22 BULLETIN 1351, U. S. DEPARTMENT OF AGRICULTURE 
condition estimates were available from 1S90 to date. A study of the 
relation between the condition estimates and the final estimate of 
production for the years 1S96-1913 showed that the condition esti- 
mates came more and more to agree with the final estimate of pro- 
duction as the harvest season approached. Since there is close 
relationship between the final estimate of production plus carry over 
and the September price, as indicated by a correlation coefficient of 
+ 0.91, the coefficients can be taken as approximately measuring the 
increasing accuracy with which the September price could be esti- 
mated from the monthly condition estimates, using condition esti- 
mate in place of production as one factor in the estimating equation. 
The monthly forecasts of production during the growing season, 
which are available since 1912, make it unnecessary to take account 
of condition estimates in predicting the September price, for the 
forecasts of production can be used directly in the estimating equa- 
tion. These production estimates would undoubtedly be more 
accurate than the condition estimates, though their relative accuracy 
as the harvest season approached would approximately compare with 
the coefficients in Table 11. The variables are expressed in terms of 
changes over preceding years. The unit employed for condition 
estimates is that used by the Department of Agriculture. 
Table 11. — Correlation of September future prices of oats during the growing 
season with cash prices in September 
September cash price correlated with- 
l 
Coeffi- 
cient 
May prices of September futures —0. 03 
June prices of September futures +. 25 
July prices of September futures +. 79 
August prices of September futures 4-. 95 
Table 12. — Correlation coefficients of preliminary condition and final production 
estimates of oats, 1896-1913 1 
Final production estimate correlated with- 
Coeffl- 
cient 
June 1 condition estimate 
July l condition estimate 
August 1 condition estimate 
September 1 condition estimate. 
+0.45 
+.77 
+.86 
+.94 
i In these correlations changes in acreage are not taken into account. When preliminary estimates of 
acreage are multiplied by September estimates of condition for the period 1896 to 1922, inclusive, and the 
result correlated with the final estimates of production for the same period, a higher coefficient is obtained, 
+0.97. All variables were expressed as changes from the corresponding months of preceding years. 
CONCLUSIONS REGARDING FUTURE PRICES 
The conclusions from this part of the study may be summarized in 
two brief statements: 
Both cash and future prices of oats are highly sensitive to changes 
in the supply of oats when the supply is known. 
Future prices, unlike cash prices, anticipate supply. Conse- 
quently, future prices which span the gap between one crop year and 
the next vary, as a rule, as widely from cash prices at the time of 
delivery as forecasts of production and carry over at the time con- 
tracts are made vary from the final measure of supply. 
