Abstract—This study examines the impact of trade
classification algorithms on estimating the probability of
informed trading (PIN). This study finds that the algorithms
themselves may not substantially influence the PIN estimates
but the poor performances of these algorithms may have the
great impact on the PIN estimates. Moreover, the new proposed
adjustment, Q-Method, seems to mitigate the bias caused by the
trade misclassification. In addition, the pattern of its estimates
responds to the important economic events. With the estimated
misclassification rate from Q-Method, this study also finds that
the performances of these algorithms are getting poor in recent
years.
Index Terms—Informed trading, market microstructure,
trade misclassification.
W.-C. Ke is with the Department of Finance and Cooperative
Management National Taipei University, No. 151, University Rd., San-Shia
District, New Taipei City 23741, Taiwan (e-mail:
wenchyan@gm.ntpu.edu.tw).
[PDF]
Cite: Wen-Chyan Ke, "The Sensitivity to Trade Classification Algorithms for Estimating the Probability of Informed Trading," International Journal of Trade, Economics and Finance vol.5, no.5, pp. 392-396, 2014.