GROSS EARNINGS IN DISTRICT I 275% 
ages for all of the banks, do the ratios similarly regress to type? 
Answer to this question is found in the types and amounts of 
difference—second year (1925) less than first year (1924)— 
set forth in the last section of Table 157. It will be seen that, 
for ratios greater and for those less than the average in the first 
year (1924), the signs in the column “second year less than first 
year” are minus (-) and plus (4), respectively, and that the 
greater the deviation, either plus (+) or minus (-), in the first 
year, the greater the amounts in the second year. Again, there is 
regression to types—types in these cases being the averages for the 
banks in the respective volume (earning assets) groups in 1924 
and in 1925. Similar regression occurs when the averages, as 
standards of reference, are those for banks classified by the sizes 
of the cities in which they are located. This fact is indicated by 
the detail in the middle section of Table 157. 
In the foregoing methods of treating the data, in order to 
determine the fact and the amounts of regression, deviations have 
been taken as percentage amounts from three types of averages. 
Will similar results be secured if, for the individual banks, in 
the first and second years, the ratios are compared without 
respect to averages as standards of reference? To answer this 
question it is necessary only to classify them into suitable groups 
in 1924, secure an average for each group in this year, and 
compare each amount with the average for the same banks in 
1925. When this is done, as summarized in Table 158 for 
the banks in each city-size group and for all groups, the answer 
is unmistakable. The ratios which increase are those which are 
low; those which decrease are those which are high. Moreover, 
the lower they are, the more they increase; the higher they are, 
the more they decrease. To these general rules, for banks by 
groups without respect to location, there are no exceptions; for 
those classified by size of city of location, the exceptions are 
few and occur within only one group. 
Two methods of study show that ratios of gross earnings 
to earning assets, for member banks in the Boston district, 
regressed to type between 1924 and 1925. Is this tendency 
significant and indicative of the persistent operation of forces, 
not only as between 1924 and 1925 but for other years as 
well, making toward the establishment of an equilibrium? It 
is believed that it is. Let us briefly review our findings with