, but there are also price shocks, money velocity changes, exchange rates crises, stock markets bubbles and crashes, financial institutions difficulties and bankruptcies, and so on. All these call for a widening of the restricted formalizations of CB targets and constraints normally used in the literature. The real experience of a wide variety of shocks also leads to a presumption of a greater flexibility in the actual operation of monetary policy. After all the recent Bundesbank conduct and record show that, contrary to the hypotheses of the Barro-Gordon model, built up credibility is not lost at once if inflation is temporarily tolerated in the event of major unexpected shock that strike the real economy and need to be accommodated in some way. Moreover the existence of a short run trade off between inflation and unemployment and the possibility of a long run trade off at low inflation rates suggests further caution towards the idea of achieving absolute price stability or reducing flexibility. In this context topics like opportunistic deflation or the effects of the adoption of a practical rule like Taylor’s (1993) deserve to be discussed and thoroughly investigated. From an institutional point of view, another useful topic for discussion seems to be whether CB independence should be most effectively guaranteed by the appointment of a single Governor or by recourse to a Board. A collective decision may grant more independence, since the Board can resist political pressures in a stronger way.

Despite the great progress made by economic theory, given the real world problems and the constraints to the actual conduct of monetary policy, especially in open economies with a commitment to fixed exchange rates, it might be wondered whether Lawrence Summers is not ultimately right in arguing that, "Even if price stability is the overriding objective of Central Banks, there are considerations bearing on long-run economic growth performance that may call for policy responses not captured by simple rules. There is, in the end, no substitute for wise discretion." (1996, p. 43).
 

References