Biased estimators and their applications in econometrics

Explanatory variables in a multiple linear regression model may be related with each other. The increasing degree of this relationship has undesired effects on classical estimators. This problem is called as multicollinearity and biased estimators can be used to overcome the problem. In this study, it?s aimed to investigate the multicollinearity and Econometric problems, as they arise together. Effects of multicollinearity for a model with heteroskedastic error terms or restricted parameters are investigated. Observation-varying restrictions are also considered in the study. Alternative biased estimators for these models are examined and their comparisons with traditional estimators are given. Comparisons are supported with Monte Carlo simulations. Application of these estimators on real data sets is investigated and estimators are compared on these data sets.