From the issue dated May 31, 2002
|
|
|
Calculated RisksHarvard professor says smokers know exactly what they're doingBy DAVID GLENN W. Kip Viscusi spends his professional life thinking about accidents and catastrophes.
Mr. Viscusi, the founding editor of the Journal of Risk and Uncertainty, tends to be risk-averse himself. He recently declined an invitation to endorse a class-action lawsuit against Osama bin Laden brought by survivors of the September 11 attacks. "I thought, Do I really want my name on a Web site somewhere where Al Qaeda might see it?" He recalls a day when workers building his Colorado vacation home invited him onto the roof to inspect their progress. "You expect me to climb up there?" he asked. He couldn't bring himself to do it. And he has never smoked. "It's a significant risk," he says. "It's not something like EPA studies of rare chemical carcinogens, where there are a lot of zeros before the number." For the past 15 years, however, Mr. Viscusi has spent much of his energy defending the rationality and judgment of those souls, more risk-tolerant than he, who choose to light up. As a chief witness for Big Tobacco, he has made a compellingly simple argument: Smokers have a very solid understanding of tobacco's health risks. In multibillion-dollar lawsuits, individual smokers and state attorneys general have claimed that tobacco companies have duped the public about the risks of smoking. But Mr. Viscusi insists that just isn't so. Even if tobacco companies have churned out misinformation and propaganda, he says, they've failed to deceive their customers. More controversially still, Mr. Viscusi declares that state governments enjoy a net fiscal gain from each pack of cigarettes sold, even before excise taxes are taken into account. Because smokers tend to die young, he argues, governments incur lower nursing-home and health-care costs for smokers as a class. So the major state and federal lawsuits against tobacco, which have sought reimbursement for smokers' Medicaid expenses, are based on a false premise. According to Mr. Viscusi, total Medicaid costs would actually be higher if there were no cigarettes. (That analysis, which one of Mr. Viscusi's critics calls "obscene," was initially disavowed by the tobacco industry itself.) In two new books -- principally Smoke-Filled Rooms: A Postmortem on the Tobacco Deal (University of Chicago Press) -- Mr. Viscusi ties his studies of smoking-risk beliefs into a broader argument about product-liability litigation, which he believes has spiraled out of control. It's unfortunate, he argues, that the tobacco industry settled the lawsuits brought by the state attorneys general, because as a result, the states' legal theories were never tested in court. In the wake of that $243-billion settlement have come similar lawsuits against manufacturers of handguns, lead paint, and even junk food. "The cigarette-litigation experience is not a single misguided venture," he writes. "The regulation-through-litigation route is now viewed more generally as a financial cash cow." The debates surrounding Mr. Viscusi's work highlight more general criticism of "law and economics" scholarship. Mr. Viscusi's critics, many of whom are social psychologists, claim that his studies of the public's smoking-risk awareness rely on a simplistic and naive model of human decision making. They assert he has force-fed his data through a predetermined ideological funnel, in which all human behavior is coolly rational, in accordance with the tenets of neoclassical economics. Skeptics also raise doubts about Mr. Viscusi's evidence -- a survey that was designed not by social scientists but by a polling firm hired by Big Tobacco's defense lawyers. "Why would any scientist in his right mind choose to base his entire theory on a survey cooked up by a tobacco-industry law firm?" asks Stanton A. Glantz, a professor of medicine at the University of California at San Francisco. "In my mind, I put Viscusi in this sort of cabal of academics who make a good living off of tobacco money." "I have a colleague who was in town for a conference," Mr. Viscusi says, "and he mentioned to a strongly antismoking health economist that he was going to have dinner with me. Well, this [health economist] reacted as if he'd heard the name Darth Vader. So there are some strong feelings out there about this. But I present this stuff in my classes, and no one walks out." His critics' approach is itself simplistic and naive, insists Mr. Viscusi. He emphasizes that most of his work (and almost every paper in the Journal of Risk and Uncertainty) concerns fallacies in decision making. Far from assuming a hyperrational Homo economicus, he says, his work focuses on human error in such venues as the jury box and the casino. A Price Tag on Human Life At a conference table in his Harvard office, Mr. Viscusi argues his points in a calm, only-going-where-the-data-take-me manner, punctuated by moments of dry wit. He is a compactly built man of 52, with thinning gray hair and a clear baritone. Mr. Viscusi was raised in Louisville, Ky., the son of two teachers -- and smokers. "We collected Raleigh coupons," he recalled during a legal deposition in February. Both of his parents stopped smoking during the early 1960s, shortly before he started high school. That was an era when the health risks of smoking were being widely publicized, and he describes his parents' decision to quit as highly analytic: They saw the data, assessed their own risk tolerance, and decided to stop. As an undergraduate economics major at Harvard, Mr. Viscusi intended to become a lawyer. (He never did; he is now one of two full-time faculty members at Harvard Law School with no law degree.) During the summers of 1970 and 1971, following his junior and senior years, he worked in one of Ralph Nader's organizations on an exposé of federal irrigation policies, which he published, at age 23, with Richard L. Berkman. The book charged the government with favoring agribusiness over small-time farmers in allocating water and with building environmentally catastrophic dams on Western rivers. His tenure as one of Nader's raiders "might be an interesting contrast," Mr. Viscusi concedes, to his current work. But he says that the experience had a big impact on his career: "That's how I got involved in all these health, safety, and environmental issues. ... We used economic analysis, cost-benefit analysis, to show that the government was doing the wrong thing from an environmental viewpoint." That is, he notes, the same toolkit he uses today. "Kip was and remains brilliant," says Mr. Berkman, now a lawyer in Philadelphia. "He had all the insights [on the irrigation policies], and I turned them into Reader's Digest language. He works very hard and writes very fast." Having gotten a taste of public-policy work, Mr. Viscusi returned to Harvard for graduate school, completing a Ph.D. in economics in 1976. His dissertation argued that workers implicitly demand a wage premium for jobs with a high risk of death or injury. That phenomenon was noted by Adam Smith as early as 1776, but Mr. Viscusi tried to advance the theory by measuring the precise size of the premium. He concluded that for each one-in-10,000 annual chance of dying on the job, workers tend to demand an extra $400 a year. Mr. Viscusi concluded his dissertation with what turned out to be a hugely influential suggestion. If workers make such a wage demand, he argued, then they're acting as if their lives were each worth $4-million -- a calculation similar to those made by insurers in pricing their coverage. Therefore the government should consider $4-million to be the statistical value of a life -- that is, for example, when the Environmental Protection Agency debates whether a given regulation is cost-effective, the answer is yes if the regulation will save at least one life per $4-million of cost. That was a much higher value than the government had theretofore used; earlier models had been crudely based on estimates of workers' lifetime earnings. From 1976 to 1996, Mr. Viscusi bounced between appointments at Northwestern and Duke Universities. He founded the Journal of Risk and Uncertainty (typical article: "Evaluating the Injury Risk Associated With All-Terrain Vehicles: An Application of Bayes' Rule") in 1987. And he worked frequently as a government consultant, preparing reports on the cost-effectiveness of various agencies. In one widely publicized study, he concluded that child-safety caps on prescription-drug bottles had been a failure. Some consumers are overconfident about the caps' ability to deter children, and so they carelessly leave the bottles where kids can easily reach them. Other consumers have so much trouble opening the caps that they simply leave the bottles open. The net effect, argued Mr. Viscusi, is that child-protection caps had done nothing to reduce the rate of child poisonings. A Forgotten Survey Mr. Viscusi became interested in tobacco in the mid-1980s, when he spent a year at the University of Chicago. There the economist Gary Becker was beginning to develop his theory of "rational addiction" -- the notion that consumers have already weighed the potential risks and costs of addiction even before they begin habits like smoking or drinking. The idea intrigued Mr. Viscusi, and he wondered: What do consumers know about the risks of cigarettes? That question was lurking in the back of Mr. Viscusi's mind a year or two later, when he was hired as a consultant by a law firm representing R.J. Reynolds, the tobacco company. He was engaged as an expert on warning labels, as he'd done research for the government on the efficacy of various kinds of chemical warnings. "They brought me to their offices," he says, "and they had a couple of file drawers full of stuff on warnings. They said, Have a look through here and see what you find." In one of those drawers Mr. Viscusi came upon a file that would change his professional life: a survey of consumers' risk awareness, financed by two law firms for tobacco companies and conducted in 1985 by Audits and Surveys Inc., a research company based in New York City. Here, Mr. Viscusi believed, was a gold mine. The survey asked respondents such questions as "Have you heard that cigarette smoking will most likely shorten a person's life?" and "Among 100 cigarette smokers, how many of them do you think will get lung cancer because they smoke?" In Mr. Viscusi's reading, the data suggested that the public is highly aware of smoking risks. The respondents surmised, for example, that 43 percent of smokers will get lung cancer because they smoke (the scientific consensus is that the actual figure is around 15 percent). The law firms had ordered the survey as part of their preparation to defend R.J. Reynolds in personal-injury lawsuits, but for some reason they had never chosen to use the data in court. Until Mr. Viscusi came along, the file had sat in its drawer, half-forgotten. The Audits and Surveys data, along with similar surveys conducted in 1991 and 1997, form the backbone of Mr. Viscusi's argument about risk beliefs. At least since the 1964 surgeon general's report, he says, Americans have been very aware of tobacco's health risks, and it's nonsense for lawyers to claim that tobacco companies have blinded the public. In Smoking: Making the Risky Decision (Oxford University Press, 1992), Mr. Viscusi wrote that he had found "important manifestations of economic rationality" in consumers' decisions to smoke. Far from being confused addicts, smokers are highly knowledgeable about the risks to which they expose themselves. To critics, the peculiar origin of Mr. Viscusi's data taints his arguments at their core. Both the survey's original role (as a potential piece of legal artillery for tobacco) and Mr. Viscusi's role when he discovered the data (as a hired-gun consultant) naturally cast doubt on his conclusions. In reply, Mr. Viscusi has encouraged skeptics to replicate his methodology, and has offered to make his raw data available. But Mr. Viscusi draws a distinction between his work as an expert witness (for which he has been paid at hourly rates of $500 and up -- a total of "over $600,000" since 1987, he said in a deposition in February) and his work as a researcher. None of his academic work, he says, has ever been financed directly or indirectly by tobacco money. "I was not going to be reimbursed for anything I did with this data," he says. "They gave it to me and turned me loose, in two senses: They weren't going to pay me a cent for my analysis, and they weren't going to be reviewing what I did with the data or interfering with my publications." (In fact, at least one lawyer at the firms did see a draft of Mr. Viscusi's first paper before they gave final approval to his use of the data.) Through a Glass Darkly Alleged conflicts of interest aside, Mr. Viscusi's critics raise several substantive questions about his approach to the public's risk beliefs. They claim that he has failed to take into account the insights of social psychology, and that he views his survey data through a crudely economistic lens. Among the principal charges is that Mr. Viscusi's surveys suffer from "third-person bias." The Audits and Surveys questionnaires asked people how many of a hypothetical 100 smokers will contract a given disease, or how many years of life a hypothetical smoker will lose. The questionnaires did not ask smokers directly how they assessed their personal risks. "Tobacco poses a special, unique challenge to people's rationality," says Paul Slovic, a professor of psychology at the University of Oregon. "A single cigarette is probably not going to harm you physically very much. And that's what you're always smoking: the next single cigarette. You don't make a decision to smoke for a lifetime. It's always, Should I take one more? Surveys suggest that people know the risks and want to quit, but they're wildly overoptimistic about the likelihood that they actually will quit within a given time frame." To that point, Mr. Viscusi replies: "It's true that for some people, quitting smoking is very, very difficult. But I'm not sure what [Mr. Slovic's questions] are actually capturing. In a heavily antismoking society, one way to shut up your critics is to say, Well, of course I intend to quit. So his results might seriously overstate the number of smokers who actually intend to quit, and how serious their intention is." Critics also charge that, contrary to Mr. Viscusi's findings, heavy smokers are foolishly optimistic about their odds of living to age 75. A 1997 study by Michael Schoenbaum, an economist at the RAND Corporation, used data from the long-running federal Health and Retirement survey to assess whether heavy smokers among the "near elderly" (that is, from 55 to 64) correctly understand their health risks. By matching respondents' answers to their eventual dates of death, Mr. Schoenbaum concluded that heavy smokers overestimate their likelihood of reaching 75. Mr. Viscusi replies that the Health and Retirement questions are unreliable, because they ask respondents out of the blue whether they expect to reach 75. Without being given baseline information about life-expectancy rates, Mr. Viscusi says, respondents will provide highly inconsistent replies that don't necessarily have the properties of true probability estimates. The debates surrounding Mr. Viscusi's risk-belief studies have been fierce, but the response to his cost-of-smoking studies has been even more high-pitched. In their late-1990s lawsuits, the state attorneys general claimed that, to atone for various sorts of bad conduct, tobacco companies should reimburse state governments for tobacco-related Medicaid costs. In a 1995 study, however, Mr. Viscusi concluded that states bear no net tobacco-related costs. True, smokers incur heavy health-care costs while they're alive, but as a class, they more than make up for that by dying relatively young. In Smoke-Filled Rooms, he estimates that in 1995, the state of Mississippi (which was one of the most aggressive in pursuing the Medicaid claim) enjoyed a fiscal benefit of as much as 12.5 cents per pack sold, depending on how one adjusts for the declining levels of tar in cigarettes. "I think this whole argument is obscene," says Dr. Glantz, the University of California medical professor. "Taken to its logical extreme, you would basically do no public health. The most cost-effective thing would be to have mandatory abortions." Mr. Viscusi replies that he did not start that line of argument. The state attorneys general made specific claims about net costs, and those claims should be evaluated on their own terms. "Is it a good thing that smokers die, because we save all this money?" he asks. "That's clearly not the message of my work. But if we accept as a given that smoking is dangerous, we should at least play this out and tally the costs correctly." Rational Smokers, Irrational Jurors In the new book Punitive Damages: How Juries Decide (Chicago), which he wrote with four colleagues, Mr. Viscusi argues that jurors behave in systematically illogical ways when they decide product-liability cases. "Juries are subject to hindsight bias," he says. "They're inclined to believe that they would have acted prudently all along. "We gave two sets of prospective jurors a test," he continues. "We asked them whether to repair a piece of [railroad] track, given a number of imaginary conditions. Then we told them that there was a train wreck at the site. When we gave that test to judges, they were consistent: Roughly the same percentage who wanted to repair the track later wanted to hit the railroad with punitive damages for the accident. But the regular jurors were inconsistent: A high number of them wanted to award punitive damages even when they themselves had said they wouldn't choose to fix the track under the circumstances." So while he consistently argues that smokers' decisions are rational, Mr. Viscusi sees jurors as prone to cognitive error. Is there a contradiction here? "The journal I edit," he says, "is almost exclusively devoted to failures and anomalies in human rationality. There are still some adherents [in law and economics] of the old-school model of the fully rational individual, but people are moving on from that. I remember when I was a visiting professor at Chicago when a guest speaker came to talk about risk and insurance. I raised my hand and asked, 'Do people really understand these risks?' And he answered, 'For purposes of this discussion, we will assume that people assess risks accurately and understand them.' Issue closed. I don't think you'd be so likely to hear that today." http://chronicle.com Section: Research & Publishing Page: A14
|
Copyright © 2002 by The Chronicle of Higher Education