Albert J. Menkveld, Anna Dreber, Felix Holzmeister, Juergen Huber, Gabriel Kaiser The Journal of Finance 79.3, 2024
In statistics, samples are drawn from a population in a data-generating process (DGP). Standard errors measure the uncertainty in estimates of population parameters. In science, evidence is generated to test hypotheses in an evidence-generating process (EGP). We claim that EGP variation across researchers adds uncertainty: Non-standard errors (NSEs). We study NSEs by letting 164 teams test the same hypotheses on the same data. NSEs turn out to be sizable, but smaller for better reproducible or higher rated research. Adding peer-review stages reduces NSEs. We further find that this type of uncertainty is underestimated by participants.
https://doi.org/10.1111/jofi.13337
This paper examines different regularization techniques in nonparametric return forecasting models to identify a reduced set of characteristics linked to expected return spreads in the cross-section of U.S. equity returns. The employed models are robust to outliers, computationally efficient, and capable of handling high-dimensional data. The number of identified firm characteristics ranges from two to thirteen, selected from a novel equity market anomaly database containing 90 variables. The key selected factors include price reversals, illiquidity measures such as share turnover (TURN) and zero trading days (ZEROTRADE), multiples like assets-to-market capitalization (A2ME), earnings yield (EP), and asset growth (AGR) for the most recent years. This study finds that in a multivariate setting both nonlinearities and temporal variations play a crucial role in explaining the cross-section of expected equity returns. Out-of-sample results indicate that Elastic Net regularization tends to overfit, whereas shrinkage via the Minimax Concave Penalty (MCP) and the Smoothly Clipped Absolute Deviation (SCAD) yields sparse nonlinear models with exceptional Sharpe ratios.
This paper proposes an experimental elicitation of individual risk attitudes in a market context. We evaluate the preferences of bidders from a first-price auction in a nested model that incorporates the constant relative risk aversion model and the dual probability of winning model as special cases. These models predict individual behavior across different market sizes for a given parametrization. We test the relevance of the elicited measures of risk aversion in an asset market setting. Our results suggest that the constant relative risk aversion model explains the behavior of the asset market well, and better than alternative risk aversion elicitation measures. We find that markets with higher levels of risk aversion have lower price levels, and that individuals with higher risk aversion submit lower limit offers to sell. We also observe that subjects who submit non-monotonous bids in the first price auction make trading losses in the asset market. Our non-monotonicity score is correlated to the score of the cognitive reflection test.
This paper explores the time-variation of expected abnormal portfolio excess returns. Explicitly, long-short equity returns subject to large quantities of arbitrage capital subsequently increase before they decay in the long-run. In order to identify periods of increased arbitrage activity, we study states with elevated abnormal net trading volume and eras that coincide with publication dates of market anomalies. The interaction of both measures is an ideal proxy for these periods. What is more, we are able to associate the pace of the time-variation of an anomaly to the demand in hedged returns as well as its turnover.
We compare the call auction designs of two major European stock exchanges. In our paper, Euronext is classified as transparent and Xetra as opaque auction mechanism. For a sample of 79 matched stocks, we find that Euronext auctions contribute more to price discovery and provide more liquidity due to lower bid-ask spreads. Nonetheless, the opaque auction attracts higher trading volume relative to daily trading volume. Moreover, we detect in both exchanges significant return reversals subsequent to the opening and closing auction, which implys positive expected profits for liquidity providers.