View original document

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



Number 95-28, September 1, 1995

Using Consumption to Track
Movements in Trend GDP
One of the goals of economic policy is to stabilize short-term fluctuations in output and
employment. Policymakers try to do this by
"leaning against the wind:' That is, they try to
stimulate aggregate demand when output and
employment are "too low" and to reduce aggregate demand when output and employment are
"too high." Whether output is "too low" or "too
high" is determined by comparing actual GDP
with its trend. Of course, trend GDP is not observable, so po!icymakers must rely on estimates.
Measuring trend GDP turns out to be a difficult
and important problem. Its difficulty stems from
the fact that accurate measurement requires high
quality data about the current state of the real
economy and a great deal of knowledge about
economic dynamics, both of which are in short
supply. Its importance stems from the fact that
policymakers cannot effectively lean against
the wind unless they know which way it blows.
Measurement errors in trend GDP are one factor
that can make policy destabilizing rather than
This Weekly Letter discusses some of the policy
problems that derive from uncertainty about
trend GDP. It also describes a proposal by John
Cochrane (1994), who suggests that consumption
can be used to track movements in trend output.
His idea is quite simple: Economic theory tells
us that consumption primarily reflects the expectations of private households about long-run
movements in income. Therefore consumption
should provide a reasonably good measure of
trend GDP.

Why good measures are important
for stabilization policy
Macroeconomic policymakers try to stabilize
short-term fluctuations in output and to maintain
low inflation over the long run. The rationale for
trying to prevent output from falling too far below
trend is a belief that this represents a waste of resources. The rationale for trying to prevent output

from rising too far above trend is a fear that this
may put upward pressure on inflation, which
would conflict with long-run objectives. Thus
policymakers seek to balance the risk of inefficient resource utilization against the risk of rising
Uncertainty about trend GDP puts policymakers
in a difficult position. If they are unsure whether
output is above or below trend, they must be
more cautious about pursuing countercyclical
policies. The reason is that a mistake in measuring the gap between actual and trend output
might cause policy to be destabilizing rather
than stabilizing. For example, suppose that GDP
were at its trend level, so that no countercyclical
action is needed. If policymakers were to overestimate trend GDp, they would conclude that
output is below trend and might choose to stimulate aggregate demand, thus increasing the risk
of a rise in inflation. Similarl~ if policymakers
were to underestimate trend GDp, they would
conclude that output is above trend and might
choose to reduce aggregate demand, thus increasing the risk of recession. In both scenarios,
policy turns out to be destabilizing, in the sense
that it pushes output away from trend rather than
toward it.
Furthermore, Milton Friedman (1953) demonstrated many years ago that it is not enough for
policymakers to guess right about as often as
they guess wrong. If policymakers bat .500, attempts at countercyclical policy turn out to be
destabilizing on average. Stabilization policy
must offset other shocks in order to be effective.
If policy actions are timed randoml~ they become an additional source of variability rather
than an offsetting factor." Therefore, in order for
policymakers to pursue countercyclical polic~
they must bat better than .500, and the more
aggressive are their countercyclical actions, the
higher their batting average must be. Bold countercyclical actions require that policymakers
guess right far more often than they guess wrong.

Estimating trend GDP
Since trend GOP is not observable, economists
must try to estimate it. One traditional approach
assumes that the economy evolves along a longrun growth path that grows at a constant rate,
with transitory shocks generating short-term
fluctuations in output and employment. This
approach implicitly assumes that all shocks are
neutral in the long run. If this were the case, it
would be a simple matter to estimate trend GOP:
Trend growth is the same as the average rate of
real GOP growth.
However, over the last 15 years, many economists
have begun to doubt the assumption that all shocks
are neutral in the long run. Indeed, some shocks,
such as technological innovations, probably have
permanent effects on output. Measuring trend
GOP is much harder when some shocks are permanent and others are transitory. The traditional
approach works very badly in this case. Although
some movements in output reflect a short-term
deviation from trend, as in the traditional model,
others reflect a shift in trend GOP itself. Since
there is no tendency for permanent shocks to
revert to the simple trend line, measurement
errors accumulate over time and ultimately grow
without bound. Some progress can be made by
allowing for occasional shifts in the trend growth
rate, but this represents an ex post adjustment
and therefore has limited value for policy
Since trend GOP is subject to random permanent shifts, it is essential to distinguish between
permanent and transitory shocks in order to track
its movements. john Cochrane (1994) proposes
a way to use information about consumption
to help sort between permanent and transitory
shocks. His idea is based on the permanent
income hypothesis, which in its simplest form
states that people save when income is unusually
high so that they will be able to sustain customary consumption levels in thefuture when income
turns out to be unusually low.
For example, an ancient version of the permanent income hypothesis appears in the Biblical
tale of joseph and the Pharaoh. The Pharaoh
dreamt of seven fat cows followed by seven lean
cows who devoured the fat ones. joseph interpreted the Pharaoh's dream in terms of the agricultural cycles to which ancient Egypt was
subject. The seven fat cows represented seven
abundant harvests, and the seven lean ones represented seven bad ones. In order to prevent starvation in the lean years, Joseph persuaded the

Pharaoh to store grain from the good harvests
and distribute it during lean years.
Modern versions of the permanent income hypothesis replace dreams with rational expectations
and incorporate risk aversion and variation in
interest rates, but they retain the same basic
insight. Consumption is determined primarily
by the expected present value of future income
rather than by current income alone.
Cochrane applies the logic of the permanent
income hypothesis to the problem of sorting between permanent and transitory movements in
income. If a shock to income is transitory, it is
likely to have little effect on the expected present
value of lifetime income. Therefore we would not
expect consumption to change much in response
to transitory shocks. However, if a shock to income is permanent, it is likely to have a greater
effect on the expected present value of lifetime
income, and thus we would expect a greater response in consumption. If we observe a movement in income that is not accompanied by a
movement in consumption, we can infer that
consumers believe that the change in income
is likely to be transitory. On the other hand, if
consumption and income move together, then
we can infer that consumers believe that the
change in income is likely to be long lasting.
Since consumption measures consumers' expectations <;>f long-run GOp, it should provide
a good measure of the trend in GOP.
The figures illustrate the results of applying
Cochrane's idea to U.S. data. Figure 1 shows per
capita real GOP along with the consumptionbased measure of trend, which was estimated by
regressing per capita real GOP onto per capita
real consumption. The fitted values from this regression illustrate movements in income that are
accompanied by movements in consumption.
These movements are likely to be long lasting,
and they approximate movements in the trend in
output. Figure 2 shows the residuals from this regression. They represent the movements in GOP
that do not correspond to movements in consumption/ and they should be transitory. The
shaded afeas mark the dates of recessions as determined by the National Bureau of Economic
Research (NBER). Troughs in the consumptionbased measure conform quite closely to the dates
of NBER recessions: They often occur in the same
quarter as the NBER trough, and they always
occur within plus or minus two quarters of the
NBER trough. Furthermore, the consumptionbased measure does not give any false signals
of recession.
Another way to understand Cochrane's idea is to
think about what it implies for forecasting output
growth. When output is below trend, it ought to

Figure 1
Per Capita Real GDP
thousands of
1987 dollars
27. 5


25. 0










\ ;j~

t ~~\;
f:j W~
15.0 +n-rh-mTlTTlTrfi-rrf'i'-rrl-rfr+h-rTT".-,-'h-rn
54 58 62 66 70 74 78 82 86 90 94

Figure 2
Cyclical Component of GDP
percent deviation
from trend














~\~ f~:~
~;~ ":;






Finally, although consumption is not an exact
measure of the trend in output, it compares favorably with a number of other commonly used
measures. For example, Cogley (1995) simulates
a variety of models of trend and cycie and finds
that the consumption-based measure is a useful,
robust indicator of the unobserved true cycle.

Measuring trend GDP is important because central banks cannot lean against the wind if they do
not know which way it blows. This is difficult
because it requires sorting between shocks that
have permanent and transitory effects on output.
Economic theory tells us that consumption may
be useful for this purpose, and there is some empirical evidence to support this idea.









relative to the path of expected future income,
because consumers will want to save in order to
prepare for less favorable times ahead. Similarly,
the income-consumption ratio should be low
when consumers bel ieve that current income is
low relative to the path of expected future income, because in this case consumers will
reduce current savings in order to maintain current consumption. It is clear from the figures that
the income-consumption ratio is an excellent
predictor of future growth. For example, the income-consumption ratio is low at the troughs of
recessions, when output growth is about to pick
up, and it is high near the end of expansions,
when output growth is about to slow down.
Cochrane uses formal statistical techniques to
confirm that the income-consumption ratio has
significant predictive power for future output



54 58 62












Desiree Schaan
Research Associate



Timothy Cogley
Senior Economist


66 70

74 78

82 86

90 94

Cochrane, John H. 1994. "Permanent and Transitory
Components of GNP and Stock Prices." Quarterly
Journal of Economics 109(1), pp. 241-265.

grow faster than average in order to catch up.
Similarly, when output is above trend, it should
grow more slowly than average in order to let the
trend catch up. The permanent income hypothesis implies that the income-consumption ratio
should be a good predictor of future output
growth. This ratio should be high when consumers believe that current income is high

Cogley, Timothy. 1995. "Evaluating Non-Structural
Measures of the Business Cycle." Unpublished
Friedman, Milton. 1953. "The Effects of a FullEmployment Policy on Economic Stability: A
Formal Analysis./I In Essays in Positive Economics. Chicago: University of Chicago Press.

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. Free copies of federal Reserve publications can be obtained from the Public Information Department, Federal
Reserve Bank of San Francisco, P.O. Box 7702, San Francisco 94120. Phone (415) 974-2246, Fax (415) 974-3341. Weekly Letter
texts and other FRBSF publications ilnd data are available on FedWest Online, a public bulletin board service reached by setting
your modem to dial (415) 896-0272.

Research Department

Federal Reserve
Bank of
San Francisco
P.O. Box 7702
San Francisco, CA 94120

Printed on recycled paper Q
with soybean inks.



Index to Recent Issues of FRBSF Weekly Letter










G Iick/Moreno

What Are the Lags in Monetary Pol icy?
Central Bank Credibility and Disinflation in New Zealand
Western Update
Reduced Deposit Insurance Risk
Rules V5. Discretion in New Zealand Monetary Policy
Mexico and the Peso
Regional Effects of the Peso Devaluation
1995 District Agricultural Outlook
Has the Fed Gotten Tougher onĀ· Inflation?
Responses to Capital Inflows in Malaysia and Thailand
Financial Liberalization and Economic Development
Central Bank Independence and Inflation
Western Banks and Derivatives
Monetary Policy in a Changing Financial Environment
Inflation Goals and Credibility
The Economics of Merging Commercial and Investment Banking
Financial Fiagil ity and the Lender of Last Resort
Understanding Trends in Foreign Exchange Rates
Federal Reserve Policy and the Predictability of Interest Rates
New Measures of Output and Inflation
Rebound in U.s. Banks' Foreign Lending
Is State and Local Competition for Firms Harmful?
Productivity and Labor Costs in Newly Industrializing Countries

The F,RBSF Weekly Letter appears on an abbreviated schedule in June, July, August, and December.