28 DEPARTMENT BULLETIN 912. 
losses paid by joint-stock fire insurance companies from 1890 to 1919, 
inclusive, will be found of interest. As indicative of the variation in 
the hail hazard from year to year, the figures for premiums and 
losses of the joint-stock companies are, of course, more significant 
than the corresponding figures for mutual companies. 1 Many of the 
mutual companies, as already stated, operate on the assessment plan, 
and hence their premium or assessment income expands or contracts 
each year in proportion to the losses incurred. In the case of joint- 
stock companies, while the rates have been frequently adjusted they 
have, however, been fixed in advance on the basis of past experience. 
Figure 6. it should be noted, compares the premium income of the 
companies as a group, not with the cost of the business to them, but 
only with the losses paid. It is doubtful if any of the companies 
represented in these figures have had an expense ratio much below 
35 per cent of the premium income. In the case of some of them, 
such ratio has certainly exceeded this figure. Assuming that on an 
average 35 per cent of the premiums has been required to cover ex- 
penses of operation, it would follow that the companies, as a group, 
have lost money during each year in which the actual losses, as indi- 
cated in the diagram, have exceeded 65 per cent of the premiums, and 
have made a profit during each year in which the loss ratio has fallen 
below this figure. The net profits in the hail business taken as a 
whole have by no means been as large as many persons have assumed. 
The last four years have been highly favorable, but many of the com- 
panies showed a net loss on their hail insurance experience at the close 
of 1916. 
The variation in the destructiveness of hail in a given State de- 
pends, of course, to some extent upon the degree to which the land 
is given over to one or two commercial crops. Thus, Kansas and 
Oklahoma, for example, with their large winter-wheat acreage in 
proportion to the total acreage in crops, Xorth Dakota and Montana 
with their similarly large spring wheat acreage, or parts of Texas 
with their cotton acreage, are likely to be subject to especially great 
variations. One or more bad hail storms occurring at a critical period 
in the development of the main crop in these States may, of course, 
ruin a relatively large percentage of the total crops on the farms 
in the territory visited by such storms. Even though during the 
next season equally severe hail storms occur, the damage will be far 
less in case such storms occur either before or after the critical 
period of the principal crop. The variation in the annual hail losses 
experienced by insurance companies in States where a single com- 
1 The figures represented in this diagram, as well as those given in the text, are as 
complete and accurate ai> it was possible to make them. An effort to supplement as well 
us to check up the figures given in the various published reports, by data secured direct 
from the home records of the companies, was only partially successful. While a large 
percentage of the companies, and among these the more important pioneers in the 
business, generously furnished all the data requested, other companies for one reason or 
another failed to do so. 
