Hello everybody,
My research nvestigates the effect of a restatement(s) on several determinants. In order to do so I use both a logit (focus of my question) and multilogit model on a panel dataset. Ofcourse I first set my panel data with xtset NAME_n Fiscalyearin order to let stata know i'm working with panel data.
The results of the Hausman test indicate that I have to use fixed effects, so those are included as well. Additionally, to account for (probably) heteroskedasticity and serial correlation I use robust standard errors (Jackknife). To obtain results the following command is used:
xtlogit Restatement Big4 LSIZE LFEE LNONFEE Tenure Merger IPO, fe vce(jackknife)
I takes a while for stata to calculate my results but the results seems in line with results of reference papers. So far so good
However things go wrong when I want to perform a robustness check. For the robustness check I want to investigate if results are different for the pre_crisis, crisis and post_crisis period. The following commands are used:
xtlogit Restatement Big4 LSIZE LFEE LNONFEE Tenure Merger IPO if (Fiscalyear>=2000&Fiscalyear<=2006), fe vce(jackknife)
xtlogit Restatement Big4 LSIZE LFEE LNONFEE Tenure Merger IPO if (Fiscalyear>=2007&Fiscalyear<=2008), fe vce(jackknife)
xtlogit Restatement Big4 LSIZE LFEE LNONFEE Tenure Merger IPO if (Fiscalyear>=2009&Fiscalyear<=2016), fe vce(jackknife)
If I want to perform this command stata keeps on running and running, without any results (also after several hours). Is this due to the fact that I also included fixed effects or is there something going wrong in my reasoning? If yes, how can I test my xtlogit fixed effect (jackknife) on different time horizons?
Many thanks,
Matthijs
My research nvestigates the effect of a restatement(s) on several determinants. In order to do so I use both a logit (focus of my question) and multilogit model on a panel dataset. Ofcourse I first set my panel data with xtset NAME_n Fiscalyearin order to let stata know i'm working with panel data.
The results of the Hausman test indicate that I have to use fixed effects, so those are included as well. Additionally, to account for (probably) heteroskedasticity and serial correlation I use robust standard errors (Jackknife). To obtain results the following command is used:
xtlogit Restatement Big4 LSIZE LFEE LNONFEE Tenure Merger IPO, fe vce(jackknife)
I takes a while for stata to calculate my results but the results seems in line with results of reference papers. So far so good
However things go wrong when I want to perform a robustness check. For the robustness check I want to investigate if results are different for the pre_crisis, crisis and post_crisis period. The following commands are used:
xtlogit Restatement Big4 LSIZE LFEE LNONFEE Tenure Merger IPO if (Fiscalyear>=2000&Fiscalyear<=2006), fe vce(jackknife)
xtlogit Restatement Big4 LSIZE LFEE LNONFEE Tenure Merger IPO if (Fiscalyear>=2007&Fiscalyear<=2008), fe vce(jackknife)
xtlogit Restatement Big4 LSIZE LFEE LNONFEE Tenure Merger IPO if (Fiscalyear>=2009&Fiscalyear<=2016), fe vce(jackknife)
If I want to perform this command stata keeps on running and running, without any results (also after several hours). Is this due to the fact that I also included fixed effects or is there something going wrong in my reasoning? If yes, how can I test my xtlogit fixed effect (jackknife) on different time horizons?
Many thanks,
Matthijs
