Bayesian Estimation of Unknown Regression Error Heteroscedasticity

We propose a Bayesian procedure to estimate heteroscedastic variances of the regression error term ƒÖ, when the form of heteroscedasticity is unknown. The prior information on ƒÖ is elicited from the wellknown Eicker-White Heteroscedasticity Consistent Variance-Covariance Matrix Estimator. Markov Chain Monte Carlo algorithm is used to simulate posterior pdffs of the unknown elements of ƒÖ. In addition to the numerical examples, we present an empirical investigation of the stock prices of Japanese pharmaceutical and biomedical companies to demonstrate usefulness of the proposed method.