Document Type


Publication Date



seasonal fluctuations, business cycles, seasonal cycles


Early scholars of aggregate fluctuations, including Jevons ( 1884) and Kemmerer ( 1910), introduced the notion that seasonal cycles were relevant to the study of other, seemingly more important, fluctuations in macroeconomic time series. Kuznets (1933) continued this approach, recognizing the tendency for seasonal variations to exacerbate the variability of employment and capital accumulation. But, as economic contractions intensified in the J 930's, economists began to discount the relative importance of the seasonal cycle. Work by Mitchell (1927), Pigou (I 929) and later, Burns and Mitchell ( 194 7), promulgated the view that fluctuations in commercial activity were relevant to the study of business cycles, while seasonal fluctuations were not. Currently, economists are calling for the reinterpretation of seasonal fluctuations as a means to understanding business cycles.2 As Miron ( 1996) explains, seasonal fluctuations account for a large portion of the overall movement in macroeconomic time series, and may be related to the behavior of non-seasonal (business) cycles. Moreover, to the extent that seasonal cycles are not Pareto Optimal, understanding how monetary and fiscal policy decisions can eliminate them may prove useful. While promising, the efficacy of such an approach is limited ultimately by economists' understanding of the seasonal cycle. Hence, in order to add to our knowledge of business cycles, the complexities of US seasonal cycles must be unraveled. One such example, for which numerous studies have been written, is the cessation of seasonal pressures on US money markets in 1914.3 While the disappearance of seasonal variations in US interest rates is compelling simply because it challenges our intuition that seasonal cycles are endemic fixtures of US money markets, the implications of this episode are all the more pressing in light of the recent literature linking seasonal and business cycles. In this paper we examine the origins of the seasonal cycle in US interest rates, beginning in the antebellum period, in order to gain a better understanding of when and why seasonal pressures emerged (and perhaps, disappeared) in early US financial markets. Seasonal tests indicate that variations in short- term interest rates are relatively unexplained by the seasonal cycle prior to the mid-1870s, and hence, are similar to their post-1914 counterparts. We propose that the absence of a seasonal cycle prior to the l 870's was most likely due to money market volatility. Prior to 1874, movements in interest rates were erratic and financial instabilities imparted relatively large shocks to money markets, particularly in the autumn months. After 1874, the effects of financial instabilities on interest rates diminished and the regularization of seasonal movements was attained. We attribute this change in behavior of short-term rates in the nineteenth century to the following institutional innovation: the introduction of futures markets and the resulting substitution away from consignment contracts in the agricultural trade shortly after 1874. The US experience with interest rate seasonals questions the view that seasonal cycles are a welldefined and predictable phenomenon in macroeconomic time series. Like business cycles, seasonal cycles can exhibit irregularities, as exemplified by their absence prior to 1874, as well as after 1914, and hence should not be dismissed as uninteresting fluctuations. Moreover, that a financial innovation contributed to the regularization of seasonal cycles in 1874 is particularly compelling, given the recent connection drawn between seasonal and business cycles, insofar as it suggests that business cycles may also be regularized by economic innovations.


Economics Research Report No. 98-2