Fixed effects and collinearity

[Brendan Clark]

Hi,

I'm running a couple of fixed effects regressions on my unbalanced panel data. I'm analyzing whether frequent, within factory (or unit) vaccination camps have an effect on vaccination status of employees. So my DV is Vaccination status and my DV is the number of vaccine camps conducted in a unit in a week, 't'.

The model I'm using for my data is a fixed effects regression and I have included a set of individual (employee id), factory level (unit) and time (week) fixed effects. (my code is presented below; my loop broke and I'm in the process of fixing it, so I'm sorry about the long lines of code) - I have attached my regression table in this post.

The issue is, I get a completely null effect when adding in employee level fixed effects yet, this does not happen when adding in factory or time fixed effects. Is this a case of collinearity? I double checked my variable consutrction but as far as i can tell, everything seems fine so i'm a little confused regarding what to do!

reghdfe vaccinated cumfreq_camp , absorb(empcode week) vce(cluster factory)
outreg2 using myreg2.doc, replace label addtext(Employee FE, YES, Time FE, YES)

reghdfe vaccinated cumfreq_camp_l1 , absorb(empcode week) vce(cluster factory)
outreg2 using myreg2.doc, append label addtext(Employee FE, YES, Time FE, YES)

reghdfe vaccinated cumfreq_camp_l2 , absorb(empcode week) vce(cluster factory)
outreg2 using myreg2.doc, append label addtext(Employee FE, YES, Time FE, YES)

reghdfe vaccinated cumfreq_camp_l3 , absorb(empcode week) vce(cluster factory)
outreg2 using myreg2.doc, append label addtext(Employee FE, YES, Time FE, YES)

reghdfe vaccinated cumfreq_camp cumfreq_camp_l1 cumfreq_camp_l2 cumfreq_camp_l3 , absorb(empcode week) vce(cluster factory)
outreg2 using myreg2.doc, append label addtext(Employee FE, YES, Time FE, YES)


reghdfe vaccinated cumfreq_camp , absorb(factory week) vce(cluster factory)
outreg2 using myreg2.doc, append label addtext(Unit FE, YES, Time FE, YES)

reghdfe vaccinated cumfreq_camp_l1 , absorb(factory week) vce(cluster factory)
outreg2 using myreg2.doc, append label addtext(Unit FE, YES, Time FE, YES)

reghdfe vaccinated cumfreq_camp_l2 , absorb(factory week) vce(cluster factory)
outreg2 using myreg2.doc, append label addtext(Unit FE, YES, Time FE, YES)

reghdfe vaccinated cumfreq_camp_l3 , absorb(factory week) vce(cluster factory)
outreg2 using myreg2.doc, append label addtext(Unit FE, YES, Time FE, YES)

reghdfe vaccinated cumfreq_camp cumfreq_camp_l1 cumfreq_camp_l2 cumfreq_camp_l3 , absorb(factory week) vce(cluster factory)
outreg2 using myreg2.doc, append label addtext(Unit FE, YES, Time FE, YES)

* Example generated by -dataex-. For more info, type help dataex
clear
input str11 factory float(week vaccinated cumfreq_camp cumfreq_camp_l1 cumfreq_camp_l2 cumfreq_camp_l3 cumfreq_vacc_l1 cumfreq_vacc_l2 cumfreq_vacc_l3) long empcode
"A39" 1 1 0 0 0 0 0 0 0 20331581
"A39" 1 1 0 0 0 0 0 0 0 20361281
"A39" 1 0 0 0 0 0 0 0 0 20424411
"A39" 1 1 0 0 0 0 0 0 0 20383342
"A39" 1 1 0 0 0 0 0 0 0 20426766
"A39" 1 1 0 0 0 0 0 0 0 20033686
"A39" 1 1 0 0 0 0 0 0 0 20121442
"A39" 1 1 0 0 0 0 0 0 0 20109266
"A39" 1 1 0 0 0 0 0 0 0 20443148
"A39" 1 1 0 0 0 0 0 0 0 20039075
"A39" 1 1 0 0 0 0 0 0 0 20439688
"A39" 1 1 0 0 0 0 0 0 0 20459184
"A39" 1 1 0 0 0 0 0 0 0 20062522
"A39" 1 0 0 0 0 0 0 0 0 20440647
"A39" 1 0 0 0 0 0 0 0 0 20458486
"A39" 1 1 0 0 0 0 0 0 0 20111069
"A39" 1 1 0 0 0 0 0 0 0 20331737
"A39" 1 1 0 0 0 0 0 0 0 20036761
"A39" 1 0 0 0 0 0 0 0 0 20455300
"A39" 1 1 0 0 0 0 0 0 0 20189180
"A39" 1 1 0 0 0 0 0 0 0 20174536
"A39" 1 0 0 0 0 0 0 0 0 20451179
"A39" 1 1 0 0 0 0 0 0 0 20206422
"A39" 1 1 0 0 0 0 0 0 0 20213773
"A39" 1 0 0 0 0 0 0 0 0 20454254
"A39" 1 1 0 0 0 0 0 0 0 20459215
"A39" 1 0 0 0 0 0 0 0 0 20458482
"A39" 1 0 0 0 0 0 0 0 0 20451226
"A39" 1 1 0 0 0 0 0 0 0 20027145
"A39" 1 1 0 0 0 0 0 0 0 20103873
"A39" 1 0 0 0 0 0 0 0 0 20432748
"A39" 1 1 0 0 0 0 0 0 0 20174251
"A39" 1 1 0 0 0 0 0 0 0 20441728
"A39" 1 0 0 0 0 0 0 0 0 20253074
"A39" 1 1 0 0 0 0 0 0 0 20031521
"A39" 1 1 0 0 0 0 0 0 0 20026593
"A39" 1 0 0 0 0 0 0 0 0 20400746
"A39" 1 1 0 0 0 0 0 0 0 20458473
"A39" 1 1 0 0 0 0 0 0 0 20028020
"A39" 1 0 0 0 0 0 0 0 0 20190820
"A39" 1 1 0 0 0 0 0 0 0 20447669
"A39" 1 1 0 0 0 0 0 0 0 20387627
"A39" 1 1 0 0 0 0 0 0 0 20061837
"A39" 1 0 0 0 0 0 0 0 0 20031625
"A39" 1 1 0 0 0 0 0 0 0 20282512
"A39" 1 1 0 0 0 0 0 0 0 20277885
"A39" 1 1 0 0 0 0 0 0 0 20193866
"A39" 1 0 0 0 0 0 0 0 0 20451237
"A39" 1 1 0 0 0 0 0 0 0 20451270
"A39" 1 0 0 0 0 0 0 0 0 20403004
"A39" 1 0 0 0 0 0 0 0 0 20438565
"A39" 1 1 0 0 0 0 0 0 0 20451167
"A39" 1 0 0 0 0 0 0 0 0 20295895
"A39" 1 1 0 0 0 0 0 0 0 20398107
"A39" 1 1 0 0 0 0 0 0 0 20085535
"A39" 1 1 0 0 0 0 0 0 0 20212132
"A39" 1 1 0 0 0 0 0 0 0 20426753
"A39" 1 1 0 0 0 0 0 0 0 20364064
"A39" 1 1 0 0 0 0 0 0 0 20358291
"A39" 1 1 0 0 0 0 0 0 0 20333624
"A39" 1 1 0 0 0 0 0 0 0 20348221
"A39" 1 1 0 0 0 0 0 0 0 20097063
"A39" 1 0 0 0 0 0 0 0 0 20031618
"A39" 1 1 0 0 0 0 0 0 0 20406845
"A39" 1 0 0 0 0 0 0 0 0 20326515
"A39" 1 1 0 0 0 0 0 0 0 20171722
"A39" 1 1 0 0 0 0 0 0 0 20031324
"A39" 1 1 0 0 0 0 0 0 0 20252998
"A39" 1 0 0 0 0 0 0 0 0 20457694
"A39" 1 1 0 0 0 0 0 0 0 20035322
"A39" 1 1 0 0 0 0 0 0 0 20365098
"A39" 1 1 0 0 0 0 0 0 0 20228508
"A39" 1 1 0 0 0 0 0 0 0 20305257
"A39" 1 1 0 0 0 0 0 0 0 20025253
"A39" 1 1 0 0 0 0 0 0 0 20451242
"A39" 1 1 0 0 0 0 0 0 0 20452215
"A39" 1 1 0 0 0 0 0 0 0 20311696
"A39" 1 1 0 0 0 0 0 0 0 20429644
"A39" 1 1 0 0 0 0 0 0 0 20027133
"A39" 1 1 0 0 0 0 0 0 0 20458469
"A39" 1 1 0 0 0 0 0 0 0 20449862
"A39" 1 1 0 0 0 0 0 0 0 20111419
"A39" 1 0 0 0 0 0 0 0 0 20178480
"A39" 1 1 0 0 0 0 0 0 0 20164987
"A39" 1 1 0 0 0 0 0 0 0 20199407
"A39" 1 1 0 0 0 0 0 0 0 20298728
"A39" 1 1 0 0 0 0 0 0 0 20329741
"A39" 1 0 0 0 0 0 0 0 0 20027354
"A39" 1 1 0 0 0 0 0 0 0 20221717
"A39" 1 1 0 0 0 0 0 0 0 20052139
"A39" 1 1 0 0 0 0 0 0 0 20459210
"A39" 1 0 0 0 0 0 0 0 0 20027372
"A39" 1 0 0 0 0 0 0 0 0 20443023
"A39" 1 1 0 0 0 0 0 0 0 20226171
"A39" 1 0 0 0 0 0 0 0 0 20304022
"A39" 1 1 0 0 0 0 0 0 0 20026951
"A39" 1 1 0 0 0 0 0 0 0 20028376
"A39" 1 1 0 0 0 0 0 0 0 20043046
"A39" 1 1 0 0 0 0 0 0 0 20430500
"A39" 1 0 0 0 0 0 0 0 0 20331687
end

[Carlo Lazzaro]

Brendan:
I find your post difficult to follow.
That said:
1) if -factory- is a time-invariant variable, I would double-check whether it causes the issue;
2) from your (not advisable, as per FAQ) screenshot, some of your regressions reports Rsq=1, which is clearly suspicious and requires a supplementary analysis;
3) in addition, it is not clear whether your Rsq is an overall or within Rsq.