Estimating PIN for firms with high levels of trading

•Estimation of the probability of informed trade can fail for actively-traded firms.•PIN values for a large fraction of market capitalization may be excluded by this.•Tests and conclusions that do not reflect PIN’s of large firms may be unreliable.•Scaling buy/sell trade counts provides PIN estimates for any actively-traded firm.

For models of the probability of informed trading (PIN), estimation can fail for firms with high levels of trading due to computer over/under-flow. Since active firms tend to have large market capitalizations, studies that use PIN have excluded as much as 40% of total market capitalization from their sample. Similarly, since trading tends to be more intense around important events, studies that use PIN may lose observations exactly during periods that are the focus of study. A simple procedure, using scaled trade counts, allows PIN to be estimated for actively-traded firms, avoiding the possible biases or false generalizations that may occur when data from large firms or important events is ignored.