Another look at breadth of ownership and asset returns

Richard Priestley and Bernt Arne Ødegaard

BI Norwegian School of Management

Abstract

This papers revisits the results of Chen Hong and Stein (2002), which shows that the breadth of ownership, implemented as the number of mutual fund owners of a firm, is important for the crossection asset returns. Their results are justified as implications of the classical model of the implications of short sales constraints by Miller (1977). Our Norwegian setting allows us to revisit this result using a better suited data set, since our database on the ownership of Norwegian equities covers monthly observations of the complete ownership structure over the period 1989 to 2003, a period of 14 years. Our data set also allows us to investigate alternative breadth measures, not just restricted to the mutual fund holdings used by Chen Hong and Stein, such as total number of owners and ownership concentration. Our results show that the Chen Hong and Stein results are really about mutual funds trading. Quarterly mutual funds trades predict next quarters returns. However, broader breadth measures, such as the change in indivdual (private) owners, have opposite effects. We thus show evidence that the channel through which the effects happen is by the mutual funds selling to (new) individual investors, and not by an increase in ownership concentration. We also find that the quarterly effect is not observed on a monthly horizon. We hypothesise that this may be due to short term effects from the mutual fund's actual trading, since we observe a strong contemporaneous link between asset returns and mutual fund trading.

The paper is downloadable as a pdf (Adobe acrobat) file