The Business Cycle E ® ects of Seasonal Shocks ¤
Yi Wen
2000-01-01
Abstract:Conventional wisdom emphasizes supply and demand shocks as the major sources of the business cycle. Yet the most visible, most synchronized, and most frequently encountered supply and demand shocks take place at the seasons. The central question to be addressed in this paper is to what extent impulses at the seasonal frequency are responsible for business cycles? Despite strong empirical evidence suggesting that seasonal °uctuations and business cycle °uctuations are closely related, questions like this are di±cult to answer because of the lack of e®ective methods for identifying seasonal versus non-seasonal shocks. Traditional methods of measurement and identi ̄cation (such as the use of seasonal dummies to isolate the seasonal and non-seasonal components) are inadequate and inappropriate because they fail to take into account the possible interactions between seasonal °uctuations and business-cycle °uctuations. In this paper, we develop a procedure that allows us to identify seasonal shocks versus non-seasonal shocks. We found that seasonal shocks account for the bulk of business-cycle °uctuations in US output (roughly 50%). The ̄nding suggests that models relying heavily on technology shocks to explain the business cycle are miss-specī ed, that using seasonally adjusted data to evaluate business cycle models can lead to incorrect conclusions, and that theories of the business cycle, as well as government policies concerning the business cycle ought to address seasonal °uctuations seriously. ¤I thank Je® Miron, Karl Shell, and especially, Peter Pedroni, as well as seminar participants at Cornell, Penn State and SUNY Albany for comments. I also thank Joseph Beaulieu for providing the seasonally unadjusted data. What shocks are responsible for economic °uctuations? Conventional wisdom emphasizes supply and demand shocks as the major sources of the business cycle. Yet the most visible, most synchronized, and most frequently encountered supply and demand shocks take place at the seasons. Most research on the business cycle, however, has worked only with seasonally adjusted data. Underlying this practice is the view that business °uctuations are generated by a fundamentally di®erent mechanism than seasonal °uctuations. Such views have been questioned by a series of empirical studies (e.g., Barsky and Miron, 1989, Beaulieu, MacKieMason, and Miron, 1992, Beaulieu and Miron, 1992, Miron, 1994, 1996, and 1998, to name just a few). These authors documented that the two types of °uctuations display striking similarities. In particular, with respect to most stylized facts about the business cycle, the seasonal cycle displays the same characteristics as the business cycle, in many cases even more dramatically than the business cycle. Further more, countries and industries with large seasonal cycles also have large business cycles. An important distinction, however, must be made between two fundamental problems regarding the business cycle: the propagation mechanism and the impulse mechanism. The propagation mechanism pertains to the endogenous properties of the business cycle, while the impulse mechanism pertains to the exogenous properties of the business cycle. To trigger the business cycle, exogenous impulses are needed to activate the economy's propagation mechanism so that recurrent °uctuations at the macroeconomic level become visible. Two distinct sets of questions, therefore, are involved in understanding the relationship between business cycle °uctuations and seasonal °uctuations. First, do they share the same underling propagation mechanism? Second, do they share the same sources of impulse? The primary interest of this paper is to address the second question. Business cycles have characteristic frequencies (e.g., the spectrum of US GDP growth has maximum power centered around the 4-6 year cycle frequency).1 If the economy is a dynamic system with endogenous propagation mechanism that determines the characteristic frequencies of the business cycle, then any external forces ought to be able to trigger dynamic responses from the system at those characteristic frequencies determined by it's internal propagation mechanism.2 The seasonal impulses are simply one particular kind of those forces which reguSee Mark Watson (1993). The post-war US investment to output ratio provides one of the most striking evidence of such kind of endogenous propagation mechanism that generates dampened oscillations (e.g., Y. Wen, 1998).