help opreg (SJ12-1: st0145_2)
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Title
opreg -- Production function estimation using Olley and Pakes' technique
Syntax
opreg depvar [if] [in], exit(varname) state(varlist) proxy(varname) free(varlist) [options]
Syntax for predict after opreg
predict [type] newvar [if] [in] [, tfp]
options Description
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
* exit(varname) varname indicates a firm's exit
* state(varlist) state variables appearing in production function
* proxy(varname) variable to proxy for unobserved productivity
* free(varlist) additional variables used in the second stage only
cvars(varlist) additional variables used in both the first and second stages
second use second-degree polynomial expansion
vce(bootstrap, bootstrap_options) specify bootstrap options
level(#) set confidence level; default is level(95)
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
* exit(varname), state(varlist), proxy(varname), and free(varlist) are required.
A panel variable and a time variable must be specified by using xtset; see [XT] xtset.
Description
opreg estimates the production function in the presence of selection bias and simultaneity by using the three-stage algorithm described in Olley and Pakes (1996).
Options
exit(varname) specifies a variable coded 0/1, where at time t for firm i, varname_it equals 1 if the firm exited at the beginning of that period and zero otherwise.
state(varlist) specifies state variables.
proxy(varname) specifies the proxy variable to be used in the equation determining exit.
free(varlist) specifies variable inputs that are not to be included in the exit equation.
cvars(varlist) specifies variable inputs that are also included in the exit equation.
second specifies that a second-degree polynomial expansion is to be used in the first and second stages. The default is to use a third-degree polynomial expansion.
vce(bootstrap, bootstrap_options) allows you to specify options to control the bootstrap process. The most commonly used bootstrap_options is reps(#), which control the number of
replications performed. The default is reps(50).
level(#); see [R] estimation_options.
Option for predict
tfp calculates the log of total factor productivity. To obtain the prediction in levels, exponentiate the results. This is the default.
Remarks
The first and second stages use a third-degree polynomial expansion. For example, with one state variable, the polynomial term is state^2 + proxy^2 + state*proxy + state^3 + proxy^3
+ state^2*proxy + state*proxy^2. With two state variables, the polynomial term is state1^2 + state2^2 + proxy^2 + state1*proxy + state2*proxy + state1*state2, state1^3 + state2^3 +
proxy^3 + state1^2*state2 + state1^2*proxy + state2^2*state1 + state2^2*proxy + proxy^2*state1 + proxy^2*state2 + state1*state2*proxy, and so on.
Examples
. opreg sales, exit(exit) state(stock) proxy(investment) free(material employment) cvars(sizedum2 sizedum3 time)
. opreg sales, exit(exit) state(stock) proxy(investment) free(material employment) cvars(sizedum2 sizedum3 time) vce(bootstrap, reps(200))
In either case, opreg would form the following equations:
exit = stock^2 + investment^2 + stock*investment + stock^3 + investment^3 + stock^2*investment + stock*investment^2 + sizedum2 + sizedum3 + time
and
sales = stock + stock^2 + investment^2 + stock*investment + stock^3 + investment^3 + stock^2*investment + stock*investment^2 + sizedum2 + sizedum3 + time + material + employment
The investment variable enters the second stage only through the polynomial term.
Reference
Olley, G. S., and A. Pakes. 1996. The dynamics of productivity in the telecommunications equipment industry. Econometrica 64: 1263-1297.
|