Wen, Yi: The Business Cycle Effects of Seasonal Shocks
World Conference Econometric Society, 2000, Seattle

Yi Wen, Cornell University
The Business Cycle Effects of Seasonal Shocks
Session: C-5-26  Sunday 13 August 2000  by Wen, Yi
Conventional wisdoms emphasize 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 fluctuations are generated by a fundamentally different source of impulses than seasonal fluctuations.
The central question to be addressed in this paper is, to what extent impulses at the seasonal frequency are responsible for business cycles? Unfortunately, despite strong empirical evidence suggesting that seasonal fluctuations and business cycle fluctuations are closely related, questions like this are difficult to answer because of the lack of effective methods for identifying seasonal versus non-seasonal shocks. Consequently, it has been difficult to assess the contributions of seasonal shocks to the business cycle. Traditional methods of measurement and identification (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 fluctuations and business-cycle fluctuations.
In this paper, we developed an econometric procedure that allows us to identify seasonal shocks versus non-seasonal shocks, and to measure the business cycle effects of seasonal shocks. Applying the procedure to the US data, we found that seasonal shocks account for the bulk of business-cycle fluctuations in output (roughly 50%), and that the larger the seasonal shocks are, the larger the business cycle is. These findings suggest that models that rely heavily on technology shocks to explain the business cycle are missspecified, that using seasonally adjusted data to test or evaluate business cycle models may lead to incorrect conclusions, and that theories of the business cycle, as well as government policies concerning the business cycle ought to address seasonal fluctuations seriously.

File created by Jurgen Doornik with eswc2000.ox on 2-01-2001