Monetary Transparency in a Time of Change

There has been much talk about structural change in the current United States economy, and our session chair has presented some of the hard facts underlying it. In keeping with the mandate of the panel-and recognizing the comparative advantage in forecasting of the chair, Monday's speakers, and many in the audience-I will stick to the general question of what monetary policy's response to structural change should be. Then I want to go a bit further and suggest what the apparently successful U.S. monetary framework at present can do to lock-in and build upon the achievements of the last few years.

My basic premise is that the appropriate response to increased uncertainty about the economy is to decrease uncertainty about the ultimate goals of monetary policy. In a word, monetary policy should be transparent. The experience of many major central banks in recent years, and my and others' research on it, confirms the importance of monetary transparency. So I want to explain what greater transparency in monetary policy would entail and what advantages it would offer, particularly in a time of change. There are those who oppose greater openness about goals by central banks, usually by claiming either that talk is cheap or that talk is dangerous. I will argue that neither is the case. But, first, I want to put structural change into a monetary policy context.

I. Change is no Shock

Leaving aside for the moment the question of whether the world really has changed-I've got to keep something to put into my book in progress-the issue is what does structural change mean for monetary policy. When we speak of structural change, and the types of factors invoked (e.g. technological progress, globalization, new mindsets), we are speaking of things unlikely to be affected by any one or a few short-run policy decisions. By definition, these are changes ultimately outside of policy or at least requiring a large and concerted initiative to reverse them. The fundamental role of monetary policy is to provide the price, financial, and payments stability necessary to letting the economy develop without undue disruptions along the way.

Yet, monetary policy critically shapes people's understanding of the changes in the economy which occur, and hence, the changes themselves. Monetary policy's effects may be thought of as the result of three factors: the preferences of the central bank, the transmission of the central bank's policy (usually interest rate) decisions into the economy, and the development of the economy in the interim until that policy takes effect. Leave aside shocks which we assume the central bank wouldn't know about or else they wouldn't be shocking. This means there are three potential sources of uncertainty-changes in what the central bank wants, changes in what the central bank can do, and changes in what (or whom) the central bank is doing it to. Both the central bank and the private sector have to look backward at the sum of these changes in order to build a model with which to look forward and predict what happens next. Thinking of it that way, it is pretty easy to discern when the total model uncertainty rises-we can tell who's doing what in the current economy, and whether that's different from what would have been expected before given initial conditions and the central bank's policy.

We should realize, however, that what I'll call forecast uncertainty, the third aspect I mentioned, is the least important source of uncertainty for monetary policy. If the source of changes in the economy is truly something structural and lasting-and not induced by a temporary shift-we'll find out pretty quickly. Policy will catch up. Say we believe that U.S. productivity growth has picked up, in an as yet unmeasured way, in recent years. If the Federal Reserve were to err on the side of counter-inflationary caution, it would be pleasantly surprised to see that we can grow faster without inflation, and it would see that (one can argue that is exactly what has happened). Inflation in the U.S. and other low inflation countries is a very inertial process, neither dead nor explosive. If the Fed only misestimates productivity, say, it will not cause that much harm. In any event, there is also little a central bank can do about this kind of uncertainty beyond collecting more data.

The most dangerous source of uncertainty for monetary policy is the second kind, change in the monetary transmission mechanism, that is change what the central bank can do. Since the primary means of transmission of monetary policy in any developed economy is through the financial system, what we're really talking about under this heading are changes in financial technologies and regulations. These luckily are pretty observable, although their implications are not always clear, and they are something over which the central bank usually can exert some influence. In addition, keeping an eye out for asset price bubbles is generally a good idea. That being said, this transmission uncertainty is the source of occasions where monetary policy truly can do damage. When people speak about monetary policy mistakes what they usually have in mind are occasions where monetary policy moves (or non-moves in the infamous case of the 1929 stock market crash) are amplified in unexpected ways by financial fragility. Thankfully, finding anyone to disagree with the statement "financial fragility is bad" is a bit difficult, so I will not belabor the point. Most importantly for today's discussion, dealing with financial fragility is something that must be done irrespective of the state of the economy or the long-run trends affecting it. This is a factor independent of structural change which I'll leave aside.

That leaves us with one remaining source of change and uncertainty-the question of the central bank's own preferences, the specific goals of policy. Clearly, the central bank has no uncertainty of its own in this regard. It's also the easiest uncertainty to remove-the central bank just has to tell the truth about what it wants and have its deeds match its stated intent. Sounds simple, does it not? You'll notice, however, that in the real world central banks disclose their goals to varying degrees, with varying explicitness in that disclosure. Leaving aside those central banks contending with undemocratic political environments or public debt in excess of 100% of GDP, it's worth asking why central banks differ along this line. Are there any reasons why it would be in the interest of good policy for businesses and households not to know in some detail the long-run goal, and therefore the intended stance and likely course of monetary policy?

Just so you don't think I'm hijacking this session away from the stated panel topic and the current events in the economy let me point out why this question is particularly compelling in a time of structural change. All of us are trying to figure out what's going to happen next in the economy in general, what monetary policy should do as a result, and what it actually will do. If we all agreed on the forecast, or at least to the broad outlines of the forecasting model to the extent we're ever able to, then a monetary policy move gives some, though not all, information about where the central bank wants to go whether or not the bank tells us explicitly what it is after. Most people making economic decisions get the same signals and these are consistent with what the central bank is expecting, and the central bank's actions in turn become better expected.

If, on the other hand, we're again planning for the future by reasoning backwards from the policy decision and the current state of the economy without either an explict knowledge of the central bank's intent or a largely agreed upon model of the economy, we don't know what to think. If the Federal Reserve makes a move in interest rates at the next FOMC meeting, or if it doesn't, what should we take it to mean? Does it mean that the Fed is confirming or denying one particular side in the "Inflation is Dead/Productivity is Dead" debate? Does it give us an insight into what the forecast of the economy should be? If different economic decision-makers see different possible forecasting models implied by the Fed's move depending on their guess of the Fed's intent, they will logically make different decisions.

II. A Time for Transparency

What can a central bank do to remove this uncertainty about its preferences? It can make its goals, and therefore its policies, more transparent. The primary way in which central banks do so is to publicly announce a numerically defined measurable target for policy, with a specified time-horizon for that target to last. In recent years, this has taken the form of multi-year inflation targets, though many countries which in the past were supposedly targeting money-including the Bundesbank and the Swiss National Bank-were actually publicly targeting inflation.

Does monetary transparency as I've just defined it offer any advantages? Does such an announcement do anything, or is this just cheap talk? The combination of a public announcement and a clear quantified measure means it cannot be merely cheap talk. There's no point in setting an unattainable or undesirable target if the central bank expects to be around when the target results come out. If deviations from the target do occur, they are visible. Either the central bank has to explain the deviation-we missed because the financial system's a wreck, there was a one-time spike in energy prices, etc.-or it has to change the target. Of course, we would expect to see the former more frequently than the latter. To get away with these deviations, the central bank has the incentive to provide enough information to make itself understood, and explain itself ahead of events.

Consider how such transparency would work by comparing two countries with equal low inflation rates. One's central bank has an announced inflation target of 2.5% to be met over the next 3 years. The other's central bank has a "pre-emptive strategy" to achieve an unspecified definition of price stability. First, think through what happens as the business cycle heats up, and each central bank sees signs of imminent inflation. The inflation targeting central bank knowing it is on record explicitly states, "If our forecasts are correct, inflation will exceed 2.5% in 18 months time unless we raise interest rates now," and raises rates. So long as the forecast is credibly explained, markets understand the stance of policy, and inflation expectations remain unchanged. If people oppose the decision to tighten, they have to argue directly either with the specifics of the forecast or explicitly for the central bank to change its known goals. This is potentially contentious, but transparent, and the conflict adds no real uncertainty.

The non-targeting central bank makes the same forecast, and its "price stability" goal defined only internally may be violated by the same uptick, so it raises rates, too. What does this rate rise tell the markets about the stance of monetary policy? Some will say, "Oh, the central bank must think inflation is rising." Some will see things differently and say, "Oh, the central bank wants to drive inflation lower than current levels." Some others might even say, "Oh, the central bank is targeting growth and slowing the economy when it gets going, and that's bad." The central bank's motivation for the move is unspecified, and the future stance of policy is unclear as a result, and each group comes up with a different inflation expectation. Of course, the central bank can issue an explanatory statement with the move, but without specifying the inflation goal, that statement is close to cheap talk. Without knowing what it is that the central bank actually wants to accomplish, and how the public will know when and if it has been accomplished, any claims by the central bank cannot be verified.

The comparison becomes even more striking in a time when we assume that the central bank temporarily cannot rely on its model of the economy, that is in a time of structural change. The central bank with the announced target sees real side activity numbers which in the past would have generated inflation, but this time through do not seem to be doing so. If the targeting central bank wants to err on the side of counter-inflationary caution and still raise rates, it can do so, but it also has to publicly indicate with respect to the numerical goal why it is raising rates. Most importantly, if the central bank turns out to be wrong about the course of the economy, it can credibly reverse its policy move. Since everyone knows that the targeting central bank is committed to the public goal irrespective of its forecasting model, so the uncertainty in the forecast and its resolution do not unduly constrain policy. The argument is symmetrical if the central bank with the announced target initially decides to hold off raising interest rates-so long as the goal is measurable and specified and eventually has to be met, the public can see this restraint as caution and experiment without seeing it as a break in counter-inflationary resolve. In fact, some countries such as Switzerland, New Zealand, and Germany have used such policy reversals in the face of change as chances to emphasize to the public that it is the transparent inflation goal which is the long-run guide to their monetary policy, and not intervening events.

What about the inexplictly pre-emptive central bank in a time of structural change? Just as before when its forecast uncertainty was low, if the central bank raises rates there is public uncertainty about the interpretation of the move's meaning, and divergence in the expectations for monetary policy and inflation. Now, however, if the central bank turns out to be wrong in its pursuit of its internally known goal, and needs to reverse its rate rise-just like the targeting central bank had to-the uncertainty is compounded. Is the reversal because the balance of opinion within the central bank on whatever goal has shifted? Is this because the central bank did not foresee the real economic costs of its actions, and its willingness to bear costs in disinflation should be reevaluated? Is this because the forecast was unavoidably wrong (the actual answer)? Is this reversal because a new shock and information came along, although the original forecast was correct at the time? There is no one response the untransparent central bank can give to answer these questions. Even if the target is then made explicit, if it is done after the fact and of unclear duration, it looks like a rationalization for a mistake.

And this compounded uncertainty about preferences and forecast can start creating uncertainty about the transmission mechanism, which presents potentially serious costs for monetary policy. If we believe that people (and especially financial markets and businesses) are forward-looking, the effect of current monetary policy actions depends in large part on what people expect future policy to be. A temporary or likely-to-be reversed policy is very differently responded to from a commitment to a policy path over the longer-horizon. When observers cannot tell what goals a central bank is pursuing, let alone whether or not those goals have been or will be achieved, they cannot deduce the likely course of policy. The untransparent central bank will find itself in turn not only having to decipher from the results of its policy moves the relevance of its own forecasting model to the changing structures, but also how much of the deviation from the forecast is due to the induced uncertainty about the effects of policy. Even keeping rates constant would present the same dilemma, because every central bank decision conveys information and affects the economy. In the world of Alan Blinder, "Even doing nothing-whatever that means-is a decision."² And a monetary policy decision with an unclear motivation is likely to have unintended effects.

I hope that comparison of scenarios I led you through illustrates the need for transparency in a time of change. I want to emphasize that these scenarios are very much the stuff of reality rather than thought exercises. In a recently published study of inflation targeting by Rick Mishkin and myself, we discuss among other things the experiences of Canadian and German monetary policy in the 1990s.³ Taking Canada first, the Bank of Canada announced an explicit numerical target for a multi-year downward path of inflation at the end of February 1991. This was a time of great structural change and uncertainty in Canada, with the effects of a major fiscal consolidation and tax reform, of free trade with the United States, and of a persistent rise in the force of Quebecois separatism all at work. By having a transparent monetary policy goal, the Bank of Canada was able to bring inflation down a great deal without inducing significant instability in either the dollar exchange rate or in domestic interest rates, both long-time sources of difficulties. When the economy slumped more and inflation fell faster than expected in 1992-94, the Bank of Canada was able to ease monetary policy without engendering credibility problems by pointing to its long-term commitment. It is important to note that the eventual 2% inflation target acted as a floor underneath inflation and real activity rather than as a ceiling on them, and forced an easing to meet the target. In addition, numerous one-time shocks from taxes, exchange rates, and energy prices passed through the Canadian economy in this period without leading to inflationary spirals as they had in the past, because the Bank could state that it would only intervene if the longer-term inflation target was endangered by second-round effects.

Germany followed the political revolution of reunification in 1989 with the economic structural shift of monetary unification in October 1990, converting Ostmarks into Deutsche Marks. The enormous demand shock and the unclear development potential of the new German states, as well as the political pressure from the Kohl government to ease integration made a rise in inflation by some undetermined amount in the early 1990s inevitable. The Bundesbank maintained its publicly announced inflation target at 2% throughout this period, even though it allowed inflation to exceed that level temporarily.⁴ That way, everyone knew that whatever might transpire in the interim, the Bundesbank would bring inflation back down to 2% in a finite number of years, but not seek to drive it any lower nor to get there immediately. Additionally, as the effects of monetary and labor market integration in Germany became clearer, the Bundesbank could change its policies in response to the accumulating information while anchoring expectations.

The recent experience of Japanese monetary policy can be seen as an example of the dangers of a lack of monetary transparency (among other things). In contrast to the Canadian and German cases, monetary policy in Japan since the structural change following the burst asset price bubble has given the public no explicit announcement of its goals. Consequently, every time the Bank of Japan has moved interest rates or left them steady, no one could tell whether the (in)action was due to political pressure from the Ministry of Finance or financial interests, a reassessment of the growth and inflation forecasts, or an extended displacement of macroeconomic goals in pursuit of renewed financial stability. The inability of markets and businesses to guess the future stance of Japanese monetary policy not only had direct negative effects on investment and spending, this uncertain response interfered with the transmission of monetary policy in an environment where expansionary monetary policy could have benefitted from a larger impact. This uncertainty is a far more likely explanation for the relative ineffectiveness of the Bank of Japan's interest rate cuts to stimulate the economy than appeals to a nominal interest rate floor at zero percent. Moreover, unlike Canada, without a target there was no clear floor for inflation below which the Japanese people could know the central bank would not allow the price-level to drop. Deflation occurred.

Beyond the historical and the theoretical stories, there is also statistical evidence for the benefits of monetary transparency in a time of change.⁵ For a simple representation, consider my Table 1. This table shows the average and standard deviation of monthy inflation and money market interest rates for the G-7 countries in two recent periods: November 1987-August 1992, from the stock market crash up to the European ERM crisis, and October 1992-December 1996, after the ERM crisis up to recent data.⁶ While all countries saw a decline in inflation and in interest rates in the 1990s, the most sizable declines by far were seen in Canada and the United Kingdom, the two countries which joined Germany in publicly announcing numerical goals for inflation. Even more importantly, the variability of short-term interest rates declined the most in these two countries except for the U.S., while the remaining countries saw no such decline. This could be consistent with a decline in uncertainty about monetary policy.

I do not by any means claim that monetary transparency in general, or inflation targeting in particular, is a cure all for whatever ails an economy. Certainly, as far as monetary policy goes, it would always help to have better understanding of the structural changes in the economy, and better forecasts. Yet, a central bank which has transparent goals can point to them as a guide to long-run expectations in a world of shocks and uncertainty. In fact, a transparent goal for monetary policy prevents some of the worst potential effects of structural change from occurring, either a dispersion of private-sector expecations about the course of monetary policy and inflation, or worse a widespread locking-in of a mistaken model of the economy thought to be validated by central bank actions.

III. No Time for Rules

Having dealt with why central bank transparency is useful rather than cheap talk, I'll now turn to the other set of arguments against clearly specifying central bank goals, those that say transparent talk is dangerous. We've all seen secrecy associated with monetary policymaking to some degree at various times. Let us assume that secrecy is not motivated by a desire on the part of central bankers to dodge accountability for their policies (which, even if it were true, would not be an argument against transparency, merely a vice to be dealt with). The main reasons why monetary transparency is opposed have to do with policy flexibility, or fear of rules, and central bank independence, or fear of politicians. Though I'm as opposed as anyone to the effect both of rules and of politicians on monetary policy, I want to explain why increased transparency is not an open door to either. In fact, the history of monetary policymaking shows that transparency does not compel single-minded pursuit of the announced goal and actually, if anything, enhances central bank independence.

The danger from arbitrary rules in monetary policy is a very legitimate concern, and any experienced central banker will tell you that some flexibility is required to keep the economy out of trouble. As Chairman Greenspan remarked at Stanford 12 days ago, "In an ever changing world, some element of discretion appears to be an unavoidable aspect of policymaking."⁷ I'll go further-in a time of structural change which is defined by the fact that the old economic relationships which could have been used as a basis for a rule are no longer dependable, rule-based monetary policy is a prescription for disaster.

Without wanting to drag you back into the academic discussion of "rules-versus-discretion," especially if you've been successful in avoiding it so far in life, let me clearly state what a rule-based monetary policy would entail. A rule is an auto-pilot mechanism for monetary policy which specifies that if and only if X event occurs, then Y policy move must be made. The theoretical prototype is Friedman's famous k-percent growth rule which states that whatever happens, keep policy such that the money supply grows at k% a year. For a rule to be a good way to run monetary policy, at least one of two things must be true: First, the rule must be robust to changes in the economy or the economy is largely stable and not subject to significant shocks or changes, because if one frequently decides to make an exception to the rule or to judgmentally identify when the rule needs to be changed, it is no longer a rule.⁸ Second, that the disadvantages of letting monetary policymakers set policy, or even letting it be known that the policymakers could use their discretion to set policy, causes greater damage than keeping policy inflexible.

It is precisely because I so strongly agree with Chairman Greenspan and others that discretion is an inevitable aspect of monetary policy in practice that I do not fear that increased transparency will bring inflexible rules. No central bank with any credibility, including those normally cited as successful examples of rule-based behavior, has ever in fact followed a rule. Under the classical gold standard, it has been demonstrated by numerous economic historians that discretionary monetary policy was used to respond to economic events and goals other than price stability, and exits and entries from gold parities should be seen as a conscious part of this effort. I have shown elsewhere in some detail that the success of the German and Swiss postwar monetary frameworks in maintaining long-run price stability cannot be attributed to behavior by their central banks as though they were following a monetary growth rule⁹-instead, they exercised what I have called "disciplined discretion" based on transparent quantified inflation goals and got low average inflation despite responding to other economic goals in the short-run. Only central banks with no choice in their monetary strategy due to their lacking credibility, country's fiscal problems, or lack of control over their country's economy actually adhere to such rules as fixed exchange rate regimes or currency boards (but the folly of fixed exchange rates is a topic for another time).

For purposes of my argument today, it is enough to demonstrate that central banks have displayed varying degrees of transparency, while all of them have similarly avoided rules. If the fear of transparency is that picking a single long-run goal of monetary policy to publicly specify numerically will force the central bank to ignore other short-run goals and be unresponsive to economic events, this fear is unfounded. The data is clear that all modern central banks have demonstrated concern for financial stability, economic growth and employment, exchange rates, among other goals, regardless of their statement or non-statement of intent. One way this can be shown is to look at the recent studies which demonstrate that a similar "Taylor Rule" reaction function predicting interest-rate moves on the basis of both inflation and real output can be used to predict not only Fed behavior, but also Bundesbank, Bank of Japan, and Bank of England policy moves. Of even more direct relevance, one can also look at the experiences of various countries which have adopted inflation targets in the 1990s and see how they have nonetheless deviated from those targets in the short-run as required to by economic events.¹⁰

I will demonstrate the lack of association between transparency and rule-like inflexibility in a reasonably direct manner. In Figure 1, I plot the average inflation rates in the 1990s against written central bank charter objectives for a group of 16 OECD countries and then for a wider set of OECD and non-hyperinflationary developing countries. Central bank charters are coded 1 if their only stated goal is price stability and it is said to be overriding, 0.8 if the only goal listed is price stability but with no mention of its precedence, 0.6 if many goals are listed along with price stability, 0.4 if the many goals listed contradict price stabiliy, and 0.2 if price stability is ranked below other goals, down to 0 if price stability is not listed as a goal.¹¹ If what central banks are required to do in the long-run actually constrains their behaving flexibly in the short-run, the countries with the narrow central bank mandates solely for price stability should have lower average inflation rates because they never loosened in response to intervening events.

We can see on these charts that there is no statistical association between central banks' charters and the countries' average inflation levels (if any patter exists it goes the opposite way). The same pattern or lack thereof would appear if one were to plot different decades or different but still valid samples of countries. Remember, the Federal Reserve currently operates under a mandate to pursue both "full employment" and "stable prices", which would earn it a 0.4 on this scale, and the Swiss National Bank has a charter listing several goals but not price stability, earning it a 0 score-yet both have produced long-run average inflation rates comparable to the single-goal mandated "safeguard the currency" Bundesbank.

Let me be very clear about what I am and what I am not saying. I am not saying that central bank laws are meaningless and are ignored by their central banks. I am saying that a variety of evidence shows that all credible central banks exercise policy flexibility in the short-run regardless of their long-run stated goals. Flexibility is inevitable because truly rule-based monetary policy is so inherently untenable that is only undertaken in the most dire circumstances, when the central back has so little credibility that no alternative is available. Thus, there is no reason to believe that increasing long-run transparency will lead to damaging single-mindedness in the short-run.

IV. Transparency and Independence

There are those who believe that even if monetary transparency does not lead to inflexibility problems as a matter of policy necessity, it will generate similar or worse disadvantages politically. Transparent talk about goals is from this view dangerous because some elected politicians might attack the central bank for any deviations from the now specified goal, no matter how unavoidable, or short-term, or meaningless, or even beneficial the deviation is. Some other elected officials might attack the central bank for its seeming lack of concern about other goals-say its disregard for unemployment if an inflation target is announced-even though the central bank in reality does have concern for those goals and responds to them in the short-run. Both types of attacks, it is feared, would not be just rhetorical, but could potentially lead to the politicization of monetary policy, or even to the removal of the central bank's independence.

Just as in the case of rules, while I agree that the politicization of monetary policy is something to be feared, I do not agree that monetary transparency will increase the danger of that fear being realized. If the removal of vagueness about goals, and the little wriggle room in public statements which that provides, is enough to bring the hounds down on a central bank and its independence, that central bank is already subject to severe political pressures. With all that is at stake in monetary policy, I do not believe that politicians and interested parties will suddenly decide to make major changes in central bank structures on the basis of one day's rhetoric. My own earlier research on central bank independence has shown that central banks are granted independence and stay independent when there already exists a politically powerful constituency which supports and insulates the central bank, and that support is based on real economic interests.¹²

This reality is reflected in the fact that no central bank in an industrialized democracy has suffered such attack in the postwar period. And during the 1990s, we know that the world has been moving towards greater central bank independence in more and more countries, inspired in part by the Federal Reserve's example, even as costly disinflations have been undertaken and as examples of central banks deviating from their targets have mounted. If anything, the link between transparency and central bank independence goes the other way, with transparency increasing independence. In the United Kingdom, the inflation targeting regime allowed the Bank of England to build up a clear record of performance linking its monetary policy recommendations to results. The fact that the Bank of England could be seen as attaining a measurable goal, and explaining the reasons for the specific goal, led directly to the new Labour government granting it independence this past May. In Canada, the protracted economic slump was one of the top two issues in the federal election campaign at the end of 1993, and frequently part of the blame for the slump was laid directly at the feet of the Bank of Canada's "single-minded fight against inflation." Even when the election resulted in the incumbent Conservative Party's virtual annihilation in parliament, the successor Liberal coalition was sufficiently won over by the Bank's explicit substantive responses to being challenged, that the Bank's target, let alone structure, were unaffected by the election.¹³

If we believe that we can draw conclusions from political and historical data, the same way that we generalize from regularities in economic data, then we can comfortably state that the link between increased transparency and fears for central bank independence is unfounded. If we believe instead that politics is an inherently unknowable and random process and such generalizations can never be made with confidence, then there is certainly no reason to be confident that whatever central bankers say or do not say will have any defensive value.

I do accept that there is a shorfall of understanding amongst some elected officials and their constituents about what monetary policy can and cannot do. As economists, we are all taught that the best way to get rid of market failures is to treat them directly, because otherwise distortions occur. Therefore, I would say let us educate politicians and voters about monetary policy by providing a clear record of what monetary policy has been and seeks to be doing. By avoiding this market failure through inexplicitness, we thereby introduce the distortion of lack of monetary transparency, with the foregoing the benefits and incurring the costs I already outlined. Perhaps this sounds naive and idealistic to think that we can educate the politicians, that their seeming lack of knowledge is remediable rather than a response to electoral incentives.

I would respond that I find it even more naive to believe that the public cares about any politician's rhetorical posturing more than actual economic results, and that it is even more idealistic to believe that every central bank head will always be endowed with the political abilities to say and not say just the right things to elected officials, to the extent that is necessary. If monetary policy is producing consistently good results over the long-term, or at least the best results that it can achieve, the central bank should say so, and the voters will learn. If monetary policy is consistently not living up to the results which can reasonably be expected of it, the central bank cannot and should not hope to elude that becoming known. Transparency and independence are compatible and perhaps even complementary.

V. Change as an Opportunity

Monetary transparency-that is the public announcement of a numerically defined goal for monetary policy-is neither cheap nor dangerous talk. It removes unnecessary uncertainty abou the stance of present and future monetary policy. In a time of structural change, it is an anchor for businesses' and individuals' beliefs about the structure of the economy, as well as a guidepost to the course of the economy over the long-run, even as the economy and monetary policy vary. By talking about long-run goals, rather than specifying rules, monetary transparency allows central banks the necessary flexibility to respond to short-run developments in the economy-a flexibility essential in a time of structural change. And in a democracy, transparency is the only appropriate monetary response to political pressures which are placed on an independent central bank.

All that being said, monetary policy can be successful, even for extended periods, with less than total transparency about its goals. I speak of "total" transparency because transparency is a matter of degree, not of complete light or dark. So even if you agree with the arguments I have presented for why monetary transparency is a good thing in principle and in practice, you may legitimately question its relevance for the current United States context. Why mess with what is not broken?

One advantage of good times is that it allows planning ahead rather than crisis management, and the opportunity to build on previous success. For central banks in particular, reputation accumulates like capital, and sometimes surplus capital should be profitably invested in new projects. The current recovery, and the credibility and prestige which the Federal Reserve has earned for its role in this economic situation, present just such an investment opportunity.

I would argue that an increase in transparency about monetary goals now might serve to lock-in low inflation expectations. As Benjamin Friedman has observed, from the late 1960s until recently, the reasonable presumption about the goal of U.S. monetary policy has been that any practical reduction in inflation was desirable. We did not need to discuss our inflation goal and the costs and benefits of achieving it in specific terms because it the goal almost always was lower. I would assert that we have been somewhat fortunate as well in recent years, and have not been confronted by any major negative macroeconomic shocks. This extended the period during which the fact that the likely direction of policy was no longer self-evident had little implication.

When another negative shock comes, however-and one will, there has not been that kind of structural change-the question is whether the difficult choices made by the Federal Reserve at that time will be properly understood. Remember my scenarios about uncertainty and monetary policy. Will a needed discretionary move in interest rates be seen as such, or as a change in goals or in the structural forecast? In particular, what happens if that negative shock hits after the extraordinarily trusted current chairman and membership of the FOMC have been succeeded in office? Will the markets and the public have enough trust in those later policymakers that uncertainty and inflation expectations will not rise even if the long-run goal of policy has not been publicly announced ahead of that shock?

A couple of millenia ago, Plato suggested that it is best to be ruled by a wise philosopher-king, but since kings do not live forever, and in succession there is regression to the mean, over the long-term we should live in a republic and be ruled by laws. Similarly, while the benefits of having monetary policy run by a platonic ideal central banker encourage complete trust, we may not always be so lucky in our monetary leaders. For the American public the best relationship to have with the Federal Reserve over the long-run is to "trust but verify." In other words, a central bank should be granted discretion in the exercise of monetary policy, but held accountable to a transparent standard of performance in its exercise of that discretion. It is in both the public's and the central bank's interest for the central bank to provide that standard and the information necessary to make reasoned asssessments of success in meeting it. That is the role of monetary transparency, and that is why it might be a means to institutionalize some of the gains we have seen in this time of structural change.

Notes

1. Prepared for the National Association of Business Economists Annual Meeting, September 14-17, 1997, New Orleans, LA, Panel on "What is the role of policy when structural change is more important than the business cycle?" The opinions expressed here are solely my own, and not those of the NABE or the Institute for International Economics.© Institute for International Economics, 1997.

2. Alan Blinder, "Central Banking in Theory and Practice - Lecture I: Targets, Instruments, and Stabilization," Marshall Lecture, Cambridge University, Mimeo, May 1996, p. 23.

3. Frederic Mishkin and Adam Posen, "Inflation Targeting: Lessons from Four Countries," Federal Reserve Bank of New York Economic Policy Review (special issue), August 1997.

4. The Bundesbank every year derives its money growth targets from an announced "normative rate of price increase," i.e. the inflation level the Bundesbank believes appropriate for that coming year. Since 1986, that level has been 2%. This is not an empty exercise, however; after the second oil shock, the Bundesbank did let the "unavoidable price increase," as it was then called, rise up to 8% before bringing it back down to 2% by 1984. For details, see Thomas Laubach and Adam Posen, "Disciplined Discretion: Lessons from the German and Swiss Monetary Frameworks," Princeton Essays in International Finance, forthcoming.

5. For extensive assessments of the effects of inflation targeting along many avenues, see Thomas Laubach and Adam Posen, "Some Comparative Evidence on the Effectiveness of Inflation Targeting," Federal Reserve Bank of New York Research Paper No. 97-14, and Alvaro Almeida and Charles Goodhart, "Does the Adoption of Inflation Targets Affect Central Bank Behaviour?", mimeo, London School of Economics, July 1996.

6. September 1992, the month of the ERM crisis, is deliberately left out because of the spike in interest rates which would distort the data. The second period considered for Canada is actually January 1991-December 1996, the time since inflation target adoption.

7. Alan Greenspan, "Remarks at the 15th Anniversary Conference of the Center for Economic Policy Research," Stanford University, mimeo, September 5, 1997.

8. For a discussion of the nature of monetary policy rules, see Benjamin Friedman and Kenneth Kuttner, "A Price Target for U.S. Monetary Policy? Lessons from the Experience with Money Growth Targets," Brookings Papers on Economic Activity, 1996:1.

9. See Laubach and Posen, "Disciplined Discretion," op cit, and Adam Posen, "Lessons from the Bundesbank on the occasion of its 40th (and second-to-last?) birthday," Institute for International Economics Working Paper No. 97-4, September 1997.

10. See Mishkin and Posen, op cit.

11. This data is from Alex Cukierman, Central Bank Strategies, Credibility, and Independence, MIT Press, 1992.

12. Adam Posen, "Declarations are not Enough: Financial Sector Sources of Central Bank Independence," NBER Macroeconomics Annual 1995, MIT Press, and Adam Posen, "Why Central Bank Independence Does Not Cause Low Inflation: There is No Institutional Fix for Politics," Finance and the International Economy: 7 (The Amex Bank Review Awards), Oxford University Press, 1993.

13. See Mishkin and Posen, op cit, for detailed accounts of these developments.

Table 1: Inflation and Interest Rates of G-7 Countries

^a Inflation rates are average monthly 12-month changes in the CPI
^b Interest rates are average monthly short-term money market rates
^c Second series dated October 1992 - September, 1996
^d Second series dated January 1991 - December 1996

Source: International Monetary Fund: International Financial Statistics.

Figure 1

y = 0.2472x + 3.1292
R2 = 0.0058

y = 4.3088x + 3.8638
R2 = 0.0982
Source: International Monetary Fund, International Financial Statistics; and Cukierman (1992)