ESTIMATING THE PROBABILITY OF INFORMED TRADING: FURTHER EVIDENCE FROM AN ORDER-DRIVEN MARKET

Nyholm (2002, 2003) [J. of Financial Research, 25, pp. 485; J. of Applied Econometrics, 18, pp. 457] has proposed a new procedure to infer the probability of informed negotiation on a trade-to-trade basis through a regime-switching model. We provide further empirical evidence about the performance of this model by using trade-related information, such as the degree of aggressiveness and the trade size, on a pure order-driven market. It is evidenced that the switching scheme of the basic model is closely related to the arrival of different types of orders and not necessarily to information. This feature also applies when controlling for market variables other than order aggressiveness (e.g., trade size). The updating process in the non-linear setting proves so complex that it is necessary to account for a number of different microstructure effects to provide probabilities related to information arrivals. This evidence casts doubts about the general suitability of the procedure.