|
Activity Number:
|
514
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Business and Economics Statistics Section
|
| Abstract - #307472 |
|
Title:
|
Effectiveness of Two-Stage Least Squares in Correcting Endogeneity Bias: a Monte Carlo Study
|
|
Author(s):
|
V. A. R. Samaranayake*+ and Xujun Wang
|
|
Companies:
|
University of Missouri-Rolla and University of Missouri-Rolla
|
|
Address:
|
202 Rolla Building, Rolla, MO, 65401,
|
|
Keywords:
|
endogeneity ; simultaneity bias ; instrumental variables ; ordinary least squares ; structural equations
|
|
Abstract:
|
A common solution to the problem of endogeneity bias in regression is the use of instrumental variables to predict the endogenous variable and the using the predicted values in place of the original variable values in the regression model. It has been pointed out that this two-stage least squares (2LS) method may not work satisfactorily when the correlation between the endogenous variable and its predicted value is weak and one or more of the instrumental variables employed is correlated, even weakly, with the error term associated with the dependent variable. We conduct a Monte Carlo study to investigate the bias of the 2SLS estimators under various conditions. The results show that the 2SLS approach does not eliminate or reduce the bias substantially in many situations.
|
