Abstract
This paper complements earlier work by the author that shows that the pattern of information arrivals into the UK stock market may explain the behaviour of returns. It is argued that delays or other systematic behaviour in the processing of this information could compound the impact of information arrival patterns. It is found, however, that this does not happen, and so it is the arrival and not the processing of news that is most important.
Notes
See, for example, Board and Sutcliffe (1988), Chang et al. (Citation1993) and Arsad and Coutts (1996, 1997).
See, Theobald and Price (Citation1984) and Choy and O’Hanlon (Citation1989). Compare to Keim and Stambaugh (Citation1984) and Rogalski (Citation1984) for the US stock market.
See, Choy and O’Hanlon (Citation1989) and Mills and Coutts (1995). Compare to Gibbons and Hess (Citation1981) for the US stock market.
Recent UK evidence in support of this argument is provided by Ryan and Taffler (2000).
Daily returns are calculated as noted. Adjustments are made to account for distortions due to ex-dividend day behaviour, see Steeley (Citation2001) for details.
This pre-holiday variable has been found to be important in these such regressions, see e.g., Ariel (Citation1990) and Kim and Park (1994), and is included here to prevent biases from misspecification.
It is possible to further refine the model to explicitly capture ARCH style heteroskedasticity, see Chang et al. (Citation1998), which is common within stock return time series. Although there was some evidence of ARCH effects within the data here, modelling it had no qualitative and only negligible quantitative impacts upon the estimated coefficients, their statistical significance, and hence the succeeding conclusions regarding day-of-the-week effects and causes.