Cycles Magazine Archives: Soybean Prices - A 1986 Extrapolation

Gertrude Shirk

Cycles Magazine Archives: Soybean Prices - A 1986 Extrapolation

Soybean Prices - A 1986 Extrapolation
by Gertrude Shirk
Editor of the Cycles magazine and Vice President of the FSC

From the Archives: This article is reposted as it was published in Cycles Volume Thirty Seven - 1986.

The 1985 extrapolation of soybean prices, first published in the January/February 1985 issue of Cycles is reproduced below as Figure 1. The "How It Came Out" section of the actual price line is completed through December 1985. The prices used in this work are the average monthly cash price of No. 1 Yellow soybeans, Illinois Processor.

The yearly average on a calendar year basis of the calculated line for 1985 was $6.07 which compares to the actual calendar year average price of $5.64 that occurred. In early 1985, low level of the extrapolated calculated line was considered by some members interested in soybean prices to be unduly pessimistic. But, as it turned out, the calculated line ran consistently above the actual prices as they developed.

The calculated/actual price chart covers only five years, and a better perspective can be gained from the entire history of soybean prices. Figure 2 shows this history by crop years, beginning with the figure for the crop year 947-48, and continuing through 1984-85.

This chart is a ratio chart, one on which equal distances measure equal percentage changes. It is obvious that since soybeans have moved to a level of 60 cents and more that the percentage changes from year to year have been much greater than during the earlier period.

The end portion of Figure 2, beginning with the figure for 1972-73 is enlarged on Figure 3, and two possible trend lines are added. The level and direction of trend are important considerations because cycles operate around — above and below — the trend. Although it is tempting to consider the straight line trend as the best description of trend, it is also possible to arrive at a different description — as shown by the bowed line. The top straight line trend gives a 1985-6 value of $6.89 compared to about $6.50 for the curved line.

In any event, actual prices through February 1986 have been well below these trend levels for many months, and the yearly average trend that was used to prepare a 1986 calculated line averaged $5.75. By historical standards (the two trends on Figure 3) the value of $5.75 is low, but it is above the last eight month's experience.

The cycles that were combined with trend were the 12-month, the 24.56-month and the 38.6-month. The record of the 24.56-month cycle is shown on Figure 4, and the 38.6-month cycle is recorded on Figure 5. Both charts are marked to show how the cycle has performed since it was first defined.

The 24.56-month cycle has an average amplitude of 10% of trend at the time of an average crest. There was an ideal crest at May 1985, and the model pattern is now on the way down to a trough at May 1986.

The 38.6-month cycle has an average amplitude of 12% of trend at the time of an ideal crest. The ideal cycle is now going up to a crest at October 1986. Since the low on the seasonal cycle occurs at October, it will be interesting to see what is actually occurring to soybean prices come October.

The combination of the assumed trend and the three cycles is shown on Figure 6 as the broken line.

The cycles used here measure as being statistically significant. That is, the numbers are good. But, even more important is the degree to which these cycles have functioned since they were first postulated.

The average seasonal pattern in soybean prices that was used in the combination is as follows: