Re: st: Clustering Standard Errors versus Dummies

[natasha agarwal <agarwana2@googlemail.com>]

Dear everyone,

Thanks Christopher for your explanation.

In book on Microeconometrics book written by Cameron and Trivedi, they
have metioned that "data may be correlated within a cluster owing to
the presence of a common unobserved cluster-specific term. For
example, it may be felt that there is an unobservable effect common to
all households in a state."

For example if am estimating an augmented production function on an
unbalanced panel dataset using a within estimator where I observe
firms over time in industry j in region r and I include firm specific
fixed effects, industry-specific fixed effects, region specific fixed
effects and time effects, wouldn't the industry-specific fixed effects
and the region specific fixed effects represent the presence of a
common unobserved cluster-specific term? In this case, it is necessary
then to cluster at the industry and the regional level?

Thanks
Natasha
*
*   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/