“Order Flow, Comovement, and Market Predictability” by Mark Seasholes
Hong Kong University of Science and Technology
If non-informational traders impact individual security returns, can the same traders affect market returns? This paper studies predictability of market returns using a mechanism — program trading on the NYSE — well suited for index traders to minimize trading costs. After a stock joins the S&P500 Index, we show the volume of program trading increases and program traders acquire a substantial fraction of shares. The stock's order imbalances and returns comove more with the order imbalances of other S&P500 stocks. Aggregate program trading order imbalances (buy minus sell volume) are positively correlated with contemporaneous S&P500 returns and negatively correlated with future S&P500 returns. Our results indicate these traders induce a common component in S&P500 returns and cause return predictability and cause excess volatility. Overall, results are consistent with frictions due to limited risk-bearing capacity operating at the market-level.