Working papers
Embedded Files
joint work with Patrick Schwarz and Patrick Weiss
The recent CRSP transition in 2025 rewrites 9.62% of monthly returns by more than one basis point and around 11.43% of all monthly long-short factor returns by more than ten basis points. Reassuringly, these differences do not affect premia nor their significance. Overall, our results suggest that CRSP changes unlikely affect asset pricing studies over long sample periods beyond unsystematic variation.
The recent CRSP transition in 2025 rewrites 9.62% of monthly returns by more than one basis point and around 11.43% of all monthly long-short factor returns by more than ten basis points. Reassuringly, these differences do not affect premia nor their significance. Overall, our results suggest that CRSP changes unlikely affect asset pricing studies over long sample periods beyond unsystematic variation.
Abstract
In January 2025, CRSP discontinued the existing stock tape used in many published papers. This transition rewrites 9.62% of monthly returns by more than 1 basis point, primarily due to a change in the dividend reinvestment assumption. Analyzing the impact for a comprehensive set of premia in several thousand sorting specifications reveals that, on average, 11.43% of all monthly long-short returns differ by more than 10 basis points — especially in early periods, NBER recessions, and return-based sorts. Reassuringly, average premia and their significance remain largely unaffected, suggesting CRSP changes mainly introduce unsystematic variation without altering key asset pricing conclusions.
Presentations: AFFI 2025 (Best paper award), AWG 2025
joint work with Rüdiger Weber
We introduce TIM, a measure of a stock's cash-flow timing. Unlike equity duration measures, TIM does not conflate discount rate with cash-flow timing information and allows for inference about equity term premia. We find unconditionally no relation between cash-flow timing and mean returns, consistent with recent findings on index dividend derivatives but contrasting with the equity duration literature. Late-minus-early timing returns exhibit a strong factor structure which helps reconcile conflicting evidence on cyclicality: In recessions, positive market beta implies high expected returns, whereas negative investment beta suggests low expected returns. The net effect suggests a downward slope in recessions.
We introduce TIM, a measure of a stock's cash-flow timing. Unlike equity duration measures, TIM does not conflate discount rate with cash-flow timing information and allows for inference about equity term premia. We find unconditionally no relation between cash-flow timing and mean returns, consistent with recent findings on index dividend derivatives but contrasting with the equity duration literature. Late-minus-early timing returns exhibit a strong factor structure which helps reconcile conflicting evidence on cyclicality: In recessions, positive market beta implies high expected returns, whereas negative investment beta suggests low expected returns. The net effect suggests a downward slope in recessions.
joint work with Rüdiger Weber and Patrick Weiss
Despite the large variation induced by making different methodological choices in portfolio sorts, estimated premia are robust regarding their sign, statistical significance, and monotonicity. To address methodological uncertainty, we propose a two-step protocol adaptable to economic motivations, for which we provide an open-source tool.
Despite the large variation induced by making different methodological choices in portfolio sorts, estimated premia are robust regarding their sign, statistical significance, and monotonicity. To address methodological uncertainty, we propose a two-step protocol adaptable to economic motivations, for which we provide an open-source tool.
Abstract
Systematically studying methodological variation in portfolio sorts reveals four key insights. (1) The average monthly non-standard error is 0.19% and exceeds standard errors. Despite this considerable variation, estimated premia are robust regarding their sign, statistical significance, and monotonicity. This alleviates concerns about replicability. (2) Decisions such as excluding firms with negative earnings or the information lag have an impact comparable to size-related choices. (3) Methodological choices induce not just orthogonal noise but add predictably non-zero returns of unclear origin. (4) To address methodological uncertainty, we propose a two-step protocol adaptable to economic motivations, for which we provide an open-source tool.
Presentations: EFA 2023, EFMA 2023, DGF 2023, AFBC 2022, PFMC 2022, AWG (2022).
The negative competition effect of patents on product rivals is around six times larger than the positive learning effect on technology peers and makes up 25% of the private patent value. Firms tend to learn from their peers by poaching inventors.
The negative competition effect of patents on product rivals is around six times larger than the positive learning effect on technology peers and makes up 25% of the private patent value. Firms tend to learn from their peers by poaching inventors.
Abstract
Technological innovations create value for the innovating firm and spillover effects on peer firms. This paper proposes a new methodology to infer technology spillover effects from patents: Around patent announcements, investors incorporate negative spillovers into stock prices of close rivals to the innovating firm (competition effects) and positive spillovers for peer firms that can learn from the patented technology (learning effects). Competition effects are six times larger than learning effects and amount to 25 percent of the private patent value. Studying patents allows for two novel insights. First, competition spillover effects have become less pronounced after 2000. I show that the American Inventors Protection Act of 1999 diminished the private patent value and thus reduced competition spillovers on close product rivals. Second, the labor mobility of inventors shows that peer firms learn by hiring new inventors.
Presentations: DGF 2024, Boca-ECGI 2024, Rotterdam School of Management, Maastricht University, CUNEF University, University of Konstanz, University of Cologne
Page updated
Google Sites
Report abuse