{smcl}
{* 19jan2004}{...}
{hline}
help for {hi:levpet}{right:(SJ4-2: st0060)}
{right:dialogs: {dialog levpet:levpet}}
{hline}

{title:Production function estimation}

{p 8 16 2}{cmd:levpet} {it:depvar} [{cmd:if} {it:exp}] [{cmd:in} {it:range}]
{cmd:,} {cmd:free(}{it:varlist}{cmd:)} {cmd:proxy(}{it:varlist}{cmd:)}
{cmd:capital(}{it:varname}{cmd:)} [ [{cmdab:va:lueadded} |
{cmdab:rev:enue}] 
{cmd:justid} 
{cmd:grid} 
{cmd:i(}{it:varname}{cmd:)} 
{cmd:t(}{it:varname}{cmd:)}
{cmd:reps(}{it:#}{cmd:)} 
{cmdab:l:evel(}{it:#}{cmd:)} ]


{title:Syntax for {help predict} after {hi:levpet}}

{p 8 16 2}
{cmd:predict} [{it:type}] {it:newvarname} {cmd:, omega}


{title:Description}

{p 4 4 2}
{cmd:levpet} estimates production functions using intermediate inputs to
control for unobservable productivity shocks.  The methodology is
described in Levinsohn and Petrin (2003) and Petrin, Poi, and Levinsohn
(2004).  All variables are assumed to be in logarithms.


{title:Options}

{p 4 8 2}
{cmd:free(}{it:varlist}{cmd:)} specifies the freely variable inputs.

{p 4 8 2}
{cmd:proxy(}{it:varlist}{cmd:)} specifies one or two intermediate inputs.
If the dependent variable is value added, one or two variables can be
specified.  If the dependent variable is revenue, only one can be specified.

{p 4 8 2}
{cmd:capital(}{it:varname}{cmd:)} specifies the capital variable.

{p 4 8 2}
{cmd:valueadded} indicates that the dependent variable represents value
added.  This is the default.

{p 4 8 2}
{cmd:revenue} indicates that the dependent variable represents gross revenues
or output.  The GMM estimator will be used.

{p 4 8 2}
{cmd:justid} requests that the GMM estimator use only present-period
capital and the first lag of the proxy variable as instruments.  This
option can only be used with {cmd:revenue}

{p 4 8 2}
{cmd:grid} requests that the GMM estimator use a grid search to minimize
the criterion function.  The default is a search method based on the
{cmd:nl} command.  This option can only be used with {cmd:revenue}.

{p 4 8 2}
{cmd:i(}{it:varname}{cmd:)} specifies the variable denoting panels.

{p 4 8 2}
{cmd:t(}{it:varname}{cmd:)} specifies the variable denoting time periods.

{p 4 8 2}
{cmd:reps(}{it:#}{cmd:)} specifies the number of bootstrap replications
to perform.  The default is 50.

{p 4 8 2}
{cmd:level(}{it:#}{cmd:)} specifies the confidence level, as a percentage, for
confidence intervals of the coefficients; see help {help level}.


{title:Examples}

{p 4 4 2}{cmd:. levpet lnva, free(lnb lnw) proxy(lne lnf) capital(lnk)}

{p 4 4 2}{cmd:. levpet lnrev, free(lnb lnw) proxy(lne) capital(lnk) revenue}


{title:References}

{p 4 8 2}
Levinsohn, J. and A. Petrin. 2003. Estimating production functions using
inputes to control for unobservables. {it:Review of Economic Studies} 70: 
317-342.

{p 4 8 2}
Petrin, A., B. P. Poi, and J. Levinsohn. Production function estimation
in Stata using inputs to control for unobservables. {it:Stata Journal} 4(2):
113-123.


{title:Authors}

      Amil Petrin
      University of Chicago
      National Bureau of Economic Research

      Brian Poi
      StataCorp

      James Levinsohn
      University of Michigan
      National Bureau of Economic Research