than that of the best ves6els--$l,795 per policy compared to $96 

 in New England, $405 against no losses in the Gulf Area, and $1,223 

 compared to $22 in California. However, this significant difference 

 is somewhat modified by the fact that the loss-and-expense-to-net- 

 preraium ratio, shown in column 2 of the table, was as high as the 

 break-even ratio of the most efficient insurers only in New England-- 

 72.5. This ratio was below the break-even ratio of the least 

 efficient insurers in the Gulf Area where it was kk.'J and in 

 California with a figure of 22.7. The difference between best and 

 worst vessels insured for protection and indemnity was highly 

 significant both in terms of paid loss per policy and loss ratio 

 in New England- -$27 per policy and 3'1 loss ratio of best vessels, 

 compared to $2,137 per policy and 111. 5 loss ratio of worst vessels. 

 In California, there were no losses among the best vessels as com- 

 pared with $1,384 loss per policy and 331-9 loss ratio for the 

 worst vessels. The above ratings will be used again later for 

 other purposes. 



Second, for verifying the proposition that insui-ance tends to 

 increase hazard, further computations were necessary. The insurance 

 participation ratio was obtained by dividing the percentage of non- 

 insured vessels with a given characteristic into the percentage of 

 insured vessels having the same characteristic. For example, in 

 New England the 51 -6 percent of captain-owned insured vessels 

 (computed from table A-6 in Appendix A) was divided by the 82. k 

 percent of captain-owned noninsured vessels to give an insurance 

 participation ratio of O.63. This ratio means that the propensity 

 to carry insurance for hull was low among captain-owned vessels 

 since the proportion of captain-owned vessels among aJLl insured 

 vessels was smaller (51-3 percent) than the proportion of captain- 

 owned vessels among all noninsured vessels (82. i+ percent). A 

 ratio equal to 1.00 means that the propensity to carry insurance 

 on captain-owned vessels was equal to the propensity not to carry 

 insurance, while a ratio greater than 1.00 would indicate a high 

 propensity to carry insurance on captain-owned vessels. 



Third, the loss experience ratio was obtained in a similar 

 fashion with the only difference that active steel vessels were 

 combined with best active wood vessels and lost wood vessels with 

 worst active wood vessels. For example, the k'J .6 percent of 

 captain-owned best vessels (computed from table A-'^ in Appendix 

 A) was divided by the i|2.2 percent of captain-owned worst vessels 

 to produce a loss experience ratio of I.13. This ratio means that 

 the loss experience of captain-owned vessels whose hull insurance 

 wa^ studied was relatively favorable compared to noncaptain-owned 



88 



