SERIES CORRELATED WITH DEPOSITS 207
the directions (but not the rates) of change from year to year
tend to agree. In other words, ratios of time deposits to total
deposits and ratios of total expense to gross earnings are posi-
tively correlated with respect to position relative to their own
district level, and also as to direction of change from year to year.
The year-to-year changes between 1919 and 1925 in ratios of
time deposits to total deposits were predominantly upward; in
ratios of investments to earning assets, they were almost evenly
divided between increases and decreases. If the upper half of
Table 128 is observed, it will be found that (1) net increases in
ratios of gross earnings, of total expense, and of net earnings are
associated with increasing ratios of time deposits and decreasing
ratios of investments, and (2) that the ratios in each of the
series show a net decrease when the ratios of both time deposits
and investments increase. Moreover, it is seen that the net per-
centage increase of 1.72 in ratios of gross earnings, for the 635
district-years in which time deposits increase, becomes 6.41%
increase for the 34 district-years in which ratios of time deposits
increase and those of investments decrease. For the 31 district-
years in which both ratios increase, the net change in ratios of
gross earnings is a decrease of 3.43%. Similar relations, accord-
ing to direction of change, obtain for ratios of total expense. For
net earnings, however, the type of association is different. The
net decrease of 1.32% obtaining for the 65 district-years in which
the ratios of time deposits increase becomes 5.20% for the 31 in
which ratios of both time deposits and investments increase. On,
the other hand, for the 34 district-years in which ratios of time
deposits increase and in which investments decrease, the net
change in ratio of net earnings is an increase of 2.22%.
On the types of association between ratios of gross earnings,
of net earnings, and of total expense and variable amounts of
time deposits in relation to total deposits, when the ratios for the
correlated series are the deviations from the yearly averages for
the country as a whole, it is unnecessary to speak at length. In
general, however, and in spite of the fact that geographical and
other factors make it difficult to isolate the functional relation-
ships, the inverse correlation between ratios of time deposits to
total deposits and ratios of net earnings to earning assets, found
for deviations from district averages and for year-to-year changes,
holds for districts which were respectively above and below the
