Modeling Non-Normality in Econometric Regression

Authors

Robert Turley, Brigham Young University
Dr. James B. McDonald, Brigham Young University

Keywords

econometric regression, hyperbolic transformation, estimating economic relationships

College

Family, Home, and Social Sciences

Department

Economics

Abstract

Traditional methods for estimating economic relationships assume the error between their predictions and their observations to be normally distributed. In other words, the random variable representing the error term is expected to follow the Gaussian distribution popularly known as the bell curve. In the case of a linear model, this is one of the justifications for the regression technique known as ordinary least squares (OLS). However, in many cases it appears the distribution of the error in economic models is peaked or skewed, suggesting that it may be inappropriate to assume normality.

Recommended Citation

Turley, Robert and McDonald, Dr. James B. (2014) "Modeling Non-Normality in Econometric Regression," Journal of Undergraduate Research: Vol. 2014: Iss. 1, Article 192.
Available at: https://scholarsarchive.byu.edu/jur/vol2014/iss1/192