Is the Tort System Costing the United States $865 Billion a Year?--Posner
A study published last month, and favorably summarized in an op-ed in the Wall Street Journal, estimates that the American tort law system costs the nation $865 billion a year. The study, entitled Jackpot Justice: The Cost of America's Tort System, was written by Lawrence J. McQuillan and other members of the staff of the Pacific Research Institute, which published the study. (The study can be downloaded at www.pacificresearch.org.) How did the authors arrive at that figure, and is it meaningful?
They begin by estimating that the nominal (that is, dollar expenditures as distinct from social costs) annual cost of the tort system, consisting mainly of attorneys' fees and other costs of administering the system plus the amount of money paid to tort claimants in judgments and settlements, is $279 billion, of which $128 billion is the amount paid out to claimants. The estimate comes from a report (U.S. Tort Costs: 2003 Update) by Tillinghast-Towers Perrin, a consulting firm for the insurance industry, with the report's estimate updated to 2006. It is impossible to determine from Tillinghast-Towers Perrin‚Äôs report what the sources for most of its data are, and so the figures I have quoted must be taken with a grain of salt; indeed, so far as I can tell, they may be completely unreliable. They are almost certainly exaggerated, given the financial connection between the firm and the insurance industry.
The authors of Jackpot Justice know the difference between a cost, which in economic terms is a reduction in the amount of valuable resources, and a transfer of wealth from one person to another that doesn't reduce the total amount of resources but merely redistributes them. The $128 billion figure is a transfer, not a cost. But as the authors point out, the opportunity to obtain a wealth transfer can generate a cost--a cost incurred to obtain the transfer (incurring costs to obtain a wealth transfer, when socially unproductive, economists call "rent-seeking"). They assume, without analysis or evidence, that the entire $128 billion is translated into a cost.
They further assume that 28 percent of the $128 billion transfer represents a deadweight cost, that is, a loss of value. They base this assumption on a study which found that increasing the corporate tax rate by $1 generates 28 cents in deadweight costs. The basis of that finding was that a tax, like a monopoly markup, causes the taxpayer, like a consumer, to substitute for the taxed item or activity something that may cost society more to provide but looks cheaper because it's untaxed, or taxed at a lower rate. The authors of Jackpot Justice do not explain why a tort transfer would have the same effect. Of course the threat of tort liability might well alter the behavior of potential injurers--indeed, it is intended to do so--but it might alter that behavior in the direction of greater efficiency, by making potential injurers internalize accident costs. That is the objective of tort law, though imperfectly achieved. Without tort liability, firms would have weak incentives to invest in safety measures to benefit potential victims of the firms‚Äô activities, unless the victims were either their employees or their customers.
With the addition of 28 percent of $128 billion ($36 billion) to $128 billion in assumed rent-seeking expenditures, the authors jack up their estimate of the annual social cost of the tort system to $164 billion. To this they add another $36 billion, on the assumption (it seems---this part of the report is none too clear) that efforts to avoid the deadweight cost will cost that amount. This appears to be double counting, based on the assumption that $36 billion is spent every year in a futile effort to avoid a $36 billion cost. It is possible, however, that some--maybe considerable--costs are being incurred to prevent that deadweight loss from rising from $36 billion to some higher level‚Äîfor example, legal-counseling costs (not included in the attorneys' fees incurred in actual litigation) or costs in curtailed new-product development (see below). But there is no basis for supposing that the sum of such costs would equal $36 billion; nor do the authors make any effort to defend the figure or estimate the actual costs.
They add to their new total of $200 billion the $128 billion transfer, for a grand total of $328 billion. The addition is improper, since the transfer is not a cost. They are adding apples and oranges.
They borrow from another study an estimate that tort liability generates some 2,000 accidental deaths a year by discouraging the introduction of risk-reducing products and services. But they fail to offset against that figure (and its monetization) the number of accidental deaths that are prevented by the deterrent effect of tort liability. In the absence of liability, potential injurers would spend less on safety, at least with regard to potential victims with whom the injurers lack actual or potential contractual relations.
The authors of Jackpot Justice cite a respectable economic study by Daniel Kessler and Mark McClellan which finds that malpractice liability increases hospital costs by 5 to 9 percent; and they treat the entire amount as social waste. But as I said earlier, the aim of liability is to induce potential injurers to spend more on safety, and so the fact that they do spend more cannot be adjudged a failure to improve social welfare. Medical-malpractice law is in its administration rife with inefficiency, but it would be surprising if eliminating it entirely would be all social benefit and no social cost (nor in fact do the authors argue for eliminating it, as we‚Äôll see).
The authors argue that products-liability law is responsible for the loss of $359 billion in new products. They base this on a study by the economists Kip Viscusi and Michael Moore which found that when the ratio of liability costs to sales revenues exceeds 5 percent, firms reduce their investment in R & D. The authors of Jackpot Justice identify product markets such as power tools, fireworks, and cigarette lighters where they believe (relying however on the questionable data in the Tillinghast-Tower Perrin reports) that this ratio is exceeded. Then, using Viscusi and Moore's estimate that 6 percent of the products of the industries that the two economists studied are new products, Jackpot Justice multiplies the output of these markets by 6 percent; with certain other adjustments, this calculation produces the estimate of $359 billion in lost sales.
This is not, however, as the authors believe, a social cost. The social cost is the consumer surplus that the sales of the new products would have produced. Suppose that a product costs $10 and is sold in a competitive market for $10, but that consumers would pay $12 for it if it were priced at $12. Then if the product is not produced, there is a loss of consumer surplus of $2. That, not the $10 in lost sales revenue, is the social cost of not producing the product.
The sum of $328 billion and $359 billion is $687 billion, which is almost $200 billion short of the authors' grand total of $865 billion. The excess malpractice costs and accidental-death costs they estimate at less than $50 billion, so there is still a big gap. I can't figure out how they fill it.
So far in the report, there is nothing about the benefits of the tort system. To estimate those benefits, the authors compare the percentage of U.S. GDP that is accounted for by our tort system with the percentage of GDP accounted for by the tort systems of other developed countries. They base all these percentages on the dubious Tillinghast-Towers Perrin report. The U.S. percentage is estimated at 2.2 percent, twice Germany's and roughly three times Japan's and the United Kingdom's. The average for the foreign countries in the comparison is 0.9 percent, so the authors of Jackpot Justice conclude that the benefits of the U.S. tort system are equal to only 0.9 percent of our GDP. The possibility that our more costly system might generate greater benefits (though not necessarily equal to the greater costs) is ignored. But a more serious weakness is the implicit assumption that a tort system generates benefits exactly equal to its costs. It might generate much greater benefits. Politics, which the authors assume lead to greater than optimal liability, might instead lead, in the countries they compare to the United States, to less than optimal liability. And even if investment in the U.S. tort system has been carried to the point at which the last dollar spent on the system generates exactly one dollar (no more) in benefits, the total costs may be far below the total benefits because average cost may be much lower than marginal cost. This is apart from the possibility that politics may have prevented our investing enough in the tort system to equate benefits and costs at the margin.
The authors' estimate of the benefits (= costs) of the average foreign tort system, when subtracted from the $865 billion "cost" of our system, results (with some further adjustments) in an estimate of an annual excess of costs over benefits of almost $600 billion. The figure, however--the authors' estimate of the net social loss created by our tort system--is, as I have tried to show, fictitious.