Interaction in reghdfe: base group

[Davide Rigo]

Hi all,

I have the following regression

reghdfe dep_var ib1.age##c.x, absorb(FEs) vce(robust)

Age equals 1,2,3,..,11. I want to specify age == 1 as the base group, but reghdfe (contrary to areg) uses the wrong base group for the interaction term. As you can see from the example below, reghdfe is taking age == 10 as the base group among the interaction coefficients. Wondering if this is a bug of the command? And how to set the base group for the interaction terms?

age
1 0 (base)
2 .2494122 .0212155 11.76 0.000 .2078298 .2909945
3 .0315153 .0221678 1.42 0.155 -.0119335 .0749642
4 -.0340088 .0235655 -1.44 0.149 -.080197 .0121795
5 -.1115884 .0291358 -3.83 0.000 -.1686946 -.0544823
6 -.1373492 .03146 -4.37 0.000 -.1990108 -.0756877
7 -.073757 .041974 -1.76 0.079 -.156026 .0085121
8 -.0449025 .0513712 -0.87 0.382 -.1455901 .0557851
9 -.3731037 .0718682 -5.19 0.000 -.5139653 -.232242
10 -.0575927 .0981525 -0.59 0.557 -.2499714 .1347861
11 .3076132 .3640556 0.84 0.398 -.4059355 1.021162

age#c.x |
1 | -34.59183 27.9904 -1.24 0.217 -89.45299 20.26933
2 | -33.87957 27.9907 -1.21 0.226 -88.74131 20.98218
3 | -32.79275 27.99093 -1.17 0.241 -87.65495 22.06945
4 | -33.51221 27.99313 -1.20 0.231 -88.37873 21.35431
5 | -34.25638 27.99369 -1.22 0.221 -89.12399 20.61123
6 | -32.55799 27.99707 -1.16 0.245 -87.43222 22.31625
7 | -32.35737 27.99757 -1.16 0.248 -87.23258 22.51783
8 | -28.59181 28.19613 -1.01 0.311 -83.85619 26.67257
9 | -64.17292 29.42105 -2.18 0.029 -121.8381 -6.507689
10 | 0 (empty)
11 | 127.3251 116.9149 1.09 0.276 -101.828 356.4782


Many thanks in advance,

Davide

[Phil Bromiley]

If you only have one "absorbed" variable, I'm not sure why you can't use xtreg instead. Being Stata written, xtreg accommodates all the conventions including letting you indicate the base group.