Journal of Mathematical Sciences: Advances and ApplicationsAuthor's: D. A. AGUNBIADE
Pages: [1] - [12]
Received Date: December 25, 2011
Submitted by:
This paper sought to examine both the intra and inter equation effects of multicollinearity on the exogeneous variables of a simultaneous equation model and also investigate the effect of the unsuspected dependence between random normal deviates used for generating the stochastic components of a simultaneous equation model. A Monte Carlo approach was used to achieve this by setting up a two-equation with five structural parameters simultaneous econometric model and applying the six different estimation techniques. It was found that while 2SLS and LIML produced identical results, the results of other techniques were at variants, which confirmed the likely presence of multicollinearity among the regressor variables.
Monte Carlo study, multicollinearity, simultaneous equation, instrumental variable estimators, exogeneous variables.
