SERIES CORRELATED WITH EXPENSES 245

limited space, to reverse the process of measuring the association
by making ratios of interest on deposits the independent, and
ratios in other series the dependent variables. The order of association
 is the same in both types of comparison; it is only the
degree which differs. Accordingly, as is the case immediately
above with respect to ratios of salaries and wages in terms of
earning assets, correlations are restricted to the individual expense
 items alone. Moreover, it is only the district deviations
which are correlated, ratios of interest on deposits being used
as the independent variable.
Table 145 shows the type and nature of the association between
ratios of salaries and wages and of interest and discounts on
borrowed money and variable ratios of interest paid on deposits,
the amounts in all cases being related to earning assets. It is
apparent, from the showing in this table, that both by direction
and by net percentage amounts of deviation, salaries and wages
are positively correlated with interest payments. Moreover, as
noted above, interest payments are positively correlated with
salaries and wages. That is, in general, whenever and wherever
high or low ratios of interest on deposits (or ratios of salaries
and wages) are found, high or low ratios of salaries and wages
(or ratios of interest on deposits) are encountered. This state-TABLE

 145

CORRELATION OF DISTRICT DEVIATIONS OF RATIOS IN PAIRED SERIES
(Percentage Deviations from the Respective District Averages, Period 1919-1925)

Pacsition

INDEPENDENT VARIABLE—
Ratios of Interest on Deposits to
Earning Assets

Distance fr~m Average

Percentage
Groups

Average
Dercentage

Number
of
District-Years


DEPENDENT VARIABLES—
"Tet Average Percentage

—
—

Salaries
and
Vages
to
"arning
Aceete

Interest and
Discounts on
Borrowed
Money to
Earning
Assets

ThAt

Abhave

rand
o tr
‘to.
Td

k
1

Jelow

cm BVM LL eYTAR

‘'atal

40

3

'3
+54.49