AQR Insight Award

Overpriced Winners


Kent Daniel, Ph.D., Columbia Business School and NBER, Alexander Klos, Ph.D., Kiel University and Kiel Institute for the World Economy, Simon Rottke, Ph.D., FCM, University of Münster


Stocks with low institutional ownership which, over a one-year period leading up to portfolio formation, earn strong positive returns and experience a simultaneous increase in short interest, subsequently earn strikingly low returns over the following five years. The authors argue that the high prices of this set of “overpriced winners” are a result of excessive optimism by a group of investors combined with arbitrage constraints for the rational investors.


The authors use data from CRSP, Compustat, and Thomson-Reuters Institutional 13-F filings over a sample period from 1989 to 2014 for U.S. common stocks. The authors first empirically confirm the literature on momentum — looking at the 20% of stocks with the highest and lowest cumulative returns over the past twelve months, excluding the most recent month, the past winners outperform the past losers, as expected. However, from that group of past winner stocks, the authors further identify a “constrained winners” portfolio consisting of those in the bottom 20% in terms of institutional ownership and in the top 20% in terms of their increase in short interest. These last two additional characteristics suggest these past winners have strong limits to arbitrage. The new finding here is that the constrained winner portfolio earns a meaningfully negative abnormal return of -2.47% in the first month after portfolio formation. Over the five years following portfolio formation, the overpriced winners lose a large share of their pre-formation gains relative to the market. Small and medium-cap momentum strategies, which exploit the fact that an asset’s relative performance tends to continue in the near future, can be enhanced by avoiding the small subset of overpriced winners.

The authors suggest that excessive optimism comes into play in that the constrained winners have experienced an irrational run up in price that implies a predictable strong decline, making them “overpriced winners.” Their argument requires an evolution of beliefs among investors. If institutional lending supply is high, then the price will adhere to the rational expected value and thus will be fairly priced, on average. However, if borrowing shares is difficult or costly, the views of a group of pessimistic investors are restricted and therefore cannot be expressed in prices. As disagreement falls over time, however, due to new information regarding the company’s financial health for example, the price converges to the rational price based on fundamentals.

Finally, the authors find that a “betting against winners” portfolio that buys a broad portfolio of past-winners and shorts a much smaller portfolio of constrained past winners earns a Sharpe ratio of 1.08 and a significant Fama-French three factor alpha of 2.71% per month over the sample period.


The authors propose a model that captures the effects of limits of arbitrage and excessive optimism or disagreement about the value of a stock. The results suggest that there is a subset of high past return stocks that earn strong negative post-formation returns. This small group of “overpriced winner” stocks has low institutional ownership, suggesting short-sale constraints, and has an increase in short interest consistent with their price increase being driven by increased optimism among a subset of investors. When new information reveals the true fundamental value, optimism wanes, and the stock price declines towards its fair value. This dynamic in beliefs among investors is consistent with the notion that short-sale constraints sideline more pessimistic market opinions, which when coinciding with excessive optimism, can result in temporary overpricing, followed by an eventual correction. Identifying these constrained or overpriced winners can help enhance a small-cap momentum strategy by either avoiding such stocks or actively trying to short them.

Read the paper