View original document

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



Number 93-33, October 1, 1993

Have Recessions Become Shorter?
Until recently, it was widely believed that business cycle fluctuations in the
were less
severe after World War II than they were before.
The improvement often has been credited to active monetary and fiscal policies that manage
aggregate demand. However, this view has been
questioned by recent studies that pay special attention to the details of the prewar economic
data that were overlooked by earlier researchers.
In fact, some studies that use newly constructed
data have found little evidence of a decline in the
variability of output in the postwar period (Walsh
1989). In addition to this research on the severity
of business cycles, others have examined possible changes in average business cycle duration. If
macroeconomic policies that manage aggregate
demand have been successful, then they should
have extended expansions and shortened


This Letter reviews the current debate on the duration of business cycles. One recent study finds
that expansion periods are longer and contraction
periods are shorter in the postwar period using
the historical National Bureau of Economic Research (NBER) business cycle reference dates.
However, a subsequent study concludes that this
finding is mainly a statistical artifact due to inconsistencies in the way business cycle peaks
and troughs were dated by the NBER in the prewar and postwar periods.

Are business cycles smoother?
The apparent stability of the macroeconomy in
the postwar period relative to the prewar period
was first seriously challenged by Christina Romer
(1986). Romer's studies focused on key macroeconomic measures such as Gross National Product (GNP). the Index of Industrial Production OP),
and the u'~employment rate. She shows that the '
quality of data for the prewar period was poor
compared to the postwar period-in terms of
both the scope of individual series collected and
the accuracy of their measurement. Romer carefully reconstructed these and other prewar and
postwar macro series to make them more comparable, and found much less difference in the
volatility between the two periods than the uncorrected data indicated.

Have postwar expansions become longer
and contractions shorter?
An alternative way of defining business cycle
stability is in terms of the duration of cyclical
phases. That is, if the postwar stabilization policies were successful, then, on average, postwar
expansions should be longer, and contractions
shorter, than those of the prewar period.
Diebold and Rudebusch (1992) provide new evidence on duration. To avoid the measurement
problems uncovered by Romer, they use the reference business cycle chronology compiled by
the NBER. This is based on not just one but several data series to determine the direction of
general business activity and therefore the turning point (a peak or trough) in a business cycle.
Durations are measured by the length of time
between a trough and the following peak for
expansions, and vice versa for contractions.
Diebold and Rudebusch use data from 1854
to 1938 for the prewar period and from 1945 to
1990 for the postwar period. Their study does
find a marked shift in the average duration of expansions and contractions across the two sample
periods. The average length of expansions increased from 26.5 months in the prewar period
to 49.9 months in the postwar period, while the
average length of contractions fell from 21.2
months in the prewar period to 10.7 months in
the postwar period. Such shifts are statistically
significant and are not very sensitive to various
changes in the prewar sample period covered.
For example, the average length of expansions for
the period from 1854 to 1929 was 25.3 months,
and it was 25.5 for the period from 1854 to 1914.
The main difference between the two sample
periods is that the second one does not include
WWI and the Great Depression. The comparable lengths of contractions were 20.5 and 22.5
months, respectively, for the same two sample

Another look
A study by Watson (1993) explored potential explanations for the postwar changes in the length
of expansions and contractions using various output, employment, and financial market data. The

basic strategy was to examine individual series
to see if they exhibit the same kind of duration
characteristics as the NBER business cycle chronology. He considered three possible explanations for the change in the duration of NBER
defined cycles. First, shocks to the economy may
have been smaller in the postwar period. Second, the composition of output may have shifted
from sectors that are cyclical (manufacturing, for
example) to those that are less cyclical (such as
services). Third, the apparent stabilization may
be due to the way prewar and postwar business
cycle dates were chosen by the N BER.
To investigate whether shocks have been smaller
in the postwar period, Watson calculated average
cyclical durations for a large number of individual data series that were measured consistently
over the two sample periods. Examining individual series eliminates the chance of a possible
bias in aggregate data due to the difference in
the number of the components included in the
aggregation in each sample period. Watson's results did not support the notion that shocks have
been smaller in the postwar period. For example,
in the construction sector, he analyzed the prewar Building Plans series for 1868-1929 and the
postwar Building Permits for 1947-1990, and
calculated their average durations for the two
sample periods. Peaks and troughs in each series
were determined by a procedure that closely imitates the NBER dating method. For the construction index, the resulting average durations of the
prewar expansions and contractions were 19.1
and 16.6, i€spectively, vvhereas for the postv"ar
period they were 18 and 19.8 months, respectively. The results were similar for the other series
Watson did find evidence of a shift in the composition of employment from the prewar to the
postwar period. The share of manufacturing employment in the
economy fell from 34.7 percent in 1948 to 17.3 percent in 1990, whereas the
share of service employment rose from 11.5 percent to 25.6 percent in the same period.


Despite the evidence of a shift, its implications
for cyclical duration are not so clear, since the
manufacturing sector was far less cyclical during
the prewar than the postwar period. Manufacturing employment grew at an average rate of 2.4
percent for the prewar period and 0.4 percent for
the postwar period. The NBER chronology identifies a downturn period when there has been a
decline in the level of a series. According to this

convention, a series with lower average growth
will appear to be more cyclical. In this sense, the
manufacturing sector has become more cyclical
in the postwar period due to its slower average
growth. This factor tends to offset the effect of the
manufacturing sector's decline as a share of overall employment in the postwar period.
After running statistical experiments that control
for the differences in the average growth pattern,
Watson found that the resulting data did not exhibit any significant shift in the average duration
of prewar and postwar expansions and contractions. Thus, the possibility that an increase in the
proportion of less cyclical sectors in the postwar
period has caused the shift in average duration is
not supported.

This leaves the third explanation, namely, that
N BER researchers chose the prewar reference
dates in a way that is notably different from the
postwar method of reference date selection. This
difference could have arisen for two reasons.
First, the prewar data may have been processed
differently from the postwar data in determining
the NBER chronology. For example, certain statistical properties of the data are sensitive to
whether a trend in the data is removed or not. It
turns out that prewar NBER researchers used detrended data in determining the reference dates,
while postwar researchers did not. This change
in procedure accounts for some of the difference
in the cyclical durations.

Second, pre\'var ,,"~BER researchers relied on sectoral data, which were more cyclically volatile
than the aggregate economy. To test this explanation, Watson repeated the NBER business cycle
dating for the postwar period using only the
same (or closely matched) series that were used
by the prewar NBER researchers. Measured in
this way, the postwar average durations of contractions and expansions turned out to be 15.6
and 28.9 months, respectively-not significantly
different from the 20.5 and 25.3 months of the
f'.~BER prevJar average contraction and expansion
durations. However, they are substantially different from 10.7 and 49.9 months of the postwar
contraction and expansion durations of the official NBER reference business cycle dates based
upon the current procedure.

It once was widely believed that policies to manage aggregate demand had reduced the volatility
of business cycles in the postwar period. Now,

however, this belief is being questioned, since
new evidence finds little difference in the volati Iity of the prewar and postwar economy.
These findings open another debate about the
nature of business cycles and about why aggregate demand policies have not been effective in
moderating them. According to one view, these
demand management policies were potentially
effective, but they were not successfully implemented. According to another view, known as
"real business cycle theory:' demand policies
could not have altered the nature of business cycles, because cycles are not so much the result of
demand shocks, but the result of supply shocks,
such as technological progress, relative changes
in input factor prices such as oil. So, while the
question of the relative volatility of prewar and
postwar business cycles seems to be moving toward a resolution, the question of the nature of
business cycles and the appropriate policies to
moderate them remains to be explored.

Diebold, Francis X., and Glenn D. Rudebusch. 1992.
"Have Postwar Economic Fluctuations Been Stabilized?" American Economic Review 82, pp.

Romer, Christina. 1986. "Is the Stabilization of the
Postwar Economy a Figment of the Data?" American Economic Review 76, pp. 314-334.
Walsh, Carl. 1989. "Postwar Stability: Fact of Fiction?" Federal Reserve Bank of San Francisco
Weekly Letter (October 13).
Watson, Mark W. 1993. "Business Cycle Durations
and Postwar Stabilization of the
American Economic Review (forthcoming).


Chan Huh

On July 20, Federal Reserve Board Chairman Alan Greenspan presented a mid-year report to the Congress
on the Federal Reserve's monetary policy objectives for the remainder of 1993. The report reviews economic
and fi.nancial developments in 1993 and presents the economic outlook heading into 1994. For single or multiple
copies of the report, write to the Public Information Department, Federal Reserve Bank of San Francisco,
P.O. Box 7702, San Francisco, CA 94120, phone (415) 974-2246 or fax (415) 974-3341.

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.




·S>lU! ueaqAos 41!M
"(!Y Jaded papJi.JsJ uo PSIUPd

Oi':L176 V::>


WLL XOll'O'd





aAJaSa~ IOJapa~

~uaw~Jodaa 4)JOaSa~

Index to Recent Issues of FRBSF Weekly Letter











Risks in the Swaps Market
On the Changing Composition of Bank Portfolios
Interest Rate Spreads as Indicators for Monetary Policy
The Lonesome Twin
Why Has Employment Grown So Slowly?
Interpreting the Term Structure of Interest Rates
California Banking Problems
Is Banking on the Brink? Another Look
European Exchange Rate Credibility before the Fall
Computers and Productivity
Western Metal Mining
Federal Reserve Independence and the Accord of 1951
China on the Fast Track
and japanese Real Interest Rates
NAFTA and U.S. jobs
Japan's Keiretsu and Korea's Chaebol
Interest Rate Risk at U.S. Commercial Banks
Whither California?
Economic Impacts of Military Base Closings and Realignments
Bank Lending and the Transmission of Monetary Policy
Summer Special Edition: Touring the West
The Federal Budget Deficit, Saving and Investment, and Growth
Adequate's not Good Enough



The FRBSF Weekly Letter appears on an abbreviated schedule in june, july, August, and December.