Can you please tell me if my interpretation is correct?
In model (1), with an estimated coefficient of -.0187 the dummy for the second year of observations is negative and significant on the one percent level. That means minus 18.7 percent post treatment change for the untreated individuals or in other words, non-adopting firms sell, on average, 18.7 percent less, ceteris paribus. The ICT adoption dummy, significant at the five percent level, is the pre-treatment difference between treated and untreated. With an estimated coefficient of 0.543, this indicates an average pre-treatment difference in sales of 54.3 percent for treated over untreated firms. This indicates that firms who adopt ICT between year 1 and 2 already had on average higher sales than those who did not, ceteris paribus. As a consequence, have I simply compared outcomes of treated and untreated firms in year 2, the estimated effect of the treatment would have been biased upwards. The interaction term between the second year dummy variable and ICT adoption is significant at the one percent level. The estimated coefficient is 0.444, this means that for ICT adopted firms between year one and year two, sales increased on average by 44 percent more than that of non-adopting firms, ceteris paribus. [Add -18.7 to 44.4 to get wave on wave % change in sales for adopting firms]
Code:
. xtreg LogSales c.dum_ICT##(c.Year2 c.LogFirmSize c.c6) c.LogFirmAge c.b2b i.Cid, vce(cluster i) Random-effects GLS regression Number of obs = 7,720 Group variable: i Number of groups = 4,404 R-sq: Obs per group: within = 0.1352 min = 1 between = 0.7384 avg = 1.8 overall = 0.6387 max = 2 Wald chi2(35) = 15608.20 corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 (Std. Err. adjusted for 4,404 clusters in i) ----------------------------------------------------------------------------------------- | Robust LogSales | Coef. Std. Err. z P>|z| [95% Conf. Interval] ------------------------+---------------------------------------------------------------- dum_ICT | .5432373 .2749938 1.98 0.048 .0042594 1.082215 Year2 | -.1868714 .0473769 -3.94 0.000 -.2797284 -.0940144 LogFirmSize | 1.089397 .0223518 48.74 0.000 1.045588 1.133206 c6 | .2488145 .0614796 4.05 0.000 .1283166 .3693123 | c.dum_ICT#c.Year2 | .4444767 .1238044 3.59 0.000 .2018246 .6871288 | c.dum_ICT#c.LogFirmSize | -.1248695 .0388428 -3.21 0.001 -.2010001 -.0487389 | c.dum_ICT#c.c6 | -.36005 .1789319 -2.01 0.044 -.71075 -.00935 | LogFirmAge | .3055828 .0500255 6.11 0.000 .2075346 .403631 b2b | .0089473 .0007782 11.50 0.000 .0074221 .0104724 | Cid | 11 | .7668476 .1964145 3.90 0.000 .3818822 1.151813 12 | -3.297935 .1730777 -19.05 0.000 -3.637161 -2.958709 13 | .7270781 .2315544 3.14 0.002 .2732398 1.180916 14 | -.6149425 .2143766 -2.87 0.004 -1.035113 -.194772 15 | .6948599 .1602769 4.34 0.000 .3807231 1.008997 16 | .9412395 .2486655 3.79 0.000 .4538641 1.428615 17 | .2757622 .1949146 1.41 0.157 -.1062633 .6577877 18 | .5553843 .1906491 2.91 0.004 .181719 .9290497 19 | -2.40611 .1583451 -15.20 0.000 -2.716461 -2.09576 20 | -.7922745 .3916108 -2.02 0.043 -1.559818 -.0247314 21 | -.8017934 .1666927 -4.81 0.000 -1.128505 -.4750817 22 | -3.343957 .2024582 -16.52 0.000 -3.740768 -2.947146 23 | -4.023811 .2263285 -17.78 0.000 -4.467406 -3.580215 24 | -.6255861 .1905041 -3.28 0.001 -.9989673 -.2522049 25 | .2832945 .1639983 1.73 0.084 -.0381362 .6047252 26 | 1.811988 .24045 7.54 0.000 1.340715 2.283262 27 | -1.585737 .1457486 -10.88 0.000 -1.871399 -1.300075 28 | 1.20013 .2273049 5.28 0.000 .7546209 1.64564 29 | .5454183 .1566247 3.48 0.000 .2384396 .8523971 30 | 1.615417 .1921696 8.41 0.000 1.238772 1.992063 31 | -2.024318 .165459 -12.23 0.000 -2.348612 -1.700024 32 | 1.527866 .1624043 9.41 0.000 1.209559 1.846173 33 | .7318168 .2407154 3.04 0.002 .2600232 1.20361 34 | 1.493808 .165864 9.01 0.000 1.168721 1.818896 35 | 2.884055 .1547119 18.64 0.000 2.580825 3.187285 36 | -4.552385 .1591759 -28.60 0.000 -4.864364 -4.240406 | _cons | 13.48519 .200877 67.13 0.000 13.09147 13.8789 ------------------------+---------------------------------------------------------------- sigma_u | 0 sigma_e | 1.7565325 rho | 0 (fraction of variance due to u_i) -----------------------------------------------------------------------------------------
I would really appreciate if you can tell me if this is right. Thanks a lot!
Kind regards
Dominik
0 Response to Interpretation Difference in Differences
Post a Comment