Working Paper, Department of Economics, Uppsala University 2000:9
Using daily data for the Swedish stock market for almost the last two decades no distinct and firm deterministic seasonal pattern for the conditional volatility for the Swedish stock market has been found. The daily turnover in the Swedish stock market has an impact on and eliminates to some extent seasonal patterns in conditional volatility. The daily turnover is a proxy variable used to test the mixture distribution model. According to this model the conditional variance of returns will display a GARCH-pattern of behaviour if the number of trades on the stock market during a day are serially correlated. We can also conclude that a feedback from the US stock market to the conditional volatility in the Swedish market exists, and trading days particularly after holidays has a positive impact on the conditional volatility. The test of the models mean equation indicates that the Swedish stock market seems to become more and more information efficient, at least in its weak form, if the 1990s are compared with the 1980s.