The Evolution of Thinking about Federal Reserve Monetary Policy

Edward M. Gramlich, The Urban Institute and the University of Michigan

The Federal Reserve was formed in 1914, just as the international gold standard was breaking down. The gold standard enforced its own constraints on economic policies and events in different countries. Money would flow into or out of a country through its trade deficits, which ultimately depended on the balance of aggregate demand and supply. But the gold standard adjustment mechanisms could be harsh, and in any event the system was collapsing, leading to an economic requirement for some other mechanism to stabilize demand and constrain policy.

While economists might think it logical to substitute one constraining mechanism, the Federal Reserve, for another, the gold standard, the truth is a bit more complicated. On one side, the more relevant cause of the birth of the Fed was a series of banking panics in the late nineteenth and early twentieth centuries – Allan Meltzer reports that the economy was in recession more than half the time between 1895 and 1912. These led to a political demand for a more stable monetary system, with enough currency to finance “the needs of trade.” On the other side, at least until World War II, the Fed was not at all interested in replacing the gold standard. Indeed, a large share of its efforts then were devoted to restoring the gold standard.

Exactly what is meant by financing the needs of trade is not clear now, and certainly was not clear to generations of earlier monetary policymakers. Meltzer gives a detailed description of how this standard has been interpreted over time, and how the generally understood mission of the Fed has evolved. I will focus on more recent developments in the evolution of thought about how to conduct monetary policy.

I will not attempt a detailed historical examination of actual Federal Reserve policies. But it is hard to talk about the evolution of policy thinking without some reference to history, and in particular to episodes where Fed policy was less than optimal. Although the Fed has evolved into a modern and highly effective organization, there are unfortunately many historical episodes of less than optimal policy – the recession of 1920-21, the Great Depression, the recession of 1937-38, excessive tightening in the early 1960s, and the Great Inflation of the late 1970s, to name a few. Meltzer is encyclopedic on the early mistakes, John Taylor on some of the more recent mistakes. Exactly how much the Fed has learned from these mistakes can be debated – generally the people who made the mistakes do not seem to have changed their opinions much. But there has nevertheless been learning in the aggregate, and the modern day discussions of Fed policy are much further along in guarding against these previous mistakes.

One issue that should be stressed at the outset is that thinking about how to conduct monetary policy does not, and has never, taken place entirely within the Fed. First off, other countries have faced similar issues in the conduct of their own monetary policies. Monetary authorities in all countries are working on these issues more or less simultaneously. For example, one of the most interesting present day ideas about the conduct of monetary policy, inflation targeting, arose primarily in other countries. United States economists have discussed inflation targeting, but it has never been official doctrine here. Secondly, unlike many other government agencies, there is an unusual amount of contact between the Fed and academic economists. Many academic economists spend part of their careers either on the Fed staff or as governors, indeed today even as the chairman of the Fed. Academic economists are routinely invited in to give seminars and discuss monetary issues, and governors, staff, and academic economists regularly interact at conferences, discussing issues of common concern. Hence the evolution of thinking about monetary policy at the Fed largely reflects the same evolution in the broader academic community, here and around the world.

To a large degree the evolution of thinking about monetary policy reflects the search for a policy rule. Perhaps not so much in the early days, but in the postwar period many participants in the policy process have regularly invoked the need for some theoretical guidance or standard on how to conduct policy – let’s bring some economic variable or set of variables into balance with some goal or target. Extremely rigid policy rules have never had much success in the policy world – there always seem to be conditions that might require some departure from almost any rigid policy rule. But there has still been a search for a general rule as an intellectual benchmark, a framework for thinking about and determining appropriate policy. A world totally without rules is a world without such a framework, where policy is likely to be without any sort of anchor whatever. So even though virtually nobody is talking about slavish adherence to a rigid policy rule, many, many policy-makers over the years have been searching for some practical guide on how to conduct monetary policy under uncertain economic conditions.

These days there are two prominent policy rules or frameworks that have come to the fore — inflation targeting and the short-term interest rate rule of John Taylor. The paper will focus on these two approaches, describing how they evolved and how they deal with the various types of policy challenges that have faced the Fed, historically and now. I also show how these two approaches, despite their apparent differences, may ultimately be complementary, and may be used together to generate improved economic outcomes.

Institutional Evolution

The Fed came into being largely as a political response to the series of banking crises in America and the world in the late nineteenth and early twentieth centuries. The predominant mission of the Fed was then stated as financing the needs of trade. That particular mission is pretty vague – lots of monetary approaches to lots of problems could be described as financing the needs of trade. Should trade be financed with stable prices or not, with trade deficits or not, with full employment or not?

Without going into the detailed changes, over time the mission statement for the Fed has become much more precise. It is now focused on keeping prices stable, keeping interest rates low, and keeping employment high. But even this more specific interpretation needs translation.

If prices are stable, the so-called inflation premium need not be built into nominal interest rates, so the first two goals actually collapse into one – keeping prices stable. And as has been recognized in standard macroeconomics for some decades now, keeping employment high can be interpreted ambiguously – are we talking about the short run, intermediate run, or long run? The vast majority of economists now believe that keeping employment too high in the long run leads to explosive inflation and the impossibility of price stability. Hence the employment goal is now generally interpreted as referring to sustainable long-term rates of employment, or unemployment. This means that the mission statement of the Fed, often described as its dual mandate, can now really be described as finding the optimal point on the long-run inflation-unemployment frontier. Of course, in an era of vertical long-run inflation-unemployment relationships, the dual mandate really means keeping inflation within a bound of price stability near the long-term sustainable rate of unemployment. Virtually every monetary policymaker in the past two decades has believed and said something like the above.

While in this sense the Fed does not control unemployment in the long run, it can have some influence in the short run. Suppose unemployment is threatening to increase, and this unemployment can be successfully combated without a rise in inflation. In such cases the dual mandate would call on the Fed to fight unemployment, or smooth the business cycle, this time confident that such could be done without increasing inflation. Hence there is still some role for smoothing cycles in unemployment and output, though the primary goal remains price stability.

On the institutional side, there has been substantial change. The Federal Reserve Board initially consisted of five appointed members, joined by the Secretary of the Treasury and the Comptroller of the Currency as ex officio members. The Treasury Secretary chaired the board. There were then and now twelve reserve banks, and in the early days the banks had the freedom to go along with or not the board recommendations. This freedom to disagree disappeared in the early years of the Fed, for example when the Fed discovered that capital mobility would trump its ability to maintain different discount rates in different Federal Reserve districts. On these and other matters, differences between the policies of different reserve banks, and between the reserve banks and the Fed, have narrowed appreciably over the years. Today the banks all work in concert.

As later experience would illustrate, there was also a potential conflict of interest in having the Secretary of the Treasury represented on the Fed board. The Treasury’s interest is in keeping interest rates down, the board’s interest at least should be in using interest rates to stabilizing the economy. The explicit connection between the Treasury and the Fed was severed when the Federal Reserve Act was modified in 1935. The 1935 Act set up the modern structure where the Federal Open Market Committee (FOMC) consisted of the chair, six governors, and five of twelve reserve bank presidents, who vote on a rotating basis. While the Treasury Secretary has not been explicitly represented on the FOMC since 1935, both in the 1930s and during World War II the interests of the Treasury have been very powerfully represented at the Fed. In part, this influence may have reflected the relative political power of the two agencies – at the time, the Treasury was by far the dominant institution. But in part the important influence of the Treasury may have also reflected the fact that even the Fed board members of the day did not believe that monetary policy was very effective in influencing the economy.

The Accord of 1951 formally freed the Fed from Treasury influence, and enabled it to grow into the powerful and independent agency it is now. But even then there have still been many attempts by the Treasury to influence Fed policy – some of the more graphic ones are described in Bob Woodward’s biography of Alan Greenspan.  This modern-day interference was largely put to rest by Lloyd Bentsen, Robert Rubin, and Larry Summers, Treasury Secretaries who steadfastly refused to lean on the Fed in any way, shape, or form. These days it would be very bad form indeed for a Treasury Secretary to even think about influencing Federal Reserve policy.

Another way in which there has been gradual evolution is in the matter of openness. In an age of vast flows of capital, market participants who can correctly anticipate Fed actions can make tons of money. Hence the question of whether and how the Fed should telegraph its policies has always been one of great delicacy. For most of its history the Fed’s approach has been one of secrecy – the FOMC would meet, members would vote for wording in a directive to the trading desk in New York, and everybody would clam up. Very occasionally there have been insider-trading scandals, but these have been exceedingly rare given the magnitudes at stake. It has always been pretty much a badge of honor at the Fed to be able to keep secrets.

Keeping secrets is still important, but as time has gone on the dominant thinking about openness has shifted. The Fed has become much more open recently, particularly in the last decade. One main reason arises largely from the urgings of academic economists – openness is said to be market-efficient. Markets more efficiently reflect true magnitudes if all traders, not just a select few who can read tea leaves, are aware of the Fed’s thinking. Another reason involves democracy – Fed policymakers are independent but they are entrusted with important national and international responsibilities. At a minimum they should be explaining what they are doing and why.

The upshot is that every meeting of the FOMC is now followed by a short public statement summarizing the Fed’s thinking on the economy, giving the target federal funds rate, and the vote on policy. Within three weeks of the FOMC meeting the minutes of the meeting are made public, and these are actually quite informative on the tone and content of the discussion. In five years the complete transcript is made available. The release of the FOMC statement and the early release of the minutes are institutional changes that have been made quite recently. Making the changes has generated moderate discussion, though no real controversy, within the current FOMC. But it is likely that many former members of the FOMC would view all these changes with great horror. There has been definite evolution from the standpoint of openness.

Policy Evolution

While in each of these institutional dimensions there has been a steady and more or less logical progression, there are other ways in which the evolution of Fed practice has been far less smooth. The most prominent illustration of the rocky, turbulent side of the Fed’s evolution has been in terms of actual policy.

As described by various authors, the policy record is not stellar. The first major historical challenge facing the Fed was the sharp recession of 1920-21. Real output was dropping sharply and the Fed responded by permitting a sharp drop in the money supply, surely aggravating the recession. The economy recovered from this shock and grew smoothly over the rest of the 1920s, but the mistake was repeated as another recession threatened in 1929. This time the economy did not recover and the serious economic weakness persisted, aggravated by the many bank failures at the time, which again greatly reduced the money supply. The passive and incorrect response to the Great Depression stands as one of the major historical mistakes of the Fed. And the mistake was repeated again in the late 1930s, as another recession threatened. In this episode there was a feeling that excess bank reserves cost the Fed leverage, and the Fed actually raised reserve requirements on the eve of a recession. In each of these three episodes there was a sharp contraction of the money supply at the start of the recession.

There is a huge literature devoted to parsing out how serious these mistakes were and why they were made. I will not attempt to summarize the literature, other than to note that the explanations can be grouped into at least four classes:

  • Many policymakers did not believe the Fed either could or had a responsibility to influence aggregate demand.
  • Many policymakers felt the fundamental need was to restore the gold standard, either to preserve the U.S. gold stock or to restore price levels to some historical value.
  • There was confusion over the role of bank reserves – many felt that they implied loss of leverage by the Fed and that they had to be reduced by policy measures.
  • There has always been confusion about the distinction between nominal and real interest rates, despite the fact that the academic literature of the day was very clear on the issue.

By the time we get to the postwar period many of these confusions had dissipated and policymaking was in many ways improved. By then the Keynesian Revolution had pretty much established that monetary policy could influence aggregate demand, at least under certain conditions. The gold standard had been replaced by the new Bretton Woods conventions. The role of bank reserves was better understood. But there was still confusion between nominal and real interest rates, and this confusion likely contributed to the outstanding postwar mistake, allowing the Great Inflation of the 1960s and 1970s to get started.

In the 1960s aggregate demand was high and rising, as the nation tried to finance the Vietnam War without a permanent tax increase. Unemployment rates were at postwar lows, and inflation was beginning to accelerate, just as the textbooks of the time were preaching. The Fed did permit nominal interest rates to rise some, but not enough that real interest rates rose in response to the excess demand pressure. Inflation continued to heat up, and it became almost a two-decade-long project to bring inflation back under control. To this day scholars debate whether overly expansionary Fed policies can be attributed to the nominal-real interest rate confusion, a misunderstanding of the inflation-unemployment tradeoff issue, or to a faulty estimate of the feasible long-term maximum rate of unemployment. Whatever the case, monetary policy was overly expansionary, inflation heated up, and the Fed failed to accomplish its key objective of preserving price stability.

While there have been other, less significant, policy mistakes, these episodes are the most prominent. Most modern day Fed policymakers have these episodes clearly imprinted in their thinking, and today’s policy decisions are still framed in terms of avoiding these mistakes, as are the policy rules I will discuss. But the focus on these historical mistakes should not obscure the good news. Through it all, monetary policy making has improved dramatically in recent years. Inflation has fallen, in the United States and elsewhere. In the last decade most developed countries have had inflation rates in the neighborhood of true price stability (more on this below), and even emerging market economies have inflation rates that are converging on price stability. This is a huge change from the experience of thirty years ago.

What is equally noteworthy is that this drop in inflation has coincided with better output performance. The average unemployment rate has dropped in most countries, and many recent authors have pointed to the improvement in aggregate output stability measures in recent years, in many countries around the world. There could be various explanations for these improvements in macroeconomic performance, but surely improved central bank policymaking has been essential, in this country and around the world. The world’s central banks must be doing something right.

In this important sense, there has also been policymaking evolution at the Fed, and at other central banks. The process has perhaps been less regular than in other areas, because there are these historical mistakes. But there is also a sense in which policymakers now understand former mistakes and are less likely to make at least these same mistakes in the future. Thinking about the two policy innovations I will discuss has played, and should continue to play, an important role in this process.

Inflation Targeting

Inflation targeting began in New Zealand in the early 1990s, and rapidly spread through the former British Commonwealth and to other countries. New Zealand, Australia, Canada, and the United Kingdom have been running successful inflation targeting regimes for many years now, and many countries in Scandinavia, Eastern Europe, Latin America, and Asia have followed suit. Altogether there are probably twenty or thirty countries that now practice some form of inflation targeting.

The prototype inflation targeting regime features a central bank that is independent of the Administration and legislature, but with an inflation target that is politically adopted by these bodies. This inflation target might aim for some low and stable rate of inflation as a permanent target, with the central bank given freedom to do what it takes to hit the target. While the central bank has the freedom to do what it takes, it is also held accountable. Generally it must provide inflation reports that forecast inflation and explain what the central bank plans to do to keep inflation on target.

One attractive feature of inflation targeting is its direct focus on the ultimate objective of monetary policy. For years academic economists have argued about whether the Fed, or other central banks, should keep the money supply, interest rates, or free banking reserves on target – each approach reduces instability in some variables and increases instability in other variables. But in some fundamental sense the appropriate long term goal for central bank monetary policy is to maintain the value of the currency, which in turn means price stability, or minimization of both the rate of inflation and the variability of this rate. If that is the goal, why not just ask the central bank to maintain price stability, and leave it up to the central bank to accomplish the objective in the most effective way? This is the straightforward approach of inflation targeting.

Inflation targeting has been a popular topic among academic economists, and there have been many quantitative evaluations. In the countries with the longest regimes, New Zealand, Australia, the United Kingdom, Canada, there has been a clear reduction in inflation from the time when the country adopted inflation targeting, a clear reduction in the variance of inflation, and a clear reduction in the inflation premium built into nominal bond yields. As mentioned previously, this was generally accomplished without output sacrifice, and indeed often an improvement in output growth and stability. Inflation targeting remains wildly popular in all of these countries – there would be virtually no political sentiment for dropping it.

Many countries in Europe, Latin America, and Asia have taken these lessons to heart and now operate their own inflation targeting regimes. While there is not as much historical experience, inflation targeting seems to be just as popular in these countries. Inflation has been reduced, and other economic outcomes are either no worse or better. In particular, many emerging market countries are very satisfied with their inflation targeting regimes. These countries have traditionally had a big problem establishing policy credibility – their past history has often featured hyperinflation, default, currency depreciations, and other events viewed with extreme distaste by international lenders. In many of these countries the central bank now targets inflation, keeping that at low levels, and lets the currency float, avoiding or moderating foreign exchange crises. The countries have needed an anchor for their monetary policy, and inflation targeting has provided that.

Of course, there is one prominent country that has not adopted any form of an inflation targeting regime, the United States. Just as the literature on inflation targeting is vast, the American literature on whether or not to adopt inflation targeting is equally vast. To keep things newsworthy, prominent proponents of inflation targeting have been Ben Bernanke, current chair of the Fed, and Rick Mishkin, a current governor. Prominent opponents have been Alan Greenspan, past chair, Don Kohn, current vice chair, and Roger Ferguson, past vice chair. Other current and past FOMC members have also been more or less split on the issue, some strongly for inflation targeting, some against, and some undecided.

Without summarizing all the issues in this debate, I give just one relevant empirical fact. While inflation targeting seems to have been quite successful in other countries, as measured by a comparison of inflation now and when inflation targeting was adopted, this particular comparison does not work for the United States. The United States too reduced its inflation rate, its variance of inflation, its inflation premium in long-term bonds, and improved its output performance over the past three decades.  But the United States, really the Fed, did all this without inflation targeting. This leads to the suspicion that what matters is the commitment to stop inflation, not the inflation targeting regime as such. There would also be a number of potential political issues in establishing an inflation targeting regime, and possibly inefficient limits on policy. The Fed might take a middle ground and just announce its long run inflation target, or target range, but even this step is controversial. This whole matter is likely to be debated for some time.

A final question relevant to inflation targeting is what the target should, in fact, be. A superficial answer would be zero – stable prices means zero inflation, period. But as policymakers have gotten used to the existence of stable prices, more and more are of the opinion that zero is too low. One reason is an apparent upward bias in the official inflation numbers. Whenever new goods are introduced, and that is often these days, one has to compute a hypothetical price for that same good in a previous period — otherwise the measurement of inflation will be biased. Such hypothetical prices have proven to be very difficult to estimate, and the result is that most economists figure that price indices have an upward bias of from a half point to a point a year.

A second reason why zero may be too low a target involves macroeconomics. When aggregate prices begin falling, the positive premium that is normally tacked on to real interest rates to determine nominal interest rates becomes negative. Hence nominal rates are below real rates. And if real rates are low to begin with, nominal rates could fall to zero, which is as low as they can fall. As nominal rates become “stuck” at zero, they cannot go lower and the central bank can no longer stabilize the economy by lowering interest rates. A country could fall into a deflationary spiral, with the central bank able to do nothing about it – arguably this has happened in Japan in recent years. There is no proven solution to this problem, though a number of measures have been suggested. But most likely the best idea is to avoid falling prices in the first place. And that can be done by targeting for a normal rate of inflation slightly above zero.

The upshot is that rather than target for zero, most monetary policy makers prefer to shoot for a low positive measured inflation rate that is still consistent with realistic, practical price stability. Hence foreign central banks that target inflation usually have a target in the neighborhood of two percent. U.S. policymakers who talk on the issue generally say they are shooting for between one and two percent. Alan Greenspan used to aim for an inflation rate low enough that it did not affect the behavior of economic actors. The approaches differ, but there seems to be general agreement that in targeting inflation, the appropriate target is low positive, not zero.

The Taylor Rule

Another interesting approach to the conduct of monetary policy is the Taylor Rule. Whereas inflation targeting is an approach that has more or less swept the world, excepting a large country in North America, the Taylor Rule is pretty much the brainchild of one particular academic economist, John Taylor. Taylor invented the Taylor Rule, he has written most of the persuasive advocacy pieces, and he is still the one to consult in fine-tuning the rule. The Taylor Rule is usually expressed as a simple equation, but that equation is really a reduced form of the way the economy is supposed to work in the presence or absence of a central bank. In an economy with constant money growth, disproportionate rises in aggregate demand raise the demand for money relative to supply and push up interest rates. This movement of interest rates can be described in terms of the response of short-term interest rates to inflation and the gap between aggregate demand and supply. Similarly with a gold standard. Disproportionate rises in demand would also lead to trade deficits and a reduction in the supply of credit: again the response of interest rates can be described in terms of their response to inflation and the gap between demand and supply. The same is true in an economy with a central bank where that central bank is leaning against the wind. Whatever the underlying macroeconomic circumstances, short-term interest rates will respond regularly to inflation and the economy’s overall utilization rate. The coefficients in the equation will vary according to the economic regime, but the basic structure of the equation will not vary.

In specific terms, the Taylor Rule can be written as:

1) i = p + ay + b(p – p*) + r*

where i is the short-term nominal interest rate, p is the inflation rate, y is the percentage deviation of real output from trend, p* is the target, or desired, inflation rate, and r* is the real short-term interest rate in equilibrium. In implementing the equation, a and b can be interpreted as coefficients and p* and r* are generally slow-moving, and often taken as constant for particular analyses. If they are taken as constant, they can be treated as part of the equation intercept. The overall equation intercept can then be expressed as r* – bp*, and the critical slope coefficients are a and (1 + b). For the economy, with or without a central bank, to be stable, both a and b should be positive. When a is positive, the central bank, or the economy, is responding in a stabilizing way to output shocks. When b is positive, (1 + b) exceeds unity, which means that nominal interest rates change more than inflation, or that real interest rates respond to inflation in a stabilizing manner.

Taylor’s own work usually finds that both a and b have been in the neighborhood of .5 in recent years. But he has also estimated the equation in other years when the coefficients were not so stabilizing.

Taylor has done simulations comparing the actual path of the federal funds rate to rules based on equation 1), assuming a is either .5 or 1.0 and b is .5. Basically he finds that the actual funds rate was well below his calculated optimal funds rate in the late 1960s. This signifies overly expansionary monetary policy, and in his view the overly expansionary monetary policy allowed the big postwar inflation to develop momentum. He also fits the equation to this period and gets a negative value of b, implying, as said above, that the Fed let nominal rates rise but allowed real interest rates to fall as inflation began to heat up – destabilizing policy. This problem has been corrected more recently – now equations fit to actual Fed behavior find coefficients a and b clearly positive. There is even some evidence that these coefficients are increasing over time, suggesting that the Fed is becoming more responsive to shocks in output and inflation.

It is possible for monetary policymakers to conduct policy in terms of this simple rule. They can compare actual funds rates with those given by the rule, they can estimate coefficients based on their own actions to determine if their policy responses are stabilizing, and they can simulate responses under various circumstances. For the central bank to be stabilizing the coefficients a and b do have to be positive, and for reasonable ranges of these coefficients, higher coefficients are more stabilizing than lower coefficients.

Complementarity

Inflation targeting and the Taylor Rule seem like very different approaches to policy. One focuses on the result of policy, inflation, and tries to keep the rate of inflation close to some target level – generally a low positive rate. The other is simply a general description of how short-term interest rates respond to inflation and resource utilization. It shows how a central bank can avoid destabilizing policy responses, but it does not show the central bank what inflation rate to shoot for.

These differences turn out to be convenient. Policymakers can use the Taylor Rule approach to shoot for a low positive rate of inflation, a high positive rate of inflation, or a negative rate of inflation. The Taylor Rule simply gives guidance on how to do it, and how to avoid destabilizing behavior. On the other side, policymakers can target inflation with a Taylor Rule, or with other rules that could be devised. Inflation targeting is simply a goal, without any real guidance on how to achieve the goal.

Since inflation targeting and the Taylor Rule do different things, there is no reason why decisions on which policies to use, or rules to follow, cannot be separated. A central bank could decide not to use the Taylor Rule and not to inflation target. Or to use one or the other of these approaches, but not both. Or to use both. The country’s inflation target can be based on underlying normative considerations, and the Taylor Rule can provide the important guidance on how to get there – in particular, how responsive short-term rates should be to inflation and output movements.

Conclusion

However the Fed, or any other central bank, chooses to operate, there has been clear evolution in thinking about monetary policy. The central bank’s basic mission has been clarified, with low and stable inflation taking primacy. Historical puzzles about the impact of monetary policy on aggregate demand have been resolved, as has the confusion between nominal and real interest rates. It is no wonder that the actual performance of monetary policy has been greatly improved in recent years.

But pride goeth before the fall, and now is no time to get complacent. Some of the nagging historical questions about the conduct of monetary policy do seem to have been resolved, but there are still plenty of new and difficult challenges to deal with. It is still hard to forecast the economy, especially around turning points, and it is particularly hard to know when to ease and when to tighten in a world of lags and uncertainty. Some of the earlier questions may have given way, but monetary policy has always generated new challenges and should continue to do so.