Classic Techniques  |  JAN 1996

Quasi-Seasonal Tendencies in Bond Futures by Scott Barrie

V.14:1 (19-25): Quasi-Seasonal Tendencies in Bond Futures by Scott Barrie Markets may have certain persistent characteristics. Here, a market analyst takes a close look at the Treasury bond futures markets to see if there are tradable patterns. Agricultural commodities traders are familiar with seasonal tendencies. Crops are planted at around the same time each year and harvested around the same time each year, so prices tend to follow the normal planting and harvesting cycle of production. Soybeans seem to rally in April, May or June based on drought scares, except in 1993, when soybeans rallied on flood scares instead. During September and October, exchange floor participants are always talking about harvest pressures. These well-documented tendencies present a sound building block for trading agricultural futures. Financial traders have no simple production cycle upon which to base trading, but quasi-seasonal biases do exist in the bond market. These tendencies may be based on the cycle of economic report releases, the refunding of the federal debt or other factors that affect interest rates. As a basis for the analysis of directional bias in the bond market, I followed the studies done by Arthur A. Merrill in his ground-breaking Behavior of Prices on Wall Street , as well as the work done by Yale Hirsch in Stock Trader's Almanac . Both works deal only with the stock market and therefore were used as guides in analytic method only. As a proxy for the bond market, I used the Chicago Board of Trade (CBOT) long bond futures contract from 1979 to 1994 unless otherwise stated. This work is intended in no way to be as complete as the previously mentioned writers' studies, simply because the number of occurrences available to study are not nearly as plentiful for bond futures as it is for the stock averages. ANALYZING BIASES In examining central tendencies, I began my study by looking for particular days of the week with high propensities to be the weekly high, weekly low or the weekly extreme close. Figure 1 reveals the outcome of this analysis. On a percentage basis, no day of the week is drastically different. In comparing the results, however, Friday has the highest probability of making both the weekly high and low. On close examination, Friday has more than twice the probability for having the extreme price of the week over Tuesday through Thursday, while Monday has the highest probability of being the low close for the week.

by Scott W. Barrie