We investigate whether the bid/ask queue imbalance in a limit order book(LOB) provides significant predictive power for the direction of the nextmid-price movement. We consider this question both in the context of a simplebinary classifier, which seeks to predict the direction of the next mid-pricemovement, and a probabilistic classifier, which seeks to predict theprobability that the next mid-price movement will be upwards. To implementthese classifiers, we fit logistic regressions between the queue imbalance andthe direction of the subsequent mid-price movement for each of 10 liquid stockson Nasdaq. In each case, we find a strongly statistically significantrelationship between these variables. Compared to a simple null model, whichassumes that the direction of mid-price changes is uncorrelated with the queueimbalance, we find that our logistic regression fits provide a considerableimprovement in binary and probabilistic classification for large-tick stocks,and provide a moderate improvement in binary and probabilistic classificationfor small-tick stocks. We also perform local logistic regression fits on thesame data, and find that this semi-parametric approach slightly outperform ourlogistic regression fits, at the expense of being more computationallyintensive to implement.
展开▼
