Detecting Information-Driven Trading in a Dealers Market

We focus on the extent of information-driven trading originating from order flows to capture the behavior of the market makers on an emerging market. We modified the classical Easley et al. (1996) model for the probability of informed trading using a jackknife approach in which trades of one particular market maker at a time are left out from the sum of all buys and sells. Using the estimates from the jackknife approach, for each market maker we test whether the order flows associated with the particular market maker behaved significantly differently from the others. Data from the Prague Stock Exchange SPAD trading platform are used to demonstrate our methodology. Finding significant differences in the probability of informed trading computed from order flows, we conclude that order flows could reveal the extent of information-driven trading and could potentially be used by regulatory authorities to identify suspicious behavior by market participants.