Formulating a Consistent Approach to Monetary Policy
Published January 1, 1996
Monetary policy is one of the principal responsibilities of the Federal Reserve, and certainly the one which receives the most attention. Simply stated, the goal of monetary policy is to achieve maximum economic performance over time. There is considerable agreement that the most significant contribution the Federal Reserve can make to this goal, characterized by sustained economic growth and improved living standards, is to achieve and maintain low inflation. However, the channels through which inflation influences growth are not clear, nor is it universally accepted that inflation even influences growth.
In addition to long-run emphasis on low inflation, there is a belief that monetary policy can improve economic performance by decreasing volatility in business activity that is, by smoothing the business cycle. However, given the greatly diminished importance of the monetary aggregates in the policy process, a major challenge currently confronting the Federal Open Market Committee (FOMC) is to guarantee that short-run decisions designed to address cyclical concerns are consistent with the long-run low inflation objective. Formerly, the monetary aggregates helped to assure that monetary policy was anchored to low inflation and was "time consistent," but these roles have not as yet been filled by other variables or changes in procedures. Several proposals address these gaps in our practices, but our knowledge is insufficient to make a selection. In my view, it is imperative that we address these issues promptly.
As suggested above, the commitment to low inflation is widely shared among the members of the FOMC. Nevertheless, a host of questions arise associated with the focus on low inflation: Will the low inflation environment contribute over time to growth and to higher standards of living? If so, how? How does low inflation contribute to financial stability? Can the Federal Reserve achieve and maintain low inflation? What weight, if any, should be given to cyclical fluctuations in unemployment and economic activity in policy determination? How should the Federal Reserve implement a low inflation policy?
Inflation and Growth
Evidence has accumulated suggesting that economies perform better, in terms of growth, employment and living standards, in low inflation environments than they do when inflation is persistently high. This evidence is principally a comparison across countries and over long periods of the association between economic performance, measured by, say, growth of output or growth of productivity, and inflation. The correlations indicate a negative relation; that is, the higher the inflation, the lower the rate of real growth. This evidence is neatly summarized in several recent academic papers.1
Evidence suggesting that low inflation promotes growth has motivated recent decisions by a number of central banks and governments, most notably New Zealand. Canada, the United Kingdom and Sweden also have moved in recent years to establish monetary policy regimes with official low inflation targets. Such actions indicate the preeminence of this goal and frequently signify increased independence for the central bank in pursuit of its policies as well. Decisions to adopt a policy objective of low inflation suggest that other policy-makers are reading the evidence pertaining to inflation and growth as we are.
An issue logical to consider next is: Why is low inflation relatively favorable for growth? After all, association does not prove causality; the relation between growth and inflation reported above may simply be fortuitous, or the causality may run the other way. This is indeed a difficult question, in part because until recently there were not well-articulated theories to explain the relationship. However, basic economic reasoning suggests that there are at least two channels through which inflation influences real economic performance. First, in contrast to high inflation, low inflation leads to improved resource allocation because price signals are more easily and more accurately interpreted. Second, low inflation contributes to financial stability.2 Let me explain.
Relative prices provide a guide in the allocation of resources. For example, a change in relative prices resulting from a change in demand patterns should shift resources and production from the activity whose price has fallen (relatively) to that whose price has risen, while a general rise in the price level inflation should not alter resource allocation in this way. But in an inflationary environment it may be difficult for individual decision makers to distinguish between inflation on the one hand and a change in relative prices on the other, and such confusion is especially likely if high inflation is correlated with variable inflation (inflation rates which fluctuate substantially from period to period), as it appears to be. Thus, resources may be seriously misallocated during inflationary periods.
In addition, inflation creates a problem in estimating the real interest rate. The real (that is, inflation-adjusted) interest rate the relative price of current to future goods is not explicitly given as a market price, but rather people deduce it from the nominal (that is, unadjusted) interest rate by taking inflation into account. When inflation becomes variable, this task of determining the true relative price becomes more difficult. The resulting misallocation of resources will adversely affect growth and living standards, because resources are not being put to their best use. Further, if the tax system is not indexed, inflation may adversely affect incentives to work and to invest. In the extreme, considerable resources may be devoted to efforts to avoid or to offset the ravages of inflation. And, without widespread indexation, inflation may well result in capricious transfers of wealth.
All of these effects diminish when inflation is persistently low. Indeed, in the Federal Reserve, our working objective is to reduce inflation to the point where it no longer is a factor in economic decision making. As we succeed, resource allocation moves closer to optimal, with attendant benefits in growth and living standards.
The second broad reason why low inflation favors growth is that it contributes, in my judgment, to financial stability. A low inflation economy is less likely to engender the sharp swings in asset prices and in expectations about such prices that have been so devastating to the financial system from time to time. Consider, for example, the damage wreaked by inflation on the savings and loan industry and some of its customers 15 or so years ago. Similarly, the "credit crunch" which inhibited the U.S. economy just a few years ago can be traced in part to capital pressures at commercial banks stemming from earlier misjudgments about inflation and asset values.
In a fundamental sense, problems associated with misjudgment of asset prices and their prospects are no different than the confusion about relative and general price changes described earlier. Investors and creditors misjudge price signals and draw incorrect conclusions, financial resources are then misallocated, and disruptions occur.
Financial stability is vital to a prosperous economy in a number of ways. Implicit in the preceding discussion, credit decisions, which determine the allocation of financial resources, are likely to be closer to optimal in a low inflation economy. This follows from the common sense notion that bankers and their customers will on average do a better job of assessing business prospects in an atmosphere of relatively stable prices.
Financial stability also enhances an economy's ability to weather shocksrun-ups in energy prices, significant technological changes, unforeseen developments in the economies of major trading partners, and so forththe bane of policymakers and forecasters alike. Such events will cause dislocations, to be sure, but a financial system which can absorb them without significant feedback to economic activity helps to limit the extent and duration of the disruption. In these circumstances, real growth will be affected less than it would be under conditions in which the financial sector magnifies and spreads the effects of the shock. Furthermore, it is likely to be easier to identify the effects of shocks and the proper responses to them in a noninflationary environment.
Policy and Inflation
To this point, I have suggested that low inflation can make significant contributions to growth and prosperity through its effects on real resource allocation and on financial stability. This suggestion is not of much moment, however, if monetary policy cannot achieve, and maintain, low inflation. On this subject, fortunately, evidence and opinion are largely of one mind.
One of the few topics about which most macroeconomists agree is that inflation is first and foremost a monetary phenomenon.3 It results from a long-term pattern of money creation which is excessive relative to the economy's ability to produce real goods and services. Further, there is widespread agreement that the supply of money is determined by the central bank in the long run. Thus, with appropriate policy, the Federal Reserve can achieve and maintain low inflation it should be expected to do so and can be held accountable for doing so. In a general sense, then, the operational responsibility of the Federal Reserve is to provide for long-run growth in money consistent with low inflation. And, as I emphasized earlier, there should be significant economic benefits to the extent the Federal Reserve achieves this objective.
The preceding policy "prescription" glosses over at least one difficult issue, namely: The mandate to avoid excessive long-term money creation permits considerable latitude in how the stock of money moves in the short term and, therefore, is not necessarily useful for short-run policy determination. In the past, monetary aggregates were used to tie the short run to the long run, but the short-run relation between money and the economy has seriously deteriorated. Likewise, confidence in the use of monetary aggregates in practical policy setting has understandably been undermined.4 Experience has convinced me that the aggregates are now of little value in short-term decision making. This leaves a significant gap in our procedures, especially when we want to calibrate a response to business cycle developments.
Business Cycle Considerations
Indeed, a critical question is: How can the Federal Reserve achieve low inflation and effectively respond to business cycle excesses in an environment in which the money supply is not a useful short-term guide to policy? Having previously addressed the money-inflation issue, I will turn to two other aspects of this question: (1) Can monetary policy influence real activity in the short run? and (2) If it can, should it?
These questions have long been debated in academia and in the Federal Reserve, and a consensus has not emerged. My position is that monetary policy has effects on real economic variables in the short term, but the magnitude of such effects is uncertain and the timing between policy action and its effects is variable. Thus, our knowledge of the short-run effects of policy is insufficient to permit us to act aggressively in most circumstances.
This is a fairly conventional position. Certainly the Federal Reserve behaves as if monetary policy has real effects, and empirical evidence supports the notion.5 But because of uncertainty about magnitude and lags, policy-makers have to be cautious in their response to unexpected deviations in economic performance. The fundamental reason is that there is a real risk of aggravating the situation that is, our actions could be destabilizing rather than constructive.
Some would go further and would maintain that even a cautious response to business cycle fluctuations is unwise, given all the reservations expressed over the years about fine tuning. But in my view it is not too difficult to argue in favor of some response. Given that monetary policy has real effects, it is only necessary to observe that short-run volatility inhibits long-run real growth. Put more forcefully, boom-bust cycles damage the economy, and therefore policy should be employed countercyclically to moderate, if possible, the tops of booms and the bottoms of contractions.6
Countercyclical policy to avoid or at least to moderate boom-bust cycles thus seems defensible but, as already suggested, such a policy should be pursued cautiously. It could be destabilizing if policy-makers are wrong about the size and timing of the effects of their actions, and thus it conceivably could inadvertently deepen a recession or stimulate inflation. Moreover, if not carefully implemented and explained, countercyclical policy might create confusion about the long-run objective of the FOMC, could disrupt private sector planning and decision making, and could add uncertainty and inflation premiums to market interest rates.
An Anchor for Policy
Money growth ranges provided a framework that reconciled the long-run commitment to low inflation with short-run reactions to economic developments. To see how, let me explain the roles the monetary aggregates formerly played in the policy process. The place to begin is with the quantity theory of money. According to the quantity theory, there is a relation between the rate of growth of the money supply, growth in real economic activity, and inflation. The linkage between money and the variables we care about inflation and real growth is provided by the velocity of money. This is precisely where recent difficulties have arisen for, as we have seen recently, money velocity has deviated from previous experience, and accurate prediction has become increasingly challenging. Over time, this deterioration in the relation between money growth and nominal business activity has afflicted virtually all of the conventional measures of money.
The quantity theory in its crude form would suggest a strict, mechanical pursuit of a precise target for money-stock growth. Although such a policy regime was approximated briefly to re-establish monetary policy credibility after the inflationary episode at the end of the 1970s, monetary aggregates have been used in a looser, more discretionary way in policy determination since that time. But it seems to me that the deteriorating relation between money growth and nominal business activity has undermined the advisability of even this looser policy regime.
To see why, consider how this regime worked. When the monetary aggregates were useful in policy determination, a rate of growth for an aggregate could be specified consistent with the FOMC's inflation objective and its understanding of the relations in question. For example, if the trend rate of growth in real GDP was estimated at 2.5 percent per annum, money velocity constant, and 3 percent an acceptable inflation outcome for the period in question, then money should expand at 5.5 percent per year. In practice, the Committee not only selected a midpoint for money growth 5.5 percent in this example but also established a range around the midpoint, recognizing both that these relations do not hold precisely on an annual basis and that flexibility to respond to unanticipated developments is desirable. When conditions turned out as expected, open market operations were conducted to keep the path of bank reserves or the federal funds rate, depending upon the short-run operating rule, consistent with desired money growth and, ultimately, inflation.
In this setup, the range for money supply growth fulfilled several functions. The midpoint of the range was typically established consistent with the Committee's inflation objective. When the aggregates were employed successfully, midpoints were reduced gradually over time, in keeping with the FOMC's desire to bring inflation down. The money supply was thus the "anchor" of policy the variable on which the FOMC focused in order to pursue a low inflation policy.
The upper and lower bounds of the ranges served as the limits within which the Committee was prepared to see money growth deviate from its midpoint. That is, the ranges defined acceptable short-run deviations in money growth perhaps for cyclical reasons which nevertheless were viewed as consistent with the Committee's commitment to low inflation. Because the ranges were relatively narrow, policy remained disciplined and, even if in error, was unlikely to be highly destabilizing.7
Consistent Policy Over Time
With the monetary aggregates no longer of significant value in the policy process, we find ourselves without an effective policy anchor that is, without a quantitative way of indicating the FOMC's long-run objective and of guiding open market operations toward that objective and without a means to define acceptable countercyclical action, whereby acceptable I mean an effective response to incipient booms or busts which does not compromise our long-run objective. What I am striving for is the concept of policy consistency over time: a countercyclical response which is consistent with, or can be reconciled with, the FOMC's long-run goal and which, furthermore, is seen as consistent by the public. To be sure, policy has been implemented effectively and largely successfully in recent years, in my view, without an explicit anchor and a method to assure time consistency. Nevertheless, the FOMC's judgment may not always be adequate, and hence it is desirable to find a more systematic way to conduct policy so as to achieve and maintain low inflation and to moderate business cycle extremes.8
There are several ways we might go about establishing a framework for a more systematic policy capable of addressing these two matters. But before describing specific proposals, we need criteria by which to evaluate the options. Based on the preceding discussion of the roles formerly played by measures of the money supply, important criteria are:
- Does the proposal contain a clear and appropriate policy anchor
which guides monetary policy to the low inflation objective? More
formally, is there a variable that bears a close and unchanging
relationship with inflation that the FOMC can influence in a predictable
- Does the proposal have an effective way of delimiting the policy response to business cycle fluctuations? Again, more formally, is there a way of putting a range around the anchor that allows responses to fluctuations, but limits those responses to ensure consistency with the FOMC's low inflation goal?
Of the three following proposals, none is especially original, and none is entirely satisfactory. Nevertheless, they are offered to illustrate a range of available approaches and to stimulate further thought about how we can best establish an anchor for, and assure time consistency in, monetary policy.
One approach to a more systematic framework for policy implementation is to resurrect the monetary aggregates, based on the following considerations. We would acknowledge that the aggregates have lost value as short-term guides to policy, but at the same time we would reaffirm that inflation remains a long-term monetary phenomenon. The responsibility of the central bank, thus, would not change: It is to keep money supply growth within bounds over long periods say 10 years so as to keep inflation low.
This proposal effectively satisfies the first criterion specified above. Money growth would once again become the anchor of policy, with exclusive emphasis on its long-run performance in view of its shortcomings as a short-term guide. Assuming past relations hold, long-run monetary control should result in long-run inflation control. The evidence as to which aggregate to select is mixed, but it appears to be a "horse race" between M2 and the monetary base; either would probably do.
The proposal to concern ourselves only with the long-run performance of a monetary aggregate fares less well against the second criterion of disciplining the response of monetary policy to changes in business conditions. On one interpretation, in the single-minded pursuit of moderate money growth and low inflation, the proposal would permit no reaction to significant deviations in business activity from what was anticipated. This is a potentially costly policy stance if earlier observations about the desirability of containing instability that is, smoothing boom-bust cycles are accurate. But another interpretation of the proposal suggests that "anything goes." This is because the proposal simply leaves open the questions of when or how to respond to unanticipated or undesired developments in the economy.
A second alternative to enhance systematic policy implementation is to focus directly on the policy objective, the rate of inflation, and adjust the instrument, say the federal funds rate, to influence the objective as desired. Intuitively, this approach is appealing, for it does not involve "extraneous" variables like intermediate policy targets. Presumably, an empirical model of the economy would be used to solve for the path of the federal funds rate consistent with the FOMC's inflation goal, and the Committee would authorize open market operations to achieve the funds rate path.
Arrayed against the two criteria specified above, the virtues and shortcomings of this proposal are evident. Since it focuses directly on inflation, it would seem to satisfy the first criterion of providing an appropriate anchor for policy. One has to be cautious, however, because our goal is optimal long-run economic performance and, as discussed above, the evidence suggests that low inflation over the long run is favorable for growth; but I am aware of no evidence which indicates that inflation control period by period is conducive to real growth. Indeed, even proposals to maintain constant growth in the money supply have recognized that there could well be a lot of period-by-period price volatility.
Although a multiperiod inflation targeting procedure would seem to ameliorate this problem, it also has problems. With a short-period horizon, the instability problems associated with a single-period horizon remain. But with a long-period horizon, there is basically no policy discipline in the short run.
A third approach to the issue of the systematic implementation of policy is for the Federal Reserve to focus on a short-term market interest rate. One question is whether the interest rate should be nominal or real. Based on stability considerations, a real interest rate seems preferable. The argument that pegging a nominal interest rate can be destabilizing is now familiar: Normally, high interest rates are associated with restrictive monetary policy, but if expected inflation rises for some reason and the Federal Reserve pegs the nominal rate, then policy actually becomes increasingly expansive. A symmetric problem occurs with nominal rate pegging when inflation expectations diminish or when deflation sets in. This problem does not apply to a real short-term interest rate, as I argue below.
Another question, then, is how to implement a real interest rate proposal. One way would be to use an empirical macro model to solve for the real rate, or the path of the real rate, consistent with the Federal Reserve's low inflation objective. Given its best estimate, or best judgment, of inflation expectations, the FOMC would then establish the nominal rate that produced the desired real rate. Presumably, the more actual and prospective inflation are above the goal, the higher the nominal rate a given real-rate target would imply. If one believes that this pattern of nominal rate setting is what the FOMC ought to be doing, then a virtue of real-rate targeting is that the FOMC would implement it in a stable and self-regulating way. Adjustment of the real-rate target would be justified when the environment of the economy changes in some fundamental way, such as an increase in the rate of return to capital investment. Advocates of a real-rate target believe that it would thus lead the FOMC to focus its deliberations appropriately on long-term considerations, without sacrificing responsiveness to short-term disturbances in the economy.
An advantage of this approach is that economic theory suggests that real rather than nominal interest rates matter for spending decisions, so the Federal Reserve would in fact be emphasizing a variable that can be expected to affect economic performance. This observation implies that this proposal could fare relatively well against the second criterion of defining the policy response to cyclical disturbances in activity.
Since there could well be a long-run correlation between real rates and inflation,9 the proposal would seem to have the potential to achieve the Federal Reserve's low inflation objective. However, the first criterion specified above calls for a clear and appropriate policy anchor; a real rate may be appropriate but its clarity is another matter. There is not agreement on measurement of the real rate nor on its controllability.
Indeed, critics of the real rate proposal assert that the Federal Reserve cannot hope to control a real rate of interest, which they view as ground out by interactions in the real economy independent of monetary policy. I hold a somewhat different view. In a world with interest-bearing and noninterest-bearing government fiat debt bonds and money there must be frictions in the marketplace which induce the public to hold them both, since bonds dominate on a rate of return basis. An implication of this observation is that monetary policy actions which alter the relative supplies of money and bonds held by the public affect real interest rates.
The Federal Reserve is committed to achieving and maintaining low inflation. This is an objective the central bank can legitimately be expected to accomplish and for which it can be held accountable.
Although important, accountability is not the reason to focus on low inflation, however. Rather, a sustained environment of low inflation should contribute over time to economic growth and to improvement of living standards. The way in which low inflation contributes to these outcomes is not entirely understood, and I have suggested in this essay that positive effects on real resource allocation and on financial stability are key.
There is only limited agreement at the moment about the systematic conduct of policy, in view of the diminution of the role of the money supply measures. I have offered several suggestions to address this issue, including exclusive focus on the long-run growth in money, on the inflation objective itself and on a real short-term interest rate. Each of these proposals has its flaws, and I do not think that we possess sufficient knowledge at present to make a selection with confidence. However, recent policy successes notwithstanding, establishing a method for the systematic conduct of policy is worthy of serious consideration and debate going forward. It is imperative that we identify an anchor for policy and a procedure which assures time consistency, so that short-term decisions are related appropriately to the long-term commitment to low inflation.
1 I am referring to “The Growth Effects of Monetary Policy,” by V. V. Chari, Larry E. Jones and Rodolfo E. Manuelli, Quarterly Review, Federal Reserve Bank of Minneapolis, Fall 1995 ; "Private Information, Money, and Growth: Indeterminacy, Fluctuations and the Mundell-Tobin Effect," by Costas Azariadis and Bruce D. Smith, forthcoming in the Journal of Economic Growth, 1996; and "Inflation and Economic Growth," by Robert J. Barro, National Bureau of Economic Research, Working Paper 5326, 1995.
2 A third channel related to distortions stemming from the interactions between inflation and financial regulations is discussed in Chari, Jones and Manuelli, 1995.
3 A recent study shows a high correlation between the rate of growth of the money supply and the rate of inflation; this correlation holds across three definitions of money and in a sampling of 110 countries. “Some Monetary Facts,” by George T. McCandless Jr. and Warren E. Weber, Quarterly Review, Federal Reserve Bank of Minneapolis, Summer 1995.
4 In "The Rise and Fall of Money Growth Targets as Guidelines for U.S. Monetary Policy," 1995, Benjamin Friedman writes: "In 1987 the Federal Reserve gave up setting a target for the narrow money stock but continued to do so for broader measures of money. In 1993 the Federal Reserve publicly acknowledged that it had 'downgraded' even its broad money growth targetsa change that most observers of U.S. monetary policy had already noticed long before."
When Lawrence J. Christiano investigated the supposed change in the relationship between money and the economy, he examined two types of models the first showed a break in the relationship, and the second revealed no break. However, in the second model the link between monetary aggregates and inflation was so weak that it would be of no practical use for short-term policy-making anyway. "Money and the U.S. Economy in the 1980s: A Break from the Past?" Quarterly Review, Federal Reserve Bank of Minneapolis, Summer 1986.
5 John H. Cochrane, in "Identifying the Output Effects of Monetary Policy," National Bureau of Economic Research, Working Paper 5154, 1995, cites a large amount of recent work which concludes that anticipated monetary policy changes can have real effects on the economy. For example, Cochrane cites David H. Romer and Christina D. Romer, "What Ends Recessions?" in NBER Macroeconomics Annual (1994), Cambridge: MIT Press; and Lawrence J. Christiano et. al., "Liquidity Effects and the Monetary Transmission Mechanism," American Economic Review 82, 346-53.
6 A recently published study marshals considerable evidence that economic growth is inversely related to volatility (Garey Ramey and Valerie A. Ramey, "Cross-Country Evidence on the Link Between Volatility and Growth," American Economic Review, 85, 1138-1151).
7 Although the ranges occasionally were violated, the FOMC had to explain the errors to Congress. The FOMC generally preferred to abide by the ranges.
8 Interestingly enough, in his recent Nobel prize acceptance speech, Robert E. Lucas Jr. grapples with similar issues. Lucas concludes that, absent a better understanding of monetary non-neutralities, it does not seem possible "to determine whether an optimal monetary policy should react in some way to the state of the economy or should be fixed on some pre-assigned objective ... In the meantime, policy must be made, nevertheless, and existing theory, empirically well-tested, offers much useful guidance."
9 The legal restrictions theory implies a long-run relationship connecting monetary policy, real interest rates and inflation (see Wallace). The broad implications of this theory seem consistent with observations (see Miller-Todd and Chin-Miller).
Neil Wallace, "A Legal Restrictions Theory of the Demand for 'Money' and the Role of Monetary Policy," The Rational Expectations Revolution (1994), Cambridge: MIT Press.
Preston J. Miller and Richard M. Todd, "Real Effects of Monetary Policy in a World Economy," Journal of Economic Dynamics & Control 19, 1995.
Dan Chin and Preston J. Miller, "Fixed vs. Floating Exchange Rates: A Dynamic General Equilibrium Analysis," Staff Report 194, Federal Reserve Bank of Minneapolis, 1995.