Statalist


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

Re: st: Olley Pakes estimation (-opreg- versus -levpet-)


From   "Rafal Raciborski" <rraciborski@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Olley Pakes estimation (-opreg- versus -levpet-)
Date   Mon, 15 Dec 2008 19:41:18 -0500

luigi,

you can modify the code from the appendix to our article (yasar,
raciborski, & poi 2008) - you can estimate the model without using the
exit variable, in which case you will lose the phat term.  this way,
you will still be controlling for simultaneity but obviously not for
the selection bias.

rafal





On Mon, Dec 15, 2008 at 6:51 PM, Luigi Pascali <Luigi.Pascali@bc.edu> wrote:
> Dear all,
>
> first of all, thank you very much for your prompt answers. They were very helpful!
>
> However I still have some doubts. I would like to estimate a gross revenue production function in a panel without entry/exit dynamics. Moreover, I would like to use "Investments" (which is not an input in the production function) as a proxy for TFP shocks and not "Intermediates" as in Levinsohn-Petrin.
>
> (Ackerberg, Caves and Frazer (2006) shows that the Levinsohn-Petrin methodology suffers of a serious collinearity problem)
>
> I cannot use -opreg- if I don't have firms leaving the markets. However, I also cannot use -levpet- since in this routine, when estimating a gross revenue production function, the proxy variable ("investments" in my case) is also considered as an input in the production function.
>
> When I use the -opreg- routine and "pretend" that a single firm (over more than a thousand) leaves the market, the routine works and I get very reasonable estimates of the production function. Since I don't know exactly how the routine works, I would like to know if you think that I am making any major mistake that could affect my estimation.
>
> Thank you so much,
> Luigi
>
>
>  |  On Mon, 15 Dec 2008 15:33:51 -0500
>  |  "Rafal Raciborski" <rraciborski@gmail.com> wrote:
>  |  luigi,
>  |
>  |  for opreg to work, some firms must exit the market.  if none does, you
>  |  may use the user-written -levpet- command which controls for
>  |  simultaneity but not the selection bias.  -search levpet- will locate
>  |  the command for you.
>  |
>  |  best,
>  |
>  |  rafal
>  |
>  |
>  |
>  |  On Mon, Dec 15, 2008 at 2:37 PM, Luigi Pascali <Luigi.Pascali@bc.edu> wrote:
>  |  > I need to estimate a production function using the olley-pakes semiparametric estimation procedure. I am using the routine -opreg- (developed by Yasar, Raciborski, & Poi).
>  |  > I don't have firms that exit the panel and this seems to be a problem when running opreg.
>  |  > The option "exit" in the command is required. For this reason I have constructed a dummy (that I called E) which is equal to 0 in each period for every firm.
>  |  > Then I run the following command:
>  |  > opreg lY, state(lK) proxy(lI) free(lL lM) exit(E)
>  |  >
>  |  > where lY=ln(sales)
>  |  > lk=ln(fixed assets)
>  |  > lL=ln(employees)
>  |  > lM=ln(intermediates)
>  |  > lI=ln(investments)
>  |  >
>  |  > I get the following message:
>  |  >
>  |  > an error occurred when bootstrap executed opreg
>  |  > r(2000);
>  |  >
>  |  > I tried to run the same command "pretending" that some firms where leaving the market and everything worked fine.
>  |  > I would really appreciate if anybody could tell me if I am making any mistake and how I can correct it.
>  |  >
>  |  > Many thanks,
>  |  >
>  |  > Luigi
>  |  >
>  |  > *
>  |  > *   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/
>  |  >
>  |  *
>  |  *   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/
>
> *
> *   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/
>
*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index