bers would appear in each prediction to present unequivocal but different-and even contradictory-findings. From one such prediction, three very favorable cost-benefit inferences were drawn to form the basis for a conclusion that federal no-fault standards (S. 354) would be a "pretty good deal" for Texas motorists. The table of predictive data underlying those three inferences has been examined and the possibility of alternative interpretations pointed out. EXHIBIT II THE JOURNAL OF RISK AND INSURANCE (FORMERLY THE JOURNAL OF INSURANCE) PUBLISHED BY THE AMERICAN RISK AND INSURANCE ASSOCIATION, INC. On the Skinning of Cats: Bob A. Hedges Perspectives on Life Insurance Financial Reporting: Robert L. Posnak; Dale R. Gustafson; Gary M. Winkle Beyond Current Reforms in Automobile Reparations: Robert E. Keeton Problems and Issues in Public Employee Retirement Systems: Thomas P. Bleakney; Kenneth H. Ross; Hugh Gillespie Auto Accidents and Alcohol in Great Britain-An Analysis: Kenneth W. Herrick Pricing Problems in Automobile Insurance: An Economic Analysis: Robert Charles Witt Massachusetts: Loss Experience Under No-Fault in 1971: Analysis and Implications: Calvin H. Brainard The Future of ARIA: Harold M. Williams; David L. Bickelhaupt Credit Life Insurance Prices: Joseph M. Belth Reviews: William K. Ghee, J. D. Hammond, William M. Howard, John F. Adams Communications: Craig R. MacPhee, Herbert S. Denenberg and J. David Cummins, Jerome Braverman and Gerald R. Hartman, Charles O. Kroncke Curricular Concepts: George L. Head MASSACHUSETTS LOSS EXPERIENCE UNDER NO-FAULT IN 1971: ANALYSIS AND IMPLICATIONS (By Calvin H. Brainard1) ABSTRACT In August 1972, the Massachusetts Rating Bureau filed with the State Insurance Department the first reliable compilations of aggregate loss experience data developed under no-fault in 1971. Later that same month, at ARIA's annual meeting, these compilations were utilized to compare the loss experience in 1971 with that in 1970, the last year prior to no-fault. A series of data exhibits (slides) was presented with accompanying commentary. This paper reproduces those exhibits and commentary. The new law extended benefits to 42,000 injury victims who would not have beeen eligible for reparation in 1970. This added $18 million to system costs. However, the law also produced an extraordinary decrease of about 75 percent in the number of tort liability claims filed in 1971 as compared with 1970. The net result was a 51 percent reduction in loss costs or $50 million. The cost results produced by no-fault in Massachusetts in 1971 have two basic dimensions: (1) the number of claims made by injury victims and (2) the total losses incurred by insurers. Division of the latter by the former provides a third derivative dimension, the average claim cost. In anticipation of the measurement of these dimensions, the Massachusetts Automobile Rating and Accident Prevention Bureau devised a new statistical plan for the 1971 coding and tabulation of claims. Some knowledge of this plan is essential to an understanding of 1971's cost results. 1 Calvin H. Brainard, Ph.D., is Professor of Finance and Insurance in the University of Rhode Island. This paper was originally prepared and presented by the author as a Report on behalf of the 1971-1972 A.R.I.A. Automobile Insurance Study Section at the 1972 Annual Meeting of A.R.I.A. CODING OF CLAIMS In 1970 and prior years, compulsory coverage A was purely bodily injury liability insurance and only tort-eligible victims, in theory the innocent victims of negligent motorists, were entitled to claim under it. The coding and tabulating of claims was therefore a relatively simple matter. All claims were coded as tort claims and were tabulated at any given valuation date either as paid and closed or as reserved and open. In 1971, with the advent of no-fault, compulsory coverage A was divided into two parts: Division 1 and Division 2. The former remained, as before, liability or third-party insurance; the latter, popularly known as PIP, provided the new no-fault first-party benefit program. This bifurcation of the old coverage A into tort and no-fault subdivisions introduced coding complexities which were aggravated by the new law's threshold effect and its provisions for subrogation and intercompany settlements.3 For example, in 1971, a tort-eligible victim could claim against his own insurer for PIP benefits and also, if his medical expenses exceeded the $500 threshold, against the tortfeasor's liability insurer for pain and suffering benefits. If each of these two insurers were to tabulate a coverage A claim, double counting would result in the sense that there would be two claims but only one claimant. Furthermore, even where the victim's medical expenses did not exceed $500, so that no claim for pain and suffering was allowable, Chapter 670 provided for intercompany settlement of the victim's PIP claim, based on subrogation under tort law principles. That is, when a tort-eligible victim received PIP benefits from his own insurer, the latter was allowed to seek reimbursement for its payment under Division 1 of the tortfeasor's coverage A insurance. To minimize double counting and tabulation confusion, the Bureau's new statistical plan provided for a four-fold classification and coding of claims. Four Claim Codes Code 01 was assigned to liability claims made by injury victims against motorists insured under Division 1 of coverage A. Code 11 was assigned to reimbursement payments made by liability (Division 1) insurers to PIP (Division 2) insurers under subrogation and intercompany settlements. Code 21 was applied to PIP (Division 2) claims and Code 31 to reimbursement recoveries received by PIP insurers from liability insurers under subrogation. A hypothetical case will illustrate the use of these four codes. Case: P, and innocent victim, is struck and injured by D, a negligent motorist, the liability of D to P being undisputed. P incurs medical expenses of $1,000 and pain and suffering valued at $2,000. (In 1970, this tort-eligible victim case would have constituted a single $3,000 coverage A claim against D's insurer.) Settlement and Coding: P receives a $1,000 PIP payment of his medical expenses from his own insured under Division 2 (code 21). He also claims and recovers $2,000 of general damages from D's insurer under Division 1 (code 01). Then, through intercompany settlement, D's insurer makes a $1,000 reimbursement payment (code 11) to P's insurer. This payment is charged to D's Division 1 coverage and is treated as a $1,000 reimbursement recovery (code 31) by P's insurer under its Division 2 coverage. Thus, P receives $3,000 in all and the entire loss lodges finally under Division 1 with D's liability insurer. a The Massachusetts no-fault law was entitled the Personal Injury Protection Act and for brevity is referred to as PIP. It is also referred to as Chapter 670, that section of Massachusetts Laws into which it was written in August 1970, to become effective Jan. 1. 1971. 3 For statutory provisions of the Massachusetts no-fault law see Mass. Laws ch. 670 (1970), as amended by Mass. Laws ch. 744 (1970). Analytical Applications The foregoing illustration presents the Bureau's strategy for tabulating claims and payments under no-fault. The total number of injury claims (one claim per claimant) can be found by adding the industrywide tabulations for codes 01, 11, and 21, and subtracting from the resulting sum the tabulations for code 31. This formula serves to develop several important aspects of aggregate claimscounting under Massachusetts no-fault in 1971. First, when all claims have finally been closed and paid, code 11 tabulations (reimbursement payments) should equal and cancel out code 31 tabulations (reimbursement recoveries). Second, after all intercompany settlements have been effected, the total payments made by all insurers on behalf of injury victims (including allocated claim tabulations under code 01 (tort payments) and code 21 (PIP payments). Third, the final tabulations under code 21 will represent primarily those claims where, either by intercompany agreement or by arbitration procedures, no liability was found to exist. That is, code 21 tabulations will provide a measure of the extent to which no-fault, in 1971, granted benefits to accident victims who, in 1970, would not have been entitled to claims and benefits under coverage A. Fourth, a comparison of code 01 tabulations in 1971 with the total coverage A tabulations in 1970 will provide a measure of the extent to which no-fault reduced the number of claims and benefit payments to injury victims where liability presumably existed. The mandatory reporting of loss experience by all insurers in accordance with a standardized statistical plan of this kind has obvious advantages for a comparison of loss experiences before and after no-fault. It was fortunate in this sense that Massachusetts became the first state to adopt no-fault. Elsewhere, the accumulation of uniform data aggregates would have been more difficult and the impact of no-fault on costs less readily determinable on an industrywide basis. Method of Presentation COST RESULTS IN 1971 Due to the still limited availability of territorial and classification data, it is too early to attempt an integrated and comprehensive analysis in depth of the loss experience developed by no-fault in Massachusetts in 1971. However, it is possible at this time to present the major elements in that experience so as to permit an accurate quantitative estimate of the impact no-fault has had on cost results. These major elements are presented in a series of data exhibits, accompanied by brief commentaries which explain the exhibits and relate them one to another in a fairly coherent sequence. No claim counts are shown for closed claims under codes 11 and 31 because reimbursements and recoveries have already been effected and are included in the tabulations shown under codes 01 and 21. (See text for meaning of codes.) Source: Bureau tabulations of Aug. 4, 1972, for issue year 1971 valued as of Mar. 31, 1972. (Data are undeveloped.) This exhibit shows the latest tabulations available from the Bureau at the time of this report. They are aggregate undeveloped data for all insurers and all motor vehicle classifications: private passenger, commercial, public, garages and dealers, and miscellaneous. For open claims, the Bureau does not segregate the tabulations for codes 01 and 11. Probably, the actual tabulations for code 11 (reimbursement payments) were less than those reported for code 31 (reimbursement recoveries), insurers of PIP claims anticipating a greater intercompany settlement of their incurred losses than insurers of tort claims were willing to concede. As only the eventual closure of disputed open claims will heal this breach, the Bureau may have decided to balance its accounts in the interim by combining Division 1 claims categories (codes 01 and 11) in a single entry. In any event, the tabulations indicate a total coverage A claim count in 1971 of about 77,000 claims and this compares with a 1970 coverage A tally of about 130,000. No-fault, or factors associated with it, reduced claims overall by about 53,000 cases or by 41 percent. Exhibit 2 shows that approximately 79,000 developed claims, with an average claim cost of $719 each, produced $56.6 million of incurred losses. While 57 percent of these claims have already been paid and closed, nearly 80 percent of the losses incurred under them remains open and outstanding. Therefore, the 1971 estimated loss cost of $56.6 million rests rather heavily on the accuracy with which case reserves have been set and development factors computed. As was to be expected, the closing of PIP claims has greatly exceeded the closing of tort claims. EXHIBIT 2.-1971 COST RESULTS ALL COMPANIES AND CLASSES DISTRIBUTED AS TO COVERAGE DIVISION AND SETTLEMENT STATUS 1 Tort refers to division 1 (liability); PIP refers to division 2 (no-fault). Source: Bureau tabulations of Aug. 4 1972 for the policy year 1971 valued as of Mar. 31 1972 and developed by factors of 1.0942 for losses (which are incurred including allocated claim expense) and 1.0240 for claims. Basic limite of $5,000/ $10 000 only. Compared with the average PIP (Division 2) claim, the average tort claim (Division 1) is expected to cost about two and a half times as much to settle. However, tort claims are only 16 percent more costly than PIP claims in the open case categories. Recalling that the PIP claims represent primarily cases of no liability, it would appear that as regards the more serious injuries, where medical expenses exceed $500, the tort-eligible victims, those entitled to pain and suffering awards, are on average not faring much better after payment of attorney fees than are the tort-ineligible (no liability) victims. This result runs counter to expectations and will require further analysis. As indicated under Exhibit 2, the 1971 estimated loss cost of $56.6 million rests heavily on the accuracy with which case reserves have been set. In this connection, as shown in Exhibit 3, the nine leading insurers in the Massachusetts automobile insurance market have produced rather curious variations in both amounts paid and reserved under tort and PIP. In view of the large numbers of earned car years written by each insurer it seems unlikely that these variations can be explained by randomness. Actuarial cost projections for no-fault plans have always allowed for an appreciable increase in claim frequency. For Massachusetts, a 30 percent increase in frequency was projected. On the other hand, a very substantial decrease in average claim cost is usually expected. For Massachusetts, the decrease was esti mated to be amount 35 percent. EXHIBIT 3.-INTERCOMPANY VARIATIONS IN 1971 COST RESULTS: 9 LEADING INSURERS Source: Bureau tabulations of May 30, 1972, for issue year 1971 valued as of Dec. 31, 1971. 1 Per 100 earned car years; tort and PIP combined, 1971 data are based on the following allocations for PPA: Earned car years 1,957,232; incurred losses $48,877,585; total claims 70,900 (estimated at 90 percent of total claims for all classifications). * Estimated. Source: Bureau tabulations for 1970 valued as of Dec. 31, 1971, and for 1971 valued as of Mar. 31, 1972. Data are developed by the appropriate factors. As Exhibit 4 shows, the actual reduction in average claim cost (18 percent) was only one half the expected reduction while 70 percentage points separated the actual (-40 percent) from the expected (+30 percent) change in claim frequency. The 51 percent reduction in indicated pure premium is primarily due to the decrease in claim frequency and only secondarily to the savings in average claim cost. Had the claim frequency increased by 30 percent as expected, that is, from .06 to .078, the 18 percent reduction in average claim cost would not have been sufficient to prevent the pure premium from rising (by six percent). Exhibit 4 demonstrated a 40 percent reduction in claim frequency under nofault in 1971. This, of course, is a net change. A major objective of no-fault is to extend reparation to all victims regardless of liability; its purpose is to pay more not fewer victims. And in an important way it did in fact achieve this purpose in 1971. EXHIBIT 5 PRIVATE PASSENGER AUTOMOBILES-TORT CLAIMS ONLY, 1971 VERSUS 1970 1 See exhibit 2 for percent allocations of total claims and losses as between division 1 (tort) and division 2 (PIP). PPA total claims are estimated to be 70,900 and total incurred losses $48,900,000 as developed. Source: Same as for exhibits 2 and 4. |