An Improved Estimation Method and Empirical Properties of the Probability of Informed Trading

Abstract

We propose a method to overcome a bias in the estimate of the probability of informed trading (PIN). This bias arises when the numerical maximization procedure generates corner solutions. We analyze the PIN estimates for about 80,000 stock-quarter pairs between 1993 and 2004, and observe a decreasing trend in PIN over this time period. The decimalization in January 2001 appears to accelerate this trend; the quarterly median PIN across stocks decreased by 20% since the first quarter of 2001 until the end of 2004. This provides direct evidence that the risk of trading against informed traders has reduced in recent years. We also find that the risk of informed trading is lower in the first quarter of a year than in the previous fourth quarter, especially for stocks whose price has gone up in previous year.

Keywords: decimalization, informed trading, market microstructure, PIN, tax-loss selling

JEL Classification: C13, C61, G12, G14