Before proceeding, let me remind
you that the interpretation of the economic outlook and the
judgments about the strategy of monetary policy I am presenting
here are my own views. I am not speaking for the Board of
Governors or the Federal Open Market Committee.
The Rationale for Monetary
Policy Tightening: Why Now?
The rationale for monetary policy tightening is, in my judgment,
quite straightforward and flows from two assessments about the
current state of the economy. First, aggregate demand has been
growing faster than potential aggregate supply, even allowing for
upward revisions to aggregate supply growth as a result of an
acceleration in productivity. Second, the level of output is
already at least at potential (in other words, the economy is
already at least at full employment) and, quite possibly, the
economy is operating beyond the point of sustainable capacity. In
a moment, I'll lay out some evidence for these claims. But the key
point is that, even though the rate of increase in nominal wages
and core measures of inflation do not yet signal that inflation
pressures are building, the balance of aggregate demand and
sustainable supply today and the distinct possibility that labor
and product markets will tighten further suggest an unacceptable
risk of overheating and, therefore, higher inflation in the
future. The move to tighter monetary policy, beginning in
mid-1999, seeks to rebalance aggregate demand and aggregate
supply. That promises to reduce the swing in interest rates that
would otherwise be necessary later on and provides the best
opportunity for containing inflation, extending the expansion, and
yielding an overall more stable and more favorable outcome.
There should be no doubt that
output has indeed been increasing at a faster rate than capacity.
The evidence is the consistent decline in the unemployment rate.
It has declined by about 0.4 percentage point per year over the
last four years, with the decline being at least 0.3 percentage
point each year. Based on historical regularities, this suggests
that output has been growing about � to 1 percentage point faster
than capacity, on average, over this interval. In a recent talk, I
reviewed estimates of trend growth in potential output from
surveys of professional forecasters and from model-based
forecasting firms. These estimates all fell short of the more than
4 percent average growth rate of real GDP over the past four years
and the 4� percent rate over the past two years. In the absence
of an appropriate degree of tightening, I believe that this
imbalance in the growth of demand and supply would persist.
There is somewhat greater
uncertainty about whether the economy is already operating beyond
full employment. Nevertheless, my survey also found that estimates
of the nonaccelerating-inflation rate of unemployment (NAIRU) were
consistently above the prevailing 4.1 percent unemployment rate,
with estimates of the NAIRU centered close to 5 percent.
To date, there is little
evidence that the rate of increase in unit labor costs or core
measures of consumer price inflation are rising. The core Consumer
Price Index (CPI), for example, has advanced at a 2.1 percent rate
over the past 12 months, only slightly above the lowest reading
for this expansion. This makes clear that the recent tightening of
monetary policy is pre-emptive--an attempt to prevent an
unacceptable rise in core inflation--not a reaction to direct
evidence of rising inflation. Given the lags in monetary policy,
such pre-emptive policy action is often essential to achieve
favorable outcomes.
Monetary policy has,
nevertheless, been adapting to heightened uncertainty about the
rate of growth in potential and the size of the gap between actual
and potential output. Policy has been somewhat more gradualist, in
my view, somewhat less pre-emptive than it otherwise might have
been, and somewhat more willing to tolerate increases in output
relative to consensus estimates of potential. But because there
are still limits to how fast the economy can grow without further
straining labor markets and to how low the unemployment rate can
go without triggering higher inflation, there are limits to
monetary policy's tolerance for above-trend growth and for further
labor market tightening.
Still, the question remains: Why
tighten now? The economy has been growing above trend for four
years. The unemployment rate has been falling for the last four
years. Core inflation remains well contained. In my view, three
developments suggest a greater risk of rising inflation going
forward and hence justify the timing of the recent tightening
moves. First, the cumulative decline in the unemployment rate has,
at the very least, pushed the economy closer to and, in my view,
likely beyond the point of full employment. The immediate threat
of overheating from continued above-trend growth is, therefore,
much greater today than previously. Second, the growth in demand
moved into a still higher gear in the second half of 1999 and
recent data suggest considerable momentum in domestic demand in
the first half of 2000, at the same time that the external drag
from declining net exports is expected to diminish. Third, some
beneficial influences on inflation are abating or reversing: The
effects of the earlier favorable relative-price shocks--including
the decline in non-oil import prices and the slowdown in health
care costs--are now dissipating or reversing; the temporary
disinflationary effect of the increase in trend productivity
growth is likely to diminish, a point I will elaborate on shortly;
and the recent sharp rebound in oil prices is now pushing overall
inflation higher. These developments give a sense of urgency to at
least slowing the economy to trend growth and encourage increased
vigilance in monitoring cost and price developments for signs of
rising inflation.
Productivity Shocks: Can They
Be Inflationary?
It has become abundantly clear that the star of the last several
years has been the remarkable rebound in productivity growth. The
rebound appears to be well beyond what could be explained by
cyclical developments, although it is hard to judge how
sustainable it will turn out to be. Because higher growth of
productivity translates mechanically into higher growth in
potential output, it might seem that, by raising aggregate supply
relative to aggregate demand, higher productivity growth would
make the economy less susceptible to higher inflation. So some
have been puzzled by the Federal Reserve's appreciation of the
importance of the increase in productivity growth, on the one
hand, and the apparent concern with the threat of higher
inflation, on the other hand. Some have even wondered whether the
Federal Reserve believes that a higher productivity trend might
actually be an adverse as opposed to a favorable development.
That would clearly be nonsense.
Higher productivity is unambiguously good. But the monetary policy
that accompanies a productivity shock, on the other hand, could be
good or bad. We are trying to combine the productivity shock with
a monetary policy that reaps the full benefits of higher
productivity growth and avoids turning what should be a favorable
event into one that threatens higher inflation and greater
economic instability over the longer haul.
Effects on Aggregate Demand,
Inflation, and Real Interest Rates
To conduct such a monetary policy, we must recognize that higher
productivity growth has at least three major effects on the macro
economy, in addition to its effect on sustainable growth rates. It
affects aggregate demand, inflation, and equilibrium real interest
rates.
Productivity and aggregate
demand. Higher productivity growth has apparently increased
aggregate demand through at least three channels.
First, the higher trend growth
in productivity likely reflects, in part, technological
innovations that have, in turn, resulted in new profitable
investment opportunities. In addition, this new technology can be
spread through the economy by being embedded in or by being used
in conjunction with new capital goods. In this way higher
productivity growth may spur an investment boom. This is
consistent with the extraordinary surge in investment in recent
years, as well as the concentration of investment spending in
computer and related high-technology equipment.
Second, the perceived
enhancement in profit opportunities has contributed to an increase
in business earnings expectations and, hence, higher equity
prices. This, in turn, has boosted household wealth and increased
consumer spending relative to disposable income. Higher equity
prices have also reduced the cost of capital and reinforced the
investment boom.
Finally, to the extent that
households expect higher productivity growth to continue, their
perceptions of the resulting higher path for future real
compensation would further boost consumer spending today. The
effect of increased wealth and expected future real compensation
on holdings of consumer durables and other tangible assets today
lead to accelerator-type effects that can be especially large
contributors to aggregate demand. This is consistent with the
exceptional strength in housing and in light vehicle sales in this
expansion.
Potentially, these three forces
are powerful enough to cause the growth in aggregate demand
initially to outpace the growth in aggregate supply, in the
absence of any offsetting tightening in the stance of monetary
policy. In fact, as I noted earlier, it appears that the growth in
aggregate demand has been exceeding the upward-revised estimate of
the growth in aggregate supply over the last few years. The
linkages I have described from higher productivity growth to more
robust growth in aggregate demand are one possible explanation for
this imbalance. But this explanation is less important than the
conclusion that we are on the mark in perceiving an imbalance in
the growth rates of demand and supply.
The FRB-US model, developed at
the Board of Governors, gives some credibility to the linkages
that I have highlighted between productivity and aggregate demand.
When the assumed productivity trend is raised in this model, the
near-term effect on aggregate demand exceeds the initial effect on
aggregate supply, as long as households and firms recognize that
there has been a sustained shift in growth.
The temporary disinflationary
effect of an unexpected increase in the productivity trend. In
the long run, inflation is a monetary phenomenon and independent
of the rate of productivity growth. In the short-run, on the other
hand, an unexpected increase in trend productivity growth can
yield a disinflationary bonus for a while. The source of the
disinflationary effect in the last few years has clearly not been
that aggregate supply is growing faster than aggregate demand.
Instead the source, in my view, is an asymmetric response of
nominal wages and prices to the productivity shock. If wages
adjust more slowly to an unexpected increase in productivity
growth than prices, the initial effect of higher productivity
growth will be a decline in the growth of unit labor costs and in
price inflation. In my judgment, econometric evidence supports
this notion of asymmetry. For a time, lower price increases feed
back into reduced pressure on nominal wages. But once wages fully
respond to the higher productivity trend--a process that appears
to take several years to complete--the disinflationary effect of
the productivity shock will dissipate. In the interim, for any
given unemployment rate, inflation will be lower than otherwise,
and, as a result, the economy can operate at a lower unemployment
rate without adverse inflationary consequences for a period of
time. Therefore, despite a decline in the unemployment rate, there
may be little urgency for an increase in nominal interest rates.
Productivity and the
equilibrium real interest rate. An increase in the trend rate
of productivity growth will also generally result in a higher
equilibrium real interest rate. Classical economic theory holds
that the economy's equilibrium real interest rate is determined by
the interaction of "productivity" and
"thrift." That is, the equilibrium real interest rate at
full employment has to balance saving (driven by the thrift
motive) and investment (responding to the productivity and, hence,
profitability of capital). Higher productivity growth increases
the profitability of investment, increasing the demand for
investment, and, hence, the equilibrium real interest rate that
balances saving and investment at full employment. This is really
an implication of the earlier discussion of the effect of a
productivity shock on aggregate demand relative to aggregate
supply. A variety of models suggest that, under reasonable
assumptions, the increase in the equilibrium real interest rate is
at least one-for-one with the pick-up in the economy's growth
rate.
A useful way of understanding
the effect of productivity on the balance between aggregate demand
and supply is in terms of the relationship between
"natural" and "market" rates--a relationship
Knut Wicksell put at the center of his analysis. The natural rate
is the equilibrium real interest rate that I described above. The
market rate is the actual (real) interest rate determined in
financial markets and affected by monetary policy as well as by
the balance between saving and investment. When the market rate is
below the natural rate, financial conditions are relatively
stimulative and aggregate demand will be boosted to a level above
potential aggregate supply. This analysis makes clear that the
effect of a productivity shock on the balance between aggregate
demand and supply depends critically on the monetary policy that
accompanies it.
Of course, to keep market rates
in line with the natural rate, we must make some judgments about
the degree to which the equilibrium real interest rate is affected
by an increase in productivity growth. This depends on, in
addition to the size of the increment in productivity growth, a
number of other considerations. For example, the change in the
equilibrium real interest rate will also depend on the prevailing
fiscal policy. The result that the equilibrium real rate rises is
typically derived under the assumption that tax rates are constant
and government spending remains a fixed proportion to output. If
nominal spending or even real government spending is fixed, on the
other hand, a productivity shock increases the government budget
surplus relative to GDP, at least partially offsetting the rise in
the equilibrium real interest rate.
The same principle holds with
respect to foreign output. The full effect on the equilibrium real
interest rate holds when the productivity shock symmetrically
affects foreign as well as domestic output growth. If the
productivity shock raises the growth rate only or predominantly in
the United States, the resulting international capital flows to
the United States in search of higher expected rates of return
will damp the effect of the productivity shock on our equilibrium
real interest rate.
As a result of the variety of
effects on the equilibrium real interest rate in this episode, it
takes some careful analysis to assess whether and how much the
real equilibrium interest rate may have increased. The observed
strength of aggregate demand relative to aggregate supply,
however, does importantly reinforce the judgment that the natural
rate has indeed increased.
Monetary Policy and
Productivity Shocks
The effect of a productivity shock on the balance between
aggregate demand and aggregate supply (or equivalently on the
balance between natural and market rates) depends importantly on
the conduct of monetary policy. Therefore, whether inflation turns
out to be higher, lower, or unchanged in response to the
productivity shock--especially once the initial disinflationary
impetus dissipates--depends on the conduct of monetary policy and
should not be attributed to the productivity shock itself.
In simulations with the FRB-US
model, for example, aggregate demand increases faster than
aggregate supply if the nominal federal funds rate is held
constant or if monetary policy is assumed to follow a Taylor Rule
and when households and firms are assumed to fairly quickly
recognize the sharp step-up in productivity growth. The Taylor
rule prescribes adjustments in the real federal funds rate in
response to deviations of output from potential and inflation from
some target rate. This formulation of policy yields something very
close to an unchanged nominal funds rate path during the first
couple of years following an increase in productivity growth. The
decline in inflation raises the real federal funds rate just about
as much as the Taylor Rule prescribes in light of the increase in
the output gap. Later, however, as the disinflationary force of
the productivity shock dissipates, the Taylor Rule will cause
increases in nominal and real interest rates that push output
toward potential and contain inflation.
Monetary policy could use such a
shock as an opportunity to temporarily move below the unemployment
rate sustainable in the long run while keeping the inflation rate
unchanged. Alternatively, monetary policy could convert the
temporary disinflationary effect into a permanent one. This would
be an example of "opportunistic disinflation": monetary
policy could take advantage of a disinflationary surprise to lower
inflation without a temporary increase in the unemployment rate.
If inflation is low at the time of the shock, the incentive is
great to take the benefits in a temporary decline in the
unemployment rate. Of course, policymakers could also choose a bit
of both of these options, and I would interpret monetary policy as
having produced this middle course during the last several years.
The task of monetary policy going forward is to avoid transforming
what could have been an opportunity to lower the underlying
inflation rate into a balance of risks that threatens to raise
inflation relative to the rate prevailing at the beginning of this
episode.
Is Monetary Policy Less
Effective Today?
The recent increases in the federal funds rate do not appear, to
date, to have slowed the momentum in demand growth. Indeed, as I
noted earlier, the economy appears to have shifted into a still
higher gear just as monetary policy turned more restrictive. As a
result, some have wondered whether recent structural changes may
have undermined the effectiveness of monetary policy.
It is well appreciated that
monetary policy affects aggregate demand with long and variable
lags. So, a limited initial effect of tighter monetary policy on
demand is to be expected. In addition, the first three
quarter-point moves simply reversed the earlier cumulative easing,
which may have added to demand in the second half of 1999. So, the
move to a more restrictive phase of policy is especially recent.
It is useful to put the most
recent episode of tightening in historical perspective. To do so,
I compare the cumulative increases in the funds rate and in other
measures of financial conditions in the current episode with the
movements in those episodes since the late 1960s during which the
funds rate increased at least about 1 percentage point.1
I have plotted in figures 1 through 4 the maximum, minimum, and
median increases in the federal funds rate and three other
financial indicators during previous episodes and the current
experience.
Figure 1 presents this analysis
for the federal funds rate. The current episode is the smallest
and most gradual tightening over the period studied, well below
the line for the minimum cumulative increase during previous
episodes. So the modest effects on aggregate demand to date
perhaps only confirm the relatively modest and extremely gradual
nature of the current tightening, along with the usual lags.
In addition, some developments
suggest that the recent policy moves may have had a more limited
effect on aggregate demand than would have been expected from even
this modest increase in the federal funds rate. The key to
understanding the effects of monetary policy on aggregate demand
is that monetary policy does not operate through the direct effect
of the federal funds rate on aggregate demand. Instead, changes in
the federal funds rate and anticipations of future movements in
the funds rate affect aggregate demand via their influence on a
broader range of financial conditions, including short and
longer-term private interest rates, equity prices, and the real
exchange rate. Financial conditions indexes that capture in one
measure the full range of relevant interest rates, asset prices,
and exchange rates have been constructed by Goldman Sachs and
Macroeconomic Advisers.
Figure 2 shows that the
cumulative increase in the Goldman Sachs financial conditions
index in the current episode is not only at the very low end of
historical experience, but is nearly unchanged over the first
three quarters of the current episode. For this measure, the data
begin in 1973. So a second conclusion about the current experience
is that the increase in the funds rate, to date, has had a smaller
effect than usual--nearly zero--on overall financial conditions
and hence on aggregate demand. In addition, the absolute level of
the Goldman Sachs index indicates that financial conditions remain
unusually stimulative, relative to historical experience.
The apparent persistence of
accommodative overall financial conditions in the face of the
recent increases in the federal funds rate is principally due to
the continued increase, on balance, of equity prices and the
failure of the real exchange rate to appreciate further. Figure 3
confirms that equity prices have been unusually resilient in this
episode, rising by an amount near the maximum during the first
nine months of significant tightenings over the period studied.
The recent decline in yields on
long-term Treasuries, on the other hand, is not in itself
relevant, because these rates do not directly affect private
borrowing decisions. Private long-term rates, in contrast, have
not declined this year, and they increased significantly over
1999. Figure 4 indicates that the increase in a representative
long-term rate--the yield on Baa-rated corporate bonds--is at the
median for episodes of significant monetary tightening since the
late 1960s. This is impressive, given that the rise in the funds
rate is well below average. The strong contribution from private
long-term rates in this episode may indicated that these rates
reflect an anticipation of future Fed tightening rather than a
lagged response to actual moves; a lagged response seems to have
been more typical of the experience earlier in the sample period.
Private long-term interest rates will rise further only if market
participants come to believe that the Federal Reserve will tighten
by more than the couple of additional moves already embedded in
the yield curve or if inflation expectations begin to rise. The
fact that long-term private rates had already risen in
anticipation of further Fed tightening has shortened the lag from
increases in the federal funds rate to the ultimate effect on
aggregate demand relative to past experience and thereby has
actually added to, rather than subtracted from, the effectiveness
of monetary policy.
No one channel of monetary
policy--in terms of the financial conditions index, no one
component--should be singled out as of controlling importance. The
appropriate adjustment in overall financial conditions can be
achieved with varying contributions from the individual
components. Still, monetary policy must take into account the
overall response of financial conditions in judging the
appropriate magnitude of any cumulative change in the federal
funds rate.
There also has been concern that
some sectors of the economy--specifically investment in high-tech
equipment, a component that has been an important contributor to
the strength of aggregate demand--may be relatively insensitive to
higher interest rates, therefore reducing the overall response to
higher interest rates in this episode of policy tightening. The
housing sector has often borne a disproportionate burden of higher
interest rates, at least initially. More flexible financing
arrangements may have reduced the interest sensitivity of this
sector as well, though it likely remains among the more
interest-sensitive sectors. The question here is whether the
economy might be less sensitive to the change in overall financial
conditions as a result of such developments.
The analysis of changes in
overall financial conditions shows that any judgment on this
question that is based on the experience in this episode would be
premature given that, to date, overall financial conditions have
changed so little. But the reality is that monetary policy always
has had uneven effects across the economy. We have a single
instrument that, by necessity, must be used to achieve balance
between aggregate demand and potential aggregate supply.
Variations in sectoral responses will be unavoidable given the
different lags and different interest sensitivities of the various
components of aggregate demand.
Another often-asked question
about monetary policy is whether, once a decision is made to
correct emerging imbalances, policy should move gradually or more
forcefully. That is, should rates be moved immediately to the
level that would be most likely to forestall the problem? Policy
responds to incoming data. To the extent that incoming data only
gradually alter perceptions of the appropriate policy stance, only
gradual policy adjustments will be called for. On the other hand,
sometimes policymakers do find themselves in a situation that
seems to call for more sizable policy changes. But when
considerable uncertainty remains about the appropriate course and
aggressiveness of the policy response--especially when policy is
moving pre-emptively against the threat of higher inflation,
without any direct corroboration from data on inflation--a more
gradualist approach allows policymakers to assess the data along
the way and adjust accordingly the desired path of interest rates.
The risk in this approach is
that imbalances become larger and more disruptive to correct if
resource utilization tightens further or inflation expectations
pick up. So it would be important to react more aggressively in
response to those developments. The appropriate speed of
adjustment might also depend on the degree to which the bond
markets reflect policymakers' expectations about likely further
increases in short-term interest rates. That is, it may be more
appropriate to move short-term interest rates slowly when
long-term interest rates have already adjusted by an amount that
policymakers believe is sufficient to achieve their objectives.
Finally, the pace of tightening may also need to be calibrated to
the degree to which broader financial market conditions are
responding to the rise in the federal funds rate.
Conclusion
The combination of favorable relative-price and productivity
shocks in this expansion are the principal sources of the
exceptional combination of robust output growth and declining
unemployment on the one hand and stable to declining inflation on
the other hand. As the disinflationary effects of relative-price
shocks and faster productivity growth dissipate, monetary policy
must be prepared to deal with more traditional concerns about the
balance between growth of demand and supply, the relationship
between output and potential, and the danger of overheating. As
these concerns have become more pressing, monetary policy has
responded in an effort to rebalance aggregate demand and supply
and contain the risk of higher inflation.
A productivity shock is
unambiguously good. But the monetary policy that accompanies it
can be good or bad. A good monetary policy is one that allows the
economy to realize the full benefits of the higher growth while
respecting the fact that limits--albeit new ones--remain and that,
if exceeded, the ultimate result will be rising inflation,
threatening the sustainability of the expansion.
Monetary policy remains fully
capable of getting the job done. The effect across sectors will
not be even, and it never has been. And the effects on a broader
range of financial conditions are always somewhat variable as
well. Policymakers have to adjust the timing and cumulative size
of tightenings to ensure that the effect on overall financial
conditions will promote a continuation of this economic expansion
and an accompanying low and stable rate of inflation.