Discussion

Henry Aaron of Brookings opened the discussion by asking what impact tort reforms in the various states have had on the malpractice insurance market. Danzon responded that the reforms have had long-lasting effects. While most were enacted in the 1970s or 1980s, effects persist even to the present. Since many caps on damages were set in nominal terms, more claims face the constraint over time. The frequency, rather than the severity, of claims seems to drive claims paid.

Steve Carney of Medical Mutual of Maryland agreed that, although there has not been a major increase in the frequency of claims, there has been an increase in the "embedded frequency," the percentage of claims requiring a payment to be closed. Maryland has seen a shift from 22 percent to nearly 30 percent of claims requiring payment. However in the past two years, the average severity of claims rose from $275,000 to about $370,000. Altogether, Medical Mutual of Maryland's total indemnity payout jumped from $43 million in 2002 to $73 million in 2003.

Neil Doherty of the Wharton School summarized the paper's findings as claiming that the roots of a hard market are in the preceding soft market and that, for lack of experience, some insurers under-reserve and are forced out of the market, but that their presence affects prices as other firms respond to them. Doherty suggested that this result seems to contradict previous work by Danzon and Cummins implying that credit risk is reflected in insurance company pricing. If so, customers should be able to differentiate between the under-reserving risk-taking insurers and the established companies with better credit risk. Doherty wanted to know if this work was based on malpractice insurance; if so that may explain the apparent discrepancy here. Danzon replied that, in revisiting the previous paper, she found that companies financially at risk could not charge high premiums and that consumers appear to have examined surplus rather than claims-specific reserves.

Terri Vaughn, Iowa state insurance commissioner, seconded the conclusion that hard markets may originate in preceding soft markets. She suggested that, in her experience with micro-level data, tracing premiums charged to specific practitioners reveals dramatic fluctuations in rates over time. Although Iowa has avoided the "crisis" designation of the American Medical Association, it saw premiums drop by half and then double in two five-year periods. She also observed that rate changes are correlated across subspecialties because insurers rely on relativity factors over time because losses for subspecialties lack credibility.

The noisiness of the data elicited several comments. Larry Cluff pointed out that the National Association of Insurance Commissioners (NAIC) and Medical Liability Monitor both rely on hospital data, which is a diminishing segment of the market (since less than 50 percent of hospitals are in the traditional insurance market). He also noted that many states have only one or two mostly doctor-owned insurance companies and thus wondered about the applicability of the Herfindahl-Hirschman index, which reported high concentrations of insurers in various markets. Scott Harrington pointed out that NAIC data were taken at the state level and were for calendar-year claims costs, which have been corrected for revisions and forecast errors of years past. As a result, these data may not be well linked to the specific medical specialties that were the subject of the paper. For these reasons, he was not troubled by the statistical insignificance of the cost-growth variable in the regressions reported in the paper. Danzon acknowledged the data difficulties as being unavoidable. She offered two competing interpretations of the higher values of the Herfindahl-Hirschman index: Are they caused by regulation, or is greater concentration insulating the market against the destabilizing influences of smaller inexperienced and overly optimistic firms? Perhaps a moderate degree of concentration is good for stability.

Cluff was unsettled by evidence that caps, but not loss experience, lowers premiums, because caps are intended to limit losses. High premiums are due to the long tail of losses. Although costs may be low for the first few years, recent entrants are later forced out.

Danzon defended the paper's use of firm-specific investment...