NORMS AND TRENDS IN EXPENSES 89 
the districts must be above and that some of them must be below 
the average. But it is not required that the same districts should 
be on a given side of the average during the seven separate years, 
more particularly when the deviations from the average, as in the 
case, for instance of St. Louis, are small. Moreover, the districts 
with ratios in a given year farthest removed, plus or minus, from 
the average for the country as a whole, tend to hold the same rela- 
tive positions throughout all of the years. This fact is more ap- 
parent in Chart 18 than in Table 61. 
Chart 18 is drawn, as are the others of this type already pre- 
sented, on a ratio basis.® While a common scale is used for the 
ratios for the various districts, the position of a solid line on the 
chart has no significance except when it is compared with the 
dotted line and with its own seven-year average. An inspection of 
this chart shows, among other things, that (1) for the country as 
a whole, and for each district® the trend of expense ratios was 
upward from 1919 to 1921, the rate tending to be greater be- 
tween 1920 and 1921 than between 1919 and 1920; (2) the ratios 
for the country as a whole and for most of the districts tended 
to fall between 1921 and 1925, the rate of decrease being most 
marked between 1921 and 1923; (3) the ratios, as a rule, were 
higher in 1924 than in 1923; (4) the average for the country was 
lower in 1925 than 1924, but the individual districts were not 
consistently either higher or lower; (5) on the whole, the rates 
of change in the respective districts do not, except in certain 
years, markedly differ from those for the country as a whole; (6) 
the ratios for certain districts are consistently above, and for 
others consistently below, those for the twelve districts combined; 
and (7) the years in which the ratios in the several districts are 
above or below their own seven-year average levels are generally 
the same. 
It should be recalled that we are here discussing the percentage 
relation between two variables—the operating expenses and the 
earning assets—of all member banks in the districts separately 
and combined. Both the expenses and the earning assets changed 
from year to year in each of the districts. Moreover, the abso- 
lute amounts and the composition of each differ markedly among 
the districts. Yet when ratios of expenses to earning assets are 
8 For discussion of the interpretation of such a chart, see page 21, 
"Except for Dallas between 1010 and 1020.