I have a monthly panel data of firm (variable: gvkey) and their turnover (variable: trn) as follow:
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input long gvkey float(modate trn) byte(ffind48 tic_newind) float(event_time mean_trn t treated) 1004 360 .073733665 41 0 . . . . 1004 361 .04401369 41 0 . . . . 1004 362 .1571562 41 0 . . . . 1004 363 .13565476 41 0 . . . . 1004 364 .1139597 41 0 . . . . 1004 365 .11191394 41 0 . . . . 1004 366 .04519458 41 0 . . . . 1004 367 .06838244 41 0 -5 .13748688 . 0 1004 368 .05052841 41 0 -4 .13370831 . 0 1004 369 .1172303 41 0 -3 .15079726 . 0 1004 370 .11515839 41 0 -2 .13852969 . 0 1004 371 .2591975 41 0 -1 .14454024 . 0 1004 372 .13647348 41 0 0 .1533158 0 0 1004 373 .1496822 41 0 1 .13952337 . 0 1004 374 .11926877 41 0 2 .1593396 . 0 1004 375 .0952426 41 0 3 .14585368 . 0 1004 376 .06534516 41 0 4 .14730963 . 0 1004 377 .07027248 41 0 . . . . 1004 378 .07071927 41 0 . . . . 1004 379 .06913976 41 0 -5 .13748688 . 0 1004 380 .05471021 41 0 -4 .13370831 . 0 1004 381 .07118324 41 0 -3 .15079726 . 0 1004 382 .06878027 41 0 -2 .13852969 . 0 1004 383 .10182425 41 0 -1 .14454024 . 0 1004 384 .08385859 41 0 0 .1533158 0 0 1004 385 .064968236 41 0 1 .13952337 . 0 1004 386 .03993835 41 0 2 .1593396 . 0 1004 387 .03755582 41 0 3 .14585368 . 0 1004 388 .0423926 41 0 4 .14730963 . 0 1004 389 .06519907 41 0 . . . . 1004 390 .04824517 41 0 . . . . 1004 391 .03668729 41 0 -5 .13748688 . 0 1004 392 .03870092 41 0 -4 .13370831 . 0 1004 393 .07271002 41 0 -3 .15079726 . 0 1004 394 .03570977 41 0 -2 .13852969 . 0 1004 395 .09181427 41 0 -1 .14454024 . 0 1004 396 .06065543 41 0 0 .1533158 0 0 1004 397 .06803372 41 0 1 .13952337 . 0 1004 398 .08417411 41 0 2 .1593396 . 0 1004 399 .03986542 41 0 3 .14585368 . 0 1004 400 .04606628 41 0 4 .14730963 . 0 1004 401 .04063267 41 0 . . . . 1004 402 .028142884 41 0 . . . . 1004 403 .02075341 41 0 -5 .13748688 . 0 1004 404 .02452676 41 0 -4 .13370831 . 0 1004 405 .05119422 41 0 -3 .15079726 . 0 1004 406 .022928976 41 0 -2 .13852969 . 0 1004 407 .026178503 41 0 -1 .14454024 . 0 1004 408 .0652122 41 0 0 .1533158 0 0 1004 409 .05849104 41 0 1 .13952337 . 0 1004 410 .04123861 41 0 2 .1593396 . 0 1004 411 .01479316 41 0 3 .14585368 . 0 1004 412 .018326418 41 0 4 .14730963 . 0 1004 413 .026970955 41 0 . . . . 1004 414 .02297102 41 0 . . . . 1004 415 .019287106 41 0 -5 .13748688 . 0 1004 416 .04034073 41 0 -4 .13370831 . 0 1004 417 .04149912 41 0 -3 .15079726 . 0 1004 418 .0207527 41 0 -2 .13852969 . 0 1004 419 .03840789 41 0 -1 .14454024 . 0 1004 420 .024877865 41 0 0 .1533158 0 0 1004 421 .016278341 41 0 1 .13952337 . 0 1004 422 .02945008 41 0 2 .1593396 . 0 1004 423 .05513439 41 0 3 .14585368 . 0 1004 424 .029302675 41 0 4 .14730963 . 0 1004 425 .04256625 41 0 . . . . 1004 426 .04514694 41 0 . . . . 1004 427 .02081584 41 0 -5 .13748688 . 0 1004 428 .02141738 41 0 -4 .13370831 . 0 1004 429 .012623002 41 0 -3 .15079726 . 0 1004 430 .020150423 41 0 -2 .13852969 . 0 1004 431 .03882795 41 0 -1 .14454024 . 0 1004 432 .05951326 41 0 0 .1533158 0 0 1004 433 .05526989 41 0 1 .13952337 . 0 1004 434 .05905772 41 0 2 .1593396 . 0 1004 435 .03972372 41 0 3 .14585368 . 0 1004 436 .063389175 41 0 4 .14730963 . 0 1004 437 .04341168 41 0 . . . . 1004 438 .10945938 41 0 . . . . 1004 439 .0552653 41 0 -5 .13748688 . 0 1004 440 .13748042 41 0 -4 .13370831 . 0 1004 441 .12149898 41 0 -3 .15079726 . 0 1004 442 .10068192 41 0 -2 .13852969 . 0 1004 443 .08416093 41 0 -1 .14454024 . 0 1004 444 .09832011 41 0 0 .1533158 0 0 1004 445 .1345781 41 0 1 .13952337 . 0 1004 446 .14719102 41 0 2 .1593396 . 0 1004 447 .067913644 41 0 3 .14585368 . 0 1004 448 .06246979 41 0 4 .14730963 . 0 1004 449 .13060866 41 0 . . . . 1004 450 .0830844 41 0 . . . . 1004 451 .08758722 41 0 -5 .13748688 . 0 1004 452 .0764337 41 0 -4 .13370831 . 0 1004 453 .08064376 41 0 -3 .15079726 . 0 1004 454 .04825445 41 0 -2 .13852969 . 0 1004 455 .06884157 41 0 -1 .14454024 . 0 1004 456 .07597327 41 0 0 .1533158 0 0 1004 457 .06794872 41 0 1 .13952337 . 0 1004 458 .09447815 41 0 2 .1593396 . 0 1004 459 .05477548 41 0 3 .14585368 . 0 end format %tm modate
To do so, I need a control group. The control group is composed by those firm that, in the moment when a firm of the same industry (I classified industry with ffind48 variable) jumps in a different one, they remain in the same industry sector. So no industry changes for them.
To be more clear: 2 firms belong to industry 1 in month 1. In month 2, one firm goes in industry 2, but the other will stay in 1.
I guess a solution is a propensity score matching based on the size (I have a variable that the name is "at"- assets total and "mkt" - market capitalization"). Nevertheless, the match should have a constraint that the firm must belong to the same industry before the event. So the matched should consider the same industry sector.
I created a treatment variable "t". If it is 1, it means that the firm belongs to a different industry compared to previous month. Is 0 for those firm that didn't change the industry. This variable can be either 1 or 0 only in january, because I don't have changes in other months. So, other months will be missing value.
Finally, as you can see from the data I also created some other variables such as event_time, that track the months before and after the event study.
If you would be so kind to give me some suggestion, I really appreciate.
Thanks in advance for your time!
0 Response to Propensity score matching with restrictions
Post a Comment