cluded to some extent in regular appropriations made for administrative expenses of the respective agencies. It is impossible to determine the amounts handled in this manner without reference to the various agencies, and the table is incomplete in this respect.

It must be emphasized that this statement must be considered only a rough compilation and analysis of the appropriations and expenditures for agricultural relief. More exact and conclusive data can be obtained only through exhaustive examination of all appropriations and expenditures by the various agencies. It is impossible to include a break-down into loans and grants, though it is likely that loan repayments in the future will reduce the net amount of expenditures to some extent. No attempt has been made to compute the amounts of interest which have been paid on loans, or which might be paid in the future.

It should be pointed out that any attempt to determine what expenditures might have been eliminated had crop insurance been in force must depend on the particular system of crop insurance under consideration, and the figures in the statement must be used with this qualification in mind.

Data for the accounts in group I were supplies by the Farm Credit Administration. The first four accounts represent activities undertaken prior to the period under consideration, but are included because the Farm Credit Administration reports that it is impossible to segregate them on a State basis from the later accounts. This discrepancy is unimportant, however, as the total amount involved is only $1,167,000.

In group III the figure reflects net expenditures up to November 30, instead of to June 30, but this difference is unimportant since expenditures between the two dates amounted to only about $250,000.

Group IV includes all of the accounts financed from funds provided by the Emergency Relief Appropriation Acts of 1935 and 1936. The figures were obtained from published reports of the Treasury Department. The account "Emergency Relief, Farm Credit Administration, emergency crop loans" is the only one in the entire statement showing expenditures in the District of Columbia; all of the amount so shown is for administrative expenses.

The qualifications described above and in the previous memorandum make it desirable to emphasize that this statement should be considered only a rough compilation of Federal expenditures for agricultural relief during the 10 years ended June 30, 1936. The data have been taken from various sources and are not strictly comparable. Nevertheless, the statement is believed to be substantially representative of the facts.

133719-37-4

FIGURE 3.-Relationship between the insurance coverage per acre and the average annual indemnities that would have been required per acre, 1930-35.1

1 Based on a sample of 75 farms with insurance coverage being 75 percent of the 6-year average yield on the insured farm.

In these charts there appears to be little relationship between the amount of insurance and the average annual indemnities required per acre. For the county in central Kansas the average insurance coverage would be about 8 bushels and the average annual indemnities something over 11⁄2 bushels. But those farms with insurance coverage in excess of 8 bushels show in general no different amount of indemnification per acre than farms with less than 8 bushels coverage.

For the county in Texas the insurance coverage would on the whole be lower and the indemnities per acre somewhat higher than for the county in central Kansas. But, as in the Kansas county, farms with coverage greater than the average appear, in general, to require about the same amount of indemnities per acre as farms with low coverage.

FIGURE 4.-Relationship between the insurance coverage per acre and the average annual indemnities that would have been required per acre, 1930-35.

2 Based on a sample of 75 farms with insurance coverage being 75 percent of the 6-year average yield on the insured farm.

In both of the above charts it appears that the average annual cost in bushels per acre of indemnifying losses is larger for those farms with the larger insurance coverage and smaller for those farms with the smaller insurance coverage. This is probably accounted for by the fact that the losses in these regions have often been complete crop failure for all farms, those with high average yields being a complete loss as well as those with low average yields. It would appear that in cases of this kind, where cost per acre is definitely related to coverage, the loss experience for the individual farm should not be combined with the simple average loss experience for the county but with a line of average relationship such as one that could be fitted to the scatter of points shown in the above charts.

FIGURE 5.-Relationship between the insurance coverage per acre and the average annual indemnities that would have been required per acre, 1930-35.3

*Based on a sample of 75 farms with insurance coverage being 75 percent of the 6-year average yield on the insured farm.

It is clearly apparent that in the county in Ohio farms with higher coverage would have required a smaller average annual indemnity per acre. This suggests that farms with higher yields also have relatively less variation in yields. Perhaps here too, as in the case of the western Kansas and North Dakota counties, the loss experience of the individual farm should be combined not with the simple average loss experience for the county but with a line of average relationship such as one that could be fitted to the points in the chart.

In the chart for the Oregon county the relationship seems to be somewhat the same as in the Ohio county although not so pronounced,

Premium for all seeded acres under plan to require payment of 25 percent of production in excess of average expectation

Premium

Premium in per

per acre

cent of coverage

All years
combined

1930

1931

1932

1933

1934

1935

All years
combined

1930

1931

1932

1933

1934

1935