Interpretation of t-1, average of 2 years, variance over 3 and 5 years

[Abdullah Baba]

URGENT

Hi, I have the following tables. Who can help me to interpret Model 3 and Model 4?

Model 1 = t-1 (last year)
Model 2 = Average of last 2 years
Model 3 = Variance over last 3 years
Model 4 = Variance over last 5 years

CEO Succession = Dependant variable	Model 1 (t-1)	Model 2 (avg. 2 years)	Model 3 (variance 3 years)	Model 4 (variance 5 years)
Net Income/Total Assets (ROA)	-0.353* (0.182)	-0.073 (0.177)	0.006*** (0.001)	0.009*** (0.001)
Buy Hold Return (BHR)	-0.284*** (0.064)	-0.356*** (0.079)	0.365*** (0.074)	0.156*** (0.084)

I don't know how to interpret the variance over 3 years of ROA and BHR. Help is much appreciated.

Thank you very much!

Clyde Schechter

[Clyde Schechter]

I have no more idea what this means than you do. Perhaps even less, since finance is not my field, and I don't even know what Buy Hold Return means. I imagine you found these results in an article, or perhaps in somebody else's Stata output. If they're from an article, read the methods section and see what they say they did for each model. If it's from somebody else's Stata output, contact that person.

A couple of tips about posting. Marking something "urgent" gains you nothing here. In this Forum people answer whatever posts they choose to answer, whenever they feel like doing so. Your urgency is your own problem, and nobody else here is in under any obligation to prioritize your request because of it. Most Stata questions, if clearly asked, and if they include adequate information, get a helpful response in a fairly short time (several hours).

It also gains you nothing to designate me (or anybody else) as the desired respondent of the post. As before, everyone here answers whatever they wish, whenever they wish. If the title of the post looks like it is out of my range of interest, I pass it by. If I read the question and I can't understand it and don't have the sense that it can be quickly clarified by asking some pointed questions, I pass it by. And then there are days, or even longer periods, when I'm not on the Forum at all. I think the use of @whoever is best reserved for the midst of a thread, where several people have responded, and you want to direct a question specifically to one of those responders because it directly asks about what that person wrote.

[Phil Bromiley]

To add to Clyde's comments, you'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. We don't even know what estimator you used to generate these parameters or whether each column reflected a single estimator or two one iv models. With turnover, it is likely you'll want a logit or duration model.

Assuming that everything is measured correctly and this is just regression, it looks like variation in ROA and variation in BHR increase the chances of turnover. But interpreting the meaning of this is really dependent on your discipline and substantive knowledge.