Do order imbalances predict Chinese stock returns? New evidence from intraday data

•We examine whether order imbalances can predict the Chinese stock market returns.•We use intraday data and a panel data predictive regression model.•Order imbalances predict stock returns from 1-minute trading to 90-minute trading.•Trading strategies reveal that profits persist during the day.

In this paper we examine whether order imbalances can predict the Chinese stock market returns. We use intraday data, a panel data predictive regression model that accounts for persistent and endogenous order imbalances and cross-sectional dependence in returns, and show that order imbalances predict stock returns from 1-minute trading to 90-minute trading. On the basis of this predictability evidence using multiple trading strategies we show that profits persist during the day. These results imply that a source of Chinese market inefficiency is order imbalances.