EXPENSES IN DISTRICTS I AND II 301 
ratios of total expense to earning assets for banks each year 
deviate from the group average for the combined years 1923 
to 1925, inclusive. According to this table, the bank ratios were 
generally low in 1923 and high in each of the other years. To 
this general rule for the combined groups, there are exceptions 
in each of the separate groups for one or more years, but 
in no year is the situation of the random type, nor is there 
serious disagreement, as between the two methods of classifying 
the banks, respecting the conditions which obtain. In general, 
and on the average, what is true for one group of banks is also 
true for other groups. The percentage amounts by which the 
yearly averages differ from the three-year averages for the respec- 
tive groups obviously vary widely. While this is to be expected 
because of the variations in the averages themselves, the varia- 
tions, as will be seen later, are not haphazard. They conform 
to a pattern observed, first, for district ratios for all member 
banks for the country as a whole, and second. for the indi- 
vidual banks in the Boston district. 
From the average yearly levels of the banks in each group, 
changes occurred between 1923 and 1924, and between 1924 
and 1925. So far as direction is concerned, were they of a 
similar type? Consistency in this respect has already been 
observed for district ratios for the country as a whole, and 
for the individual member banks in the Boston district. Do 
similar conditions obtain in the New York district for the years 
in question? From Table 176 it is seen that the directions of 
change, according to the two bases of classifying the banks, were 
upward between 1923 and 1924, and downward between 1924 
and 1925 for all groups except two in each of the respective 
yearly comparisons. Moreover, as is shown by comparing Tables 
165 and 176, the ratios were prevailingly higher in 1924 than 
in 1923 in the respective groups of banks, as classified, both in 
the Boston and in the New York districts. 
But directions, as well as percentage amounts of change in 
ratios of total expense to earning assets, are found to be func- 
tions of the positions of the ratios relative to their own average 
levels. This is true for district ratios for the country as a whole 
and also for group ratios for banks in the Boston district. Is 
it true for the sample banks in the Second district? The answer 
to this question, so far as direction is concerned. is found in