FRBSF Economic Letter
2000-08; March 17, 2000
 Economic
Letter Index
Uncertainty and Monetary Policy
Uncertainty is pervasive in the policy environment the Federal Reserve
faces as it strives to promote economic stability and low inflation. The
economic situation in the U.S. today shows that, even in the best of times,
making monetary policy isn't easy. Is the low unemployment rate a signal
inflation is likely to rise soon? Is the stock market boom that seems
to be fueling consumption simply a bubble destined to burst? Has the economy
entered a period of faster productivity growth? Will a quarter point rise
in the funds rate be enough to achieve the Fed's objectives? The Fed's
policymaking committee, the Federal Open Market Committee (FOMC), must
deal with these and other uncertainties in reaching its policy decisions.
In recent years, economists have studied extensively the impact of uncertainty
on the conduct of policy. This research has focused primarily on how our
imperfect knowledge of the economy affects policy choices, but economists
also have studied how public uncertainty about the Fed's intentions can
influence economic developments. This Economic Letter discusses
some of the kinds of uncertainties and their implications for monetary
policy.
Going with your best
guess
One form of uncertainty that affects policy choices is imperfect knowledge
of the economy. A good example of this is the concept of the natural rate
of unemployment--the level of measured unemployment that is associated
with a balance between demand and supply in labor markets with steady
inflation. As actual unemployment falls below this level, the tight labor
market leads to wage growth in excess of productivity growth. Firms recoup
these higher labor costs by raising prices, which boosts inflation. When
unemployment rises above the natural rate, wage growth falls below productivity
growth and inflation declines. A central bank with a mandate to keep inflation
low and stable will generally want to tighten policy to cool the economy
whenever unemployment dips below the natural rate and loosen policy to
promote faster growth whenever it rises above the natural rate.
Until the late 1990s, most economists estimated that the natural rate
of unemployment was somewhere in the range of 5.5% to 6%. But estimates
are the best we can do--we cannot measure the natural rate directly; so
our imprecise knowledge about its value is one important source of the
uncertainty policymakers face.
If earlier estimates of the natural rate are right, our current low unemployment
rate of around 4% is a sign that inflation is likely to start rising.
The appropriate response of the Fed in this case is to tighten policy
to prevent inflation from rising. But suppose the "new economy" alternative
is correct and the economy's natural rate of unemployment is no longer
in the 5.5% to 6% range, but has fallen, perhaps to 4%. In this case,
there is no need for the Fed to tighten policy. Doing so would needlessly
slow the economy and lead to an increase in unemployment.
How should this uncertainty about the natural rate affect the Fed's decisions?
One approach says it shouldn't. The Fed should just form its best estimate
of the natural rate and then ignore the uncertainty surrounding that estimate--it
should act as if it knew the natural rate with certainty. This approach--called
certainty equivalence--implies that the Fed should act the same way if
it believes the natural rate could be anywhere between 4% and 5% as it
would if it believed the natural rate could be anywhere between 2% and
7%. Even though there is much more uncertainty in the latter case, in
both cases the certainty equivalence approach says the Fed should act
as if it knows with certainty that the natural rate is 4.5%.
Most popular discussions assume certainty equivalence is the way to deal
with uncertainty--discussions focus on what our best guesses are for the
natural rate, future inflation, or the economy's trend growth rate while
ignoring how much uncertainty surrounds these estimates.
Certainty equivalence allows the policymaking process to take place in
two stages. In the first stage, the Fed's staff can prepare their best
forecasts of economic conditions and report these to the members of the
FOMC. The members of the FOMC can then make their decisions, acting as
if they were completely certain about economic conditions (Clarida, Gali,
and Gertler 1999). This approach provides a clean separation between the
preparation of forecasts and the making of decisions. But it's an approach
that incorporates an important assumption about the costs of being wrong:
certainty equivalence makes sense only if the costs of being wrong do
not depend on whether one overestimates or underestimates the natural
rate. If the staff tells the FOMC that the best estimate of the natural
rate is, say, 4.5%, the FOMC can ignore the uncertainty surrounding that
estimate only if they will be equally unhappy if the true natural rate
turns out to be 7% or if it turns out to be 2%.
Are costs symmetric?
In many situations, the costs of making a mistake are not symmetric.
Consider your actions when you have a plane to catch and the flight is
scheduled to depart at 6 p.m. The time it takes to get to the airport
is uncertain because of possible traffic delays. Arriving twenty minutes
early does not have the same consequences as arriving even five minutes
late. In one case, you have a few extra minutes to kill at the airport;
in the other, you may miss an important meeting or have to stay over an
extra night. In this case, most people leave a bit earlier for the airport
than they believe is absolutely necessary--if your best guess is that
it takes an hour to get to the airport, you may decide to leave yourself
an hour and fifteen minutes to get there. Uncertainty about the drive
to the airport affects your decision about when to leave.
How does this apply to the Fed's decision? The Fed must assess the costs
that would arise if it underestimated or overestimated the natural rate.
Consider the costs of underestimating the natural rate. Suppose the Fed's
best guess of the natural rate is that it has fallen to around 4%, equal
to the current actual unemployment rate. What are the costs if the Fed
has underestimated the natural rate versus overestimating the natural
rate? If the natural rate is really still much higher than 4%, a failure
to tighten policy now risks a significant increase in inflation. If the
natural rate has fallen even lower, to, say, 3.5%, then a tightening leads
in the short run to unnecessarily high unemployment and its associated
lost income. Are these two outcomes equally costly? That is the question
policymakers need to grapple with. Only if the costs of errors are symmetric
can policymakers ignore the degree of their uncertainty about the natural
rate.
Cautionary policy
While uncertainty about the natural rate provides a useful illustration
of one type of uncertainty policymakers face, it isn't the only type the
Fed faces. The Fed also cannot be sure of exactly how its actions will
affect the economy. Milton Friedman described one aspect of this uncertainty
in terms of the "long and variable" lags between a change in policy and
the time at which it finally influences the economy. Will a quarter point
increase in the funds rate, such as the FOMC decided upon in February,
be sufficient to prevent the economy from overheating? Or will that increase
have too small an impact, so that further rate increases will be needed?
And how does this sort of uncertainty affect the implementation of policy?
Over 30 years ago, William Brainard of Yale University (Brainard 1967)
provided an answer. Imagine you are driving a car. Sometimes
when you turn the steering wheel, the car barely responds; at other times,
the slightest adjustment of the steering wheel produces a sharp change
of direction. How should you drive? Very cautiously, according to Brainard.
When a policy action's impact on output and inflation is hard to predict,
it becomes optimal to respond cautiously to new developments.
Alan Blinder, a former Vice Chair of the Board of Governors, has argued
(Blinder 1998) that Brainard's result accorded closely
with what Blinder felt was a reasonable approach to policy--calculate
the best change in the federal funds target as if one faced no uncertainty,
and then change the funds rate a little less.
Can it pay to be aggressive?
Uncertainty can lead to caution. Can it ever lead to more aggressive
policy actions? Recent research has suggested the answer to this question
is yes.
Suppose you are driving along a narrow ridge, buffeted by gusting winds,
and the linkage between movements in your steering wheel and the car's
response is hard to predict. A strong blast of wind could push you over
the edge. You don't know where the next gust will come from. Does it make
sense to respond cautiously if the wind suddenly pushes to the left? No--respond
too cautiously and you might end up over the edge if the car fails to
respond when you turn the wheel slightly. Better to risk over-steering
than to find yourself over the edge.
In this example, uncertainty about the response of the car to your attempts
to steer would lead you to respond strongly rather than cautiously. In
doing so, you might have the best chance of avoiding a really bad outcome.
Similarly in monetary policy, aggressive interest rate movements might
be called for in order to avoid a major economic disruption (Giannoni
1999).
Can policy cause uncertainty?
The examples so far have involved uncertainty about the economy or about
the impact on the economy of policy actions. But policy decisions themselves
also can contribute to economic uncertainty. During the 1980s, when central
banks in many countries were striving to reduce inflation, the public's
expectations about inflation often remained stubbornly above the inflation
rates the central banks claimed they were going to achieve. The public
was uncertain about whether policymakers really were committed to getting
inflation down. This uncertainty led the public to continue to expect
inflation, and that raised the unemployment costs of bringing inflation
down.
In today's low-inflation environment, people still face uncertainty about
policymakers' actions. Will the Fed wait until inflation starts to rise
before tightening? Will it act preemptively to ensure low inflation is
maintained? Will recent interest rate increases be followed by further
ones over the next few months? To reduce this type of uncertainty, many
central banks have begun to issue detailed inflation forecasts or to announce
inflation targets publicly. The intention is to provide more information
about policy objectives to the public.
Conclusions
Economists have studied in detail the implications of uncertainty for
the conduct of monetary policy. Unfortunately, uncertainty comes in more
than one flavor, and the appropriate policy response depends on the costs
it imposes. In the certainty equivalence approach, uncertainty can be
ignored. At the same time, in some cases, policy should respond less strongly,
while in others, it should respond more strongly. And finally, some of
the uncertainty in the economy may even be traced to the public's uncertainty
about the future course of policy. Uncertainty is pervasive, but there
is no one answer about how to deal with it.
Carl E. Walsh
Professor of Economics, UC Santa Cruz
Visiting Scholar, FRBSF
References
Blinder, Alan. 1998. Central Banking in Theory and Practice.
Cambridge: MIT Press.
Brainard, William. 1967. "Uncertainty and the Effectiveness of Policy."
American Economic Review 57, pp. 411-425.
Clarida, Richard, Jordi Galí, and Mark Gertler. 1999. "The Science
of Monetary Policy: A New Keynesian Perspective." Journal of Economic
Literature (December) pp. 1,661-1,707.
Giannoni, Marc. 1999. "Does Model Uncertainty Justify Caution? Robust
Optimal Monetary Policy in a Forward-Looking Model." Mimeo. Princeton
University, November.
Opinions expressed in this newsletter do not necessarily reflect
the views of the management of the Federal Reserve Bank of San Francisco
or of the Board of Governors of the Federal Reserve System. Editorial
comments may be addressed to the editor or to the author. Mail comments
to:
Research Department
Federal Reserve Bank of San Francisco
P.O. Box 7702
San Francisco, CA 94120
|
