We use machine learning to capture nonlinearities and interactions in the relation between trades and holdings of multiple market participants and future stock returns. Our predictor yields a long-short portfolio with significant out-of-sample alpha, forecasts firm fundamentals, and assigns stocks on the right side of most anomalies. Predictability is stronger for smaller or illiquid stocks and stocks with lower analyst coverage or higher idiosyncratic volatility. A factor model based on our predictor achieves higher Sharpe ratio than existing models. Our findings suggest that incorporating nonlinear interactions between trades and holdings of various participants reveals valuable information for price discovery.

This paper explores the information diversity between two distinct investor groups — hedge funds and off-exchange retail investors — who may access different sources of stock-level information. We find strong and persistent return predictability from stocks that exhibit consistent trading between the two groups. Such consistent trading significantly predicts firm fundamentals and contributes to the correction of stock-level mispricing. Conversely, when the two groups trade in opposite directions, neither achieves significant excess returns. The overall findings highlight the importance of considering trading signals from heterogeneous investor groups to uncover valuable information: only when these groups agree with each other. Our empirical evidence also echos with Goldstein and Yang (2015) by showing that aggregating information held by different investor types can improve price efficiency. 

We examine the economics of financial scams by analyzing the market for initial coin offerings (ICOs). Using data snapshots of 5,873 ICOs, we find that irregularities in ICO characteristics across listing websites predict higher scam risk. These patterns are consistent with a framework where malicious issuers maximize profits by using irregularities to screen for naïve investors. Almost half of the ICOs in our sample may be scams, amounting to more than U.S. $6 billion in losses. Our results draw attention to the frequent use of screening mechanisms in financial scams.

This paper examines the trading patterns of retail investors following insider trading. Retail investors promptly follow opportunistic instead of routine purchases by insiders. The abnormal retail searches of the Form 4 filings increase for opportunistic insider purchases. Neither attention nor common information drives the results. Moreover, price efficiency is improved for stocks bought by retail investors following opportunistic insider purchases. The effect is mostly driven by the information component of retail trades, rather than by liquidity provision or price pressure. The evidence is consistent with retail investors learning from informed insider purchases, and their trading helping expedite price discovery.

Using detailed bank lending data from Chinese listed firms to identify bank-financing linkage, we find strong evidence of return predictability across bank-financing-linked firms. A long-short portfolio formed on the past returns of linked firms can generate risked-adjusted returns of 5.5%–7.8% annually for focal firms. This cross-firm predictability is distinct from industry or geographic momentum. It is more pronounced for focal firms that receive lower investor attention, exhibit higher arbitrage cost, or have closer relationships with banks. Overall, our findings suggest a unique channel of stock market spillover, where the bank-financing linkage among firms, coupled with limited investor attention, leads to sluggish information diffusion.

Using data from Binance, we find strong evidence of cross-cryptocurrency return predictability. The lagged returns of other cryptocurrencies serve as significant predictors of focal cryptocurrencies. The results are robust across various methods, including the adaptive LASSO and principal component analysis. Furthermore, a long-short portfolio formed on the past returns of cryptocurrencies can generate a sizable return out-of-sample after accounting for transaction costs. Overall, our findings corroborate cross-cryptocurrency return predictability and are consistent with the spillover effect mechanism, where common shocks among cryptocurrencies coupled with the limited attention of investors lead to slow information diffusion across coins.