Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: very small sample comarsion
From
Gordon Hughes <[email protected]>
To
[email protected]
Subject
Re: st: very small sample comarsion
Date
Sun, 27 Feb 2011 11:11:01 +0000
Dr. Conroy is entirely correct that the interest
in this data lies in whether there is some change
in trend after the entry of a new CEO relative to
the performance of either the industry or the
company before the change. But you need to think
about one additional feature: the fact that there
has been a change in CEO is not random. A few
CEO's may leave because of age or ill-health, but
many will leave because the performance of the company was disappointing.
In practice, what you are really trying to
isolate is whether a change in CEO leads to a
change in the performance of the company
conditional on the economic environment, its
competitors and the fact that the change was
often prompted by a belief that a newcomer ought
to be able to do better. I would not be
confident that 50 companies and ~200 changes of
CEO is sufficient to do this, but that is the
basic character of your sample. So your model
might express growth in real assets or profits as
a function of current & (variously) lagged GDP
growth, industry growth, other industry
characteristics plus whether there was a change
in CEO 1, 2, .. years ago. You could treat that
as either being endogenous (using instrumental
variables) or using a sample selection model.
Gordon Hughes
[email protected]
------------------------------
Date: Sat, 26 Feb 2011 16:23:36 +0000
From: Ronan Conroy <[email protected]>
Subject: Re: st: Re: very small sample comarsion
On 26 Feb 2011, at 13:51, ajjee wrote:
> Thanks Mr.Ronan for reply
>
> I've panel data of near 50 firms for 15 years. I want to anlyse the
> performance of CEO of each firm.
Are you sure? I don't think that the individual
performance of each CEO is interesting. I think
that the interesting thing is the arrival of a new CEO.
So your data structure is that you have, as your
predictor variable, the time since the new CEO
took over, and as a predicted variable you have
the change in CAR for that year. Or something
like that I'm not an economist (and clearly
no-one in Ireland is!). Your data are clustered
by company. Or so it seems from your description
of your hypothesis. Certainly, analysing each
individual CEO won't get you any nearer to it.
> Each CEO has 3 or 5 years in office. So I
> have many indicators from the company's balance sheet and I want to check
> improvement/distortion in these indicators
before and after a new CEO comes.
> Suppose I want to analsye capital to assets ratio (CAR), for this purpose I
> computed 3/5 year average of CAR and then compare these averages before and
> after the new CEO appointment. But it is very crude measure to evaluate the
> performance. About the independence of the samples, the firms are
> independent, of course, but within firm, sample are not independent.
Ronán Conroy
[email protected]
Associate Professor
Division of Population Health Sciences
Royal College of Surgeons in Ireland
Beaux Lane House
Dublin 2
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/