« "Is There a Crisis in Medical Malpractice? New Evidence from Texas" |
More Simplistic Tort Reform Research »
July 08, 2005
CJD's med-mal math
By James R. Copland
Back in January, I responded to a Bob Herbert New York Times column that attacked President Bush's medical malpractice liability reform plan. Herbert had relied heavily on the "Center for Justice and Democracy" (CJD), a Naderite shill for the trial bar that counts on its board of advisors such luminaries as Michael Moore and Erin Brockovich. (For more on the CJD, see Ted's dissection of their "Zany Immunity Laws Awards" at Overlawyered and my challenge to their misleading statements about the Tillinghast study here.)
Yesterday, the "paper of record" was at it again -- this time on the news page -- trumpeting the results of a new CJD study (PDF) purporting to show "that doctors have been price-gouged for several years as insurance industry profits have ballooned to unprecedented levels." Can this be right? As Ted has argued here, such claims make little economic sense: new entrants would take advantage of the abnormal profit opportunity and enter the medical malpractice market. Instead, though, what we've seen in recent years is medical malpractice insurers losing money and exiting the market. Something doesn't add up.
A look deeper into the numbers shows that, as usual, CJD is "creatively" using statistics to mislead its readers:
(1) The study mismatches cash flows by linking current premiums payments to current claim payments. The Times notes that the study claims that "the increase in premiums collected by the leading 15 medical malpractice insurance companies was 21 times the increase in the claims they paid" from 2000-2004. And indeed, the bulk of the report focuses on the relationship between current premiums and claims. But such year-to-year comparisons make no sense, since the average claim takes about 4.5 years to resolve; indeed, about 12 percent of claims take at least 8 years to resolve. It is thus hardly surprising that premium growth might outstrip current claim growth, substantially, over a short period of time. (The CJD report does give some attention to the relationship between incurred losses and premiums, the only sensible comparison from an accounting perspective. Indeed, the CJD admits that "insurers and regulators typically use the incurred loss ratio as a measure of profitability." Nevertheless, the study argues that "many malpractice insurers have in the past posted incurred loss estimates that ultimately proved to be overstated." While that is the case, such variations are not only inevitable given the inexact science of predicting liability exposure, but loss reserves must be reconciled in the accounting statements each year -- a point that might not be self-evident to those with no background in accounting. Also, it's worth noting that in its discussion of incurred loss ratios, CJD mysteriously (or not so mysteriously, see point 3 below) shows only 2003-2004 dollar total comparisons, otherwise sticking to company-by-company ratios.)
2. The study is seriously skewed by "survivorship bias" effects. The CJD study inevitably generates a disconnect between premiums paid and claims paid (or losses incurred) because it fails to acknowledge and account for the exit from the market of major medical malpractice insurers, which guarantees a skewed result that shows premiums growing much more quickly than they actually are and claims/losses not growing as quickly as they actually are. How so? Well, in 2001, the then-largest medical malpractice insurer, the St. Paul Companies, exited the market. In 2001 alone, St. Paul had collected about $530 million in premiums (and generated an underwriting loss of $940 million--out of over $3 billion in underwriting losses that year for the industry, see here (PDF), p. 13, exh. 3). In 2002, the physician-owned Pennsylvania insurer PHICO went bankrupt; by the end of last year, PHICO still had over $1 billion in claim liabilities. Then, in 2003, The Farmers Insurance Group exited the malpractice market; Farmers had written $231 million in premiums as recently as 2002. The exit of these and other major insurers from the medical malpractice market not only contradicts the basic theme of the CJD study--that insurance companies are scoring unprecedented profits by "gouging" their customers--but the sheer volume of these companies' premiums and claims completely distorts the numbers presented by CJD, whose study includes only the 15 largest surviving companies. Let's see how:
Table 2 on page 7 of the CJD report shows premiums growing from 2001 to 2004 79 percent, an annual growth rate of 21 percent. But wait--CJD only shows the $3 billion in 2001 premiums collected by these 15 companies, and does not account for the $500 million+ collected by St. Paul, not to mention those collected by PHICO and Farmers, among others. Add back in ~$1 billion, and all of a sudden the annual growth rate in premiums from 2001-04 is 10 percent--well above inflation, but completely in line with health care inflation. Moreover, it's important to remember that St. Paul and Farmers are still paying claims, and PHICO has $1 billion in claim liabilities outstanding, so the growth in claims paid presented by CJD, at 2 percent per annum from 2001-04, grossly distorts the real picture.
It is hardly surprising that companies like Lexington (a subsidiary of AIG) and MedPro (a subsidiary of GE Insurance) have had such rapid growth in premiums written in recent years after the exodus of such major players as St. Paul, PHICO, and Farmers; nor is it surprising that claims paid (which, remember, take about 4.5 years to resolve on average) have not yet caught up with that rapid premium growth.
3. The study hand-picks its time frame to generate its results. CJD's study picks as its beginning year 2000. At first glance, such a selection might seem arbitrary and innocuous, but in 2000, the industry generated a record $1.8 billion in underwriting losses, followed up by a $3 billion loss in 2001 (see here (PDF), p. 13, exh. 3). These losses followed naturally from an unanticipated and unprecedented rise in the expected value of med-mal claims: the median jury verdict in med-mal cases rose from $500,000 in 1997 to $712,000 in 1999 to $1 million in 2000, where it has since remained. Thus, CJD's use of ratios (e.g., page 14, Table 3) is inherently skewed. Is 2004's 51% incurred loss ratio out of line with what the med-mal insurance industry saw in the 1990s or before? We don't know, because CJD only begins with the beginnings of the current crisis in 2000.
That CJD's choice of years is not merely arbitrary is also evidenced by its "surplus analysis" (see pages 17-19). Without explanation, Table 5 on page 18 shows insurers' surpluses only for years 2002-04. Why would CJD omit years 2000 and 2001, which it had included in its earlier analysis? Well, because the med-mal insurers lost a ton of money in those years and had to draw down their surpluses, which for the industry topped out at $4 billion in 1999 and then were drawn down some $600 billion by 2002 (see here, slide 11). The increase in surpluses since that time is, therefore, merely a return to normalcy for the industry (plus, probably, a little extra cushion given the recent crisis and heightened concerns about potential future claims). (I also note that CJD's insistence that any surplus over the level deemed "adequate" by NAIC is "excess" and thus somehow unneeded is poppycock, akin to arguing that any bank which holds reserves above the minimum level prescribed by the Federal Reserve is somehow acting improperly. There's nothing wrong with management being risk averse, particularly in light of recent industry insolvencies; remember too that 9 of the 12 monoline insurance companies in CJD's analysis are mutual companies, thus owned by their (doctor) policyholders, not outside shareholders.)
Similarly, CJD's "note about medical malpractice stock performance" (pages 19-20) shows similarly manipulative date selection. It tracks the 3 publicly held monoline med-mal insurers' stock performance from May 17, 2002 through May 17, 2005, as compared to the Dow Jones (see Table 7, page 20). But why 2002? Because, of course, in May 2002 med-mal stocks would understandably be depressed, since the industry had had enormous losses ($4.8 billion) in the two preceding years. Remember that St. Paul had exited the market in December 2001 and PHICO was declared insolvent in February 2002. Let's go back further--to 1999, before the 2000-01 med-mal insurance crisis--and check things out:
Of the 3 monoline public companies CJD examines, only FPIC has had publicly traded stock since 1999. On May 17, 1999, FPIC's stock closed at $45.19, which means that it subsequently declined 34 percent by May 17, 2005. On May 17, 1999, the DJIA closed at 10,853, which means that it subsequently declined 5 percent by May 17, 2005.
When you don't use CJD's hand-picked dates, the situation for med-mal insurers doesn't look so pretty does it? Have med-mal stocks grown a lot since 2002? Of course, but only because regulators have permitted premium increases to cover the increases in expected liability, which were generated by the rapid rise in verdicts in the late 90s through 2000. Again, the change is only natural, and has not yet even returned to the status quo ante.
4. The study ignores all costs insurance companies incur apart from payments to claiments to create the illusion of profitability. Finally, it deserves mentioning that CJD seems to assume that insurance companies' only costs are incurred losses, i.e., payments made to claimants. But insurance companies also have allocated loss adjustment expenses (e.g., their own defense costs, including attorneys' fees and expert witnesses) and underwriting expenses (the administrative cost of writing policies). It's not as if an insurance company with an incurred loss ratio of 50 or 60 percent is making huge profit margins (though we can be sure that when the industry has an incurred loss ratio of 100 percent--as CJD shows that it did in 2001, see Table 3, page 14--it's losing a lot of money). Unfortunately, the costs of insuring tort liability are very high; there's just tons of administrative expense in the system.
In failing to take account for the market exit of some of the industry's largest players, mismatching premiums and losses, hand-picking dates to skew results, and painting a deceptive picture of the insurance industry's profitability, CJD's research is at best shoddy and at worst intentionally misleading. It's somewhat tiresome to rebut these studies, but as long as mainstream media sources pick them up rather uncritically, I think it's a useful exercise. For more information on medical malpractice and the misuse of statistics by the trial bar and its supporters, see my rebuttal to a similar Public Citizen report released a couple of months back.
Medicine and Law