Estimation of Heteroscedasticity Effects in a Classical Linear Regression Model of a Cross-Sectional Data

This paper investigates the effects of heteroscedasticity in the Classical Linear Regression Model (CLRM) of auditor's remuneration. Several efforts of building a realistic econometric model for Auditor's Remuneration with regards to core banking activities have been undertaken. The work involves the use of White heteroscedasticity and Newey-West test techniques to examine the presence of heteroscedasticity, which shows that heteroscedasticity is an inherent feature of cross-sectional data. The superiority of Weighted Least Squares (WLS) on Ordinary Least Squares (OLS) was put to test in estimating the parameters of Auditor’s Remuneration model designed as: hspace*{10mm} $ AR_i = heta_0 + heta_1 T A_i + heta_2 T E_i + heta_3 C D_i + heta_4 P B T_i + varepsilon$ hspace*{6mm}And it was established that OLS is not appropriate for estimation if heteroscedasticity is present in research data, and that the model fitted using WLS is the most appropriate that is deemed fit for proper review of auditor's remuneration in banking industry.