AN OVERVIEW ON COMPARATIVE METHODOLOGY OF CLASSICAL OLS AND TWO-STAGE TECHNIQUES IN REGRESSION ANALYSIS MODEL

Abstract This article provides a methodological overview for the classical Ordinary Least Square (OLS) and Instrumental variable (or Two-stages) technique by discussing the conditions that satisfy the use of the method. The Ordinary Least Squares (OLS) estimator in the classical linear regression model is considered the Best Linear Unbiased Estimator (BLUE) if the model assumptions hold. When these assumptions are met, OLS provides linear, unbiased estimates with the smallest variance compared to other estimators. Although alternative methods exist, especially for specific issues like multicollinearity or heteroscedasticity, OLS remains the most efficient estimator, particularly in large samples. Even when assumptions are slightly violated, OLS generally performs well, maintaining its reliability due to the central limit theorem.