282

BANKING STANDARDS

Confidence in the significance of the results is secured from
these different approaches to the problem so far as the member
banks in the Boston district are concerned. It is true that the
data refer to but two years, but, as indicated in Table 132 for
district differences and in Table 134 for percentage changes from
year to year, analogous relations hold for all districts, for the
Boston district alone for the years 1919-1925, and for the respective
 ‘districts in 1924 and 1925. These results suggest the probability,
 if they do not demonstrate the fact, that the phenomenon
observed in the Boston district for each of the years, 1924 and
1923, and for the two years combined, is not peculiar to this
period and to this district.
It is unfortunate that data are not available for the individual
Boston member banks with which to duplicate, for gross earnings
 ratios, the types of analyses presented in Chapters IX and
X, in which district ratios of gross earnings are correlated with
variable operating conditions. It will be recalled that, according
 to three methods of measurement, such ratios were positively
correlated with ratios of loans and discounts to earning assets.
Moreover, they increased most from year to year when ratios
of loans and discounts to earning assets were high—that is, above
the respective district levels—and increased, and they decreased
most from year to year when such ratios were low—that is, below
the district levels—and decreased from year to year. On the
other hand, gross earnings ratios were generally negatively correlated
 with total deposit ratios and also with ratios of demand
deposits to earning assets and to total deposits. Again, when
total deposit and demand deposit ratios increase or decrease,
ratios of gross earnings decrease or increase, respectively. On
the other hand, ratios of time deposits to earning assets and to
total deposits are positively correlated with gross earnings ratios.
Such are a few of the types of relations found to obtain between
district ratios of gross earnings and variable operating conditions.
But district ratios of total expense are correlated in much the
same way as are district ratios of gross earnings with each of
the above series. Accordingly, gross earnings ratios and total
expense ratios tend to be positively correlated with each other.
This relation is undoubtedly causal so far as interest is common
to both series.
District differences in gross earnings ratios depend primarily