Journal of
Economics and International Finance

  • Abbreviation: J. Econ. Int. Finance
  • Language: English
  • ISSN: 2006-9812
  • DOI: 10.5897/JEIF
  • Start Year: 2009
  • Published Articles: 363

Full Length Research Paper

On multicollinearity in nonlinear econometric models with mis-specified error terms in large samples.

Isaac D.  Essi1*, E. C. Chukuigwe2 and Nathaniel A. Ojekudo3
  1Centre for Econometric and Allied Research (CEAR), University of Ibadan, Ibadan, Nigeria. 2Department of Agricultural Economics and Extension, Rivers State University of Science and Technology, Port Harcourt, Nigeria. 3Department of Mathematics, Rivers State College of Education Port Harcourt, Nigeria.
Email: [email protected]

  •  Accepted: 03 December 2010
  •  Published: 28 February 2011

Abstract

 

The Cobb-Douglas model is a common occurrence in econometrics and other areas of research. Earlier results show that the consequence is more serious when a multiplicative error plagued data set is fitted with an additive error based model than vice-versa. In this study, involving large samples, we investigate the impact of multicollinearity in nonlinear econometric models with mis-specified error terms. .As it was in small samples, we observe that in large samples the above result and trend also hold in the presence of multicollinearity. It is also observed that the effect of multicollinearity is not purged by large sample size.

 

Key words: Cobb-Douglass model, mis-specified error terms, multicollinearity.