Does Simultaneity Matter?

A simultaneous equations model is developed which endogenizes economic growth, income inequality, corruption, and political instability while using initial GDP, average schooling, investment share of GDP, population growth, educational inequality, trade share of GDP, and ethnolinguistic fractionalization as exogenous variables. Twenty-two different specifications of the model are then estimated using ordinary least squares (OLS), two-stage least squares (2SLS), and three-stage least squares (3SLS) estimation methods. Both the structural and reduced form estimates exhibit changing results for several variables as we switch from OLS to simultaneous methods. Therefore, it is inappropriate to use only OLS estimators when there may be feedback effects between variables. The OLS results may exhibit simultaneous bias and lead to incorrect policy recommendations if they are used without being compared to the results of simultaneous estimation methods. However, simultaneous estimators may also experience certain biases. Therefore, both OLS and simultaneous estimators must be used by researchers in order to determine the best course of policy.

Autorentext

Carol Tallarico completed her Ph.D. in Economics at the University of Illinois at Chicago in 2004. She is currently an Associate Professor of Economics and Exelon Research Fellow in the Brennan School of Business at Dominican University in River Forest, IL USA. Her published research has been focused in the area of environmental economics.