View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

For release on delivery
12:45 p.m. EST
February 20, 2004

The Great Moderation

Remarks by
Ben S. Bernanke
Board of Governors of the Federal Reserve System
at the
Meetings of the Eastern Economic Association
Washington, D.C.
February 20, 2004

One of the most striking features of the economic landscape over the past twenty
years or so has been a substantial decline in macroeconomic volatility. In a recent article,
Olivier Blanchard and John Simon (2001) documented that the variability of quarterly
growth in real output (as measured by its standard deviation) has declined by half since
the mid-1980s, while the variability of quarterly inflation has declined by about two
thirds.! Several writers on the topic have dubbed this remarkable decline in the
variability of both output and inflation "the Great Moderation." Similar declines in the
volatility of output and inflation occurred at about the same time in other major industrial
countries, with the recent exception of Japan, a country that has faced a distinctive set of
economic problems in the past decade.
Reduced macroeconomic volatility has numerous benefits. Lower volatility of
inflation improves market functioning, makes economic planning easier, and reduces the
resources devoted to hedging inflation risks. Lower volatility of output tends to imply
more stable employment and a reduction in the extent of economic uncertainty
confronting households and firms. The reduction in the volatility of output is also closely
associated with the fact that recessions have become less frequent and less severe. 2

I Kim and Nelson (1999) and McConnell and Perez-Quiros (2000) were among the first
to note the reduction in the volatility of output. Kim, Nelson, and Piger (2003) show that
the reduction in the volatility of output is quite broad based, affecting many sectors and
aspects of the economy. Warnock and Warnock (2000) find a parallel decline in the
volatility of employment, especially in goods-producing sectors.
2 The United States has experienced only two relatively mild recessions since 1984,
compared with four recessions--two of them quite deep--in the fifteen years before 1984.
Indeed, according to the National Bureau of Economic Research's monthly business
cycle chronology, which covers the period since the Civil War, the 120-month expansion
of the 1990s was the longest recession-free period the United States has enjoyed, and the
92-month expansion of the 1980s was the third longest such period.


Why has macroeconomic volatility declined? Three types of explanations have
been suggested for this dramatic change; for brevity, I will refer to these classes of
explanations as structural change, improved macroeconomic policies, and good luck.
Explanations focusing on structural change suggest that changes in economic
institutions, technology, business practices, or other structural features of the economy
have improved the ability ofthe economy to absorb shocks. Some economists have
argued, for example, that improved management of business inventories, made possible
by advances in computation and communication, has reduced the amplitude of
fluctuations in inventory stocks, which in earlier decades played an important role in
cyclical fluctuations. 3 The increased depth and sophistication of financial markets,
deregulation in many industries, the shift away from manufacturing toward services, and
increased openness to trade and international capital flows are other examples of
structural changes that may have increased macroeconomic flexibility and stability.
The second class of explanations focuses on the arguably improved performance
of macroeconomic policies, particularly monetary policy. The historical pattern of
changes in the volatilities of output growth and inflation gives some credence to the idea
that better monetary policy may have been a major contributor to increased economic
stability. As Blanchard and Simon (2001) show, output volatility and inflation volatility
have had a strong tendency to move together, both in the United States and other
industrial countries. In particular, output volatility in the United States, at a high level in

3 McConnell and Perez-Quiros (2000) and Kahn, McConnell, and Perez-Quiros (2002)
make this argument. McCarthy and Zakrajsek (2003) provide an overview and
evaluation of this literature; they conclude that better inventory management has
reinforced the trend toward lower volatility but is not the ultimate cause. Willis (2003)
discusses structural changes that may have contributed to reduced variability of inflation.


the immediate postwar era, declined significantly between 1955 and 1970, a period in
which inflation volatility was low. Both output volatility and inflation volatility rose
significantly in the 1970s and early 1980s and, as I have noted, both fell sharply after
about 1984. Economists generally agree that the 1970s, the period of highest volatility in
both output and inflation, was also a period in which monetary policy performed quite
poorly, relative to both earlier and later periods (Romer and Romer, 2002).4 Few
disagree that monetary policy has played a large part in stabilizing inflation, and so the
fact that output volatility has declined in parallel with inflation volatility, both in the
United States and abroad, suggests that monetary policy may have helped moderate the
variability of output as well.
The third class of explanations suggests that the Great Moderation did not result
primarily from changes in the structure of the economy or improvements in policymaking
but occurred because the shocks hitting the economy became smaller and more
infrequent. In other words, the reduction in macroeconomic volatility we have lately
enjoyed is largely the result of good luck, not an intrinsically more stable economy or
better policies. Several prominent studies using distinct empirical approaches have
provided support for the good-luck hypothesis (Aluned, Levin, and Wilson, 2002; Stock
and Watson, 2003).
Explanations of complicated phenomena are rarely clear cut and simple, and each
of the three classes of explanations I have described probably contains elements of truth.
Nevertheless, sorting out the relative importance of these explanations is of more than

Using more formal econometric methods, Kim, Nelson, and Piger (2003) also found that
structural breaks in the volatility and persistence of inflation occurred about the same
times as the changes in output volatility.



purely historical interest. Notably, ifthe Great Moderation was largely the result of good
luck rather than a more stable economy or better policies, then we have no particular
reason to expect the relatively benign economic environment of the past twenty years to
continue. Indeed, ifthe good-luck hypothesis is true, it is entirely possible that the
variability of output growth and inflation in the United States may, at some point, return
to the levels of the 1970s. If instead the Great Moderation was the result of structural
change or improved policymaking, then the increase in stability should be more likely to
persist, assuming of course that policymakers do not forget the lessons of history.
My view is that improvements in monetary policy, though certainly not the only
factor, have probably been an important source of the Great Moderation. In particular, I
am not convinced that the decline in macroeconomic volatility of the past two decades
was primarily the result of good luck, as some have argued, though I am sure good luck
had its part to playas well. In the remainder of my remarks, I will provide some support
for the "improved-monetary-policy" explanation for the Great Moderation. I will not
spend much time on the other two classes of explanations, not because they are
uninteresting or unimportant, but because my time is limited and the structural change
and good-luck hypotheses have been extensively discussed elsewhere. s Before
proceeding, I should note that my views are not necessarily those of my colleagues on the
Board of Governors or the Federal Open Market Committee.

The Taylor Curve and the Variability Tradeoff
Let us begin by asking what economic theory has to say about the relationship of
output volatility and inflation volatility. To keep matters simple, I will make the strong


Stock and Watson (2003) provide a recent overview of the debate.


(but only temporary!) assumption that monetary policymakers have an accurate
understanding of the economy and that they choose policies to promote the best
economic performance possible, given their economic objectives. I also assume for the
moment that the structure of the economy and the distribution of economic shocks are
stable and unchanging. Under these baseline assumptions, macroeconomists have
obtained an interesting and important result. Specifically, standard economic models
imply that, in the long run, monetary policymakers can reduce the volatility of inflation

only by allowing greater volatility in output growth, and vice versa. In other words, if
monetary policies are chosen optimally and the economic structure is held constant, there
exists a long-run tradeoff between volatility in output and volatility in inflation.
The ultimate source of this long-run tradeoff is the existence of shocks to
aggregate supply. Consider the canonical example of an aggregate supply shock, a sharp
rise in oil prices caused by disruptions to foreign sources of supply. According to
conventional analysis, an increase in the price of oil raises the overall price level (a
temporary burst in inflation) while depressing output and employment. Monetary
policymakers are therefore faced with a difficult choice. If they choose to tighten policy
(raise the short-term interest rate) in order to offset the effects of the oil price shock on
the general price level, they may well succeed--but only at the cost of making the decline
in output more severe. Likewise, if monetary policymakers choose to ease in order to
mitigate the effects of the oil price shock on output, their action will exacerbate the
inflationary impact. Hence, in the standard framework, the periodic occurrence of shocks
to aggregate supply (such as oil price shocks) forces policymakers to choose between

- 6-

stabilizing output and stabilizing inflation. 6 Note that shocks to aggregate demand do not
create the same tradeoff, as offsetting an aggregate demand shock stabilizes both output
and inflation.
This apparent tradeoff between output variability and inflation variability faced by
policymakers gives rise to what has been dubbed the Taylor curve, reflecting early work
by the Stanford economist and current Undersecretary of the Treasury John B. Taylor. 7
(Taylor also originated the eponymous Taylor rule, to which I will refer later.)
Graphically, the Taylor curve depicts the menu of possible combinations of output
volatility and inflation volatility from which monetary policymakers can choose in the
long run. Figure 1 shows two examples of Taylor curves, marked TCI and TC2. In
Figure 1, volatility in output is measured on the vertical axis and volatility in inflation is
measured on the horizontal axis. As shown in the figure, Taylor curves slope downward,
reflecting the theoretical conclusion that an optimizing policymaker can choose less of
one type of volatility in the long run only by accepting more of the other. 8 A direct
implication of the Taylor curve framework is that a change in the preferences or
objectives of the central bank alone--a decision to be tougher on inflation, for example--

Strictly speaking, according to standard models, policymakers face a tradeoff between
volatility of inflation and volatility of the output gap, the difference between potential
output and actual output. If the economy's potential output grows relatively smoothly,
variability in the output gap will be closely related to variability in actual output.
7 Chatterjee (2002) provides an overview of the Taylor curve and its implications. For an
exposition by Taylor himself, see Taylor (1998).
8 The policy tradeoff between the variability of inflation and the variability of output
implied by the Taylor curve is reminiscent of an older proposition, that policymakers
could achieve a permanently higher level of output (and thus a permanently lower level of
unemployment) by accepting a permanently higher level of inflation. However, for both
theoretical and empirical reasons, this older idea of a long-run tradeoff between the levels
of inflation and output has been largely discredited, and the Taylor curve tradeoff is in
some sense its natural successor.


-7 cannot explain the Great Moderation. Indeed, in this framework, a conscious attempt by
policymakers to try to moderate the variability of inflation should lead to higher, not
lower, variability of output.
How, then, can the Great Moderation be explained? Figure 1 suggests two
possibilities. First, suppose it were the case, contrary to what we assumed in deriving the
Taylor curve, that monetary policies during the period of high macroeconomic volatility
were not optimal, perhaps because policymakers did not have an accurate understanding
of the structure of the economy or of the impact of their policy actions. Ifmonetary
policies during the late 1960s and the 1970s were sufficiently far from optimal, the result
could be a combination of output volatility and inflation volatility lying well above the
efficient frontier defined by the Taylor curve. Graphically, suppose that the true Taylor
curve is the solid curve shown in Figure 1, labeled


Then, in principle, sufficiently

well executed policies could achieve a combination of output volatility and inflation
volatility such as that represented by point B, which lies on that curve. However, less
effective policies could lead to the economic outcome represented by point A in Figure 1,
at which both output volatility and inflation volatility are higher than at point B. We can
see now how improvements in monetary policy might account for the Great Moderation,
even in the absence of any change in the structure of the economy or in the underlying
shocks. Improvements in the policy framework, in policy implementation, or in the
policymakers' understanding ofthe economy could allow the economy to move from the
inefficient point A to the efficient point B, where the volatility of both inflation and
output are more moderate.


Figure 1 can also be used to depict a second possible explanation for the Great
Moderation, which is that, rather than monetary policy having improved, the underlying
economic environment may have become more stable. Changes in the structure ofthe
economy that increased its resilience to shocks or reductions in the variance of the shocks
themselves would improve the volatility tradeoff faced by policymakers. In Figure 1, we
can imagine now that the true Taylor curve in the 1970s is given by the dashed curve,
TC I , and the actual economic outcome chosen by policymakers is point A, which lies on
TC I . Improved economic stability in the 1980s and 1990s, whether arising from
structural change or good luck, can be represented by a shift of the Taylor curve from

TC I to TC2, and the new economic outcome as determined by policy is point B. Relative
to TCI. the Taylor curve TC2 represents economic outcomes with lower volatility in
output for any given volatility of inflation, and vice versa. According to the "shifting
Taylor curve" explanation, the Great Moderation resulted not from improved practice of
monetary policy (which has always been as effective as possible, given the environment)
but rather by favorable structural change or reduced variability of economic shocks. Of
course, more complicated scenarios in which policy becomes more effective and the
underlying economic environment becomes more stable are possible and .indeed likely.
With this bit of theory as background, I will focus on two key points. First,
without claiming that monetary policy during the 1950s or in the period since 1984 has
been ideal by any means, I will try to support my view that the policies of the late 1960s
and 1970s were particularly inefficient, for reasons that I think we now understand.
Thus, as in the first scenario just discussed (represented in Figure 1 as a movement from
point A to point B), improvements in the execution of monetary policy can plausibly

- 9account for a significant part of the Great Moderation. Second, more subtly, I will argue
that some of the benefits of improved monetary policy may easily be confused with
changes in the underlying environment (that is, improvements in policy may be
incorrectly identified as shifts in the Taylor curve), increasing the risk that standard
statistical methods of analyzing this question could understate the contribution of
monetary policy to the Great Moderation.
Reaching the Taylor Curve: Improvements in the Effectiveness of Monetary Policy
Monetary policymakers face difficult challenges in their efforts to stabilize the
economy. We are uncertain about many aspects of the workings of the economy,
including the channels by which the effects of monetary policy are transmitted. We are
even uncertain about the current economic situation as economic data are received with a
lag, are typically subject to multiple revisions, and in any case can only roughly and
partially depict the underlying economic reality. Thus, in practice, monetary policy will
never achieve as much reduction in macroeconomic volatility as would be possible if our
understanding were more complete.
Nevertheless, a number of economists have argued that monetary policy during
the late 1960s and the 1970s was unusually prone to creating volatility, relative to both
earlier and later periods (DeLong, 1997; Mayer, 1998; Romer and Romer, 2002).
Economic historians have suggested that the relative inefficiency of policy during this
period arose because monetary policymakers labored under some important
misconceptions about policy and the economy. First, during this period, central bankers
seemed to have been excessively optimistic about the ability of activist monetary policies
to offset shocks to output and to deliver permanently low levels of unemployment.

- 10-

Second, monetary policymakers appeared to underestimate their own contributions to the
inflationary problems of the time, believing instead that inflation was in large part the
result of nonmonetary forces. One might say that, in terms oftheir ability to deliver good
macroeconomic outcomes, policymakers suffered from excessive "output optimism" and
"inflation pessimism."
The output optimism of the late 1960s and the 1970s had several aspects. First, at
least during the early part of that period, many economists and policymakers held the
view that policy could exploit a permanent tradeoff between inflation and unemployment,
as described by a simple Phillips curve relationship. The idea of a permanent tradeoff
opened up the beguiling possibility that, in return for accepting just a bit more inflation,
policymakers could deliver apermanentiy low rate of unemployment. This view is now
discredited, of course, on both theoretical and empirical grounds. 9 Second, estimates of
the rate of unemployment that could be sustained without igniting inflation were typically
unrealistically low, with a long-term unemployment rate of 4 percent or less often being
characterized as a modest and easily attainable objective. 10 Third, economists of the time
may have been unduly optimistic about the ability of fiscal and monetary policymakers to
eliminate short-term fluctuations in output and employment, that is, to "fine-tune" the

Friedman (1968) provided a major theoretical critique of the idea of a permanent
tradeoff. Scholars disagree about when and to what degree U.S. monetary policymakers
absorbed the lessons of Friedman's article.
10 Orphanides (2003) has emphasized the importance of poor estimates of potential output
and the closely associated concept of the natural rate of unemployment for explaining the
inflationary policies of the 1970s. He notes the difficulty that policymakers of the time
faced in distinguishing the productivity slowdown of the period from a cyclical decline in
output. Analytical support for the view that confusion between the cyclical and secular
aspects of the 1970s' slowdown had inflationary consequences is provided by Lansing
(2002) and Bullard and Eusepi (2003)

- 11 -

What I have called inflation pessimism was the increasing conviction of
policymakers in the 1960s and 1970s, as inflation rose and remained stubbornly high, that
monetary policy was an ineffective tool for controlling inflation. As emphasized in
recent work on the United States and the United Kingdom by Edward Nelson (2004),
during this period policymakers became more and more inclined to blame inflation on socalled cost-push shocks rather than on monetary forces. Cost-push shocks, in the
paradigm of the time, included diverse factors such as union wage pressures, price
increases by oligopolistic firms, and increases in the prices of commodities such as oil
and beef brought about by adverse changes in supply conditions. For the purpose of
understanding the upward trend in inflation, however, the most salient attribute of costpush shocks was that they were putatively out ofthe control of the monetary
The combination of output optimism and inflation pessimism during the latter part
of the 1960s and the 1970s was a recipe for high volatility in output and inflation--that is,
a set of outcomes well away from the efficient frontier represented by the economy's
Taylor curve. Notably, the belief in a long-run tradeoff between output and inflation,
together with an unrealistically low assessment of the sustainable rate of unemployment,
resulted in high inflation but did not deliver the expected payoff in terms of higher output
and employment. Moreover, the Fed's periodic attempts to rein in surging inflation led to
a pattern of "go-stop" policies, in which swings in policy from ease to tightness
contributed to a highly volatile real economy as well as a highly variable inflation rate.
Wage-price controls, invoked in the belief that monetary policy was ineffective against
cost-push forces, also ultimately proved destabilizing.

- 12 Monetary policymakers bemoaned the high rate of inflation in the 1970s but did
not fully appreciate their own role in its creation. Ironically, their errors in estimating the
natural rate and in ascribing inflation to nonmonetary forces were mutually reinforcing.
On the one hand, because unemployment remained well above their over-optimistic
estimates of the sustainable rate, they were inclined to attribute inflation to outside forces
(such as the actions of firms and unions) rather than to an overheated economy (Romer
and Romer, 2002; Nelson, 2004). On the other hand, the view ofpolicymakers that
exogenous forces largely drove inflation made it more difficult for them to recognize that
their estimate of the sustainable rate of unemployment was too low. Several years passed
before policymakers were finally persuaded by the evidence that sustained antiinflationary monetary policies would actually work (primiceri, 2003). As you know,
these policies were implemented successfully after 1979, beginning under Fed Chainnan
Better known than even the Taylor curve is John Taylor's famous Taylor rule, a
simple equation that has proved remarkably a rule-of-thumb description of
monetary policy (Taylor, 1993). In its basic fonn, the Taylor rule relates the Federal
Reserve's policy instrument, the overnight federal funds interest rate, to the deviations of
inflation and output from the central bank's desired levels for those variables. Estimates
of the Taylor rule for the late 1960s and the 1970s reflect the output optimism and
inflation pessimism of the period, in that researchers tend to find a weaker response of the
policy rate to inflation and (in some studies) a relatively stronger response to the output

- 13gap than in more recent periods. I I As I will shortly discuss further, an insufficiently
strong response to inflation let inflation and inflation expectations get out of control and
thus added volatility to the economy. At the same time, strong responses to what we
understand in retrospect to have been over-optimistic estimates of the output gap created
additional instability. As output optimism and inflation pessimism both waned under the
force of the data, policy responses became more appropriate and the economy more
stable. In this sense, improvements in policymakers' understanding of the economy and
the role of monetary policy allowed the economy to move closer to the Taylor curve (or,
in terms of Figure 1, to move from point A to point B).

Improved Monetary Policy or a Shifting Taylor Curve?
Improvements in monetary policy that moved the economy closer to the efficient
frontier described by the Taylor curve can account for part of the Great Moderation.
However, several empirical studies have questioned the quantitative importance of this
effect and emphasized instead shifts in the Taylor curve, brought about by structural
change or good luck. For example, in a paper presented at the Federal Reserve Bank of
Kansas City's annual Jackson Hole conference, James Stock and Mark Watson (2003)
use several alternative macroeconomic models to simulate how the economy would have
performed after 1984 if monetary policy had followed its pre-l 979 pattern. Although
inflation performance after 1984 would clearly have been worse if pre-1979 monetary

II See, for example, Judd and Rudebusch (1998), Taylor (1999), Clarida, Gali, and Gertler
(2000), Cogley and Sargent (2002), and Mehra (2002). Orphanides (2003) argues that, if
one takes account ofpolicymakers' mis-estimates of the output gap in the 1970s, the
same Taylor rule that describes policy after 1979 applies to the 1970s as well. The
debate is an important one, but it may bear more on what policymakers actually thought
they were doing--and thus on the history ofideas--then on the question of whether
monetary policy was in fact inefficient or even destabilizing during the period. There
seems to be little doubt that it was.

-14policies had been used, Stock and Watson find that output volatility would have been
little different. They conclude that improved monetary policy does not account for much
of the reduction in output volatility since the mid-1980s. Instead, noting that the variance
of the economic shocks implied by their models for the 1970s was much higher than the
variance of shocks in the more recent period, they embrace the good-luck explanation of
the Great Moderation. Interesting research by Timothy Cogley and Thomas Sargent
(2002) and by Shaghil Ahmed, Andrew Levin, and Beth Anne Wilson (2002) likewise
find a substantial reduction in the size and frequency of shocks in the more recent period,
supporting the good-luck hypothesis.
Both the structural change and good-luck explanations of the Great Moderation
are intriguing and (to reiterate) both are no doubt part of the story. However, an
unsatisfying aspect of both explanations is the difficulty of identifying changes in the
economic environment large enough and persistent enough to explain the Great
Moderation, both in the United States and abroad. In particular, it is not obvious that
economic shocks have become significantly smaller or more infrequent, as required by
the good-luck hypothesis. Tensions in the Middle East, often blamed for the oil price
shocks of the 1970s, have hardly declined in recent years, and important developments in
technology and productivity have continued to buffet the economy (albeit in a more
positive direction than in the 1970s). Nor has the international economic environment
become obviously more placid, as a series of financial crises struck various regions of the
world during the 1990s and the powerful forces of globalization have proceeded apace.
In contrast, following the adverse experience of the 1970s, changes in the practice of

- 15 monetary policy occurred around the world in similar ways and during approximately the
same period.
Certainly, stability-enhancing changes in the economic environment have
occurred in the past two decades. However, an intriguing possibility is that some of these
changes, rather than being truly exogenous, may have been induced by improved
monetary policies. That is, better monetary policies may have resulted in what appear to
be (but only appear to be) favorable shifts in the economy's Taylor curve. Here are some
examples of what I have in mind.
First, monetary policies that brought down and stabilized inflation may have led
to stabilizing changes in the structure of the economy as well, in line with the prediction
of the famous Lucas (1976) critique that economic structure depends on the policy
regime. High and unstable inflation increases the variability of relative prices and real
interest rates, for example, distorting decisions regarding consumption, capital
investment, and inventory investment, among others. Likewise, the high level,
variability, and unpredictability of inflation profoundly affected decisions regarding
financial investments and money holdings. Theories of "rational inattention" (Sims,
2003), according to which people vary the frequency with which they re-examine
economic decisions according to the underlying economic environment, imply that the
dynamic behavior of the economy would change--probably in the direction of greater
stability and persistence--in a more stable pricing environment, in which people
reconsider their economic decisions less frequently.
Second, changes in monetary policy could conceivably affect the size and
frequency of shocks hitting the economy, at least as an econometrician would measure

- 16those shocks. This assertion seems odd at first, as we are used to thinking of shocks as
exogenous events, arising from "outside the model," so to speak. However,
econometricians typically do not measure shocks directly but instead infer them from
movements in macroeconomic variables that they cannot otherwise explain. Shocks in
this sense may certainly reflect the monetary regime. For example, consider the costpush shocks that played such an important role in 1970s' thinking about inflation.
Seemingly unexplained or autonomous movements in wages and prices during this
period, which analysts would have interpreted as shocks to wage and price equations,
may in fact have been the result of earlier monetary policy actions, or (more subtly) of
monetary policy actions expected by wage- and price-setters to take place in the future.

In an influential paper, Robert Barsky and Lutz Kilian (2001) analyze the oil price shocks
of the 1970s in this spirit. Barsky and Kilian provide evidence that the extraordinary
increases in nominal oil prices during the 1970s were made feasible primarily by earlier
expansionary monetary policies rather than by truly exogenous political or economic
Third, monetary policy can also affect the distribution of measured shocks by
changing the sensitivity of pricing and other economic decisions to exogenous outside
events. For example, significant movements in the price of oil and other commodities
continued to occur after 1984. However, in a low-inflation environment, with stable
inflation expectations and a general perception that firms do not have pricing power,
commodity price shocks are not passed into final goods prices to nearly the same degree
as in a looser monetary environment. As a result, a change in commodity prices of a
given size shows up as a smaller shock to output and consumer prices today than it would

-17have in the earlier period. Likewise, there is evidence that fluctuations in exchange rates
have smaller effects on domestic prices and economic activity when inflation is less
volatile and inflation expectations are stabilized (Gagnon and Ihrig, 2002; Devereux,
Engel, and Storgaard, 2003).
Fourth, changes in inflation expectations, which are ultimately the product of the
monetary policy regime, can also be confused with truly exogenous shocks in
conventional econometric analyses. Marvin Goodfriend (1993) has suggested, for
example, that insufficiently anchored inflation expectations have led to periodic
"inflation scares," in which i~flation expectations have risen in an apparently autonomous
manner. Increases in inflation expectations have the flavor of adverse aggregate supply
shocks in that they tend to increase the volatility of both inflation and output, in a
combination that depends on how strongly the monetary policymakers act to offset these
changes in expectations.
Theoretical and empirical support for the idea that inflation expectations may
become an independent source of instability has grown in recent years.


As I mentioned

earlier, a number of researchers have found that the reaction of monetary policymakers to
inflation has strengthened, in that the estimated coefficient on inflation in the Taylor rule
has risen from something less than 1 before 1979 to a value significantly greater than 1 in
the more recent period. If the policy interest rate responds to increases in inflation by
less than one-for-one (so that the real policy rate does not rise in the face of higher
inflation), economic theory tells us that inflation expectations and the economy in general
can become unstable. The problem arises from the fact that, if policymakers do not react


See Bemanke (2003, 2004) for more extensive discussions.

- 18 sufficiently aggressively to increases in inflation, spontaneously arising expectations of
increased inflation can ultimately be self-confirming and even self-reinforcing.
Incidentally, the stability requirement that the policy rate respond to inflation by more
than one-for-one is called the Taylor principle (Taylor, 1993, 1999)--the third concept
named after John Taylor that has played a role in this talk. The finding that monetary
policymakers violated the Taylor principle during the 1970s but satisfied the principle in
the past two decades would be consistent with a reduced incidence of destabilizing
expectational shocks. 13
Support for the view that inflation expectations can be an independent source of
economic volatility has also emerged from the extensive recent literature on learning and
macroeconomics (Evans and Honkopohja, 2001). For example, Athanasios Orphanides
and John C. Williams (2003a, 2003b) have studied models in which the public must learn
the central bank's underlying preferences regarding inflation by observing the actual
inflation process. 14 With learning, inflation expectations take on a more adaptive
character; in particular, high and unstable inflation will beget similar characteristics in the
pattern of inflation expectations. As Orphanides and Williams show, when inflation
expectations are poorly anchored, so that the public is highly uncertain about the long-run
rate of inflation that the central bank hopes to achieve, they can become an additional
source of volatility in the economy. An analysis that did not properly control for the

13 In a similar spirit, Stefania Albanesi, V.V. Chari, and Lawrence Christiano (2003) have
shown that when the central bank's commitment to fighting inflation is perceived to be
weak, as may have been the case during the 1970s, self-confirming increases in expected
inflation are possible and will tend to destabilize the economy.
14 See Bemanke (2004) for additional discussion.

- 19expectational effects of changes in monetary policy might incorrectly conclude that the
Taylor curve had shifted in an adverse direction.
The Great Moderation, the substantial decline in macroeconomic volatility over
the past twenty years, is a striking economic development. Whether the dominant cause
of the Great Moderation is structural change, improved monetary policy, or simply good
luck is an important question about which no consensus has yet formed. I have argued
today that improved monetary policy has likely made an important contribution not only
to the reduced volatility of inflation (which is not particularly controversial) but to the
reduced volatility of output as well. Moreover, because a change in the monetary policy
regime has pervasive effects, I have suggested that some of the effects of improved
monetary policies may have been misidentified as exogenous changes in economic
structure or in the distribution of economic shocks. This conclusion on my part makes
me optimistic for the future, because I am confident that monetary policymakers will not
forget the lessons of the 1970s.
I have put my case for better monetary policy rather forcefully today, because I

think it likely that the policy explanation for the Great Moderation deserves more credit
than it has received in the literature. However, let me close by emphasizing that the
debate remains very much open. Although I have focused on its strengths, the monetary
policy hypothesis has potential deficiencies as well. For example, although I pointed out
the difficulty that the structural change and good-luck explanations have in accounting
for the rather sharp decline in volatility after 1984, one might also question whether the
change in monetary policy regime was sufficiently sharp to have had the effects I have

-20attributed to it. IS The consistency of the monetary policy explanation with the experience
of the 1950s, a period of stable inflation during which output volatility declined but was
high in absolute tenns, deserves further investigation. Moreover, several of the channels
by which monetary policy may have affected volatility that I have mentioned today
remain largely theoretical possibilities and have not received much in the way of rigorous
empirical testing. One of my goals today was to stimulate further research on this
question. Clearly, the sources of the Great Moderation will continue to be an area for
fruitful analysis and debate.

Stock and Watson (2003) make this point. Supporting their argument, in Bemanke
(2004) I present evidence that even today inflation expectations may not be anchored as
well as we would like.

- 21 -




\,A (1970s)



........ ,


-- .. TC1


Figure 1
Monetary Policy and the Variability of Output and Inflation

- 22-

Ahmed, Shaghil, Andrew Levin, and Beth Anne Wilson (2002). "Recent U.S.
Macroeconomic Stability: Good Policies, Good Practices, or Good Luck?" Board of
Governors of the Federal Reserve System, International Finance Discussion Paper 2002730 (July).
Albanesi, Stefania, V.V. Chari, and Lawrence Christiano (2003). "Expectation Traps and
Monetary Policy," Federal Reserve Bank of Minneapolis, Research Department Staff
Report 319 (August).
Barsky, Robert, and Lutz Kilian (2001). "Do We Really Know That Oil Caused the
Great Stagflation? A Monetary Alternative," in Ben Bernanke and Kenneth Rogoff, eds.,
NBER Macroeconomics Annual, Cambridge, Mass.: MIT Press for NBER, pp. 137-82.
Bernanke, Ben (2003). '''Constrained Discretion' and Monetary Policy," before the
Money Marketeers of New York University, New York, New York, February 3.
Bernanke, Ben (2004). "Fedspeak," at the meetings of the American Economic
Association, San Diego, California, January 3.
Blanchard, Olivier, and John Simon (2001). "The Long and Large Decline in U.S.
Output Volatility," Brookings Papers on Economic Activity, 1, pp. 135-64.
Bullard, James, and Stefano Eusepi (2003). "Did the Great Inflation Occur Despite
Policymaker Commitment to a Taylor Rule?" Federal Reserve Bank of Atlanta, working
paper 2003-20 (October).
Chatterjee, Satyajit (2002). "The Taylor Curve and the Unemployment-Inflation
Tradeoff," Federal Reserve Bank of Philadelphia, Business Review, (Third Quarter), pp.
Clarida, Richard, Jordi Gali, and Mark Gertler (2000). "Monetary Policy Rules and
Macroeconomic Stability: Evidence and Some Theory," Quarterly Journal of Economics,
lIS, pp. 147-80.
Cogley, Timothy, and Thomas Sargent (2002). "Drifts and Volatilities: Monetary
Policies and Outcomes in the Post-wwn U.S.," working paper, Arizona State University
and New York University (August).
DeLong, J. Bradford (1997). "America's Peacetime Inflation: The 1970s," in Christina
Romer and David Romer, eds., Reducing Inflation: Motivation and Strategy, Chicago:
University of Chicago Press for NBER.

- 23Devereux, Michael, Charles Engel, and Peter Storgaard (2003). "Endogenous ExchangeRate Pass-through when Nominal Prices Are Set in Advance," National Bureau of
Economic Research working paper 9543 (March).
Evans, George and Seppo Honkopohja (2001). Learning and Expectations in
Macroeconomics, Princeton, N.J.: Princeton University Press.
Friedman, Milton (1968). "The Role of Monetary Policy," American Economic Review,
58 (March), pp. 1-17.
Gagnon, Joseph, and Jane Ihrig (2002). "Monetary Policy and Exchange-Rate
Passthrough," Board of Governors of the Federal Reserve System, International Finance
Discussion Paper 2001-704 (latest version March 2002).
Goodfriend, Marvin (1993). "Interest Rate Policy and the Inflation Scare Problem, 19791992," Federal Reserve Bank of Richmond, Economic Quarterly, 1 (Winter), pp. 1-23
(on the Federal Reserve Bank of Richmond web site).
Judd, John, and Glenn Rudebusch (1998). "Taylor's Rule and the Fed: 1970-1997,"
Federal Reserve Bank of San Francisco, Economic Review, 3, pp. 3-16.
Kahn, James, Margaret McConnell, and Gabriel Perez-Quiros (2002). "On the Causes of
the Increased Stability of the U.S. Economy, "Federal Reserve Bank of New York,
Economic Policy Review, 8, pp. 183-202.
Kim, Chang-Jin, and Charles Nelson (1999). "Has the U.S. Economy Become More
Stable? A Bayesian Approach Based on a Markov-Switching Model of the Business
Cycle," Review of Economics and Statistics, 81, pp. 608-16.
Kim, Chang-Jin, Charles Nelson, and Jeremy Piger (2003). "The Less Volatile U.S.
Economy: A Bayesian Investigation of Timing, Breadth, and Potential Explanations,"
Journal ofBusiness and Economic Statistics, forthcoming.
Lansing, Kevin (2002). "Learning about a Shift in Trend Output: Implications for
Monetary Policy and Inflation," Federal Reserve Bank of San Francisco, working paper
Lucas, Robert, Jr. (1976). "Econometric Policy Evaluation: A Critique," CarnegieRochester Conference Series on Public Policy, 1, pp. 19-46.
Mayer, Thomas (1998). Monetary Policy and the Great Inflation in the United States:
The Federal Reserve and the Failure ofMacroeconomic Policy, 1965-79, Cheltenham,
United Kingdom: Edward Elgar.

- 24McCarthy, Jonathan, and Egon Zakrajsek (2003). "Inventory Dynamics and Business
Cycles: What Has Changed?" Federal Reserve Bank of New York and Board of
Governors of the Federal Reserve System, working paper (June).
McConnell, Margaret, and Gabriel Perez-Quiros (2000). "Output Fluctuations in the
United States: What Has Changed since the Early 1980s?" American Economic Review,
90, pp. 1464-76.
Mehra, Yash (2002). "The Taylor Principle, Interest Rate Smoothing, and Fed Policy in
the 1970s and 1980s," Federal Reserve Bank of Richmond, working paper 02-3
Nelson, Edward (2004). "The Great Inflation of the Seventies: What Really Happened?"
Federal Reserve Bank ofSt. Louis, working paper (January).
Orphanides, Athanasios (2003). "The Quest for Prosperity Without Inflation, "Journal of
Monetary Economics, 50 (April), pp. 633-63.
Orphanides, Athanasios, and John C. Williams (2003a). "Imperfect Knowledge, Inflation
Expectations, and Monetary Policy," forthcoming in Ben Bernanke and Michael
Woodford, eds., Inflation Targeting, Chicago: University of Chicago Press for NBER.
Orphanides, Athanasios, and John C. Williams (2003b). "Inflation Scares and ForecastBased Monetary Policy," Board of Governors of the Federal Reserve System, Finance
and Economics Discussion Series 2003-41 (August).
Primiceri, Giorgio (2003). "Why Inflation Rose and Fell: Policymakers' Beliefs and
U.S. Postwar Stabilization Policy," Princeton University, working paper (November).
Romer, Christina, and David Romer (2002). "The Evolution of Economic Understanding
and Postwar Stabilization Policy," Rethinking Stabilization Policy, Federal Reserve Bank
of Kansas City, pp. 11-78.
Sims,. Christopher (2003). "Implications of Rational Inattention," Journal of Monetary
Economics, 50 (April), pp. 665-90.
Stock, James, and Mark Watson (2003). "Has the Business Cycle Changed? Evidence
and Explanations," prepared for the Federal Reserve Bank of Kansas City symposium,
"Monetary Policy and Uncertainty," Jackson Hole, Wyoming, August 28-30.
Taylor, John B. (1993). "Discretion versus Policy Rules in Practice," CarnegieRochester Conference Series on Public Policy, 39, pp. 195-214.
Taylor, John B. (1998). "Monetary Policy Guidelines for Employment and Inflation
Stability," in Benjamin Friedman and Robert Solow, eds., Inflation, Unemployment, and
Monetary Policy, Cambridge, Massachusetts: MIT Press.

- 25 )

Taylor, John B. (1999). "A Historical Analysis of Monetary Policy Rules," in John B.
Taylor, ed., Monetary Policy Rules, Chicago: University of Chicago Press for NBER, pp.
Warnock, M.V. Cacdac, and Francis Warnock (2000). "The Declining Volatility of U.S.
Employment: Was Arthur Bums Right?" Board of Governors of the Federal Reserve
System, International Finance Discussion Paper no. 677 (August).
Willis, Jonathan (2003). "Implications of Structural Changes in the U.S. Economy for
Pricing Behavior and Inflation Dynamics," Federal Reserve Bank of Kansas City,
Economic Review (First Quarter), pp. 5-27.