26 BULLETIN 1224, U. S, DEPARTMENT OF AGRICULTURE. 
from region to region. They must, then, be accounted for by under- 
lying economic causes. 
SUMMARY AND CONCLUSIONS REGARDING ACCURACY AND NORMALITY OF DATAo 
The problem was to find the relation of land income, as measured 
by cash rent, to land value for any given year. Before attempting 
to offer an explanation of the percentage relationship in any one 
area, or the variations of this relationship from region to region, it 
was necessary to determine: (1) Degree of accuracy of the data, and 
(2) whether the relationship between cash rent and land value in 
1920 was typical of other years. In considering the accuracy of the 
data it was found that the ratios of rent to value were not very 
reliable for individual counties, but for groups of counties they were 
highly accurater In regions, however, where the land boom had been 
most active, the census valuations of land were biased above actual 
sale prices. The greatest error resulting from this bias, for the county 
groups used in this study, was 0.4 per cent. (See p. 12.) 
Another problem was to find out whether the ratios of rent to 
value in 1920 were typical of other years, or whether they were to be 
explained by fluctuations of either cash rents or land values from 
their normal trends. It was found that from year to year the ratio 
of rent to value was remarkably consistent in four important areas, 
but that in 1920 land values advanced more rapidly than cash rents, 
thus making the ratios of rent to value in that year lower than normal, 
by as much as 0.5 per cent in Iowa, and it is likely that the spread 
between cash rents and land values was greater in this area than 
any other. It was also found that the relative variations in the 
ratios of rent to value for the different parts of the country were not 
due to abnormal fluctuations in rents or values in those regions. 
This is brought out by comparing Figure 3 and Figure 5. In other 
words, the regional variations in the ratios as shown by the census 
data were typical of other years. 
While it must not be forgotten that some inaccuracies in the data 
do exist, the explanation of the ratios of rent to value in any one 
area or of the variations in the ratios from area to area must be sought 
elsewhere than in inaccuracies or abnormal fluctuations in the data. 
NET CASH RENTS AND THEIR RELATION TO LAND VALUE. 
The cash-rent data which have been dealt with so far are gross 
cash rents. Land values, of course, are based on the net earnings of 
land, so that it is necessary to reduce gross cash rents to a net basis 
if they are to be used to measure land income. To reduce gross 
rents to a net basis the average tax per acre and depreciation and 
repair of improvements must be deducted. 
Data necessary to obtain net cash rents were available in only 9 of 
the 48 groups. These groups are indicated in Table 8. The tax 
data were obtained from the county tax rolls. A figure showing the 
percentage of the building value which must be set aside for depre- 
ciation and repairs was obtained from farm-management surveys 
located in each of these 9 groups or very near them. For all the 
northern groups and for the 2 in California this rate was 3 per cent, 
except in group 9, where it was 3.6 per cent. In the southern groups 
the average rate was 6.6 per cent. The average value per acre of the 
