Magazine article Insight on the News

Cato's Levy Challenges Federal Tobacco Myths

Magazine article Insight on the News

Cato's Levy Challenges Federal Tobacco Myths

Article excerpt

Research conducted by businessman and lawyer Robert Levy shows that the dangers of smoking tobacco have been greatly exaggerated by the government, raising important questions.

For a quarter-century Robert Levy was an enormously successful businessman as founder and chief executive officer of CDA Investment Technologies Inc. In 1991, Levy left business and went to law school, graduating in 1994. Since 1997 the 58-year-old attorney has been at the Cato Institute in Washington, where he has written on issues such as the tobacco wars, Microsoft and the war against the gun industry.

Levy is no stranger to statistical analysis. His doctoral dissertation, completed for his degree in business at American University, took up quantitative modeling of the stock market. In the fall 1998 issue of Regulation, The Cato Review of Business and Government, Levy published "Lies, Damned Lies & 400,000 Smoking-Related Deaths," an article he cowrote with mathematician and scientist Rosalind Marimont, examining the government's method of coming up with the claim that 400,000 die annually from tobacco-related disease.

The Regulation article was attacked as "unscientific and inflammatory" by the American Council on Science and Health, with which Levy has been having a running debate since the piece appeared. A nonsmoker, Levy has no doubt that smoking is dangerous and can cause diseases, including lung cancer. What he argues, though, is that the threat of tobacco has been vastly exaggerated and that we need to acquire a more balanced notion of just what that threat is.

Insight: In "Lies, Damned Lies, & 400,000 Smoking-Related Deaths" you concluded that the government's estimate of 400,000 annual deaths due to cigarette smoking is unreliable. What's wrong with that figure?

Robert A. Levy: According to the Centers for Disease Control and Prevention, tobacco-related diseases are those in which the rate of risk among smokers is higher than among nonsmokers. But epidemiologists will tell you without exception, I think, that "simply higher" is not enough. In most studies, the requirement to show a correlation is that the risk be three or four times as high.

The reason for requiting a relative risk rate among smokers of three or four times what it is among nonsmokers before categorizing a disease as smoking-related is that epidemiological studies are subject to all sorts of statistical problems.

There's the problem of sampling error. There is the problem of bias. The third problem is what epidemiologists call "compounding variables": that is the failure to take into account variables that are correlated both with the disease and with smoking. The obvious one in this case is socioeconomic status, as smokers tend to be less affluent than nonsmokers. So because of the problems of sampling, bias and of compounding variables, epidemiologists insist that to categorize a disease as tobacco-related the disease has to have a relative risk among smokers that is three or four times that among nonsmokers.

The relative risk of smoking for many types of heart disease is less than 2-to-1, and if you eliminate even those that are just below 2-to-1 you reduce the estimated number of tobacco-related deaths by about 55 to 60 percent.

We know that smokers are poorer than nonsmokers, have worse nutritional intake, typically have less exercise and less education. Those factors contribute to the contraction of various diseases described as smoking-related. To suggest that the entire incidence among smokers is because these people smoke is to ignore that they share exposure to many other characteristics that also impact health.

Insight: What's the mason for the distortion?

RAL: Underlying such manipulation of statistics one often finds a political agenda, a public-policy agenda, which seeks to convince the public that something is a terrible scourge -- and to do so even if the polemic violates the standards and principles of statistics and epidemiology. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.