Dear Forum,
I am researching the effects of index inclusion on the investor recognition of small cap companies and using therefore a difference in difference approach. However, I do not have a fixed date where the treatment happens, it is different for every company (multiple period DID). I am currently stuck at the propensity score matching process of ccontrol and treatment group. Therefore, I try to match only for the quarter before the treatment happens and extend the matching afterwards to the whole sample. I also set no replacement, one-on-one match. My question is that the no replacement criteria only applies to the firm-year observation level. For example, if control company A in quarter 2004q3 matched to one of the treated companies in the quarter 2004q3, it can still be matched again to another company in e.g. 2009q2. I want the firm-level one-on-one match, which means one treated firm will have one unique control firm that matches with it. And then I will keep all the firm-year observations for both the treated and control companies. I am already matching exactly for the time period and industry. The other variables are matches based on the propensity score.
logit dummy_treat ln_tot_assets profitability tangibility mtb_ratio
predict ps
egen exactmatch = group( timevar ICBIC)
gen pscore = 10*exactmatch + ps
psmatch2 dummy_treat , pscore(pscore) cal(0.05) common neighbor(1) noreplacement
Can anyone please advise on how to achieve that? Many thanks.
Code:
* Example generated by -dataex-. For more info, type help dataex clear input int c_id float(timevar ln_tot_assets profitability tangibility mtb_ratio) byte dummy_join float dummy_treat double _weight int(_id _n1) 193 243 16.802202 .10005046 .2314329 9.003746 0 0 1 27664 . 208 243 12.00873 -.04992021 .4600139 7.113888 0 0 1 27665 . 169 243 12.30863 .18140794 .1556859 6.029928 0 1 1 27665 27665 302 243 12.143713 .14730002 .8396511 2.559861 0 0 1 27784 . 158 243 13.424753 .2460771 .4534546 5.314976 0 1 1 27784 27784 143 244 12.53439 .10639983 .16025147 2.5588746 0 1 1 27831 27831 213 244 12.48605 .15771295 .20859346 3.713875 0 0 1 27831 . 100 244 15.037695 .08367095 .3745413 1.4208206 0 1 1 28072 28072 241 244 13.73277 .08712249 .8889005 39.22791 0 0 1 28072 . 131 244 14.13156 .1303492 .3942553 2.530993 0 1 1 28117 28117 220 244 12.744736 .18055803 .54411614 2.706573 0 0 1 28117 . 213 245 12.48605 .15771295 .20859346 3.713875 0 0 1 28184 . 109 245 13.149642 .13692157 .13714732 28.7388 0 1 1 28184 28184 155 245 16.185017 .08183612 .2994641 3.025999 0 1 1 28224 28224 175 245 15.770484 .10087815 .020241147 .8113729 0 0 1 28224 . 193 245 16.854015 .14865384 .2311488 6.451898 0 0 1 28367 . 71 245 15.70876 .03490321 .01868133 1.2961305 0 1 1 28367 28367 213 246 12.48605 .15771295 .20859346 3.713875 0 0 1 28538 . 119 246 12.370604 .0962811 .13141732 11.00779 0 1 1 28538 28538 214 246 13.089137 .0864781 .10118265 5.538691 0 0 1 28539 . 145 246 14.938308 -.017067846 .006352816 3.336922 0 1 1 28539 28539 146 246 15.49898 .22059405 .1585224 2.3237605 0 1 1 28606 28606 253 246 18.566174 .06230276 .1096999 1.3983625 0 0 1 28606 . 17 246 12.330513 .1973793 .12675329 12.30748 0 1 1 28780 28780 241 246 13.73277 .08712249 .8889005 39.22791 0 0 1 28780 . 140 246 13.868586 .1776808 .3540425 2.3460033 0 1 1 28781 28781 265 246 14.947557 .07865893 .14036106 .9359225 0 0 1 28781 . 66 247 12.992812 .13954854 .12826326 7.561958 0 1 1 28891 28891 213 247 12.73605 .13770127 .1760821 2.2743716 0 0 1 28891 . 193 247 16.854015 .14865384 .2311488 6.451898 0 0 1 29073 . 160 247 15.781055 .07878075 .3564885 .6431261 0 1 1 29073 29073 76 247 14.561218 .081857 .2991022 3.005537 0 1 1 29132 29132 241 247 13.73277 .08712249 .8889005 39.22791 0 0 1 29132 . 3 248 13.385486 .1402982 .2602872 4.086562 0 1 1 29243 29243 213 248 12.73605 .13770127 .1760821 2.2743716 0 0 1 29243 . 265 249 14.919476 .05818773 .13905375 .9311196 0 0 1 29814 . 79 249 14.905082 .08980828 .0877901 3.848259 0 1 1 29814 29814 end format %tq timevar
------------------ copy up to and including the previous line ------------------
Thanks for your responses!
I am researching the effects of index inclusion on the investor recognition of small cap companies and using therefore a difference in difference approach. However, I do not have a fixed date where the treatment happens, it is different for every company (multiple period DID). I am currently stuck at the propensity score matching process of ccontrol and treatment group. Therefore, I try to match only for the quarter before the treatment happens and extend the matching afterwards to the whole sample. I also set no replacement, one-on-one match. My question is that the no replacement criteria only applies to the firm-year observation level. For example, if control company A in quarter 2004q3 matched to one of the treated companies in the quarter 2004q3, it can still be matched again to another company in e.g. 2009q2. I want the firm-level one-on-one match, which means one treated firm will have one unique control firm that matches with it. And then I will keep all the firm-year observations for both the treated and control companies. I am already matching exactly for the time period and industry. The other variables are matches based on the propensity score.
logit dummy_treat ln_tot_assets profitability tangibility mtb_ratio
predict ps
egen exactmatch = group( timevar ICBIC)
gen pscore = 10*exactmatch + ps
psmatch2 dummy_treat , pscore(pscore) cal(0.05) common neighbor(1) noreplacement
Can anyone please advise on how to achieve that? Many thanks.
Code:
* Example generated by -dataex-. For more info, type help dataex clear input int c_id float(timevar ln_tot_assets profitability tangibility mtb_ratio) byte dummy_join float dummy_treat double _weight int(_id _n1) 193 243 16.802202 .10005046 .2314329 9.003746 0 0 1 27664 . 208 243 12.00873 -.04992021 .4600139 7.113888 0 0 1 27665 . 169 243 12.30863 .18140794 .1556859 6.029928 0 1 1 27665 27665 302 243 12.143713 .14730002 .8396511 2.559861 0 0 1 27784 . 158 243 13.424753 .2460771 .4534546 5.314976 0 1 1 27784 27784 143 244 12.53439 .10639983 .16025147 2.5588746 0 1 1 27831 27831 213 244 12.48605 .15771295 .20859346 3.713875 0 0 1 27831 . 100 244 15.037695 .08367095 .3745413 1.4208206 0 1 1 28072 28072 241 244 13.73277 .08712249 .8889005 39.22791 0 0 1 28072 . 131 244 14.13156 .1303492 .3942553 2.530993 0 1 1 28117 28117 220 244 12.744736 .18055803 .54411614 2.706573 0 0 1 28117 . 213 245 12.48605 .15771295 .20859346 3.713875 0 0 1 28184 . 109 245 13.149642 .13692157 .13714732 28.7388 0 1 1 28184 28184 155 245 16.185017 .08183612 .2994641 3.025999 0 1 1 28224 28224 175 245 15.770484 .10087815 .020241147 .8113729 0 0 1 28224 . 193 245 16.854015 .14865384 .2311488 6.451898 0 0 1 28367 . 71 245 15.70876 .03490321 .01868133 1.2961305 0 1 1 28367 28367 213 246 12.48605 .15771295 .20859346 3.713875 0 0 1 28538 . 119 246 12.370604 .0962811 .13141732 11.00779 0 1 1 28538 28538 214 246 13.089137 .0864781 .10118265 5.538691 0 0 1 28539 . 145 246 14.938308 -.017067846 .006352816 3.336922 0 1 1 28539 28539 146 246 15.49898 .22059405 .1585224 2.3237605 0 1 1 28606 28606 253 246 18.566174 .06230276 .1096999 1.3983625 0 0 1 28606 . 17 246 12.330513 .1973793 .12675329 12.30748 0 1 1 28780 28780 241 246 13.73277 .08712249 .8889005 39.22791 0 0 1 28780 . 140 246 13.868586 .1776808 .3540425 2.3460033 0 1 1 28781 28781 265 246 14.947557 .07865893 .14036106 .9359225 0 0 1 28781 . 66 247 12.992812 .13954854 .12826326 7.561958 0 1 1 28891 28891 213 247 12.73605 .13770127 .1760821 2.2743716 0 0 1 28891 . 193 247 16.854015 .14865384 .2311488 6.451898 0 0 1 29073 . 160 247 15.781055 .07878075 .3564885 .6431261 0 1 1 29073 29073 76 247 14.561218 .081857 .2991022 3.005537 0 1 1 29132 29132 241 247 13.73277 .08712249 .8889005 39.22791 0 0 1 29132 . 3 248 13.385486 .1402982 .2602872 4.086562 0 1 1 29243 29243 213 248 12.73605 .13770127 .1760821 2.2743716 0 0 1 29243 . 265 249 14.919476 .05818773 .13905375 .9311196 0 0 1 29814 . 79 249 14.905082 .08980828 .0877901 3.848259 0 1 1 29814 29814 end format %tq timevar
------------------ copy up to and including the previous line ------------------
Thanks for your responses!
