I am trying to estimate a fixed-effect model to evaluate the effect of program intervention. I have daily data 12 months before the intervention and 12 months after the intervention.
As a benchmark model, I first estimate the following, and the ATE is identified by the coefficient of the interaction between the treatment and post variables (b3),
yit =b1TREATi + b2POSTit +b3TREATi*POSTit +controls +fixed effects+eit
Code:
gen TREAT_POST=el_TREAT*POST_treat xtset newid new_day xtreg daily_water i.POST_treat TREAT_POST i.month Mtemp Sun_D Rain_D away i.dow i.year,fe cl(newid)
After this, I want to estimate and plot the ATEs in each month after the treatment period. I use the following code to generate the ATEs. From this model, I presume that the coefficients for the interaction between month_treat and POST are ATEs for each month after the treatment period.
Code:
gen month_treat=month*el_TREAT xtreg daily_water i.month_treat##i.POST_treat Mtemp Sun_D Rain_D away i.dow i.year,fe cl(newid)
I put the sample data and model results below.
Many thanks for your help in advance!
Code:
Fixed-effects (within) regression Number of obs = 262,435 Group variable: newid Number of groups = 415 R-sq: Obs per group: within = 0.1783 min = 260 between = 0.1623 avg = 632.4 overall = 0.1639 max = 731 F(36,414) = 32.03 corr(u_i, Xb) = 0.0656 Prob > F = 0.0000 (Std. Err. adjusted for 415 clusters in newid) Robust daily_water Coef. Std. Err. t P>t [95% Conf. Interval] month_treat 1 -14.05208 4.755825 -2.95 0.003 -23.40066 -4.703508 2 -13.3265 4.876675 -2.73 0.007 -22.91263 -3.740371 3 -.1929066 5.669033 -0.03 0.973 -11.33659 10.95077 4 -2.404938 5.429825 -0.44 0.658 -13.0784 8.268526 5 4.115936 5.508886 0.75 0.455 -6.712939 14.94481 6 4.275163 5.780306 0.74 0.460 -7.087246 15.63757 7 3.327438 5.874876 0.57 0.571 -8.220867 14.87574 8 4.405579 5.138175 0.86 0.392 -5.694587 14.50574 9 8.603835 4.601088 1.87 0.062 -.4405733 17.64824 10 4.321167 3.383359 1.28 0.202 -2.329539 10.97187 11 -1.052435 3.101057 -0.34 0.734 -7.148215 5.043346 12 0 (omitted) 1.POST_treat -4.729409 3.57058 -1.32 0.186 -11.74814 2.289318 month_treat#POST_treat 1 1 8.122836 7.652072 1.06 0.289 -6.918924 23.1646 2 1 8.110173 7.870883 1.03 0.303 -7.361705 23.58205 3 1 -9.648266 7.453082 -1.29 0.196 -24.29887 5.002335 4 1 -7.793852 7.782507 -1.00 0.317 -23.09201 7.504304 5 1 -9.078844 7.394379 -1.23 0.220 -23.61405 5.456366 6 1 -7.826685 7.525708 -1.04 0.299 -22.62005 6.966679 7 1 -9.695505 6.62882 -1.46 0.144 -22.72585 3.334836 8 1 -11.36933 6.432636 -1.77 0.078 -24.01403 1.275367 9 1 -14.44174 7.03736 -2.05 0.041 -28.27515 -.6083234 10 1 -10.67377 6.98552 -1.53 0.127 -24.40528 3.05774 11 1 -2.222318 7.668682 -0.29 0.772 -17.29673 12.85209 12 1 -7.011918 6.918561 -1.01 0.311 -20.61181 6.587971 Mtemp -.4811931 .0891859 -5.40 0.000 -.6565069 -.3058794 Sun_D .0053095 .0103424 0.51 0.608 -.0150206 .0256396 Rain_D -.2029803 .0736975 -2.75 0.006 -.3478483 -.0581123 away -165.7579 5.145493 -32.21 0.000 -175.8725 -155.6434 dow 1 -19.45075 1.844929 -10.54 0.000 -23.07735 -15.82416 2 -18.78981 1.754284 -10.71 0.000 -22.23823 -15.3414 3 -18.33057 1.733601 -10.57 0.000 -21.73832 -14.92281 4 -18.02998 1.639566 -11.00 0.000 -21.25289 -14.80706 5 -19.4266 1.868959 -10.39 0.000 -23.10043 -15.75277 6 -.6209718 1.246531 -0.50 0.619 -3.071291 1.829348 year 2016 10.58831 3.006831 3.52 0.000 4.677755 16.49887 2017 7.982201 3.661211 2.18 0.030 .7853198 15.17908 _cons 192.2065 2.591394 74.17 0.000 187.1125 197.3004 sigma_u 78.626063 sigma_e 104.21317 rho .36274481 (fraction of variance due to u_i)
[CODE]
* Example generated by -dataex-. To install: ssc install dataex
clear
input byte month float(newid daily_water el_TREAT POST_treat)
3 2 234 1 0
3 2 70 1 0
3 2 175 1 0
3 2 77 1 0
3 2 89 1 0
3 2 124 1 0
3 2 262 1 0
3 2 225 1 0
3 2 135 1 0
3 2 143 1 0
3 2 8 1 0
3 2 1 1 0
3 2 6 1 0
3 2 1 1 0
3 2 1 1 0
3 2 1 1 0
3 2 5 1 0
3 2 50 1 0
3 2 67 1 0
3 2 229 1 0
3 2 306 1 0
3 2 96 1 0
3 2 74 1 0
3 2 100 1 0
3 2 335 1 0
3 2 115 1 0
3 2 204 1 0
3 2 144 1 0
3 2 99 1 0
3 2 74 1 0
3 2 89 1 0
4 2 59 1 0
4 2 156 1 0
4 2 396 1 0
4 2 117 1 0
4 2 171 1 0
4 2 163 1 0
4 2 252 1 0
4 2 66 1 0
4 2 53 1 0
4 2 112 1 0
4 2 229 1 0
4 2 211 1 0
4 2 77 1 0
4 2 83 1 0
4 2 114 1 0
4 2 101 1 0
4 2 101 1 0
4 2 319 1 0
4 2 237 1 0
4 2 104 1 0
4 2 79 1 0
4 2 108 1 0
4 2 61 1 0
4 2 125 1 0
4 2 265 1 0
4 2 175 1 0
4 2 85 1 0
4 2 125 1 0
4 2 85 1 0
4 2 91 1 0
5 2 241 1 0
5 2 134 1 0
5 2 205 1 0
5 2 74 1 0
5 2 101 1 0
5 2 134 1 0
5 2 141 1 0
5 2 116 1 0
5 2 259 1 0
5 2 223 1 0
5 2 134 1 0
5 2 86 1 0
5 2 65 1 0
5 2 209 1 0
5 2 302 1 0
5 2 183 1 0
5 2 100 1 0
5 2 65 1 0
5 2 131 1 0
5 2 298 1 0
5 2 109 1 0
5 2 269 1 0
5 2 264 1 0
5 2 82 1 0
5 2 68 1 0
5 2 482 1 0
5 2 57 1 0
5 2 239 1 0
5 2 254 1 0
5 2 282 1 0
5 2 150 1 0
6 2 327 1 0
6 2 114 1 0
6 2 93 1 0
6 2 251 1 0
6 2 55 1 0
6 2 134 1 0
6 2 240 1 0
6 2 101 1 0
0 Response to How to estimate a monthly treatment effect in fixed effect panel data model ?
Post a Comment