Econometrics For Dummies

Chapter 7

The Normality Assumption and Inference with OLS

In This Chapter

Understanding what the normal distribution implies

Deriving hypothesis testing procedures for regression coefficients

Determining whether regression results are statistically significant

Using the normal distribution to determine forecast/prediction error

When you use ordinary least squares (OLS) regression for hypothesis testing and/or prediction and forecasting, you always assume that the distribution of the unobserved error is normal. However, the idea of assuming a normal distribution is often misunderstood. That’s what this chapter clears up. It helps you understand precisely how a normal distribution is used in econometrics and the importance of the normality assumption for tests of statistical significance and calculations of forecast error. You also get to check out some example scenarios in which the assumption is likely to be reasonable and others for which it’s likely to fail.

Note: In this chapter, I assume that you already have a basic understanding of regression mechanics and are familiar with ...