Dear all,
I need some help with running the 'cgmwildboot' command. I intend to utilize a Control Function approach in my subsequent analyses but decided to first test the procedure by checking with the results from an IV (2SLS) regression that worked.
When I run the command lines below (with 'cgmwildboot'), I obtain the same coefficients as I had obtained in a 2SLS approach (using the partialling-out procedure in STATA). However, as you will see in the picture below, I do not obtain standard errors in my regression output. Moreover when I outreg2 the output, I also find that the excel file has asterisks only for coefficients which are significant at a 1% level. What could be the issue?
****************************************Command lines***********************************
global controls1 age agesq male
global controls2 HHsize bmaritalstatus1 bmaritalstatus2 bmaritalstatus3 bsch_grade1 boccupation1 boccupation2 boccupation3 boccupation4 blandownership1 bincome2 bincome3
reghdfe ln_welfare ln_ref50 $controls1 $controls2 [pw=wgt] if district_code!=102, absorb(year ea_code) cluster(ea_code)
gen smpl_AJ_A1=0
replace smpl_AJ_A1=1 if e(sample)==1
foreach x in ln_welfare ln_ref50 $controls1 $controls2 {
reghdfe `x' if smpl_AJ_A1==1 [pw=wgt], absorb(year ea_code) residuals(`x'_A1)
}
global A1controls1 age_A1 agesq_A1 male_A1
global A1controls2 HHsize_A1 bmaritalstatus1_A1 bmaritalstatus2_A1 bmaritalstatus3_A1 bsch_grade1_A1 boccupation1_A1 boccupation2_A1 boccupation3_A1 boccupation4_A1 blandownership1_A1 bincome2_A1 bincome3_A1
reghdfe ln_ref50 ln_iv_dist $controls1 $controls2 [pw=wgt] if district_code!=102 & smpl_AJ_A1==1, absorb(year ea_code) cluster(ea_code) residuals(control)
outreg2 ln_ivdist using "$ivresults/Partial_1st_cluster_cf.xls", stats(coef se) asterisk(coef) dec(5) nor2 nocons adds("Root MSE", e(rmse)) replace
cgmwildboot ln_welfare_A1 ln_ref50_A1 control $A1controls1 $A1controls2 [pw=wgt] if district_code!=102, cluster(ea_code) bootcluster(ea_code) seed(999)
outreg2 ln_ref50 using "$ivresults/Partial_2SLS_cluster_cf.xls", stats(coef se) asterisk(coef) dec(5) nor2 nocons replace
****************************************Output from the cgmwildboot command line***********************************
Array
So in summary, 2 questions:
1. What could be the issue causing Null Standard errors and also missing t-stats?
2. How can I ensure that the exported coefficients also have asterisks if significant at 5% and 10% level? (Only those significant at 1% level have asterisk (***)!)
Thanks a lot in advance.
Related Posts with Using 'cgmwildboot' _ Standard errors "Null"
SVAR estimation with growth and inflation - Impulse responsesHello I am estimating an SVAR model with growth, taken as the first difference of log(GDP) and infl…
Splitting YearDear, I am quite new to STATA and I need t seprate the year from tthe months and days, i.e. right no…
How to keep ONLY individuals who respond every year in a household panel dataHello everyone, I am new to Stata and as a part of my bachelor thesis, I have to do a fixed effect …
How to model persistence using survival analysis techniquesDear members, I am trying to model persistence in exporting on a sample of new firms using survival…
how to do fixed effects regressionHey my regression does not work, how do I do a fund and time fixed effectes regression? What do I ha…
Subscribe to:
Post Comments (Atom)
0 Response to Using 'cgmwildboot' _ Standard errors "Null"
Post a Comment