CORN AND HOG CORRELATIONS 57 
tinuance of the general trends of the various quantities. Under 
ordinary conditions one is not likely to go far astray in projecting a 
trend for one or two years into the future and measuring the expected 
fluctuations from it. Conditions may arise, however, which make 
any prediction of this sort of little value. The World War, for 
example, caused a radical change in the price level as well as a dis- 
turbance of the normal relations among all the factors. Conditions 
which, as we have seen, had previously determined 70 or 80 per cent 
of the fluctuation of the hog and corn variables, suddenly became of 
minor importance, as new conditions came to dominate the situation. 
The war conditions have not yet ceased to act. For the present, 
then, the system of relations and the prediction formulas are probably 
of little more than historic interest. In time, however, it is to be 
expected that conditions will again reach the stability essential for the 
success of predictions based on correlations. Formulas similar in 
form to those given here but with due regard to the current amount of 
fluctuation in the various quantities, should then become useful. 
Even under these conditions, however, caution is necessary. The 
system of relations among the variables is not a mechanically inevit- 
able one. Fundamental changes may take place. 
Hog prices are determined by a broader vision of the conditions 
than the mere number of hogs which are being received by the 
packers. The amount of breeding, on the other hand, seems to be 
determined more by the memory of past profits than by a sound 
judgment of those in prospect. The result is the cycle of successive 
overproduction and underproduction which is one of the dominating 
features of the hog situation. If means can be devised for obtain- 
ing a sounder estimate of prospective conditions the whole system of 
relations will need revision. The general use of prediction formulas 
would itself be a factor that would tend to modify the system of 
relations on which the predictions were based. This, indeed, would 
be the greatest good which they could bring about. 
The variables in Tables 12 and 13 are represented by the following 
symbols, primes being used to represent those pertaining to a pre- 
ceding year. The year is considered as beginning with the summer 
pack, March 1. 
Table 12. — Multiple regression equations 
Western Hog Pack 
Corn acreage (A) __ Summer price (S) Summer pack (F) Western annual pack (J) . 
Corn yield (Y) Winter price (W) Winter pack (G) Eastern annual pack (K). 
Corn crop (C) Summer weight (D) Summer pork (H) Farm price (Jan. 1.) (L). 
Corn price (Dec. 1.) (P) Winter weight (E) Winter pork (I) 
Summer Hog Price 
Per cent 
Corn crops S=-0.48C -0.07C" 22 
Corn prices, — - S=+0.64P' +0. 24 P" _ 50 
Crop and summer weights S=-0.40C -0. 59 D' -0. 32 D" 69 
Corn price and summer weights S= +0. 53 P ' -0. 44 D' -0. 36 D"_... 76 
Winter Hog Price 
Corncrops W=-0.32C -0.35C -0. 11 C" +0.22C'"... 24 
Corn prices W=+0.39P +0. 48 P' +0. 23 P" -0.27P'" 55 
Crops and summer weights W= -0. 25 C -0. 08 C -0. 38 D -0. 63 D' 67 
Corn prices and summer weights W=+0.47P +0. 27 P' -0. 63 D' 71 
