Two Stage Least Squares Regression (2SLS)
A two-step application of the instrumental variables (IV) technique to correct for the correlation of a suspected endogenous explanatory variable with the error term in the equation of interest. In the first step, the suspected endogenous variable is regressed on all the exogenous (predetermined) variables in the model. The values of the suspected explanatory variable predicted from the first step then are used as instruments for the suspected endogenous explanatory variable in the equation of interest. Inclusion of a predicted value of an explanatory variable in the equation of interest requires adjustment to the standard errors of the regression coefficients. Most statistical software packages do this adjustment as part of the 2SLS package, but if the steps are run one at a time, the standard errors in the second stage must be corrected manually.
Wooldridge, J.M. 2003. Introductory Econometrics, Mason, OH: Thomas South-Western.
Kennedy, P. 1998. A Guide to Econometrics, Cambridge, MA: The MIT Press.
|