Comments on Peltzman

Abstract

As does all of his work, Peltzman's article combines thoughtful analysis and empirical investigation to produce a useful, interesting product. But boldness is also a hallmark of a Peltzman job: He always tries to say something important and new. This time some of his ventures do not quite come off. But even his nonsuccesses provide the fruitful basis for future work. I do not question Peltzman's main conclusion that FTC regulation is effective. However, this effort to evaluate the consequences of those regulations has some real difficulties. For the moment, accept Peltzman's general method for determining the possible positive returns to eliminating deceptions from advertising, namely, the increase in customer loyalty produced thereby. There is a serious problem with Peltzman's technique for establishing that increase in loyalty. Peltzman calls any increase in average loyalty among all customers a social return from eliminating deception in advertising. But there is another important reason for variation in average customer loyalty which Peltzman recognizes elsewhere in his article. Repeat-purchase probabilities for experimenters will be substantially lower than this probability for those who have been consuming the brand for a long time. Changing the mix of these two groups will generate most of Peltzman's results, even when there is no deception in advertising. Suppose that the FTC eliminates an effective advertisement and, in consequence, the brand's rate of growth in its market share declines through a reduction in the number of first-time users of the brand. Then there will necessarily be an increase in average customer loyalty because there will be a change in the distribution of customers by the amount of their previous purchases. Since experimenters will be reduced, so, too, will be number of nonrepeaters. Clearly, the evidence that customer loyalty goes up with the elimination of a so-called deceptive advertisement is not sufficient proof that the