Dear Statalist members.
For my thesis, I am running a tobit regression with the DV being the ownership percentage in acquisitions.
I have observations from 4 years (but not all companies have an acquisition every year and some have multiple acquisitions per year). As a consequence, the point in time cannot be used as a panel identifier which leaves only the company grouping: xtset CompanyID. As it was done in similar studies, I used Time as a categorical variable instead.
I got recommended that it will be important to decide between tobit and xttobit:
. tobit acquisitionstake IVs i.Year, ll(0.05) ul(1)
. xtset CompanyID
. xttobit acquisitionstake IVs i.Year, ll(0.05) ul(1)
Therefore, I used the likelihood-ratio (LR) test that compares the panel regression with the pooled tobit regression with the null hypothesis being the absence of panel effects. The chi-bar-square of 13.97 (p-value < 0.01) indicates the null hypothesis can be rejected and that xttobit is the appropriate regression (rho =0.43 >0).
However, given the data structure, I have my doubts about whether xttobit should be used in this scenario at all. As xttobit executes a random effects (RE) Tobit regression assuming RE on the panel level, can I use the model with only the companyID as panel identifier?
In my research, I came across a study that has a very similar setup to my thesis (Chari & Chang, 2009). They did not test between tobit and xttobit and used but used a model like the one following. The clustering accounts for acquisitions made by the same companies.
. tobit acquisitionstake IVs i.Year, ll(0.05) ul(1) vce(cluster CompanyID)
To summarize my question, I have some doubts about which approach to follow. Given the data structure and the use in the existing study, I would rather opt for the second option. However, as the LR test considers xttobit to be superior, I am not sure about the right procedure.
Therefore, I am very thankful for any help regarding the choice among the models.
Best regards,
Pascal
Source:
Chari, M. D., & Chang, K. (2009). Determinants of the share of equity sought in cross-border acquisitions. Journal of International Business Studies, 40(8), 1277-1297.
Related Posts with categorical timeID in tobit regression and LR test to decide between tobit and xttobit
How to use hp filter on different groupsI have GDP data for each state from the last 10 years, and I want to use hp filter on each state's t…
Adjusted/Within/Between R squaredHi, I would be very grateful if someone could confirm whether my understanding is correct. I ran a …
Failed ramesy RESET testHi, Im fairly new to stata, i've just manually run a RESET test on my panel data and it failed. I u…
Multinomial logit or -gsem- which is best for simultaneous choiceHello, I am looking for some suggestions to choose an appropriate model for my research. I am tryin…
.ado Programming in Stata without invoking mata: Errors while passing arguments to functionLevel: Beginner Trying to write an .ado file. This is how it looks. When I was debugging this line …
Subscribe to:
Post Comments (Atom)
0 Response to categorical timeID in tobit regression and LR test to decide between tobit and xttobit
Post a Comment