Chapter 5

The Multiple Regression Model

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 1
Edition
 Chapter Contents

 5.1 Introduction
 5.2 Estimating the Parameters of the Multiple
Regression Model
 5.3 Sampling Properties of the Least Squares
Estimators 5.1, 5.5 and 5.8 are
 5.4 Interval Estimation important; new concepts
are introduced.
 5.5 Hypothesis Testing 5.2 – 5.4 are repeat of first
 5.8 Measuring Goodness-of-fit two lectures; technical
details are not needed for
assessment

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 2
Edition
 5.1
Introduction

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 3
Edition
 5.1
Introduction

An economic model in which sales revenue is

related to price and advertising expenditure
Eq. 5.1
(Andy.wf1):
SALES 1   2 PRICE   3 ADVERT

The corresponding econometric model is referred

Eq. 5.2
to as a multiple regression model:

SALES E SALES   e  β1  β 2 PRICE  β3 ADVERT  e

Most of the results for SLR can be extended to

MR
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 4
Edition
 5.1
Introduction

5.1.1
The Economic
Model
β2 is the change in monthly sales SALES ($1000) when the
price index PRICE is increased by one unit ($1), and
advertising expenditure ADVERT is held constant
SALES SALES
2  
PRICE  ADVERT held constant  PRICE

Similarly, β3 is the change in monthly sales SALES ($1000)

when the advertising expenditure is increased by one unit
($1000), and the price index PRICE is held constant
SALES SALES
3  
ADVERT PRICE held constant  ADVERT
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 5
Edition
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 6
Edition
 5.1
Introduction ASSUMPTIONS of the Multiple Regression Model

5.1.2b
The

MR1. yi 1 2 xi 2    K xiK  ei , i 1, , N

Assumptions of
the Model

MR2. E ( yi ) 1 2 xi 2   K xiK  E (ei ) 0

MR3. var( yi ) var(ei ) 2
MR4. cov( yi , y j ) cov(ei , e j ) 0

MR5. The values of each xtk are not random and are not
exact linear functions of the other explanatory variables
– otherwise, we would have collinearity problem

MR6.
yi ~ N  (1 2 xi 2    K xiK ),  2   ei ~ N (0,  2 )
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 7
Edition
 5.2
Estimating the Parameters of the
Multiple Regression Model

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 8
Edition
 Table 5.1 Observations on Monthly Sales, Price, and Advertising in Big
Andy’s Burger Barn

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 9
Edition
 5.2
Estimating the
Parameters of
the Multiple
Regression
Model
5.2.1
Least Squares The least squares estimators b1, b2, and b3 are obtained by
Estimation
Procedure
minimizing the sum of squares function S(β1, β2, β3), which
is given below: N
S β1 ,β 2 ,β3    yi  E  yi 
2

Eq. 5.5 i 1
N
  yi  β1  β 2 xi 2  β3 xi 3 
2

i 1

Estimates along with their standard errors and the

equation’s R2 are typically reported in equation format as:
Eq. 5.6 
SALES 118.91  7.908 PRICE  1863 ADVERT R 2 0.448
( se) (6.35) (1.096) (0.683)

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 10
Edition
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 11
Edition
 5.2
Estimating the
Parameters of
the Multiple
Regression
Model
5.2.2
Least Squares Interpretations of the results:
Estimates Using
Hamburger
Chain Data
1. The negative coefficient on PRICE suggests that
demand is price elastic; we estimate that, with
advertising held constant, an increase in price of $1
will lead to a fall in monthly revenue of $7,908
2. The coefficient on advertising is positive; we estimate
that with price held constant, an increase in
advertising expenditure of $1,000 will lead to an
increase in sales revenue of $1,863
Using the model to predict sales if price is $5.50 and
advertising expenditure is $1,200:
SALES  118.91 - 7.908 PRICE  1.863 ADVERT
118.914 - 7.9079 5.5  1.8626 1.2
77.656
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 12
Edition
 5.2
Estimating the
Parameters of
the Multiple
Regression
Model
5.2.2
Least Squares
Estimates Using
A word of caution:
Hamburger
Chain Data – Estimated regression models describe the
relationship between the economic variables for
values similar to those found in the sample data
– Extrapolating the results to extreme values is
generally not a good idea
• E.g. the model predicts SALES = β1 =118.9 even if
PRICE=ADVERT=0
– Predicting the value of the dependent variable
for values of the explanatory variables far from
the sample values invites disaster
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 13
Edition
 5.3
Sampling Properties of the Least
Squares Estimators

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 14
Edition
 5.3
Sampling
Properties of the
Least Squares
Estimators

For the multiple regression model, if assumptions

MR1–MR5 hold, then the least squares estimators
are the best linear unbiased estimators (BLUE) of the
parameters.

If the errors are not normally distributed, then the

least squares estimators are approximately
normally distributed in large samples
– Frequently, N – K = 50 will be large enough

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 15
Edition
 5.4
Interval Estimation

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 16
Edition
 5.4
Interval
Estimation
Interval Estimation for a Single Coefficient
5.4.1
Interval
Estimation for a
For the hamburger example, we need:
Single
Coefficient
Eq. 5.15 P ( tc  t(72)  tc ) .95

Using tc = 1.993, we can rewrite (5.15) as:

 b2  2 
P   1.993  1.993  .95
 se(b2 ) 

From which, the 95% confidence interval is

Eq. 5.16  b2  1.993 se(b2 ), b2  1.993 se(b2 ) 

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 17
Edition
 5.4
Interval
Estimation
Interval Estimation for a Single Coefficient
5.4.1
Interval For price, we have b2 = -7.908 and se(b2) = 1.096, so that:
Estimation for a
Single
Coefficient  7.9079  1.993 1.096,  7.9079  1.993 1.096   10.093,  5.723

– A decrease in price by $1 will lead to an increase in

revenue somewhere between $5,723 and $10,093.

Similarly for advertising, we get:

1.8626  1.9935 0.6832, 1.8626  1.9935 0.6832  0.501,3.225

– An increase in advertising expenditure of $1,000 leads to

an increase in sales revenue of between $501 and $3,225
• Extra advertising expenditure could be unprofitable if
the revenue increase is less than $1,000
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 18
Edition
 5.5
Hypothesis Testing

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 19
Edition
 5.5
Hypothesis Answer economic questions using testable hypotheses
Testing

5.5.2a
Testing for
Elastic Demand

Two economic questions

1. Is demand price-elastic or price-inelastic?

2. Would additional sales revenue from

additional advertising expenditure cover the
costs of the advertising?

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 20
Edition
 5.5
Hypothesis Is demand price elastic?
Testing

5.5.2a
Testing for
Elastic Demand
To see if demand is price-elastic, set price-
inelastic as null:
1. The null and alternative hypotheses are:
H 0 : 2 0 (demand is unit-elastic or inelastic)

H1 : 2 < 0 (demand is elastic)

2. The test statistic, if the null hypothesis is true,

is: t b2 se(b2 ) ~ tN  K 

3. At a 5% significance
t.05,72 level, we reject H0 if t
≤ -1.666 (= ) or if the p-value ≤ 0.05
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 21
Edition
 5.5
Hypothesis Is demand price elastic?
Testing

5.5.2a
Testing for
Elastic Demand
Hypothesis test (Continued) :
4. The test statistic is:
b2  7.908
t   7.215
se b2  1.096

and the p-value is:


P t72   7.215 0.000 
5. Since -7.215 < -1.666, we reject H0: β2 ≥ 0
and conclude that H0: β2 < 0 (demand is
elastic)
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 22
Edition
 5.5
Hypothesis
Testing
A marketing question

5.5.3
Hypothesis
Testing for a
Linear
Combination of
Coefficients
The marketing adviser claims that dropping the
price by 20 cents will be more effective for
increasing sales revenue than increasing
advertising expenditure by $500
– In other words, she claims that -0.2β2 > 0.5β3,
or -0.2β2 - 0.5β3> 0
– We want to test a hypothesis about the linear
combination -0.2β2 - 0.5β3

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 23
Edition
 5.5
Hypothesis
Testing
A marketing question

5.5.3
Hypothesis
Testing for a
As before:
Linear
Combination of
Coefficients
1. The null and alternative hypotheses are:
H 0 :  0.2β 2  0.5β3 0 (marketer's claim is not correct)

H1 :  0.2β 2  0.5β3  0 (marketer's claim is correct)

2. The test statistic, if the null hypothesis is true,

is:  0.2b2  0.5b3
t ~ t72 
se(  0.2b2  0.5b3 )
3. At a 5% significance level, we reject H0
if t ≥ 1.666 or if the p-value ≤ 0.05
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 24
Edition
 5.5
Hypothesis
Testing
A marketing question

5.5.3
Hypothesis We need the standard error:
Testing for a
Linear
Combination of

se(−0.2𝑏 2 −0.5𝑏
Coefficients

4. test statistic is:

 0.2b2  0.5b3 1.58158  0.9319
t  1.622
se  0.2b2  0.5b3  0.4010
5. Since 1.622 < 1.666, we do not reject H0

 
Consistent with P t72   1.622 0.055, a p value > 0.05

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 25
Edition
 5.8
Measuring Goodness-of-fit

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 26
Edition
 5.8
Measuring
Goodness-of-fit

The coefficient of determination is the proportion of

variation in y explained by x’s
 i 1  yˆi  y 
N 2
2SSR where
R  
SST 
N
 yi  y 
2
yˆi b1  b 2 xi 2    bK xiK
i 1

SSE
1  implies
SST 0 ≤ R2 ≤ 1

N 2
ˆ
e
Eq. 5.31 1  i 1 i

 i 1  yi  y
N 2

the hamburger example:

 i 1 i
N
2
2
ˆ
e 1718.943
R 1  1  0.448
 i 1  yi  y 
N 2
3115.482
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 27
Edition
 5.8
Measuring
Goodness-of-fit

Interpretation
– 44.8% of the variation in sales revenue is
explained by the variation in price and by the
variation in the level of advertising expenditure
– 55.2% of the variation in revenue is left
unexplained and is due to variation in the error
term

Principles of Econometrics, 4th

Chapter 5: The Multiple Regression Model Page 28
Edition
 Summary
In multiple regression, 2 or more x variables are used
to explain y
βk (k ≠ 1) can be interpreted as the rate of ∆y in
response to ∆xk when the other x variables are held
constant
Most properties and results of SLR are applicable to
MR
If we hope to find A not B, then set B as null and A as
alternative.
Interesting economic hypothesis can be tested by
linear combination of β’s
R2 gives a measure of how much y is explained by x’s
Principles of Econometrics, 4th
Chapter 5: The Multiple Regression Model Page 29
Edition