Abstract: This paper develops a two-state asymmetric logarithmic ACD model for price duration，and incorporates informed trading indirect measurement variables including bid-ask spread , trading volume, trading size and order flow to describe the variation of conditional expected price duration asymmetrical depending on price increase and decrease states. At the same time it explores the information transmission mechanisms of price durations and tests market microstructure related hypotheses. The empirical analysis on the choosing sample shows that the lagged bid-ask spread and trading volume have a significantly negative correlation with the conditional expectation price duration; the lagged orders behind the best bid and ask prices have a significant correlation with the conditional expectation price durations, and the sign is determined by current price state; the large trading size has a more significant impact on price duration than the medium. The results support the view that the trade durations decrease as informed trading increase, but don’t support stealth trading hypotheses.

Key words: price duratio, asymmetric logarithmic ACD model, information transmission, stealth trading hypothese