Graduate Econometrics

Lecture 3: The Linear Regression Model cont.

Yonas Alem

Department of Economics
University of Gothenburg

November 17, 2014

1/4
Yonas Alem Graduate Econometrics Lecture 3: The Linear Regression Mode
The Simple Linear Regression Model
Example

Consider a statistical wage model

wagei = 1 + 2 male + i (1)

Where wagei denotes the hourly wage rate of individual i and

malei = 1 if i is male and 0 otherwise
If E{ i } = 0 and E{ i |malei } = 0
E{wagei |femalei } = 1
E{wagei |malei } = 1 + 2
2 is the expected wage differential between an arbitrary male
and female

2/4
Yonas Alem Graduate Econometrics Lecture 3: The Linear Regression Mode
The Simple Linear Regression Model
Example cont.

Using a subsample of the US national Longitudinal Survey

(NLS) 1987 data we estimate the above wage model as follows

Table: 2.1: OLS results wage equation

Variable Estimate SE
Constant 5.1469 0.0812
male 1.1661 0.1122
Notes: s=3.2174; R2 = 0.0317; F = 107.93

= the wage differential between males and females is about

$1.17 with a SE of $0.11.
Is this wage differential real or it happened by chance? To
answer this, we should undertake a hypothesis testing that
H0 : 2 = 0
3/4
Yonas Alem Graduate Econometrics Lecture 3: The Linear Regression Mode
The OLS Model
Hypothesis Testing

The Gauss-Markov Assumptions again

[A1] : E{ i } = 0, i = 1, ..., N
[A2] : E{xi i } = 0
[A3] : V { i } = 2 , i = 1, ..., N
[A4] : Cov{ i , j } = 0, i 6= j
For exact statistical inference from a given sample, of N
observations one needs the additional distributional
assumption (normality), i.e.,

i NID(0, 2 ) [A5] (2)

[A1] [A5] = b N ( , 2 (X0 X)1 ) and bk N ( k , 2 ckk )

Where ckk is the (k, k) element in (X0 X)1
4/4
Yonas Alem Graduate Econometrics Lecture 3: The Linear Regression Mode