QA 
BANKING STANDARDS 
TABLE 168 
CoMPARATIVE MEASURES OF REGRESSION TO TYPE FOR RATIOS OF TOTAL 
ExPENSE To EARNING AssETS, MEMBER BANKS, BosToN 
FEDERAL RESERVE DISTRICT, 1922-1925 
(Percentage Differences from Yearly Averages for all Banks. See page 293.) 
aerate 
—_— 
PERCENTAGE DIFFERENCES: 
{First Year of Each Pair 
of Years) 
All Pairs ' 1922 and 1923 |} 1923 and 1924 1024 and 1925 
Second Second Second 
wear year year 
less mbes less |Number| Jess 
an than than 
rst First First 
| yom. Rm, 
— an 
DIFFERENCES BY PAIRS OF YEARS 
rem 
Number 
Position 
Groups ! 
Average 
¢o and over 24 
30 and under 40 34 
20 and under 30 ! 108 
10 and under 20 102 
TTnder To ® 5 
Above 
-~11.8 } 
- 7.11 3 
-— e 4 
TF 
-, 2 
—-—0 7 
-7.% 
le 
2 
s | 
' 
—14.3 
—- 5.0 
—- 6.6 
- 1.7 
- 0.06 
Under 10 
ro and under 2¢ 
20 and under ar 
30 and under - 
40 and over 
r » - 
~o . 
5 . 
+ 2.4 
+ 3.8 
+ 2.3 
+ 3.9 
+ 4.3 
42.0! 107 | +2.0) 210 | + 2.9 
Below 
Aversa 
0 
10% 
*The signs relate to the prevailing changes in the ratios themselves compared with the averages in the 
frst and second vears, minus {(—) indicating that the ratios decrease, and plus (+) that they increase. 
did the ratios of the same banks in the first year deviate from the 
averages for that year. The signs show the fact, and the figures 
the amounts of regression to type—type in each year being the 
average for the entire group of banks. Third, that the greater 
the deviation, plus or minus, of the ratios from the general aver- 
age in the first of each pair of years, the greater the amount of 
regression to type in the second of each pair of years. The lat- 
ter fact is indicated by the sizes of the differences “Second year 
less than first year,” which, it will be noted, increase without 
exception for all pairs of years, and generally so for each separate 
pair, with the dispersion of the groups in the first years. 
Briefly, and in summary, Table 168 shows that the percentage 
amounts of change are positively correlated with the percentage 
amounts of dispersion from type. A similar phenomenon was 
observed in Part II, in the analysis of district ratios. It was there 
shown that district ratios of total expense which, in a given year, 
are above their own seven-year average level tend to decrease, 
and that those which, in a given year, are below their own seven-