Inferring trade directions in fast markets

The reliability of trade classification algorithms that identify the liquidity demander in financial markets transaction data has been questioned due to an increase frequency quote changes. Hence, this paper proposes a new method. While established rely on ad hoc assignment trades quotes, proposed full-information (FI) algorithm actively searches for matches trade. FI outperforms existing ones, particularly at low timestamp precision: For timestamped seconds misclassification is reduced by half compared popular Lee-Ready algorithm. These improvements also carry over into empirical applications such as estimation costs. recently interpolation method and bulk volume do not offer improvements.