LAUSR

Abstract This paper aims to explore the predictive ability of skewness on CNY-CNH spread. We first prove positive effect of skewness with and without control variables under in-sample analysis. We further demonstrate the predictive power of skewness exists in the out-of-sample forecasts. This predictability becomes stronger when using longer skewness calculation windows but tends to decline at longer forecast horizon. And it is also robust even though adopting different measures of skewness and changing time horizons. Moreover, this predictability is realized through driving the changes in market liquidity. Our study illustrates skewness carries useful information for predicting CNY-CNH spread.