Incomplete and Preliminary Draft

We estimate the incentives to get ahead by hurting rivals in the context of comparative advertising. To do this, we watched all ads broadcast by the US OTC analgesics industry for a 5-year period and coded them according to which brands target which rival brands in comparisons. Data on how much was spent airing each ad then allows us to determine the dollar amounts spent in these attacks. We take these data to a structural model of targeting in which comparative advertising has a direct effect of pushing up own brand perception along with pulling down the brand images of targeted rivals. Brands’ optimal choices of advertising mix yield simple oligopoly equilibrium relations between advertising levels (for different types of advertising) and market shares. These we estimate by using as instruments the prices of equivalent generic drugs; and we use medical news shocks as further explanatory variables. We estimate that each dollar spent on comparative advertising has the same direct effect as 50 cents spent on non-comparative (purely direct) advertising: the remainder is attributable to pulling down rivals. There is strong evidence of damage to targets: each $1 spent against a target needs at least 40 cents to rectify