Hello,

I have panel data looking like the following:

Code:
* Example generated by -dataex-. For more info, type help dataex
clear
input float date_b long ticker float(bp_return returnD1 holdingratio)
  1 1  65.239365   99.48805  -.005976806
  2 1   99.48805   58.20886  .0016789314
  3 1   58.20886 -2.9673014 -.0026526705
  4 1 -2.9673014  112.80904   -.01497696
  5 1  112.80904   107.1486   .007075301
  6 1   107.1486   85.68056   .003154348
  7 1   85.68056  -34.55693    .02158222
  8 1  -34.55693   -969.513    .11129766
  9 1   -969.513  -89.60022     .3750472
 10 1  -89.60022   266.3871     .3035616
 11 1   266.3871  114.79784    .03285338
 12 1  114.79784  191.23187   -.06439177
 13 1  191.23187 -108.31528    -.0557736
 14 1 -108.31528  -47.80962   -.07505196
 15 1  -47.80962  -63.53572   -.06179261
end
format %tbCalendar date_b
label values ticker ticker
label def ticker 1 "A", modify
returnD1 is the lead variable of bp_return, so the subsequent return in basis points.

What I want to show now: I want to consider the days when the bp_return is negative. I then want to show that the probability of the subsequent return (returnD1) being larger (less negative or even positive) than the bp_return increases IF the holdingratio was positive on the day 0 of negative bp_return.

I thought of the logistic command but I am not sure how to implement this or if this is even possible in Stata.

Thank you in advance!