This chapter discusses how properties of OLS estimators change as the sample size increases. It states that OLS estimators are consistent, meaning they converge in probability to the true population parameters, as long as the explanatory variables are uncorrelated with the error term (MLR.4'). It also describes how OLS estimators are asymptotically normally distributed without requiring the errors to be normally distributed (MLR.6), allowing t-tests, confidence intervals, and F-tests to be valid for large samples. Examples are provided to illustrate consistency and asymptotic normality.

1. Chapter 5
Multiple Regression
Analysis: OLS Asymptotics
Wooldridge: Introductory Econometrics:
A Modern Approach

2. So far we have focused on properties of OLS that hold for any sample
Properties of OLS that hold for any sample/sample size
Expected values/unbiasedness under MLR.1 – MLR.4
Variance formulas under MLR.1 – MLR.5
Gauss-Markov Theorem under MLR.1 – MLR.5
Exact sampling distributions/tests under MLR.1 – MLR.6
Properties of OLS that hold in large samples
Consistency under MLR.1 – MLR.4
Asymptotic normality/tests under MLR.1 – MLR.5
Without assuming nor-
mality of the error term!
Multiple Regression
Analysis: OLS Asymptotics

3. Consistency
Interpretation:
Consistency means that the probability that the estimate is arbitrari-
ly close to the true population value can be made arbitrarily high by
increasing the sample size
In words, intuitively: the estimate will be close to the true population
value if the sample size is big
Consistency is a minimum requirement for sensible estimators
An estimator is consistent for a population parameter if
for arbitrary and .
Alternative notation: The estimate converges in proba-
bility to the true population value
Multiple Regression
Analysis: OLS Asymptotics

4. Theorem 5.1 (Consistency of OLS)
Ex. simple regression model
Assumption MLR.4’ The slope estimate is consistent if the explanatory
variable is exogenous, i.e., uncorrelated with the
error term.
The explanatory variables are uncorrelated with the error
term. This assumption is weaker than the zero conditional
mean assumption MLR.4.
Multiple Regression
Analysis: OLS Asymptotics

5. OLS is consistent under the weaker MLR.4’
Asymptotic analog of omitted variable bias
True model
Misspecified
model
There is no omitted variable bias if the omitted variable is
irrelevant or uncorrelated with the included variable
Bias
Multiple Regression
Analysis: OLS Asymptotics

6. Asymptotic normality and large sample inference
In practice, the normality assumption MLR.6 is often questionable
If MLR.6 does not hold, and the sample size is small, OLS estimators
are not necessarily normally distributed and t- or F-tests may not be
valid
Fortunately, OLS estimators are approximately normally distributed
and F- and t-tests still work if the sample size is large enough
Multiple Regression
Analysis: OLS Asymptotics

7. Theorem 5.2 (Asymptotic normality of OLS)
Under assumptions MLR.1 – MLR.5, when the sample is large, the
standardized estimators are approximately normally distributed:
Similarly:
Multiple Regression
Analysis: OLS Asymptotics

8. Converges to n ∙
Converges to
Practical consequences
In large samples, the t-distribution is close to the N(0,1) distribution
As a consequence, t-tests are valid in large samples without MLR.6
The same is true for confidence intervals and F-tests
Why large samples are better
Asymptotic analysis of the OLS sampling errors
Converges to a fixed
number
Multiple Regression
Analysis: OLS Asymptotics

9. Example: Standard errors in a birth weight equation
Uses only half of the sample
Multiple Regression
Analysis: OLS Asymptotics