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]
st: GMM and Asset Pricing
From
Adrian Stork <[email protected]>
To
"[email protected]" <[email protected]>
Subject
st: GMM and Asset Pricing
Date
Sat, 5 Apr 2014 11:42:20 +0200
I'm currently trying to implement a GMM estimation in an asset pricing
context where I want to determine the risk premium of a factor (i.e.
the market factor).
Usually I would estimate this with a Fama-MacBeth procedure (i.e. a
two step procedure where I run time-series regressions to get the beta
coefficients and then
run cross-sectional regressions each month to get the risk premium of
that beta-coefficient). Furthermore, I'm using a panel of individual
stocks which is why Errors-in-Variables
is a big issue here.
Now, I read that GMM is:
1.) a way to reduce the Errors-in-Variables problem (the alternative
with the Instrumental-Variables solution is clear to me and I already
have implemented it)
2.) able to this two-step approach at the same time.
So my key question is how to implement this time-series and
cross-sectional approach for GMM in Stata?
Sor far the only thing I came up with is:
. gmm (return - ({xb:beta}+{b0})), instruments(beta)
but here I already had to estimate my beta in a time-series regression
and didn't estimate it at the same time in the GMM framework and even
then I think it's not correctly specified although it works.
Does anyone know how to do this GMM-estimation in an asset pricing context?
Any help would be highly appreciated and would save my weekend.
P.S. regarding the data: I have a panel of monthly individual stock
returns over 20 years as dependent variable and I have the market
factor (e.g. the return of the S&P500) as independent variable.
Best wishes,
Adrian
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/