)2 BANKING STANDARDS their seven-year averages, as is done in Table 63, the following generalizations may be made. The amounts were low—that is, below the averages—in 1919, 1920, and 1921; high—that is, above the averages—in 1923, 1924, and 1925; and mixed— high and low—in 1922. All districts were low in 1920; all but one, Philadelphia, in 1919; and all but three, Richmond, Atlanta, and Dallas, in 1921. Similarly, all districts were high in 1924; all but one, Atlanta, in 1925; and all but three, Philadel- phia, Cleveland, and Richmond, in 1923.8 In 1922 the districts were neither prevailingly high nor low. Speaking broadly, yet with comparative accuracy, then, the years in which ratios of operating expenses to gross earnings in the respective districts were greater or less than the typical amounts—the respective seven-year average levels—were essentially the same for all the districts. Prior to 1922, operating expenses were relatively low; subsequent to 1922, they were relatively high. Dissimilar as are the ratios in the various districts, they have these common yearly characteristics. While the district ratios in all cases but one were low in 1919 relative to the seven-year level, the direction of change between 1919 and 1920 was downward for the twelve districts combined and for all but three of them individually. Between 1920 and 1921, the general direction was upward—that is, operating ex- nenses were encroaching on gross earnings—and to this rise the banks in all but two of the districts contributed. While the aver- age ratio for all districts in 1922 was the same as that for 1921, ratios in Boston and the districts west of the Mississippi River registered a rise, and those in the other districts a fall, between these two years. Beginning in 1922 and lasting to 1924, the av- erage and prevailing tendency was upward; from 1924 to 1925, it was generally downward. The directions of change from year to year are summarized in Table 64. 8 In two districts, New York and Atlanta, the ratios were the same as the seven- year average. NUMBER OF DISTRICTS WITH RATIOS OF Tora. EXPENSE TO Gross EARNINGS ABOVE OR BELOW THEIR RESPECTIVE AVERAGES FOR THE PERIOD 1Q10-102% r