I use simulated data to test different PIN estimation algorithms in SAS. I conclude that the algorithms do a good job estimating PIN when applied to series generated from high-PIN data generating process (PIN>0.2). However, applied to series generated from zero-PIN or low-PIN (PIN<0.2) DGP, the algorithms yield overstated PIN estimates. Supplying initial values has strong effect on estimated PINs. Factorization of the likelihood function plays less important role. Not supplying initial values dramatically overstates PINs for zero-PIN and low-PIN data: zero-PIN data shows PIN as high as 0.26; low-PIN data tend to have PINs overstated by 0.05 to 0.10.
|Publication status||In preparation - 2016|
- probability of informed trading