THEORY AND HISTORY BEHIND BUSINESS CYCLES ...Theory and History Behind Business Cycles: Are the 1990s the Onset of a Golden Age? Victor Zarnowitz NBER Working Paper No. 7010 March
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
NBER WORKING PAPER SERIES
THEORY AND HISTORY BEHINDBUSINESS CYCLES: ARE THE 1990S
THE ONSET OF A GOLDEN AGE?
Victor Zarnowitz
Working Paper 7010http://www.nber.org/papers/w70 10
NATIONAL BUREAU OF ECONOMIC RESEARCH1050 Massachusetts Avenue
Cambridge, MA 02138March 1999
The views expressed in this paper are those of the authors and do not reflect those of the National Bureauof Economic Research.
Theory and History Behind Business Cycles:Are the 1990s the Onset of a Golden Age?Victor ZarnowitzNBER Working Paper No. 7010March 1999JEL No. E3, E30, E31, E32
ABSTRACT
The disputes over the prospects for the current U.S. expansion reopen the issue of the causes
of business cycles. A recurrent concern about the present is that expectations of business profits and
market returns may be outrunning the economy's potential to deliver.
The theory presented in this paper ties together profits, investment, credit, stock prices,
inflation and interest rates. I discuss new estimates of profit and investment functions with important
roles for growth of demand and productivity, price and cost levels, risk perception, credit volume
and credit difficulties. The relationships among these endogenous variables are viewed as
constituting an enduring core of business cycles, the exogenous shocks and policy effects as more
transitory and peripheral.
The U.S. upswing of the past three years provides a vivid example of how profits,
investment, and an exuberant stock market can reinforce each other. Long business expansions
benefit society in several ways but they generate imbalances and are difficult to sustain. Recent
events in Asia demonstrate how investment-driven booms can give way to a protracted stagnation
with tendencies toward deflation and underconsumption or to severe depressions.
After a deterioration in the 1 970s and early I 980s, U.S. business cycles moderated again, as
in the first two post-W WIT decades. But globally recessions became more frequent and more severe
in the second half of the postwar era. The arguments in favor a new Golden Age are generally not
persuasive.
Victor ZarnowitzFoundation for International Business and Economic Business122 East 42nd Street, Suite 1512New York, NY 10168and NBERvictor.zarnowitzgsb.uchicago.edu
2
Some have viewed the current business expansion in the United
States as the onset of a Golden Age in which the long-time evils of
inflationary booms followed by recessions with high unemployment
will never return. It is not surprising that the excellent
economic conditions in the mid- and late 1990s -- substantial real
economic growth, falling unemployment rate, low inflation, and a
persistent bull market in stocks -- have led to a widespread
euphoria. During such earlier notable economic expansions as the
1920s and the 1960s, consumer and investor confidence rose to high
levels, as growing numbers of people came to believe that a seismic
shift has taken place and great new opportunities were opening up
at remarkably low costs and risks. The vision of endless and
uninterrupted expansion of total employment, output, real income
and wealth is, of course, immensely attractive not only to
economists, but to all people of good will.
The happy prophecy of a growing recession-free economy has
been ascribed to a number of different changes in the economy, but
none of these suggested reasons is fully persuasive.1 Some of the
arguments seem to make the dubious assumption that factors which
raise productivity growth must also lead to greater economic
stability. Others exaggerate the reasons why the economy may be
more stable now than in the past into a claim that economic
instability is now obsolete. All leave ample room for
3
counterarguments.
First, the U.S. economy is allegedly much more stable because
of the successes of the recent "downsizing" or rationalization
efforts of business management. However, layoffs, cost-cutting,
corporate reorganizations and factor reallocations have long been
part and parcel of the cyclical growth process; for example, Davis,
Haltiwanger and Schuh (1996) discuss in rich detail how job
destruction varies greatly over the business cycle, rising strongly
in recessions, while job creation varies much less. Clearly,
effective labor cost reduction might at first raise unemployment
and the share of profits, but later enhance productivity and
growth. The same companies that become more efficient through
downsizing will then need or want to grow in the future, and so
will turn to "upsizing," from layoffs to hires, from downward to
upward wage adjustments. It is not clear why such long-standing
frictions associated with cyclical or irregular supply and demand
shifts should permanently alter the cyclical growth process, or why
any allocative shocks due to changes in business policy should have
more than mixed and temporary effects.
Second, some have claimed that the technological breakthroughs
in computer hardware and software will assure greater economic
stability. Clearly, technological advances have been indispensable
throughout the modern era in promoting productivity, economic
growth, rising standards of living, and even effective systems of
government. But the models that rely on generally unidentified
4
exogenous productivity shocks as the primary explanation of the
"real" business cycles, are, I believe, generally lacking in
plausibility and evidence (Zarnowitz, 1992, ch. 1 and 2). After
all, most technical changes are localized and gradual, with long
half-lives in their adoption and diffusion. In the 1990s, the
notable progress in computer technology has certainly contributed
to the recent sharp rise in business investment and profits. But a
strong productivity-enhancing effect of computers is yet to be
documented, and it is not at all clear why and how this particular
technological advance should perpetuate the present U.S. business
expansion.
Third, inventory control is said to have improved greatly, in
a way that will make the economy more stable. This claim has some
truth. Movements in inventories do tend to propagate economic
fluctuations; for example, an economic slowdown causes a build-up
of inventories which then becomes a secondary cause of business
output weakening further. The ratio of manufacturing and trade
inventories to sales has followed a gradual downward trend in the
1990s, probably thanks to the widespread adoption of just-in-time
inventory control systems, which tend to reduce the average stocks
of purchased materials and finished products on hand. Leaner
inventories are likely to have smaller macroeconomic effects.
However, it is also true that business inventory investment in
constant dollars was about as volatile and as cyclical in the 1990s
as it had been in the past, and volatility in inventories certainly
5
remains large enough to play a substantial role in propagating
economic cycles.
A fourth argument is that the share in total U.S. employment
of the relatively volatile goods-producing sectors like
manufacturing and construction has declined in favor of the
presumably more stable services such as trade, finance,
transportation, entertainment, education and government. This
shift does tend to moderate business cycles, largely by reducing
the weight of cyclically volatile inventory investment. But many
services appear to be becoming more cyclical, as they confront
growing competition at home and abroad. For example, business and
consumer services actually declined in the recessions of 1981-82
and 1990-91, while their growth had merely slowed down in earlier
downturns (Fosler and Stiroh, 1998).
Fifth, it is argued that deregulation of financial and other
industries has helped to stabilize the economy. One common example
harks back to the time from the mid-1960s to the early 1980s, when
the Regulation Q ceilings on the interest that could be paid on
bank and savings and loan accounts were in place. During this
time, when short-term interest rates rose above the ceilings, funds
poured from banks and thrifts into direct money-market instruments,
which in turn severely reduced the quantity of available consumer
and mortgage credit. Clearly, excessive or wrongheaded regulation
can harm efficient resource allocation and aggravate economic
instability. Also, one can make a strong case that more free
6
competition in banking, airlines, trucking, and other industries
has increased productivity growth; and to the extent that relative
price flexibility is enhanced, less instability in quantities may
be expected. But it seems farfetched to think that more
deregulation will deliver large benefits in stabilizing the United
States economy.
A sixth claim is that we have learned how to use discretionary
government macroeconomic policies in a way that reduces or ends
cyclical instability. The Federal Reserve has allegedly learned
how to forecast inflation and how to avert it by timely increases
in short-term interest rates. Fiscal policy is no longer used for
discretionary stabilization purposes for which it is unsuited; in
the past, it was more often than not misapplied or mistimed,
creating at least as much economic instability as it reduced.
However, there is no clear support in the data for assertions that
government can anticipate business recessions or financial crises,
nor that it can avert such events with preemptive action. True,
policy-makers can and sometimes do ameliorate recessions, but
wrong, mistimed, or bungled policies can also destabilize the
economy.
Finally, globalization is argued to have reduced cyclical
instability. Global markets diminish the economy's dependence on
domestic demand by creating new markets abroad for U.S.-produced
goods and services. They also open up new sources of supply for
raw and intermediate materials, final consumer and producer goods,
7
and labor, with the consequence of reducing domestic inflationary
pressures on prices and wages. Capital markets have become
increasingly global, too, which was supposed to make them broader
and more liquid and to reduce the risk of market bubbles and
crashes. But while greater openness of economies can surely bring
benefits, it also can bring increased vulnerability. Since the
economic debacle in East Asia over the last few years, the
attendant risks of financial meltdowns in Russia and Latin America,
and the manner in which these dangers have contributed to
volatility in stock markets everywhere, talk that globalization
will put an end to business cycles has understandably dwindled.
Where and How to Proceed
The disputes over the prospects for the current U.S. expansion
are far from being merely academic: they reopen the fundamental
yet unresolved issue of the underlying causes of business cycles.
One widespread and recurrent concern about the present, which has
long roots in the past, is that expectations of business profits
and market returns may be outrunning the economy's potential to
deliver. Up to a point, high levels of consumer and investor
confidence are self-confirming in their positive consequences.
However, high confidence can easily shade into overconfidence,
which breeds misdirected or excessive investment. Eventually, the
balance of expectations shifts, as people realize that the market
fundamentals no longer support the euphoria. The expansion slows,
8
then ends, as spending, employment and output turn down. All of
this has occurred repeatedly in the past and there is no compelling
reason why it should not happen again.
Believers in the inherent stability of market economies
attribute recessions to policy errors and external disturbances.
Thus, misguided stimulation by excessively easy credit causes
inflation, the belated curtailment of which causes business
activity to turn down. Some analysts consider the reactive nature
of government actions and other possible shocks but give no
attention at all to any endogenous theories (for example, Temin,
1998). On the other hand, those who suspect more systematic
instability doubt that the story of the Fed killing each of the
recent U.S. expansions is the right and full one. They can point
to many domestic and foreign recessions that originated mainly in
market developments. Just to take the latest few years,
overconfidence, overborrowing, and overinvestment contributed to
severe business downturns, financial crises, and incipient
deflation overseas, presenting the U.S. economy with new
challenges.
However, it is not with the specific problems and prospects of
the current economic situation that I propose to deal in the body
of this paper. We do not know just when and how the present U.S.
business expansion is going to end: although the business cycle
contains strong self-sustaining elements, it is not predetermined.
Instead, I shall focus on broader theoretical and historical
9
perspectives. The next, central part builds upon selected older
ideas about business cycles, which I believe still have
considerable relevance today when given some new features and
interpretations. The theory ties together profits, investment,
credit, stock prices, inflation and interest rates, treating their
interactions as central to the process of economic fluctuations and
growth. It purposely limits itself to essentially endogenous
Current Cycle 7.9 4.6 8.0 22.0 6.8 6.5 5.5 4.7Past Average 17.3 9.1 6.1 35.8 8.9 6.2 4.8 4.6
Nonresidential Structures CPI Inflation Rate
Current Cycle -13.5 9.8 11.2 5.7 4.2 2.5 3.0 1.1Past Average 13.8 2.3 -0.6 15.6 2.6 2.5 3.6 4.1
Producers Durable Equipment Corporate Bond Yield
Current Cycle 26.9 22.7 30.8 103.6 9.1 7.6 7.6 6.6Past Average 37.8 8.6 9.7 64.2 8.2 7.8 7.6 7.9
S&P 500 Stock Price Index Corporate Profit Margin
Current Cycle 22.0 39.5 66.4 183.2 6.0 8.0 10.0 9.0Past Average 26.4 22.9 27.6 98.1 7.4 8.4 9.3 8.1
Dating: According to the NBER chronology, the current U.S. business cycle expansion began in March 1991 (for monthly data) or first quarter of 1991 (for quarterly data). Hence, for the “Current Cycle”, the years denote: 1-3, 3/1991-3/1994; 4-5, 3/1994-3/1996; etc. The months denote: 1, 3/1991; 36, 3/1994; 60, 3/1996; 84, 3/1998. “Past Average” refers to U.S. expansions of February 1961-December 1969 and of November 1982-July 1990. Here the years and months are dated analogously starting in February 1961 and November 1982.
Sources: The Conference Board, Business Cycle Indicators:
42
Real GDP, series 55, nonresid. structures, 87; prod. dur. eqpt., 88; stock price index, 19; unemployment rate, series 43; CPI inflation rate, 320; corp. bond yield, 116; corp. profit margin, 81.
43
44
References
Blanchard, O.J. 1993. "Consumption and the Recession of
1990-91." American Economic Review, 83, 270-74 (May).
Black, F. 1986. "Noise." Journal of Finance. 41(3).
529-43 (July).
Blatt, J.M. 1978. "On the Econometric Approach to Business-
Parameters, and Pooling Techniques". Journal of Econometrics 49,
275-304.
1. For an affirmative version of the arguments summarized here, seeWeber (1997); Klein and Cullity (1998) is a rebuttal, and Weber(1998) is a reply. For extensions of the debate, see theproceedings of meetings of the Eastern Economic Association in 1998and the American Economic Association in 1999, with participationof Blinder, Boldin, Klein, Stock, Watson, and Zarnowitz. Otherrelated articles include Dornbusch (1998); Zarnowitz (1998a); andseveral unsigned contributions to selected 1998 issues of theConference Board's Business Cycle Indicators.
2. Mitchell (1913) saw the fluctuations in profits arising fromcost-price imbalances as critical for understanding the changinginvestment and production decisions of firms and hence forexplaining business cycles. For Schumpeter (1912), profits are thereward for innovations adopted by pioneering entrepreneurs. In vonNeumann's classical production model (1937), the ultimate source ofprofits is exogenous technology that permits growth. In earlyKeynes (1930), profits disappear in equilibrium defined by theequality of investment and saving. In Kaldor (1955/56), profitspersist because investment contributes to growth of income andcapital alike. In most of these models profits include allnonlabor (property) income, and in some pricing is via a markuprule, implying the prevalence of imperfect competition or oligopoly(Kalecki, 1954).
3. Of course, any individual firm under imperfect competition maychoose to expand its market share by accepting a lower profitmargin for a time (Wood 1975). By choosing to do so, the firmwould give rise to an inverse relationship between growth andprofits. But this occasional counterexample does not vitiate themore general connection that growth increases profits as a whole.
4. The data used are based on the following series in the BusinessCycle Indicators published by the Conference Board: profits areseries 81; economic growth, 55; productivity growth, 358; the pricelevel, 311; employment cost level, 63; inflation rate, 320;interest rate on long-term Treasury bonds, 115; and interest rateon new high-grade corporate securities, 116.
5. The correction (AR(1)=0.844) helped explain the large localtrends (up in 1953-65, down in 1966-74, and up again since 1980),which show up in ð in addition to pronounced cycles and in absence
of any overall trend
6. It is worth noting that these observations are broadlyconsistent with certain ideas of economists as original andinfluential as Wicksell, Schumpeter, Hayek, and Keynes (1930). Their theories, though deeply different otherwise, agreed that inexpansions the demand for investment rises above the supply ofsaving and is financed by an effectively endogenous process ofcreation of credit money. The classical exposition of the two-ratemodel is Wicksell (1901-06). Here rising demand for money drivenby the firms' perceived profit opportunities is met by risingsupply of bank credit at costs low enough to permit the profits tobe realized. This is described as an endogenous "cumulativeprocess," which gradually reduces the excess of investment demandover saving supply (and of the natural over the market rate).
7. See Cowen (1997) for further, mainly theoretical discussion ofthe role of risk in business cycles inspired by old Austrian ideasas well as modern financial literature. However, his discussion isgenerally silent on the other early line of ideas about riskincreasing in investment upswing, from Kalecki (1937) to Shackle(1970); on which, see Courvisanos (1996).
8. Nevertheless, shocks such as a sudden market crash, outbreak ofa war and fears of shortages, rationing, or credit restrictionscould and on infrequent occasions probably did produce "autonomous"destabilizing shifts in consumer demand. For further discussion,see Temin (1976, 1998), Hall (1986), and Blanchard (1993).
9. The original theory stems from the time when the gold standardruled and inflation alternated with deflation so that price levelstability in the long run was widely expected (and reflected inremarkably stable and low government bond yields). In theWicksellian two-rate model, prices were taken to be perfectlyflexible and changing so as to maintain full employment. However,changes in real incomes can be readily added to the cumulativeprocess, with quantity adjustments either replacing priceadjustments (Laidler, 1972) or, more generally, complementing them. It is worth noting that in historical periods of declining orstable prices nominal aggregates reflected business cycles well andwere widely used, whereas when inflation prevailed, as in the lasthalf-century, business cycles were best measured in real terms.
10. The data used in the investment regressions of this sectioninclude the following series from Business Cycle Indicators (inaddition to some listed in note 4 above): real business fixedinvestment, series 86; real corporate profits, 18; stock price
51
index, 19; exchange value of U.S. dollar, 750; funds raised incredit markets, 110; and current liabilities of business failures,14.
11. More generally, I believe that the analysis of economicfluctuations derived substantial benefits from methods developed bythe NBER, including the historical reference dates, turning-pointidentification, emphasis on short unit periods, and efforts attime-series decomposition, deseasonalization, and detrending. Manyrecent studies of business cycles suffer from ignoring the abovematters and relying instead on general methods only. Some useannual data which reveal little of what happens during shortrecessions, for example. What is desirable are combinations ofbest modern statistical and econometric techniques with insightsfrom cyclical indicator analysis. For examples of how leading andother indicators can be used along with univariate and multivariatetime-series models, Bayesian forecasting methods, probability,econometric, and nonlinear models, the interested reader mightbegin with the essays in P.A. Klein (1990), Lahiri and Moore(1991), and Stock and Watson (1993) and follow with Zellner, Hongand Min (1991) and Montgomery, Zarnowitz, Tsay and Tiao (1998).