"The main difference between a scholar and a practitioner is the former's acquired ability to take delight in discursive and complicated errors"
(Anatole France).

"Economics is not nearly as much of a science as the free use of allegedly accurate figures would seem to indicate"
(Oscar Morgenstern).

"Practical men, who believe themselves to be quite exempt from any intellectual influences, are usually the slaves of some defunct economist"
(John Maynard Keynes).

"Hence no one can claim to have acquired practical knowledge in a particular discipline and at the same time despise theory without revealing that he is an ignoramus in his field"
(Immanuel Kant).
 

Monetary Theory as a Basis for Monetary Policy: Reflections of a Central Banker*)

Otmar Issing

Member of the Board of the Deutsche Bundesbank

 
1. The relationship between theory and practice

This conference is considering the relationship between monetary theory and monetary policy. Even if that particular relationship is of a rather special nature, I would like to begin with some more general observations. In considering the general relationship between theory and practice, one of course runs the risk of sliding into the vastness of philosophical speculation, for philosophers have been pondering this question for more than 2000 years, which takes us all the way back to Aristotle and his system of metaphysics.

Do such philosophical ponderings have any bearing on the issue we are examining? Well, consider the following postulation: monetary policy, just like other fields of human activity, always involves (in part) a mismatch between accumulated learning and the need to take specific action. In other words, the necessity to act invariably exceeds the measure of our knowledge. But in the field of monetary policy, in particular, it is not easy to define "action". Let me put it another way: if a "passive" monetary policy stance is called for rather than an "activist" approach, it is very hard to decide what such a policy should look like. The level of interest rates set by the central bank permits no direct conclusions. For example, the Bundesbank has not altered the discount and lombard rates since April 1996, and the securities repurchase rate has similarly been held at 3.0 % since August 1996 (final draft of Sept. 1997). But the effect which this constellation of central bank interest rates has had on the economy has changed substantially during this period, not least owing to the depreciation of the D-Mark. In this context it is not surprising that there is almost always disagreement about the monetary policy stance: how tight - how ease, or even, tight or ease?

That inevitably creates a certain gulf between theoreticians and practitioners. Depending on his point of view, the practitioner or policy maker - in this case the central banker - may regard the lack of certain knowledge either as a welcome measure of freedom from the constraints of theory or as an overbearing dilemma. Jörg Niehans put this problem in a nutshell (Niehans 1978, p. 294):

"... economics should be under no illusion that central banking will ever become a science. Academic critics love to chide central bankers for their lack of a fully articulated doctrine of monetary policy, based on testable - and perhaps even tested - hypotheses. These critics mistake central bankers for what they are themselves, namely teachers and intellectuals. In fact, a good central banker is a doer and a politician, for whom even ambiguity and inconsistency may sometimes serve his purposes. His field of action is the ever-changing stream of economic history, where every day may pose new problems requiring new solutions ... This treatise may thus end on a role of humility: However far monetary theory may progress, central banking is likely to remain an art."

In the following, I shall first put forward arguments in favour of Niehans' hypothesis. On the other hand, we must guard against placing exaggerated emphasis on the importance of "praxis". In this context, questions concerning the organisation of the decision-making process and of institutional arrangements also need to be discussed.
 

2. Theory and empirics: uncertain basis?

Anyone looking for differences of opinion among economic theorists as a justification for rejecting academic knowledge as a rigorous guideline for action does not need to look far. That is true of macroeconomics in general and of monetary theory in particular. One recent example was the NAIRU symposium as reported in the Journal of Economic Perspectives, winter 1997. (And the debate in the Journal does not even cover the full range of opinion).

For a time it seemed that there was a certain consensus, which was termed the "neoclassic Keynesian synthesis". This consensus, which from the outset was built on shaky foundations, is now shattered. Macroeconomics is once again split into various schools. E. Phelps (1990) counts as many as seven different ones, which is a greater plurality than ever.

The narrower field of monetary theory likewise yields extremely contentious answers and leaves key questions unanswered (Hellwig 1993); even today the discipline can "at best be regarded as scaffolding and not as the building" (Hahn 1982, p. 106). I should add, however, that not all unresolved questions which are of great theoretical interest are also directly of great importance for central bank policy.

A more serious concern is the unsatisfactory nature of the empirical studies. One need only think of the works devoted to money demand, which continue to provide a wide array of results. The task of trying to evaluate such studies objectively for the purposes of central bank policy is rendered all the more difficult by the fact that, in many cases, a glance at the author's name suffices to predict what findings he will come to (see Cooley and LeRoy 1981). It’s an open question to me if this period is forever behind us, as some seem to argue.

Over and above such differences of opinion, economic science and economic policy alike face general problems of their own. Economic science seeks to analyse a subject which is in a constant state of flux. Inevitably, therefore, it has a less tenuous grasp of up-to-the-minute economic reality, and is not infrequently plain wrong.

Let me point out just two particular difficulties. There may be considerable time lags before new insights are sufficiently widely disseminated among and accepted by economists (see Gans and Sheperd 1994). Secondly, we all know - at the latest since the publication of Thomas Kuhn's theses on the paradigm shift - how powerful the innate resistance to innovation is within the scientific process.

Besides this problem of cognition, economic theory and practice, in this case monetary theory and monetary policy, are confronted with the difficult problem of the interdependence between policy actions and the reactions of the general public. This problem is encompassed in broad terms in the Lucas Critique (Lucas 1976). Policy makers must pay particular attention to this interlinkage in respect of the financial markets, with their growing "expectation bias".

In a way, Goodhart's Law is an example of a specific application of the Lucas Critique. It postulates "that any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes" (Goodhart 1975, p. 5). If this theorem were generally valid, which Goodhart himself definitely does not claim, central bank policy would be faced by an additional and virtually insurmountable difficulty. The vicious circle posited by Goodhart's Law can under certain preconditions be broken in the case of monetary targeting. By pursuing a steadfastly stability-oriented policy, the central bank can establish an anchor for inflation expectations and positively influence the stability of money demand (Issing 1997).

I would just also like to mention the difficulties for practical monetary policy which arise from measuring problems in important areas. The Boskin Report reminded us of the difficulties associated with the traditional measurement of inflation rates - it is not our task today to consider whether it is on the right path to solving the problems. We are all aware of the substantial uncertainties that are involved in measuring GDP/GNP and of the time lags in recording. The same applies to estimates of the output gap. For monetary policy makers, the significance of uncertainty depends not least on the strategy chosen. For instance, the considerable time lag in the availability of data is undoubtedly one aspect that must be taken into consideration in opting for a GNP rule.

3. The gap between theory and reality, and the information gaps hampering practical policy

The discrepancy between economic theory and practical policy often stems from the highly abstract nature of the economic models, i.e. the fact that they are divorced from concrete reality. This is also one of the sources of the criticisms levelled by theorists at central bank policy.

Let me illustrate this with a typical example. Many publications claim that the central bank can control the monetary base at all times. Thus if the central bank abides by the philosophy of monetary targeting, any deviations from the predefined target path must surely, it is claimed, be due to targeting errors, i.e. they are an expression of a faulty policy.

In the context of the specific institutional framework in Germany, the following constellation applies. For the sake of simplicity, we shall disregard the question of minimum reserves. As the volume of discount lending is limited by quotas - and the quotas are virtually always fully utilized - this question focuses essentially on the Bundesbank's open market policy. In principle, this exclusively takes the form of securities repurchase agreements, which in recent years have invariably had a maturity of (around) two weeks.

If the Bundesbank wanted to stringently control the monetary base, it could achieve this fairly simply in principle. All the Bundesbank would need to do would be to make an independently determined amount of central bank money available each week via securities repurchase transactions. Setting the lombard rate at an exorbitant level could discourage banks from resorting to lombard borrowing.

Such a regime of stringently controlling the monetary base would entail extreme fluctuations in money market rates. Even this fact would probably fail to make some theoreticians change their mind - they would point to learning effects etc. and say that fluctuations in money market rates are simply a consequence of controlling the money stock and have to be put up with. The extent to which interest rate fluctuations should be tolerated may be an interesting topic of debate. But what the theorists more or less totally ignore are the purely "technical" problems associated with targeting the monetary base.

These technical problems include abrupt changes in the liquidity position of the public sector. Substantial temporary fluctuations may also arise from deposits which international organisations maintain at the Bundesbank for very short periods. But it is not least unpredictable fluctuations in the float which lead to short-term changes in the banking system's need for central bank money.

The greatest impact, however, is attributable to changes in currency in circulation. Although the seasonal pattern is broadly predictable, there are always some surprises. Finally, the volume of D-Mark notes in circulation abroad constitutes a major uncertainty factor. This cannot be monitored directly, for obvious reasons. For a long time the Bundesbank had only vague information about the changes in D-Mark circulation outside Germany. Following a Bundesbank study (Seitz 1995), the Bundesbank now estimates that between 30 and 40 % of the stock of D-Mark notes are circulating abroad; that currently amounts to around DM 100 billion. We have no information at all on the short-term changes.

Any approach to controlling the stock of base money which fails to take account of such developments would give rise to interest rate fluctuations that are completely unproductive. Moreover, a purely random raising or lowering of money market rates might be wrongly interpreted as "signalling" changes in the monetary policy stance. All in all, the repercussions on the financial markets - which are closely integrated with one another worldwide - would be extremely detrimental.

One can discuss the appropriate response to such influences at length. But the point I want to make is that theoreticians in general do not bother to concern themselves with such trivial "technicalities". That means, however, that criticism of central bank policy may be based on a system of references which does not accord with reality.

For rule-based theoretical approaches, these considerations need to embrace the entire gamut of institutional arrangements, which for the central bank are generally an exogenous factor. These range from the set of instruments stipulated by law, the structure of the banking system and of the entire financial system to the question of external safeguards (foreign exchange controls, exchange rate regime).

The discrepancy I have just described applies above all to rule-based approaches. However, discretionary monetary policy is in no way better placed. There is a huge gap between theory and reality in the knowledge of the structure of the economy and in the availability of all the necessary information.
 

4. Expectations, credibility and precommitment

The theory of rational expectations, and the discussion to which it gave rise, has had a major impact on monetary policy - in both theory and practice. Amid this illustrious gathering, I would not presume to retrace the theoretical debate. But I do think that, as far as practical monetary policy is concerned, one conclusion is incontrovertible: central banks which rely solely on their information lead and which ignore expectations and policy-induced changes in behaviour are doomed to fail.

And the learning process is by no means confined to the general public. Central banks, too, learn lessons from the past - at least they ought to. It is an interactive learning process. Central banks can influence this process positively through their policy, not least also through a corresponding information policy. This "teaching by doing" (King 1996) can ultimately help to minimize the output losses in the event of a disinflationary process and a subsequent safeguarding of price stability. (For the interaction between the measures taken by monetary policy makers and the learning curve of market players, see also Caplin and Leahy 1996).

The general dictum of game theory - "don't be too clever" - (Axelrod 1984, p. 120) has particular validity for monetary policy. On the one hand, globalized financial markets, with their high "expectation bias", could represent a great temptation for the central bank to manipulate the operations of market participants - through selective "signals" - for example. On the other hand, the rapid dissemination of information and the markets' ability to learn quickly would soon envelop the central bank in a tangle of contradictory developments. In such circumstances the central bank must take particular care not to further magnify the uncertainty prevailing in the markets but to stabilize expectations through giving a clear orientation. Regardless of what strategy the central bank pursues, it can only provide orientation sucessfully on the basis of public confidence in the stability of the currency.

Monetary theory has made important contributions in this field under the headings dynamic inconsistency, credibility and precommitment. (For a different assessment see Blinder 1997.) It is moot point whether, in this instance, the theorists have merely supplied the formal proof for insights which central bankers gained much earlier. "The game theoretic approach to policy makes it possible to model notions such as reputation and credibility that have long been staples of policymakers' own discussions of their actions" (Fischer 1988, p. 330). For my part, at least, I regard this development as the most important contribution made by monetary theory for a long time to a sustained stability-oriented monetary policy.
 

5. Monetary policy as an art - a carte blanche for ignoramuses?

The hope expressed at the end of the previous section presupposes that central bankers will in fact accept the role assigned to them by the more recent theoretical approaches. Niehans directs his call for humility at the economic theorists. But his dictum that, in the end, monetary policy is always an "art" could easily be misunderstood and could open the door to a regime in which the economic "practitioners" completely rule the roost (- certainly not his intention). Important insights gained by the theorists could then be crowded out, either because the decision makers fail to take note of them at all or because such insights do not fit in with their strategy. The desire of central banks for secrecy and ambiguity has been demonstrated e. g. by Cukierman and Meltzer (1986).

Karl Brunner pointed to this danger with particular insistence; the direction of his warning is thus diametrically opposed to that of Niehans (Brunner 1981). Brunner takes the view that central bankers have deliberately fashioned monetary policy into an "esoteric art". As a result, central banking is traditionally surrounded by a "peculiar and protective political mystique". This aura forms a veil behind which incompetence and ignorance can successfully be concealed.

The theory of public choice then added to this critique the presumption that central bankers by no means see themselves primarily as custodians of public welfare. Instead, they pursue their own interests in the first instance, just like other economic agents. The associated quest for prestige could easily come into conflict with the public interest. Furthermore, say proponents of the public choice theory, central bankers attempt to hinder a critical assessment of their own policy by incorporating ambiguities into their measures and by pointing the finger of blame at policy makers in other fields.

It would be surprising if central bankers, of all people, were immune to such pitfalls. The wish to avoid such dangers takes us to the heart of the debate for which H. Simons (1936) coined the classic formulation "rules versus authorities in monetary policy". Independence of the central bank - or, more especially, of its decision makers - would be completely the wrong answer, according to the advocates of rules. Instead, it was suggested, any risk of policy being determined by vested interests should be eliminated by committing policy makers to a strict rule.

With regard to this argument, one can say that the economic theorists have been unable (to date) to devise a rule that satisfies the requirements. To put it another way, there is no acceptable approach which could completely "automate" monetary policy once and for all on the basis of our present knowledge. The search for the optimal institutional arrangement continues. Economists have come up with many useful ideas. These range from feedback rules, the appointment of "conservative" central bankers to a penalty-carrying incentive-compatible mechanism to encourage achievement of the predefined objective.

The common feature of all these approaches is that they are difficult to put adequately into practice. In the light of the past experience of inflation, stagflation and disinflation, a debate is now been conducted worldwide not only on monetary policy strategies but also on the optimal institutional arrangements. Theoretical insights and practical experience intermingle in this debate. This is shown not least by the question of the central bank's independence. The worldwide trend towards granting independence to central banks, or to confirming this status where it already exists, undoubtedly also reflects the outcome of the scientific debate.

But practical experience has made a contribution, too. And I think the Deutsche Bundesbank can claim to have played a major role in that. Its statute served not least as a model for the wording of the relevant sections in the Maastricht Treaty. On the other hand, the arrangements for exchange rate policy and for the length of office of the members of the Central Bank Council - which are far from being fully satisfactory from the theorists' point of view - reflect resistance to solutions that are as far as possible "non-political".

The statute of the European System of Central Banks provides two key elements by stipulating the objective - the priority of price stability - and independence. But a considerable leeway remains. How will this be used for a monetary policy which is scientifically based? How great is the danger that, in the end, political influences (via the decision makers) will lead to wrong decisions?

Questions concerning the organisation of the decision-making process play a major role in this context. Notwithstanding all the associated difficulties, a body like the one envisaged for the European Central Bank is more likely to guarantee that objective arguments will prevail.

In this context, the call for the "transparency" of monetary policy is likewise of major importance. Economic theorists generally demand fully documented and analysable decisions. But if the Niehans argument can claim even a modicum of plausibility, this desire for transparency faces insurmountable barriers. That by no means frees the central bank from the obligation to explain its policy course continuously. Its paramount concern must be to convince the general public of the appropriateness of its policy. In terms of detail, purists may nevertheless still complain of a deficit of transparency.

 
6. Conclusions

These few, simple thoughts devoted to a major topic confirm the presumption suggested by the quotations with which I began. The relationship between monetary theory and monetary policy, between academic analysis and concrete practice, is and remains ambivalent.

Surveys of developments in macroeconomics have been marked for some time by eclecticism (Fischer 1988). That does little to alleviate the decision-making dilemma of the central banks as this merely boils down to the rejection of extreme approaches by common consent. Monetary policy practitioners expect concrete help from the economic theorists in selecting the approaches that are relevant for their purpose - depending on their specific central bank statute (final objective, instruments etc.) - i.e. above all consideration of the specific setting for the operation of monetary policy, such as the exchange rate, fiscal policy and the state of the labour market.

The less likely it is that sound concepts will be developed in these fields, the more important it is for a central bank which has a statutory mandate to ensure monetary stability to concentrate on the basic policy stance. In this respect economic theory seems to provide a broad degree of certainty: the link between output and inflation, or between fluctuations in output and the degree of variance of the rate of price increases, holds out the promise of monetary stability as a "free lunch". This consensus is deceptive, however. On the issue of short-run trade-offs (and their long-run consequences) controversy flares up immediately. And the current debate is dominated by the different answers to the question of what inflation rate should be regarded as being compatible with stability efforts (see Akerlof, Dickens and Perry 1996 and Feldstein 1996; for the Federal Republic of Germany see Tödter and Ziebarth 1997).

On the one hand, monetary theory is the basis for the entire activity of a central bank, from the choice of strategy to decisions on the deployment of the instruments. On the other hand, economic theory still leaves a "residual uncertainty". If it were able to supply complete certainty - given an exogenous objective - about which policy course is "correct" at any given time, we would have to appoint the most renowned theoretical expert to the board of governors of the central bank, and possibly make him the supreme and sole decision maker. I fear that such a suggestion might not meet with much support. Moreover, the objections raised by public choice theory would presumably apply equally to a "rocket scientist" imported from the academic world.

In weighing up alternative explanations and in assessing the measure of uncertainty in diagnosing both the situation and the effectiveness of the measures taken, a role will always be played by elements to which pure theory cannot give completely satisfactory answers.

But precisely this realisation demands modesty from monetary policy makers, too. They must take great care not to overestimate their own abilities. Milton Friedman's famous admonition to "avoid major mistakes" (Friedman 1968, p. 12) still holds. The statutorily defined objective must provide the bias which is binding for monetary policy measures - for example, the primacy of price stability.

A well formulated rule can deliver guidance for current policy measures. Within the framework of the Bundesbank’s monetary policy the monetary target has always played this role and restricted pure discretion. I think the different approaches to monetary policy in theory and practice can in short be best described by two quotations from leading experts in this field. „I believe in judgement within the framework of a rule." (A. Meltzer in a letter to the author.) - „Enlightened discretion is the rule." (Blinder 1996).

The crucial requirement, as hitherto, is an ongoing, fruitful dialogue between monetary theory and monetary policy. Monetary policy makers need the critical analysis of economic theorists. It would be very helpful if the theorists did not merely concentrate on the big picture but also devoted themselves frequently with the same diligence to analysing concrete decision-making situations and measures. Central banks, for their part, need to be open to this dialogue. Internally this means that they need to incorporate the results of analysis and research appropriately into their decision-making process. Externally central banks need to follow developments in monetary theory, digest them and assess their practical usefulness - without making themselves, to modify the Keynes metaphor, the slaves of some living economist. This „warning" does in no way imply any disrespect for the importance of monetary theory as the basis for monetary policy.

 

References
 

Back