Research Interests

Innovation economics, q theory of investment, behavioral finance


Working Papers

Patent Thickets and Mergers and Acquisitions (with Logan Emery

We find that patent thickets are important determinants of target selection because the commercialization costs created by patent thickets affect the realization of synergies.  Firms are less likely to be acquired when their patents are in thickets that expose firms to these costs. Conversely, firms are more likely to be acquired when their patents create thickets that insulate firms from these costs. When a firm occupies the same thicket as the acquirer, acquisition probability depends on each firm's ability to impose costs on the other. Consistent with thickets affecting synergies, we find that targets' thickets are associated with post-acquisition profitability.

Price-Path Convexity, Extrapolation, and Short-Horizon Return Predictability (with Huseyin Gulen)

We document a striking negative relation between price-path convexity and future short-horizon returns. Regardless of cumulative return in the period during which convexity is estimated, stocks with the least convex price paths subsequently outperform stocks with the most convex price paths. Our results are not driven by firm size, the bid-ask bounce, illiquidity, or other short-term return predictors. We find similar results in international samples and even at the aggregate market level. We provide evidence that this relation is consistent with investor over-extrapolation of recent price changes at short horizons.

Work in Progress

Disagreement in Collateral Valuation (with Jordan Martel)

We develop a model in which borrowers and lenders agree to disagree about the value of collateral. We establish conditions under which returns on collateralized assets exhibit momentum and investigate how default risk affects momentum strength. To illustrate the model's empirical relevance, we test the model's predictions in the market for U.S. single-family homes and find evidence consistent with those predictions.