Applied Econometrics 4rd edition
Dimitrios Asteriou
and
Stephen G Hall
Applied Econometrics 4rd edition
SIMPLE REGRESSION
1. Introduction to the Classical Linear Regression
Model
2. The OLS Method of Estimation
3. The Overall Goodness of Fit
4. Hypothesis Testing
5. How to Estimate a Simple Regression in
Eviews
6. Applications and Examples
Applied Econometrics 4rd edition
Learning Objectives
• Compute the equation of a simple regression line from a
sample of data, and interpret the slope and intercept of
the equation.
• A full understanding of the simple OLS method of
estimation and discussion of the properties of estimated
coefficients.
• Computation of a standard error of the estimate and
interpretation of its meaning and its use in Hypothesis
Testing.
• Understanding and interpretation of the R2
Applied Econometrics 4rd edition
Introduction
• Regression analysis is the process of constructing a
mathematical model or function that can be used to
predict or determine one variable by another variable.
• Key issue here is direction of causation of the two
variables, or which variable depends on the other.
• Therefore we have two cases of variables
dependent variables (usually denoted by Y)
independent or explanatory (usually denoted by
X)
Applied Econometrics 4rd edition
Regression Models
• Deterministic Regression Model: Y=b0+b1X
• Probabilistic Regression Model: Y=β0+β1X+u
• b0 and b1 are population parameters
• β0 and β1 are estimated by sample statistics
b0 and b1
Applied Econometrics 4rd edition
Regression Models
Probabilistic Regression Model: Yt=a+βXt+u
• Why probabilistic?
E(Yt)=a+βXt
Yt=E(Yt) +ut
Yt=a+βXt+u
Applied Econometrics 4rd edition
Regression Models
Why probabilistic?
1. Omission of explanatory variables
2. Aggregation of variables
3. Model specification
4. Functional misspecification
5. Measurement errors
Applied Econometrics 4rd edition
The Scatter Plot
X
300
250
200
X
150
100
50
0
60 80 100 120 140 160 180
Applied Econometrics 4rd edition
Four Ways of Fitting a Line in the Data
• By eye
• Connecting the first with the last observation
• Take the average of the first two and the
average of the two last and connect
• Apply Ordinary Least Squares
Applied Econometrics 4rd edition
The Scatter Plot
Applied Econometrics 4rd edition
Equation of the Regression Line
Yˆ b0 b1 X
where : b0 = the sample intercept
b1 = the sample slope
Yˆ = the predicted value of Y
Applied Econometrics 4rd edition
Slope and Intercept of the
Regression Line
X Y
X X Y Y XY nXY XY
n
b
X X X n X
2 2 2
X
1 2
X 2
n
Y X
b Y b X n b n
0 1 1
Applied Econometrics 4rd edition
Least Squares Analysis
SSXY X X Y Y XY
X Y
n
2
SSXX X X
2
X
2
X
n
SSXY
b1 SSXX
Y X
b Y b X n b n
0 1 1
Applied Econometrics 4rd edition
Least Squares Analysis
Why OLS?
1. Elimination of the effect of the sign (positive and
negative residuals offset each other)
2. More weight is given to larger residuals and so we
work harder to reduce the very large errors
3. It chooses a method that follow some numerical and
statistical properties (remember unbiasedness,
efficiency etc)
Applied Econometrics 4rd edition
Assumptions of the Least Squares Analysis
Applied Econometrics 4rd edition
Properties of the OLS estimators
Linearity
Unbiasedness
Efficiency
BLUEness
Consistency
Applied Econometrics 4rd edition
Properties of the OLS estimators
• BLUE - What does the acronym stand for?
• “Estimator” - is an estimator of the true value of .
• “Linear” - is a linear estimator
• “Unbiased” - On average, the actual value of the α’s and β’s will be
equal to the true values.
• “Best” - means that the OLS estimator has minimum variance among
the class of linear unbiased estimators. The Gauss-Markov
theorem proves that the OLS estimator is best.
Applied Econometrics 4rd edition
Example: The Keynesian
Consumption Function
Applied Econometrics 4rd edition
Example: The Keynesian
Consumption Function
C2=B2*A2
D2=B2*B2
A22=SUM(A2:A21)
B22=SUM(B2:B21)
and so on!
Excel
Applied Calculations
Econometrics 4rd edition
b0=(C22-(A22*B22)/20)/(D22-((B22ˆ2)/20))=0.601888903
b1=AVERAGE(A2:A21)-G2*AVERAGE(B2:B21)=15.116408
Applied Econometrics 4rd edition
Excel Calculations (the easy way!)
• Step 1: go to the menu Tools/Data Analysis and choose the command
regression.
• Step 2: We are then asked to specify the Input Range, Output Range, and
a choice of including or not labels in the first row.
• Step 3: The Input Range is the columns that contain the data for Y and X
(i.e. we enter ‘$A$1:$B$21’ or simply select this area using the mouse).
• Step 4: The Output Range can be either a different sheet (not
recommended) or any empty cell in the current sheet (i.e. we might specify
cell F5).
• Step 5: Since we have chosen the labels in our selection we tick the box.
• Step 6: By clicking <OK> we obtain the display shown in Table 4.4.
Applied Econometrics 4rd edition
Excel Results
Applied Econometrics: A Modern Approach 22
using Eviews and Microfit © Dr D Asteriou
Applied Econometrics 4rd edition
The Regression Line
300
X
250
200
X
150
Linear (X)
100
50
0
60 80 100 120 140 160 180
Applied Econometrics 4rd edition
Regression in Eviews (1) (1)
Step 1 Open EViews.
Step 2 Choose File/New/Workfile in order to create a new file.
Step 3 Choose Undated or Irregular and specify the number of
observations (in this case 20). A new window appears which
automatically contains a constant (c) and a residual (resid) series.
Applied Econometrics 4rd edition
Regression in EViews (2)
Step 4 In the command line type:
genr x=0 (press enter)
genr y=0 (press enter)
which creates two new series named x and y that contain zeros
for every observation.
Open x and y as a group by selecting them and double clicking
with the mouse.
Step 5 Either type the data in EViews or copy/paste the data
from Excel®. To edit the series press the edit +/− button. After
finishing with editing the series press the edit +/− button again
to lock or secure the data.
Applied Econometrics 4rd edition
Regression in EViews (3)
Step 6 Once the data have been entered into Eviews, the
regression line (to obtain alpha and beta) may be estimated either
by typing:
ls y c x (press ‘enter’)
on the command line, or by clicking on Quick/Estimate
Equation and then writing your equation (i.e. y c x) in hte new
window.
Applied Econometrics 4rd edition
Reading the Eviews Regression Output
Applied Econometrics 4rd edition
The Coefficient of Determination
The proportion of variability of the dependent
variable accounted for or explained by the
independent variable in a regression model.
It is called R2 and it takes values from 0-1.
TSS=ESS+RSS
R2 = ESS/TSS or R2 = 1 – (RSS/TSS)
28
Applied Econometrics 4rd edition
The Coefficient of Determination
Problems associated with R2
1. Spurious regression problem
2. High correlation of Xt with another varible Zt
3. Correlation does not necessarily imply causality
4. Time Series vs Cross Sectional equations
5. Low R2 does not mean wrong choice of Xt
6. R2s from equation with different forms of Yt are
not comparable
29
Applied Econometrics 4rd edition
Hypothesis Tests for the Slope
of the Regression Model
H 0: 1 0 b 1
t 1
S
H 1: 1 0
b
S
where: S e
H 0: 1 0
b
SSXX
SSE
S
H 1: 1 0 n2
e
SSXX
2 X
2
H 0: 1 0 X
n
the hypothesized slope
H 1: 1 0 1
df n 2
Applied Econometrics 4rd edition
Hypothesis Tests for the Slope
of the Regression Model
Applied Econometrics 4rd edition
Hypothesis Tests for the Slope
of the Regression Model
We don’t like statistical Tables? – good!!!
We don’t need statistical Tables
A rule of thumb | t | > 2
The p-value approach (<0.05)