The informational content of technical analysis : An econometric study of "chartism" on the Swiss stock exchange market

We apply the automatic pattern recognition algorithm of Lo et al. (2000) to 25 stocks listed on the Swiss stock exchange, over the last 15 years. We systematically identify exogenously specified technical patterns and then compare the empirical unconditional distributions of daily returns to the empirical distributions conditioned on the occurence of each of the patterns. We find that several technical indicators provide valuable information. This is the first study of technical analysis on the Swiss stock exchange market. Moreover, for the first time in the literature, we test for stochastic dominance using the new methodology by Linton et al (2003), which allows for serially and mutually correlated observations. In a first attempt to learn the statistical properties of technical indicators, we find evidence that many patterns may be indicators of ARCH behaviour in the price process