Detecting Information from Directors' Trades: Signal Definition and Variable Size Effects

Publication Date
Financial Markets Group Discussion Papers DP 245
Publication Authors

There have been three empirical studies examining the share price reaction following trades by directors of UK companies (King and Röell, 1988; Pope, Morris and Peel, 1990; Gregory, Matatko, Tonks and Purkiss, 1994). To some degree these papers have produced conflicting evidence. This evidence is important because it reveals not only whether potentially profitable trading rules may be formulated on the basis of directors' trading activity, but also because it is indicative of whether directors have the ability to predict mis-pricing in the market.

All three of these UK studies used different definitions of "buy" and "sell" signals resulting from the transactions of directors and employ different controls to detect the presence of any "size effects". In this paper we use the most comprehensive data set available, consisting of all directors' trades in UK companies over the period 1986 to 1992. We investigate whether the signal definition explains the different conclusions drawn by these earlier studies, and examines whether or not any observed abnormal returns are explicable by the small companies effect. We also investigate trading strategies based on holding a long portfolio of shares purchased or a short portfolio of shares sold by directors held until the end of the study period or until a "reversing event" (e.g. a sale following a purchase by director[s]) is observed.

We show that it is essential to take into account of company size effects (either alone or in combination with a measure of risk) in calculating abnormal returns. Allowing for these factors, for "signal" definitions, except those filtered to use only large trades, significant positive returns follow purchases by directors and these appear to persist for up to 24 months following the "buy" signal. These results are confirmed when a benchmark which controls only for size is employed. For large trades, generally insignificant negative returns follow purchases by directors, whilst for large sales significant abnormal returns can occur although these are sensitive to the benchmark employed. Following sales by directors, significant negative abnormal returns are earned for up to 6 months after the month of sale using both size-control benchmarks. We also show the abnormal returns for portfolios which are held until the next reversing trade by directors is observed and find that the returns earned vary according to the definition of reversing events.

Download is not available