NORMS AND TRENDS IN EXPENSES 105

ratios essentially agree as to the
years which are “low,” “moderate,”
 and “high,” the percentage
 deviations from such levels
for the respectives ratios being
positively correlated; (3) the
year-to-year changes in the ratios
 for the respective groups of
banks are coincident with respect
 to time, and positively
correlated with respect to direction
 and amount; and (4) the
positions of the district ratios
for the two classes of banks,
relative to their respective
yearly levels for the country as
a whole, coincide as to time and
vary directly as to amount.
A graphic picture of the rates of change from year to year of
the district ratios for all member banks, for national banks and
for state bank members, is given in Chart 21. The manner in
which it is drawn is as follows:

TABLE %%

Disagree

1. The vertical positions of the lines for the different types of
banks and for the total membership for each district, and also for
the several districts, have no significance when compared with
each other, the respective lines being drawn so as to reduce as
much as possible any overlapping. This was accomplished by
beginning the lines for each district equally distant apart in 1919.
Accordingly, the relative size of the ratios is #zo¢ indicated. It
is the relative rates of change, indicated by the slopes of the lines,
which are shown for the banks in each district, the percentage
amounts of increase and of decrease being drawn to the scales
shown on the right-hand side of the chart.

2. The short lines at the left of each district indicate the sevenyear
 average ratios for the types of banks to which they apply—
the distances above or below such lines, for the respective types
of banks, indicating the percentages by which the ratios for the
different years are greater or less than such averages.
From the chart it is apparent that (1) there is marked uni-