|
The following is in response to our article two weeks ago, The New Ptolemains, by C. Thomas Howard.
It is impossible to determine whether a robustly healthy 95-year-old man who smoked all his life owes his health to the cigarette brand he smoked, or only to luck. The statistical fact of his having lived to be 95 is not enough to prove that the cigarette was responsible. Similarly, it is impossible to determine whether an investment manager who performed well over a long period was skillful, or just lucky.2
. . .
Suppose that you’ve stumbled on a remote but populous country of 70 million people in Central Asia, where 99 percent of the population smokes, averaging four packs a day. Let’s call the country Four-Packistan. Its tobacco industry is enormous; there are over 100 companies selling cigarette brands. Each company offers an average of 100 brands, for a total of more than 10,000 brands. Sales of cigarettes are well over $100 billion a year, though the whole industry—as we shall see—is much larger than that.
The Four-Packistani Ministry of Health requires every cigarette company to publish the cancer rate each year for smokers of its brands. As a result, an enormous quantity of statistical data exists on cancer rates for smokers of the thousands of cigarette brands.
Cancer Prevention in Four-Packistan
The industry is highly competitive. Cigarette companies compete by making implied claims that their brands prevent cancer. Though the vast majority of the population smokes, the cancer rate in the nonsmoking population has been determined from the half-million nonsmokers. The tobacco companies’ advertising implies that smoking their brands reduces cancer risk below this nonsmoker level—in other words, that smoking their brand prevents cancer. The companies back their claims with the carefully compiled statistics on cancer rates required by the Four-Packistan Ministry of Health.
Most people in Four-Packistan believe that smoking prevents cancer, provided you smoke the right brands. Many researchers at universities, who are funded by grants from the huge tobacco companies, perform studies to try to determine what characteristics of cigarettes most help them to prevent cancer. The researchers do this by plowing through the mountains of existing data on cancer rates for the ten thousand cigarette brands, running multiple regressions and other “sophisticated” statistical procedures to try to find factors that correlate with low cancer rates.
1Excerpted and condensed from The Big Investment Lie by Michael Edesess, Berrett-Koehler 2007.
2 Even active management advocate Tom Howard, in his recent article in Advisor Perspectives, “The New Ptolemains”, acknowledges, “A manager who has produced good returns could just be lucky rather than skillful.”
Display article as PDF for printing.
Would you like to send this article to a friend?
Remember, if you have a question or comment, send it to
. |
