180 
BANKING STANDARDS 
TasLE IV 
SuMMARY OF RESULTS FOR SEPARATE GROUPS 
GrOUP® 
CHARACTER OF REGRESSION 
—_— 
Net on Total ! Net on Gross 
— — th? I — 
—0.7¢ =a 22 
+ 
+ 
=n 
~r 
—0.40 =*=0.08 
== 29 =0.0¢ 
N74 0.04 
~ esp 1 oa 
; Bh 
=0, 10 
Lo 
.3I 
=0.Y 
*See Table I of Appendix I. 
- ProBABLE ERROR OF 
A SINGLE 
DBSERVATION 
Net on 
Gross 
=2.8 
=2.0 
=3.1 
x1.6 
Groups 2 and 4) and the difference between any group value of 
E.’ and the value for all groups, equation (6), (with the possible 
exception of the value for Group 1 and the value based on all 
groups)—these differences are significant in indicating different 
laws of variation of net earnings between any two groups and 
between the respective groups and the combined groups, with 
the exceptions noted. It is this variation of influence of total 
expense on net earnings which accounts for the distribution of 
the residuals in Tables II and III, and, from the evidence in 
Table IV, it is clear that proposition 1 on page 379 is much more 
true than proposition 2 on the same page. A third outstanding 
fact from Table IV is that gross earnings is the dominant factor 
in determining net earnings in every group, although, relatively, 
it is more important in some groups than in others. The relative 
influences are shown in Table V. 
From this table it appears that annual changes in gross earn- 
ings ratios have from 24% to 133% more influence than do 
annual changes in total expense ratios in causing annual changes 
in net earnings ratios, and that the order of the relative influence 
in the different groups is—Ilargest first—Groups 4, 2, 1, and 3. 
Further inspection of this table shows that for banks having 
gross earnings ratios below and total expense ratios above their 
respective averages in 1924 (Group 4), the relative influence of 
changes in gross earnings ratios and of changes in total expense 
ratios from 1924 to 1925 was Cn =) 5.5 times as great as 
were the relative influences of these changes in Group 3—banks 
having gross earnings ratios above, and total expense ratios 
below, their respective 1924 averages.