NORMS AND TRENDS IN DEPOSITS 45

tions of change from year to year are studied for ratios classified
according to their positions with respect to their levels in the first
of each pair of years, it is seen that (1) those which are high
in a first tend in the following year to fall, and those that are
low in a first, tend in the following year to rise; and (2) the net
percentage amounts of rise or fall vary directly with the percentage
 amounts of dispersion of the ratios from the seven-year
levels in the first years. The first fact for the 72 year-to-year
changes—six for each of the twelve districts—is summarized in
Table 30. The predominant tendency is for percentages which
are high to decrease; for those which are low, to increase. That
is, there is regression to type, all but 12 of the 72 ratios following
this rule.

TABLE 31

COMPARATIVE PERCENTAGE AMOUNTS OF DEVIATION FROM DISTRICT
LEVELS AND NET YEAR-TO-YEAR CHANGE IN RATIOS OF
DEMAND Drrosits To EARNING ASSETS

DIFFERENCES FROM DISTRICT AVERAGES, 1919-192§

Sign

Number of
Cases

a

Percentage
Groups

Average

5.00 and over
4.00 to 6.00
2.00 to 4.00
Inder 2 an

Under 2.00
2,00 t0 4.00
4.00 to 6.00
6.00 and over

Average

Average
Percentage

NET PERCENTAGE
CHANGE FROM
YEAR TO YEAR

T

—~ 16
06
81
2 ob

3

t. 05
46.00

+2 .60

The tendency in the ratios for the net percentage amounts of
change to vary directly with their percentage amounts of dispersion
 from the district averages is shown by direction and amount
in Table 31. This tendency, so apparent for the ratios of demand
deposits to earning assets, was not found for ratios of total
deposits to earning assets, although it held for the ratios of loans
and discounts and for investments to earning assets, the trend in
the former case being upward from 1919 to 1921 and downward