Professional investors often “talk up their book.” That is, they not only disclose their positions and express opinions, but to back their assertions with data on the allegedly mispriced assets. Examples range from hedge funds presenting their buy or sell recommendations on individual stocks at conferences attended by institutional investors to small investigative firms (like Muddy Waters Research, Glaucus Research Group, Citron Research and Gotham City Research) shorting companies while publishing evidence of fraudulent accounting. Empirical studies find that this advertising activity is typically associated with abnormal returns.
In this paper, we present a model showing that professional investors who detect mispriced securities (“arbitrageurs”) have the incentive to advertise their information about mispriced assets to rational investors with limited attention, as they can profit from it by exploiting the resulting price correction. Insofar as advertising catches the attention of other investors, it reduces the risk incurred by the arbitrageur in liquidating his position due to noise traders. This risk reduction in turn enhances the arbitrageur’s willingness to bet on his private information, engendering a complementarity between advertising and investing in the advertised securities.
The model yields several predictions. First, even when an arbitrageur identifies a number of mispriced assets, he will concentrate his advertising on a few: diluting investors’ attention across too many assets would reduce the number of informed traders for each, diminish price discovery and leave a large liquidation risk for all. Indeed, in practice hedge fund managers who advertise their recommendations typically pitch a single asset at a time.
Second, concentrated advertising produces portfolio under-diversification. By lowering liquidation risk, advertising a given mispriced asset raises the arbitrageur’s risk-adjusted expected return, inducing him to overweight that asset in his portfolio.
Thirdly, if an arbitrageur has private information about a number of assets, he will get the most out of his advertising if he pitches those for which mispricing is largest, his private information is most precise, and there is least noise trading. Moreover, the arbitrageur will prefer to advertise assets whose information is easy to process by investors, as the corresponding ads require less attention by investors.
Fourthly, multiple arbitrageurs with common information will exhibit “wolf pack” behavior, advertising and trading the same assets: no individual arbitrageur has the incentive to deviate and divert investors’ attention to assets not advertised by others. Hence, in equilibrium each arbitrageur mimics the others. However, this “piggybacking” tends to generate multiple equilibria, some of which are inefficient: arbitrageurs may get collectively trapped in a situation where they all advertise assets that are not the most sharply mispriced. This may explain why at times the market appears to pick up minor mispricing of some assets and neglect much more pronounced mispricing of others, such as RMBSs and CDOs before the financial crisis.