By Walter Olson
On Tuesday, readers will recall, the New York Times business section ran an article entitled "Behind Those Medical Malpractice Rates" under the byline of Joseph B. Treaster and Joel Brinkley. Yesterday I criticized at length the article's assertion that "legal costs do not seem to be at the root of the recent increase in malpractice insurance premiums". Toward its end, the article goes to make some even more tendentious and misleading comments on the relationship between damage caps and insurance rates. Those will serve as our topic today.
Now, it happens that the impact of damages caps on malpractice insurance rates has been extensively studied by researchers whose opinions on other aspects of the subject diverge vigorously, and the consensus is to confirm the common-sense notion that when set relatively low and not riddled with exceptions, caps do keep insurance rates markedly lower than they would be otherwise. (This isn't to say that the researchers necessarily support caps as a policy matter -- many don't -- just that they find that the impact on insurance rates is in the expected direction and of significant magnitude.) See, for example, Zuckerman, Bovbjerg & Sloan, 1990 (caps on physician liability "significantly lower premiums"), Congressional Office of Technology Assessment, 1993 (at pp. 64-65, summarizing five studies suggesting link between caps on damages and reduction in insurance premiums); Kessler & McClellan, 1997 (finding "substantially and statistically significant lower trend growth in [doctors'] real malpractice insurance premiums" within three years of reforms); HHS, March 2003 (sec. 6, "States with Realistic Limits on Non-Economic Damages Are Faring Better"); Congress's Joint Economic Committee, 2003 (many references); General Accounting Office, Aug. 2003 (see p. 30, "Premium Growth Was Lower in States with Noneconomic Damage Caps Than in States with Limited Reforms"); American Academy of Actuaries testimony, 2003; Congressional Budget Office, 2004 (summarizing, at footnote 11 and accompanying text, yet another such finding). For more summaries of research, see the AMA's Dec. 2004 position paper (pp. 23 et seq.), and Daniel Kessler's comments at pp. 6-8 of his working paper here.
How does the Times handle the task of fairly conveying all this empirical work to its readers? It just ignores the whole pile -- even the 1990 study co-authored by Duke economist Frank Sloan, interviewed himself in the article, which found a "significant" such effect.
Instead, after contrasting the predictably opposed views of insurers and "consumer groups" as to caps' efficacy, it turns to the views of "some researchers" -- the only researchers whose views it will bother to convey on the subject:
Some researchers are skeptical that caps ultimately reduce costs for doctors. Mr. Weiss [Martin D. Weiss] of Weiss Ratings and researchers at Dartmouth College, who separately studied data on premiums and payouts for medical mistakes in the 1990's and early 2000's, said they were unable to find a meaningful link between claims payments by insurers and the prices they charged doctors.
"We didn't see it," said Amitabh Chandra, an assistant professor of economics at Dartmouth. "Surprisingly, there appears to be a fairly weak relationship."
Let's start with the research by Dartmouth assistant professors Katherine Baicker and Amitabh Chandra, available here to SSBN subscribers or for purchase (it's an NBER working paper from last August). Baicker and Chandra say (p. 14) they found "a fairly weak relationship between malpractice payments (for judgments and settlements) and premiums". (Martin Grace yesterday offered some thoughts on the possible weaknesses of their way of looking for such a correlation.) As the authors explain on p. 15, they did in fact find a correlation between premiums and changes in both the number of payments and the size of the average payment: "Premiums seem roughly equally responsive to each of these components. While premiums do respond to increases in payouts, they do not increase dollar for dollar, suggesting that other factors are at work as well." Since I am aware of no one on either side of the issue who contends that premiums track payouts dollar for dollar with no other factors involved, this part of the paper may not count as being as controversial as all that.
It gets worse, though. The Dartmouth/NBER authors emphasize that the point of their paper was not to evaluate the effects of tort reform, but also say (p. 24):
In our data, we do see lower premium and payment growth in states that had adopted tort reforms or damage caps before the mid-1990s, but we do not find that such states have higher physician growth.
In other words, the authors expressly contradict the position for which the Times invoked their authority on Tuesday, namely one of "skeptic[ism] that caps ultimately reduce costs for doctors". (Incidentally, other researchers, unlike those in this paper, have indeed found a link between physician numbers and damage caps.)
So maybe we'd better set aside the Dartmouth paper. What's left? Only Martin Weiss of Weiss Ratings and his ATLA-flogged, thoroughly discredited paper, on which see posts here (Medical Liability Monitor and National Practitioner Data Bank characterize as misleading Weiss's use of their data), here (Kevin Drum, stalwart Litigation Lobby defender though he is, says Weiss's use of medians rather than averages "is so obviously the wrong statistic to use in this case that there must be some kind of axe to grind here"") and here.
Time doesn't permit exploring all the other things wrong with the piece; for example, there's the Times's indulgent treatment of the oft-heard Litigation Lobby claim that the anti-insurer Proposition 103, as opposed to MICRA, accounts for California's favorable premium experience. The paper provides critics with no chance to contradict that assertion, but as the AMA points out on p. 48 of its position paper, about half of medical liability coverage in California is provided by entities that aren't even subject to Prop 103; and the measure's provisions requiring hearings following challenges of any rate increases above 15 percent have seldom kicked in because insurers haven't mostly pursued rate increases that high. (Update: more on 103).
But there's no need to be completely comprehensive in our critique to reach the question: what were the Times editors thinking when they ran this piece?
(Some websites that gave uncritically favorable coverage to the Times piece: Kevin Drum, Thomas Mayo, Tom Kirkendall, Legal Reader, SlowpokeBlog, Honest Partisan, ACS Blog, ThinkProgress.org, Igor Volsky, Words and Numbers Over Tea, Robot-Invasion, No More Mister Nice Blog, Saheli (who seems tempted to construe the gap between the two chart lines as markup, just as we warned some readers would do), and Generic Spot. Peter Nordberg takes a more complicated view. Thanks to Martin Grace, David Giacalone, Paul Winston (Business Insurance), Jonathan Adler (National Review Online) and E.L. Eversman for their kind words. For Part I of our critique, again, go here.)