Market Microstructure and the Dynamic Relation of Stock Returns and Trading Flows
Terence Lim, Goldman Sachs Asset Management, formerly at Dartmouth College

Market micro structure theories of trading predict that lead-lag relationships exist between stock returns and trading flows. This study examines the dynamic relationship between daily returns and signed trading flow for a large sample of individual NYSE stocks between 1993 and 1999. The empirical evidence suggests that, at the forecast horizons studied in this paper, the reversal of stock returns following trading flow innovations is more consistent with the predictions of non-informational than with informational trading models. Also, the response of trading flows following return innovations suggest that feedback trading is more important at short horizons of up to one week, while value-motivated trading is more important at longer horizons. Economically, the magnitude of predictability is not large, but may be useful to investors interested in minimizing trading costs. (Accepted Spring 2002.)