Printer Friendly

The seasonal cycle and the business cycle.

The Seasonal Cycle and the Business Cycle

Most recent research on macroeconomic fluctuations ignores the seasonal variation in the economy by working with seasonally adjusted or annual data. There are several reasons why this approach may be desirable. Seasonal fluctuations might be small enough relative to business cycle fluctuations that removing them has little effect on empirical results. Alternatively, seasonal fluctuations might be generated by fundamentally different mechanisms than the ones that produce cyclical fluctuations, in which case the two kinds of fluctuations can be addressed separately. Finally, seasonal fluctuations might be natural and desirable while business cycle fluctuations traditionally are thought to involve significant welfare losses.1 (1) Undoubtedly one further reason for the general preference for seasonally adjusted data is that seasonally unadjusted data often are difficult to obtain. In most cases, however, seasonally unadjusted data (or reasonable proxies for such data) can be constructed. See the appendixes to the papers cited below for sources of seasonally unadjusted data.

The research that I have carried out over the past five years challenges a number of the most frequently cited reasons for ignoring seasonality, and it suggests that a unified approach to studying seasonal and cyclical fluctuations may be warranted. This line of research resumes an older tradition of NBER analysis of fluctuations, exemplified by Simon Kuznets, in which both seasonal and business cycle fluctuations were regarded as important topics of investigation.2 Here I summarize the main findings of my research to date and discuss possible implications of the results for understanding economic fluctuations. (2) S. Kuznets, Seasonal Variations in Industry and Trade, New York: NBER, 1933; W.S. Woytinsky, Seasonal Variations in Employment in the United States, Washington, DC: Social Science Research Council, 1939; J.P. Bursk, Seasonal Variations in Employment in Manufacturing Industries, Philadelphia: University of Pennsylvania Press, 1931; E.W. Kemmerer, Seasonal Variations in the Relative Demand for Money and Capital in the United States, National Monetary Commission, 1910; and F.R. Macaulay, Some Theoretical Problems Suggested by Movements of Interest Rates, Bond Yields, and Stock Prices in the United States since 1856, New York: NBER, 1938.

The Patterns and Importance

of Seasonal Fluctuations

The first goal of my research has been to examine the seasonal patterns in standard macroeconomic variables and to document their quantitative importance.3 Seasonal fluctuations account for more than 85 percent of the fluctuations in the rate of growth of real GNP, and seasonal fluctuations are present in every major type of economic activity, including consumption, investment, government purchases, industrial production, retail sales, unemployment, and the money stock. Seasonal movements are not a quantitatively important feature of either prices or interest rates, however. (3) R.B. Barsky and J.A. Miron, "The Seasonal Cycle and the Business Cycle," NBER Working Paper No. 2688, August 1988, and Journal of Political Economy, forthcoming; and J.A. Miron, "A Cross-Country Comparison of Seasonal Cycles and Business Cycles," manuscript, University of Michigan, November 1988.

The seasonal pattern in quarterly real GNP consists of large increases in the second and fourth quarters, a large decrease in the first quarter, and a mild decrease in the third quarter. In the fourth quarter boom, output on average is 8 percent higher than it is every winter. The seasonal patterns in consumption purchases and output are similar, and the seasonal pattern in government purchases also is dominated by a first quarter decline and a fourth quarter increase. Fixed investment grows strongly in the second quarter, grows slightly in the third quarter, declines weakly in the fourth quarter, and declines strongly in the first quarter.

Monthly data show that industrial production falls strongly one to three months before Christmas, recovers in February, declines dramatically in July, and then rebounds strongly in August. Retail sales grow substantially in December but then decline tremendously in January. Likewise, the money stock exhibits a large, positive growth rate in December and a large, negative growth rate in January. The movements in labor market variables mirror those in output for the most part, although the fluctuations are smaller in magnitude.

The seasonal patterns just described are characteristic of most developed countries. It is particularly note-worthy that Australia and New Zealand exhibit seasonal patterns strikingly similar to those in most countries in the Northern Hemisphere. These results therefore shed light on the reasons for particular seasonal movements in aggregate data. The large booms in consumption in the fourth quarter and in retail sales in December obviously are caused by Christmas spending, especially since these occur in Southern Hemisphere countries. The July-August troughs in industrial production, which do not appear in the Southern Hemisphere, presumably represent preferences for summer vacations. The first quarter trough in all economic activity, which is more pronounced in the Northern Hemisphere, plausibly reflects both the end of the Christmas season and the comparatively poor weather in the first quarter.

In addition to demonstrating the quantitative importance of seasonal fluctuations, the results on the patterns of seasonal variation provide some easily identifiable examples of shifts in demand or supply at the aggregate level. That is, there are identifying restrictions available for the seasonal cycle that are not available for the conventional business cycle. One important reason for studying seasonal cycles is that we can use the seasonal information to identify relationships that may not be identifiable from business cycle information alone.

The Similarity of the Seasonal Cycle

and the Business Cycle

The second phase in my examination of seasonal fluctuations consists of a comparison of the properties of seasonal cycles and business cycles. That is, rather than assuming that seasonal cycles and business cycles are generated by different mechanisms, I ask to what extent the two types of fluctuations display similarities. The key finding of this investigation is that business cycles and seasonal cycles are surprisingly similar.

The first important similarity between seasonal and cyclical fluctuations is simply that there is an aggregate seasonal cycle. Just as with business cycles, there are sufficient similarities in seasonal cycles across sectors that a large seasonal cycle is present in aggregate output. The presence of this cycle is surprising. Seasonals in technology imply that production in certain sectors should be seasonal (for example, it is not surprising that construction falls in the winter), but it is far from obvious what aggregate seasonals in technology might be. Instead, it may be necessary to explain the seasonal bunching of output by means of increasing returns or other synergies combined with relatively small shifts in the productive technology.

A second similarity between cyclical and seasonal fluctuations is the tendency for output to be produced at the last minute--the absence of production smoothing.4 The seasonal evidence against production smoothing consists of the finding that seasonals in production and sales are virtually identical in two-digit manufacturing industries.5 This evidence is perhaps even more striking than the business cycle evidence against production smoothing because the anticipated, transitory fluctuations in demand represented by seasonals are precisely the ones that should be smoothed most easily by firms via inventory accumulations. (4) This fact has been emphasized for business cycle fluctuations by A. S. Blinder in "Can the Production-Smoothing Model of Inventories Be Saved?" NBER Working Paper No. 1257, January 1984, and Quarterly Journal of Economics 101 (1986). (5) J.A. Miron and S.P. Zeldes, "Seasonality, Cost Shocks, and the Production-Smoothing Model of Inventories," NBER Working Paper No. 2360, August 1987, and Econometrica 56, 4 (July 1988).

The third important similarity between seasonal and cyclical fluctuations is that increases in labor input are associated with more than one-for-one increases in output: labor productivity is procyclical. The fact that much of this procyclicality occurs between the third and fourth calendar quarters (output increases substantially without a significant increase in labor input) suggests that labor hoarding in the face of the Christmas demand shift is the most likely explanation for procyclicality at the seasonal frequencies. It seems harder to account for the seasonal variation in labor productivity by relying solely on changes in technological opportunities, which is the explanation found in real business cycle models.

A final important similarity between seasonal and cyclical fluctuations is a strong correlation between movements in nominal money and real output. The presence of this correlation over the seasonal cycle is a prime example of the endogeneity of money with respect to real output fluctuations, since it is implausible that seasonal fluctuations in output could be driven by monetary factors. The fourth quarter peak in output is more likely the result of the impulse to real spending associated with Christmas, and the comovement of money with output reflects active accommodation of the seasonal variation in output by the Fed. Indeed, a desire to eliminate the seasonal movements in interest rates was probably a primary motivation for the founding of the Fed, in addition to being one of the Fed's dominant objectives since its inception.6 The Fed responds to the fourth quarter surge in spending by letting the nominal money supply increase just enough so that the money market clears without changes in the price level or nominal interest rates. (6) J.A. Miron, "Financial Panics, the Seasonality of the Nominal Interest Rates, and the Founding of the Fed," American Economic Review 76, 1 (March 1986), pp. 125-140; N.G. Mankiw and J.A. Miron, "The Changing Behavior of the Term Structure of Interest Rates," NBER Reprint No. 734, July 1986, and Quarterly Journal of Economics 101, 2 (May 1986); N.G. Mankiw, J.A. Miron, and D.N. Weil, "The Adjustment of Expectations to a Change in Regime: A Study of the Founding of the Federal Reserve," NBER Reprint No. 915, October 1987, and American Economic Review 77, 3 (June 1987); and R.B. Barsky, N.G. Mankiw, J.A. Miron, and D.N. Weil, "The Worldwide Change in the Behavior of Interest Rates and Prices in 1914," NBER Reprint No. 1115, March 1989, and European Economic Review 32, 5 (June 1988).

Interactions Between Seasonals

and Cyclical Fluctuations

In the third main phase of my research on seasonality I have documented that countries with substantial amounts of seasonal variation also have substantial amounts of business cycle variation. For example, countries in which the seasonal variability of industrial production is high are also countries in which the nonseasonal variability of industrial production is high. This kind of cross-sectional relationship also holds for retail sales, the price level, the money stock, and nominal interest rates.

One issue that arises immediately is whether these cross-sectional correlations might be explained by some exogenous third factor. Consider, for example, an economy that has two sectors, one seasonal and one nonseasonal. Assume that the seasonal sector (for example, construction) displays greater cyclical variation than the nonseasonal sector (for example, services). In this setting, there will be a positive correlation between the amount of seasonal variation displayed by a particular economy and the amount of nonseasonal variation displayed by that economy. If one accounts for the share of output originating in the construction sector, however, the correlation between the amounts of seasonal and nonseasonal variation should disappear.

The cross-sectional relationship between seasonal and cyclical variation does not appear to be explained easily by some mechanism such as the one just described. I have analyzed the potential role of the degree of economic diversity, of the industrial composition of output, and of the level of economic development, and I find that there is a strong correlation across countries between the amount of seasonal variation exhibited by a country and the amount of business cycle variation exhibited by that country, even after controlling for these other country characteristics.

The cross-sectional correlation between the amount of seasonal and cyclical variation most likely results instead from having the same economic mechanism operative in producing both seasonal and cyclical fluctuations. If this is the case, then it is possible to learn about the business cycle by studying what occurs over the seasonal cycle. Since, as mentioned above, there are identifying restrictions available for the seasonal cycle that are not available for the business cycle, this is a useful new perspective on aggregate fluctuations.

Future Research

The most important results of my research to date are basic stylized facts about aggregate fluctuations: seasonal cycles are quantitatively important; their properties are similar to those of business cycles; and countries with significant seasonal cycles are also ones with significant business cycles. Some of these facts tend to refute or support particular models of aggregate fluctuations, but for the most part they are not conclusive by themselves. Therefore, the most important implication of these results is to suggest that explicitly accounting for seasonality in macroeconomic modeling and empirical work is likely to improve our understanding of the economy.

The next step in my research is to develop specific, economic models that are capable of accounting simultaneously for the seasonal and business cycle facts about the economy as well as for possible connections between seasonal and cyclical fluctuations. Perhaps the most ambitious and far-reaching question is to understand why (indeed if) we should care about business cycle fluctuations but not about seasonal fluctuations.
COPYRIGHT 1989 National Bureau of Economic Research, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1989 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Miron, Jeffrey C.
Publication:NBER Reporter
Date:Mar 22, 1989
Previous Article:Asset-price expectations.
Next Article:First quarter 1989.

Related Articles
Economic fluctuations.
Economic fluctuations.
Political parties and the variable duration of business cycles.
Warming reaps earlier spring growth.
Seasonal fluctuations and business cycles.
Seasonal Variation in Host Susceptibility and Cycles of Certain Infectious Diseases.
Dietary Software.
Seasonal variations in reproductive activity and biochemical composition of the cockle Fulvia mutica (Reeve) from the eastern coast of China.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters