Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


From   Markus Eberhardt <>
Date   Tue, 22 Feb 2011 11:53:20 +0000

The canned levpet command for Levinsohn & Petrin (2003, REStud) is
relatively restrictive in what you can and cannot include in the
estimation equation (e.g. you are still looking at a Cobb-Douglas
form). You can of course code the estimator yourself, but this may be
a bridge too far for many Stata users. Together with my buddy
Christian Helmers I've written a detailed introduction to these
production function estimators ("Untested Assumptions and Data
Slicing", see:
for which we're also disseminating the Stata code... just not yet (we
need to clean the code and especially the ACF needs to be implemented
much more efficiently).

I think it'd be better to familiarise yourself with the literature on
these estimators before rushing in and running some pre-coded
estimation command...

Markus Eberhardt
ESRC Post-doctoral Research Fellow, Centre for the Study of African
Economies, Department of Economics, University of Oxford
Stipendiary Lecturer, St Catherine's College, Oxford

mail: Centre for the Study of African Economies, Department of
Economics, Manor Rd, Oxford OX1 3UQ, England

On 22 February 2011 11:44, Neil Shephard <> wrote:
> On Tue, Feb 22, 2011 at 10:38 AM, <> wrote:
>> Hi everybody,
>> I need to run a regression using Levinshon&Petrin
>> estimators.
> I'm not familiar with this method at all.  The Statalist FAQ (linked
> from the bottom of every post) requests that full citations are
> provided for those not familiar with the work you are citing.
>> I have a panel of data for about 20.000 italian industries
>> divided in three samples (North, South, Middle). I have also created a
>> dummy variable for each year and I would like to run a regression for
>> each sample (using Lev-Pet) with the dummy-year variable. I would like
>> to know the stata commands to do that.
>> Alternatively, is it possible
>> to run a regression (using Lev-Pet) year by year in order to see how
>> the coefficients vary across the years??
> -findit petrin- returns one hit that may be of interest to you which
> is a package that was published in the Stata Journal...
> Petrin, A., B. P. Poi, and J. Levinsohn. Production function
> estimation in Stata using inputs to control for unobservables. Stata
> Journal 4(2):113-123.
> Neil
> --
> “Truth in science can be defined as the working hypothesis best suited
> to open the way to the next better one.” - Konrad Lorenz
> Email -
> Website -
> Photos -
> *
> *   For searches and help try:
> *
> *
> *

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index