Probability of Informed Trading

This package allows to compute the probability of informed trading from bilateral trades.

The probability of informed trading (PIN) denotes that probability that a counterparty in the trading process has superior information on the value of the asset exchanged. This is a key concept in empirical studies of market microstructure.

The parameters needed to compute the PIN are obtained from the estimation of a theoretical model of the trading process. The details are discussed in Easley et al. (1996), among others.

This package consists of a Matlab function that computes the empirical likelihood for the trading game arising from the theoretical model of asset market. The parameters that maximize this function are then used to compute the PIN.

A detailed readme file in pdf explaining the theoretical model behind the PIN is attached.

The package also presents a sample file that uses data from an actual asset market. The data file includes the number of buy-initiated and sell-initiated trades, as well as trades that cannot be classified denoted as 'no trades' for a number of trading days.

NOTE: The example uses the fminsearchbnd package for constrained objective functions. So far, I have tested with fminsearchbnd by John D'Errico, release 4 dated 7/23/06

Reference:

Easley, D., N. M. Kiefer, M. O'Hara and J. B. Paperman, 1996, "Liquidity, Information, and Infrequently Traded Stocks", Journal of Finance, 51(4), 1405-36.