Does Algorithmic Trading Affect Analyst Research Production?

Abstract

We document a causal negative relationship between algorithmic trading (AT) and analyst research production, as captured by a decreased frequency of earnings forecasts and stock recommendations and lower analyst coverage. This is consistent with AT increasing the speed of price discovery, reducing the profitability of trades on analyst research by non-algorithmic traders and, consequently, their demand for analyst investment advice. Supporting evidence shows that the effect of AT on analyst research production is stronger for stock recommendations, which institutions follow primarily for investment decisions, and for forecasts issued before earnings announcements when analysts’ information discovery dominates the information interpretation role. We also find a negative relationship between AT and investment-focused institutional investors such as transient and non-monitoring investors. Our analysis demonstrates that AT can have long-lasting consequences on capital markets, beyond microstructure effects, through its negative effect on firm’s information environment.