lines amounts to a comparison of the California experience with the ten other states included in the NAIC study and the entire United States. As shown in Table XI, the five California loss ratios were several points below the CNU and national averages. Although permissible loss ratios were not available to the writer and the problem of possible variations in permissible loss ratios among rate filers in each jurisdiction remains, it is unlikely that the weighted permissible loss ratios for the CNU states and the United States would vary from California's as much as their actual loss ratios vary from California's.

According to the second NAIC report, 59 percent of the insurers expressing opinions as to the rate levels for commercial fire insurance in Kentucky and New York felt they were inadequate. Missouri had the third highest percentage of insurers feeling that such rates were inadequate. The New York experience, producing the second highest loss ratio for the four-year period, appears to be consistent with insurer belief; however, five of the eleven states including New York had higher four-year loss ratios than Kentucky and Missouri. It must be remembered that the loss ratios in Tables XI and XII are based upon all types of fire insurance whereas the insurers' opinions were to have been based upon commercial fire only.

G. Homeowners - Multiperil - All Insurers

As shown in Table XIII, California's homeowners multiperil loss ratios were lower than both those of the ten CNU states and the entire United States for three out of four years and for the four years combined. However, as shown in Table XIV, Delaware, North Carolina, and New York had lower four-year loss ratios than California. The homeowners rate filings of the Multi-Line Insurance Rating Bureau, the major rating bureau for this line of insurance, generally have been based upon either a 54 percent expected loss ratio or a 60 percent expected loss and loss adjustment expense ratio. Although deviation and independent filings may be based upon different expected loss ratios, it seems unlikely that the weighted average expected loss ratio would be sufficiently above California's four-year average of .55 to make the state's rate level excessive.

H. Commercial Multiperil - All Insurers

In only one year, as indicated in Table XV, was California's commercial multiperil loss ratio less than either the CNU or the U. S. weighted average. In fact, of the eleven sample states, California's four-year loss ratio (.56) was only exceeded by Florida's (.73). This result is consistent with Mr. Kai-Kee's analysis of California's loss ratios. While countrywide underwriting gains for this line were not excessive during the four-year period under review, the results were consistently better than those for fire, homeowners multiperil, automobile bodily injury and property damage liability insurance.

I. Workmen's Compensation - All Insurers

Of the lines included in this study workmen's compensation was the only one for which no state allowed rates to be used before they were filed. Possible reasons for this are the fact that it is a form of compulsory social insurance and the fact that there is considerably less price competition in the form of deviations and independent filings in this line as opposed to all of the other lines included in this study. In fact, in most states all insurers belong to the same bureau. Thus, in theory at least greater scrutiny of compensation rates by government is justified because of monopolistic pricing power and/or the social and compulsory nature of the coverage. The extreme of government involvement in pricing, selling, and servicing insurance is typified by government operation of a monopolistic state fund. In this line, minimum government participation occurs under an ail-industry committee type of law, with a waiting period and a deemer. In many states statutory actual prior approval is required - even in California.

Workmen's compensation was the third line for which insurer opinion as to rate adequacy was reported upon in the second NAIC report. Of the three lines it had the smallest percentage of insurers (19 percent for all eleven states combined) believing that its rate levels were inadequate. In Florida, however, 58 percent of the responding insurers thought that workmen's compensation rate levels were inadequate. This is fairly consistent with the loss ratios indicated in Tables XVII and XVIII, since Florida had the second highest four-year loss ratio. Missouri had the second highest percentage of insurers (24 percent) indicating inadequate workmen's compensation rate levels. While its four-year loss ratio was the second lowest, there was an upward trend in the loss ratio in Missouri to the point where it was the third highest of the eleven sample states. This may partially account for the relatively high percentage of insurer dissatisfaction with workmen's compensation rate levels in Missouri.

Comparisons of Assigned Risk Plans

Tables XIX and XX reveal that the percentages of premiums earned in the automobile assigned risk plans of the two CU states were smaller than those in both the nine CNU states and the country as a whole for each of the three years and for the three years combined. The percentages for the nine CNU states were affected considerably by New York and North Carolina, which in 1965 had the third highest and highest percentages of assigned risk premiums earned in the country, respectively. The three-year percentage for these two compulsory insurance states combined (9.60) was over 2.6 times the national average. When the data for these two states are removed from the CNU states, their three-year percentage (2.19) falls below that of the two CU states (2.61). Since New York and North Carolina are compulsory states, it would appear that something other than the type of rate regulatory law is a major determinant of rate of assignment to assigned risk plans. However, Massachusetts, which also is a compulsory state, has a rate of assignment less than the national average. Of the eleven states, Missouri, a CU state, consistently had the smallest percentage of assigned risk premiums whereas California, the other CU state, ranged from the fourth to the fifth largest percentage.

Comparisons of Price Competition as Reflected in Market Concentration

A. Automobile Bodily Injury and Property Damage Liability

Because of time limitations and the fact that the concentration percentages appeared to be relatively stable over three and four-year periods, only the 1965 data are presented. Column A in Tables XXI and XXII indicates the percentages of each state's total automobile liability premiums that were written by the company or affiliated group of companies writing the largest amount of those premiums in the state. The leading writers in the two CU states wrote 12.03 percent of the total automobile liability premiums written in those states whereas the leading writers in the nine CNU states wrote only 11.16 percent of the business. However, the U. S. weighted average of 12.69 percent is larger than the CU states. So, while one might conclude, on the basis of the eleven-state experience, that the leading writer tends to capture a larger share of the market in states in which rate filers appear to have more flexibility, the national data do not wholely support this conclusion. In fact, the two sample states, North Carolina and Texas, which require all insurers to charge the same rates, are at opposite ends of the eleven-state concentration spectrum shown in Column A of Table XXII.

In the eleven states there was a range of over six percentage points in the percentages written by the leading insurers. The national range was over seventeen percentage points based upon a low of 8.38 percent in Rhode Island and a high of 25.89 percent in Wyoming. Factors which account for such differences, other than the type of rating law, may be the number of insurers doing business in a state, the size of the premium volume in the state (which may be highly correlated with the first factor), and the importance of domestic or regionally operating insurers. Other factors which also should be considered in an analysis of market concentration are the length of time the major national writers have been active in the state and the relative rates of growth of the different types of insurers.

Column B in Tables XXI and XXII shows the percentages of each state's total automobile liability premiums that were written by the NBCU member company or group writing the largest amount of these premiums of the top ten companies or groups. The zero percentage in North Dakota indicates that no NBCU member or group was among the state's top ten writers. In this category, the two CU states had a lower percentage than both the nine CNU states and the entire U. S. Thus, it would appear that where insurers have more rate freedom, the leading NBCU member er group in the state nets a smaller share of the market. Column D indicates that the same conclusion appears to apply for the share of the market written by all NBCU members and subscribers as a group, since in the two CU states they wrote only 30.51 percent of the premiums whereas in the nine CNU states and the entire U. S. they wrote 33.01 and 31.70 percent, respectively. Of all the states, Massachusetts, which is a "statemade uniform rate" state for the majority of automobile liability premiums, was the one in which NBCU affiliates wrote their largest share of the market (53.46).

Column C demonstrates that premium writing concentration, as measured by the share of the top ten, is greater in the two CU states (57.86) than in both the nine CNU states (50.05) and the entire nation (53.23). These figures give an even stronger indication than those in Column A that market concentration tends to increase in competitive price flexibility.

39-280 71 - pt. 17A 5

D. Automobile Physical Damage

The concentration picture for automobile physical damage is very similar to that for liability. As shown in Column E of Table XXI, the leading bureau members (this time NAUA rather than NBCU) in the top ten wrote a larger percentage of the market both countrywide (6.18) and in the nine CNU states (6.90) than they did in the two CU states (3.63). A similar situation prevailed for the share of NAUA members and subscribers combined, as shown in Column G, where the 1965 percentages were 29.40, 43.71 and 41.24 for the two CU, the nine CNU, and fifty-one jurisdictions, respectively. Thus, the data seem to correlate with the historical emphasis given to non-price, as opposed to price, competition by bureau insurers.

Column F of Table XXI also reveals greater concentration by the top ten writers in the two CU states (54.27) than in the nine CU states (47.54) and the fifty-one national jurisdictions (51.09). Although the concentration percentages in each state were higher for liability than for physical damage, the state rankings for these two categories were similar. For example, the smallest amounts of concentration for both categories were found in Texas, Florida, and New York in that order.

Comparisons of Product Innovation

A. Automobile Insurance

In the automobile insurance field three examples of "product innovation" are examined the Special Automobile Policy (SAP), Safe Driver Insurance Plan (SDIP) and the New Classification and Rating Plan (NCP). Missouri was one of the first four states to approve the SAP for use in September, 1959. The SAP, however, did not become effective in California until February, 1962. Only two states approved the plan for use after February, 1962, but of the six states which still have not approved the SAP, half were CÑU states included in the NAIC sample-New York, North Carolina and Texas.

California and Missouri were the first two states in which the SDIP became effective. Texas is one of the three states in which this plan is not in effect. The two CU states also were among the twenty-seven jurisdictions which approved the NCP for use effective January, 1965. However, Florida, Kentucky, New York, North Carolina and Texas are among the seventeen states which have not approved this plan.

Thus, it appears on the limited basis of these three innovations, that the CU states are more receptive to experimentation than are the CNU states.

B. Fire Insurance

Two examples of "product innovation" in_fire insurance are the Dwelling Loss Constant Plan (DLCP) and the All Perils Dwelling Deductible (APDD) both of which were intended to help alleviate the deteriorating residential loss experience. The DLCP was designed primarily to increase premiums on low valued properties while the APDD was designed to curb the pressures on the rate level by eliminating first dollar coverage on losses due to any peril-not just losses due to wind and hail for which deductibles already were applicable. Of the eleven sample states, California, the only CU state for fire insurance, was the seventh to adopt the DLCP; however, this plan has yet to be approved in three of the ten CNU states-Delaware, New York and North Dakota.

California was the fourth state, of the eleven, in which the APDD became effective; however, this deductible is not available in two of the CU states. The extent to which these innovations were needed because of loss experience may have varied considerably among the states, and therefore, very little weight can be given to their acceptance as a measure of the degree to which innovation is dependent upon the type of rate regulatory law.

Summary and Conclusions

A comparison has been made of underwriting variables reflecting rate adequacy, price competition, product innovation, and use of assigned risk plans for the eleven states being studied by the NAIC. Comparisons also were made for certain groupings of states, e.g., the group of states within the eleven which prohibited the use of rates before they were filed (CNU) and those which did not have such a prohibition (CU). The CU states consisted of California and Missouri for the automotive lines and California alone for the fire and multiperil lines. However, not a single state could be included in this category for workmen's compensation insurance. From an insurer viewpoint the average experience was more favorable in

CU than in CNU states. Yet rate levels did not appear to be excessive in the CU states. Also in the one line, commercial and multiperil, for which the national underwriting result was generally profitable, California's loss ratios were higher than average.

While the CU "group" generally had lower loss ratios. lower percentages of earned premiums derived from assigned risk plans, greater percentages of market concentration when bureau affiliation was disregarded, less market concentration for the major bureau affiliates, and no less acceptance of product innovations than the CNU states, one should be reluctant to generalize as to the degree to which these relationships would continue to hold if more states actually fell within the CU "group." This conclusion appears to be supported by the variation in results among the more numerous CNU states. In all nine of the tables in which four-year combined loss ratios were arrayed, a CNU state was found at both the top and the bottom of the array. In two of these tables, New York, a CNU state, had the lowest loss ratios while in two other tables it had the highest.

The results of this study generally were consistent with the results indicated in the two NAIC reports on rate regulation in the eleven-state sample. For example, in the first report more insurers were said to have cited Florida as the state in the Union with which they had the most serious problems.24 Florida had the highest four-year loss ratio in four of the nine tables of arrayed loss ratios. In the other five tables Florida had the second and third highest loss ratios two and three times, respectively. On the other hand, the apparently greater than average insurer dissatisfaction with New York does not appear to be entirely justified by data presented in this study. In two tables it had the lowest four-year loss ratio, in one the second lowest, in two the third lowest, in one third highest, in one second highest and in two it had the highest loss ratios. Because of differences in premium volume insurers may well be more dissatisfied with a 5 percent underwriting loss on one line in New York than they would be with larger percentage losses on more than one line in Delaware or North Dakota, for example.

In the first NAIC report it also was stated that "a large number of insurers commented that the administration of the law is far more important than the law itself. Indeed, this was the major theme of many responses."25 When the results for individual states having the same type of rating law are compared rather than the results for groups of states, the same theme seems to be applicable. Notwithstanding the fact that opportunities for arbitrary and capricious actions (at least from an insurer's viewpoint) by state insurance departments probably would be diminished under no file laws, it is difficult to believe that differences in the administration of such laws would not materially influence underwriting results.

Please indicate effective date of law changes, if any.

Key: AIC-All Industry Committee

APA-Actual Prior Approval

FU-File and Use

MB-Mandatory Bureau

MPA-Modified Prior Approval

MSF-Monopolistic State Fund
ND No Deviations Allowed
NF-No Filing Required
SM-State Made Rates

Questions

1.

2.

3.

4.

5.

Does your department encourage use of the deemer provision where it is applicable? Yes

No

For what kinds of insurance, if any, do all insurers belong to the same bureau even though the statutory law does not require this?

Do all insurers belong to the same bureau at the insurance department's behest or because of industry preference?

For what kinds of insurance, if any, do all insurers charge identical rates (ie., no deviations and/or independent filings are made) even though the statutory law does not require this?

Are identical rates charged at the insurance department's behest or because of industry preference?