RDP 8605: On Some Recent Developments in Monetary Economics 7. Policy in a Changing Economy

With innovation and financial deregulation changing the money stock – money income relationship, it is not surprising that there has been substantial modification or even suspension of the 1970s practice of monetary targeting. Australia suspended its M3 target in January 1985. Canada has done likewise. In the U.S. and the U.K., the monetary authorities have de-emphasised their money stock targets and resorted to a wider range of indicators.

This has forced a return to explicit recognition of the role of discretion in policy-making. Even in the U.K., where monetary policy rules are still stated, the role of discretion is explicit. The Governor of the Bank of England has noted: “the significance of a departure from monetary targets is that it establishes an important presumption of the need for policy action. This presumption is, more often than not, likely to be confirmed by careful examination of all the other available evidence … But it is only a presumption, and, where there is justification to override, it would be perverse and damaging to the economy not to do so.”[18] Chairman Volcker of the U.S. Fed has explained the current U.S. approach in similar terms.[19] In Australia the authorities now consider a “check list” of relevant indicators, including: a range of monetary aggregates; interest rates; the exchange rate; the external accounts; the current performance and outlook for the economy, including movements in asset prices; inflation, the outlook for inflation and market expectations about inflation. The signals from these indicators are balanced before a judgement is made whether policy should be tightened or loosened and by how much.[20]

These approaches formalise operationally the need for monetary policymakers to take into account the whole range of economic information. The use of a rigid monetary growth rule assumes that only the targeted monetary aggregate convey information which is useful as an indicator of policy. It has been argued that this is not the case, even when the money stock – money income relationship remains stable. One proponent of this view is B. Friedman (1977), who shows that a rigid monetary target can be informationally inefficient, even in a small aggregated model of a stable economic structure.[21]

When the relationship between the money stock and the ultimate targets of policy is changing, due to deregulation or whatever reason, the case for utilising additional information is strengthened. But this does not mean that there is no useful information in the monetary aggregates. Judgments about whether or not the monetary aggregates are growing too quickly will be harder than in a more stable regulatory framework but the information content of the aggregates will not be zero. They will have lost, if only temporarily, the special role that targeting implied.

There is, of course, plenty of room for debate about such issues as: the current and prospective state of each indicator; what the state of each indicator implies about the setting of monetary policy; how to weigh the influence of each indicator; and even more intangible matters such as the interaction amonmg policy instruments and the state of the economy.

The use of a broad range of information in the setting of monetary policy brings clearly into focus this interrelationship of the macroeconomic policy instruments to a much greater degree than did targeting. For example, a rate of inflation that is currently “too high” will point to the need for a firmer rather than an easier monetary policy.[22] But monetary policy is not the only influence on inflation. Consideration of prospects for inflation may indicate that a firmer monetary policy would be desirable; it may also suggest the need for action in other areas of policy, for example wages policy.

This point applies with special force with respect to the external indicators. What is the implication for the setting of monetary policy of a large, indeed unsustainably large, current account deficit? Tighter monetary policy may have the effect – at least over some horizon – of putting upward pressure on the current account deficit by producing a higher exchange rate. In this case, changes to other policies – for example fiscal policy – may more clearly be indicated. But a further question can arise, as in the United States in the early 1980s. What should be the desirable course for monetary policy if fiscal policy is not sufficiently firm? This raises questions of the policy mix, and generally speaking these are not questions which have received much attention in the profession.[23]

The approach to the making of monetary policy based on the systematic monitoring of a number of indicators, in addition to the monetary aggregates, does not represent a radical departure from that of the “targeting” regimes practised in the past. Central banks have always paid close attention to a wide range of economic indicators. Even those central banks which have paid closest attention to the achievement of pre-specified monetary targets have deviated from this approach when the non-monetary signals gave a sufficiently contrary reading.

Nor would the essential task of determining monetary policy vary much if the formulation of policy were based on other proximate indicators. There are proponents of a range of alternative possible “targets” for monetary policy – including the exchange rate (real or nominal), interest rates, even the growth of nominal GDP. Any single indicator can suffer from the same sorts of technical weaknesses as the practice of targeting a particular monetary aggregate.[24] Whatever the proximate targets of monetary policy, however, it is necessary to interpret performance in relation to this in the light of all the available evidence. At the end of the day, a whole “check list” of indicators will be consulted.

Recent developments have heightened debates about the formulation of monetary policy. They have also impacted upon, without clearly changing, many of the practical problems of monetary analysis. In particular there remains much uncertainty about the choice of indicators, about lags in response to changes in policy of credit demands and interest-sensitive items of spending, and so on. During the current transitional period, it may be difficult to isolate the effect of structural change from the effect of policy settings when looking at individual indicators. Figuring on the exact location of economic constraints remains rough – no one can quantify precisely and in advance the size of a sustainable external deficit or a sustainable rate of non-inflationary growth.

A possible objection to use of a “check list” approach is that it fails to provide a sufficient amount of discipline on the monetary authorities.[25] Those who believe this generally recognise the technical difficulties with strict “targeting” in a rapidly changing financial system. However, advocates of continued targeting usually conclude that the correct response is to have frequent changes of target, accompanied by (if necessary extensive) technical justification for the latest change. This is, in effect, the approach which has been used in the U.S.A. by the Federal Reserve Board. Some proponents of “rational expectations” would take issue with even this approach. They would argue that a fixed monetary rule should be maintained in the face of all shocks – private agents will determine the nature of the shocks and adjust their behaviour accordingly. It is important to remember that these conclusions are derived from models which depend crucially on a number of restrictive assumptions, not least of which are the absence of imperfections in labour and product markets and the absence of inside money. In models where information and markets are not perfect, monetary targets will not be optimal in the face of shocks from all sources.[26] A major increase in the demand for money, for example as a result of financial deregulation, should be accommodated.

This may be an area in which large macroeconometric models have a role to play, since they can be used to “fingerprint” shocks from different sources. The cost of building and maintaining models is high, and the approach is open to the objections noted above that the underlying structure may change too quickly for the model-builders to keep up.

Model specification issues aside, there is still the difficulty of reading the fingerprints – e.g. distinguishing financial shocks from real shocks. The authorities, however, are in a position to tackle such a task and it may be that it can be done at a reasonable cost.[27] Therefore, the discretion of the policymaker to alter the approach to policy, even only by changing targets, may be an efficient way for the economy as a whole to adjust to shifts in the sources of shocks; that is, to structural change in a general sense.

Whether or not to use monetary targets (however frequently changed) is, in our view, difficult to resolve in any definitive way. What is widely accepted is that monetary policy should be clearly explained as part of a stable and well understood general approach to economic policy. At least from the perspective of academic economics, this can be seen as the enduring residue of the rational expectations revolution.


Leigh-Pemberton (1985), p 534. [18]

Volcker (1985). [19]

This approach was explained by the Governor of the Reserve Bank in Johnston (1985). [20]

Similar arguments may be found in Burns (1980) and Waud (1972). [21]

Defining what is “too high” will present the usual difficulties. Some define this as anything above zero; others may make reference to what is deemed to be attainable in the period ahead or to the benchmark provided by inflation in major overseas countries. [22]

One Australian example is the work of Perkins (1982), which has advocated the explicit analysis of the policy mix on both a theoretical and a practical level. [23]

For example, “targeting” exchange rates or interest rates has, in the past, helped to produce cumulative inflations or deflations. Targeting nominal GDP suffers from the problem that its “controllability” is less certain than that of monetary aggregates or other financial variables, and it is observable only with a long lag. [24]

This is, for example, the conclusion in Laidler (1986). Poole (1986) also provides an interesting discussion. [25]

The standard analysis is that of Poole (1970); for an Australian example, using an empirical model, see Jonson and Trevor (1981). [26]

This argument has been made by Fischer (1980) and Tobin (1985). In the Australian case, the debate following the abolition of the “conditional projection” for M3 in January 1985 would seem relevant. [27]