Multiple regression Example

Example 1 - Suppose the quantity supplied of commodity is assumed to be a linear function of the price of
the commodity itself & the wage rate of labor used in the production of the commodity- If the supply
equation is given by
Q 1=α + β 1 P X 1 + β 2 W +u
Where Q1 = quantity supplied. Px, price of commodity X & W is wage rate
Using the following sample data
a) Estimate the parameters using OLS
^ β^ ¿ at 5% significant level
b) Test the statistical significance of the individual coefficients (α^ , β∧ 2
c) Test the overall significance of the coefficients (F-test)
d) Compute the price elasticity of supply
Y Q1 20 35 30 47 60 68 76 90 100 105 130 140 125 120 135
X Px 10 15 21 26 40 37 42 33 30 38 60 65 50 35 42
Z W 12 10 9 8 5 7 4 5 7 5 3 4 3 1 2
From these we will have the following values.
Let Qs =Y, Px = X & W=Z
Ȳ = 85.4 x̄ =36.27 ∑ =7207.4
yx ∑ Zy =-1553 ∑ xz =-514.667 ∑ y i2=23211.6 ∑ x i2
=3192.93 ∑ Z 2=135.3 ∑ e i2=4386.49 N=15
Answer
a) Estimate the parameter using OLS

^β = ∑ x1 yi ∑ z −∑ xz ∑ zy
2

1 2
∑ x i2 ∑ z 2−(∑ xz )
( 7207.4 ) ( 135−3 ) — 514−667
(−1553 ) (3192.93)(135.33)−¿
❑
173668
¿ = 1.16
162,712
^β = ∑ zy ∑ x −∑ Xz ∑ xy
2

2 2
∑ x i2 ∑ Z 2−(∑ xz )
(−1553 ) (3192 ) — 514
¿ (7207 ) 162,712
❑
(−495,7176 ) — 3704398 −1,252778
¿ 162712= 162,712 ≈−7.69
❑
α^ =Ȳ − β1 x̄− ^β 2 Z̄
^
α^ =85.4- (1.06x36.2) -(-7.699x5.6)
α^ =85.4-38.372-(-42.946) =89.974
Y^ =α^ + ^β 1 x− β^ 2 Z̄
Y^ = 89.974 + 1.06X - 7.69Z. This equation will be read as follow
 α^ = 89.974 means if the price of the commodity & the wage rate is zero the supplier will
supply 89.974 units of goods. But it is meaningless to interpret the constant term ( α^ ). (In
some analysis it doesn’t give sense.)
 ^β is the coefficient of price of the commodity. The value 1.06 signifies that if the price of the
1
commodity is increasing by one birr given the price of wage is constant quantity supplied will
increase on the average by 1.06 units.
 ^β is the coefficient of the wage rate. If the wage rate is increased by 1 birr keeping constant
2
the price of the commodity quantity supplied will decrease on the average by 7.69.
 ^β & ^β are coefficients of the explanatory variables & they are containing the marginal
1 2
values. If you take the first derivative of the equation, you will have marginal values.
Ex. Y^ i=¿ 89.974+16px-7.697Z

1
 ∂ yi 1.16∧2 Y^ i
= =−7.697
∂ px 2Z
^ β^ ¿.
b) Test the statistical significance of (α^ , β∧ 2
Standard error test
^ β^ ¿.The variance of
To test this, we have to have variance & the standard error of the parameters ( α^ , β∧ 2
^ β^ will have a value of σ 2 (see equation number 4.16 – 4.18).
α^ , β∧ 2

[ ]
2 2 2 2

1 x̄ ∑ x + x̄ ∑ x −2 x̄ x̄ ∑ x x
1 2 2 1 1 2 1 2
( α^ )=¿ + 2 2 σu
2
∑ x ∑ x −( ∑ x x )
n 2
Var ( 1 2 1 2
. . .. . . . . . .. 4.16

[ ∑ x2
]
2

( ^β1 )=¿ σu ∑ ∑
2

x x −( ∑ x x
2 2
2
2)
Var ( 1 2 1
. . . . . . . . . . . . . . . . . . . . . 4.17

∑e
[ ]
2
∑ x1
2

^β ¿ σu 2 σu
2 1
2¿ 2 2

Var = x 1 x 2−( x 1 x 2 ) . . . . . . . .. . . . . . . . . . .. . . . 4.18. Where = n−k 1 , n is

∑ ∑ ∑ 2

number of samples, k is number of parameters which are estimated.

Where σ u2 = ∑
e i2
from our example n=sample size of 15
n−k
K is number of estimated parameters which is 3 (α^ , β∧ ^ β^ ¿
2

σ u2 =
∑ e i = 4386.49 =¿ 365.5
2

N−k 15.3
Using equation number 4.16 we can calculate Var( α^ )
1
Var( α^ ) = +¿
15
¿¿
Var( α^ )=¿ = 1,110.54
Var( ^β1 ) again we can calculate using equation number 4.18

Var( ^β1 ) = 365.5

135.33
162,712
= 0.303 [ ]
Var( ^β1 ) = the Var of ^β 1can be calculated using equation number 4.19

Var( β^ 2 ) = 365.5
3192.933
162,712
= 7.17 [ ]
From the above values we can calculate S.E ( α^ ), S.E ( ^β1 ) & S.E ( ^β2 ) as follows
S.E ( α^ ) = √ Var α^ =√ 1110.54 ≈ 31.748
√
S.E ( ^β1 ) = Var ( β^ 1) =√ 0.303 ≈0.521
√
S.E ( β^ 2 ) = Var ( β^ 2 )=√ 7.17 ≈ 2.529
Having calculated S.E of the coefficients of the variables ( α^ , β∧ ^ β^ ¿ we can undertake S.E. tests – as
2
follows
α^
- If S.E(α^ )> we can accept the null hypothesis & reject the alternative
2
31.76∧^α 89.974
S . E ( α^ )= = =¿ 44.987
2 2
α^
Then S.E( α^ ) =i.e 31.76 is less than which is 44.987. Therefore, we can conclude that α^ is statistically
2
significant
β^ 1 1.16
( ^β1 )
S.E = 0.521 2 2 =0.58
2
 ^β
From this we can see that S.E ( ^β1 ) is less than 2 then we can conclude that ( ^β2 ) is significant. Lastly S.E
2
^β 7.697 ^β
( ^β2 ) = 0.529 & 2 = 2 = 3.848 Again here S.E ( ^β2 ) is less than 2 . All the estimators are statistically
2 2
significant or we reject the null hypothesis that H 0=α^ =0 , H0= ^β i=0 (we reject the hypothesis which
^ β^ ¿ are equal to zero) & accept the alternative that H1= α^ ≠ 0 , ^β ≠ 0 H1= ^β i ≠ 0 (α^ , β∧
says (α^ , β∧ ^ β^ ¿ are
2 i 2
different from zero.)
The economic interpretation of rejecting the null hypothesis and accepting the alternative states the
following
i) The estimators are statistically significant
ii) The explanatory variables X1 & Z (price of commodity X & wage rate) influence the supply of
commodity (Y)
Student –t –test
In the t-test analysis we compare the calculated t with table value of t. How to get calculated t-value
α^ 89.974
t= Computed value of t = =2.23
S . E ( α^ 1 ) 31.76
^β
t= 1 1.16
Computed t = =2.83
^
S . E ( β1 ) 0.521
^β
t= 2 −7.697
Computed t = =3.043
S . E ( ^β2 ) 2.529
How to get t-value from t- table
Given 5% significance level, the total sample used to calculate the estimators are 15 & the number of
estimators is 3( α^ , ^β∧ ^β 2 ). If we take two tail test
α
t , ( n−k ) α =¿ Significance level
2
α
Show two tail test
2
n = sample size
k = estimators
n-k = degree of freedom
0.05
t , ( 15−3 ) = t0.25,12
2
From the t- table in the top of the raw find 0.025 & in the first column of the table find 12 then when these
two values are intersecting with each other, that point will give you the table value of t. From our t 0.025,12
the table value is 2.179. Compare this table value with the computed value & if the computed value is
greater than the table value, we reject the null hypothesis & accept the alternative. Again, if the computed
value is less than the table value, we accept the null hypothesis & reject the null hypothesis.
Compare computed with table value
 Compared t value of α^ is 2.83 & the table value is 2.179 here the computed t value is greater than the
table value for α^
 Again, the computed t value for ^β 1∧ ^β 2 is 2.23 & 3.043 respectively they are greater than the table
value.
Since in all this cases α^ , β∧ ^ β^ computed t-value is greater than the t-table value. We will have the following
2
interpretation
i. α^ , β∧ ^ β^ are statistically significant
2
ii. The quantity supplied is influenced by the price of the commodity & wage rate
Coefficient of determination /R2/

β^ ∑ yi x + β^ ∑ yzi
1 1 2
( 1.16 ) ( 7207.4 ) + (−7.47 )(−1553 )
23211.6
R2= ∑ yi2 =
3
 ( 8,360.58+11,600.91 ) 19961.49
R2= = = 0.8599
23,211.6 23,211.6
This means 85.99% of quantity supplied is explained by price of the commodity & wage rate.
Adjusted R̄2
2 2 n−1 15−1
R̄ =1-(1- R ) = 1-(1-0.8599) = 0.8365
n−k 15−3
F - test
The overall significance of the explanatory variables can be tested using F-test. Just like t- test in the case of
F test we will have computed & table value of F. Calculated value of F * can be obtained using the following
formula
2
R 0.8599 0.8599
k −1 3−1 2
F* = 2
= =
1−R 0.1401 0.1401
N −k 15−3 12
0.42995
¿ =36.826
0.11675
Table value of F can be obtained by taking the enumerator degree of freedom (k-1) & the denominator
degree of freedom (n-k). Then F (k-1), (n-k). Given the level of significance of 5% we can get F 2, 12 i.e.
k-1 = 3-1 = 2 & n-k = 15-3 = 12. Then from the table we find the enumerator degree of freedom (k-1) in
the top row of the table & the denominator degree of freedom (n-k) in the first column of the table at the
intersection point of these values you will get the table value. From our example F 2, 12 at 5% level is 3.89.
Comparison of calculated F & table value of F. If the calculated value of F is greater than the table value
then we can reject the null hypothesis & accept the alternative. From our example the calculated F value is
36.826 is greater than the table value 3.89. The economic interpretation of this is
i) All the estimators are significant or statistically different from zero
ii) Quantity supplied is affected by the price of the commodity & wage rate
Presentation of regression results: - Different books used different presentation methods but the most
commonly is the one which we write under here using our previous example.
Ȳ =¿89.974 + 1.16PX - 7.697Z
S.E (31.748) (0.521) (2.529)
t (2.83) (2.23) (3.043)
R2 = 0.8599 R̄2= 0.8365 F = 36.826
N = 15 ∑ e i2 = 4019.714
Example 2: Consider the model: Y =α + β 1 X 1 i+ β2 X 2 i +U i
On the basis of the information given below answer the following question
2 2
Σ X 1=3200 Σ X 1 X 2=4300 Σ X 2=400Σ X 2=7300 Σ X 1 Y =8400 Σ X 2 Y =13500

ΣY =800 Σ X 1=250 n=25 Σ Y 2i =28,000

a. Find the OLS estimate of the slope coefficient

b. Compute variance of ^β 2
c. Test the significant of β 2 slope parameter at 5% level of significant
d. Compute R2and R̄2and interpret the result
e. Test the overall significance of the model
Solution:
a. Since the above model is a two explanatory variable model, we can estimate ^β 1∧ ^β 2using the formula
in equation (3.21) and (3.22) i.e.

4
 2
^β = Σ x 2 yΣ x 2−Σ x 2 yΣ x 1 x 2
1 2 2
Σ x 1 Σ x2 −¿ ¿
2
^β = Σ x 2 yΣ x 1−Σ x 1 yΣ x 1 x 2
2 2 2
Σ x 1 Σ x2 −¿ ¿
Since the x’s and y’s in the above formula are in deviation form, we have to find the corresponding
deviation forms of the above given values.
We know that:
Σ x 1 x 2=Σ X 1 X 2 −n X̄ 1 X̄ 2
¿ 4300−( 25 ) ( 10 ) ( 16 )=300
Σ x 1 y=Σ X 1 Y −n X̄ 1 Ȳ
¿ 8400−25 ( 10 ) (32 )=400
Σ x 2 y=Σ X 2 Y −n X̄ 2 Ȳ
¿ 13500−25 (16 ) (32 )=700
2 2 2
Σ x 1=Σ X 1 −n X̄ 1
¿ 3200−25 ¿
2 2 2
Σ x 2=Σ X 2 −n X̄ 2
¿ 7300−25¿
2 2 2
Σ y =ΣY −n Y
¿ 28,000−25 ¿ ¿ 2,400
Now we can compute the parameters.
2
^β = Σ x 2 yΣ x 2−Σ x 2 yΣ x 1 x 2
1 2 2
Σ x 1 Σ x2 −¿ ¿
( 400 ) ( 900 ) −( 700 ) ( 300 )
¿
(900)(700)−¿ ¿
2
^β = Σ x 2 yΣ x 1−Σ x 1 yΣ x 1 x 2
2 2 2
Σ x 1 Σ x2 −¿ ¿
( 700 ) (700 )−( 400 )( 300 )
¿
(900)(700)−¿ ¿
The intercept parameter can be computed using the following formula.
^
β 0=Ȳ − ^β 1 X̄ 1− ^β2 X̄ 2
¿ 32−( 0.278 )( 10 )−( 0.685 ) ( 16 ) =18.26
σ^ Σ x 2
2 2

b. var (¿ ^β 1)= 2 2
¿
Σ x 1 Σ x 2−¿ ¿
2
2 Σ ei
⇒ σ^ = Where k is the number of parameters
n−k
In our case k=3
5
 2
2Σ ei
⇒ σ^ =
n−3
Σ e1 =Σ y − β^ 1 Σ x 1 y − ^β 2 Σ x 2 y
2 2

¿ 2,400−0.278 ( 400 ) −( 0.685 ) ( 700 )=1809.3

2
Σ ei 1809.3
2
σ^ u = ¿ =82.24
n−3 25−3
( 82.24 ) ( 900 )
⇒ var (¿ ^β 1)= =0.137 ¿
540,000
β 1 )= √ var (¿ ^β 1)= √0.137=0.370 ¿
SE ( ^

σ^ Σ x 1
2 2
var (¿ ^β 2)= 2 2
¿
Σ x1 Σ x 1−¿ ¿

β 2 )= √ var (¿ ^β 2)=√ 0.1067=0.327 ¿

SE ( ^

c. ^β 1can be tested using students t-test

This is done by comparing the computed value of t and critical value of t which is obtained from the table at
α
level of significance and n-k degree of freedom.
2
t∗¿ 0.278
Hence; = =0.751¿
^
SE ( β1 ) 0.370
α 0.05
The critical value of t from the t-table at = =0.025 level of significance and 22 degree of freedom is
2 2
2.074.
t c =2.074 , t∗¿ 0.755 ⇒ t∗¿ t c
The decision rule if t∗¿ t c is to reject the alternative hypothesis that says β is different from zero and to
accept the null hypothesis that says β is equal to zero. The conclusion is ^β 1is statistically insignificant or the
sample we use to estimate ^β 1is drawn from the population of Y & X 1in which there is no relationship
between Y and X1(i.e. β 1=0).
d. R2can be easily using the following equation
2 ESS RSS
R= =¿ 1-
TSS TSS
We know that RSS=Σ e 2i
and TSS=Σ y 2 and ESS=Σ ^y = ^β1 Σ x 1 y+ β^ 2 Σ x 2 y+ … …+ ^β k Σ x k y
2

For two explanatory variable model:

RSS 1809.3
2
R =¿1- =1−
TSS 2400
¿ 0.24

6
⇒24% of the total variation in Y is explained by the regression line Y^ =18.26 +0.278 X 1 +0.685 X 2) or by the
explanatory variables (X1 and X2).
2
Σ ei
2 n−k ( 1−R2 ) ( n−1 )
Adjusted R =1− 2
=1−
Σy n−k
n−1
( 1−0.24 ) ( 24 )
¿ 1− =0.178
22
e. Let’s set first the joint hypothesis as
H 0 : β 1=β 2=0
against H 1 : at least one of the slope parameters is different from zero.
The joint test hypothesis is testing using the F-test given below.
0.8599 0.8599
3−1 2
F* =¿ =
0.1401 0.1401
15−3 12
0.42995
¿ =36.826
0.11675

ESS
k−1
F ¿ [ (k−1 ) , (n−k )] =
RSS
n−k
R2 0.24 0.24
k −1 3−1 2 0.12
¿ 2= = = =¿3.478
1−R 0.76 0.76 0.0345
N −k 25−3 22
From (d) R2=0.24∧k=3
F ¿ (2 ,22)=3.478 this is the computed value of F. Let’s compare this with the critical value F (2,22) at 5%
level of significance of 3.44.
F*(2,22) = 3.478 > Fc(2,22)=3.44,⇒ F*>Fc, the decision rule is to reject H 0 and accept H1. We can say
that the model is significant i.e. the dependent variable is, at least, linearly related to one of the explanatory
variables.
Application of Multiple Regression in valuation
Hedonic Pricing Model Example 1. A real estate researcher collected the following data on housing
characteristics and sale prices.
House Price (000 Birr) (Y) Size (m²) (X₁) Bedrooms (X2)
1 700 100 2
2 850 120 3
3 1100 150 4
4 650 90 2
5 900 130 3
The relationship between price and property characteristics follows the Hedonic Pricing Model.
7
Required
A. Estimate a regression model: Price=β0+β1Size+β2Bedrooms
B. Interpret the estimated coefficients.
C. Explain how housing attributes affect market value.
D. Discuss the advantages of hedonic models in property valuation.
Solution: Hedonic Pricing Model
Regression model:
Price=β0+β1Size+β2Bedrooms n=5, number of estimated parameters k=3 (including intercept).
Ȳ = 840 X¯1=118 X¯2 =2.8 ∑ Y =4200 ∑ X 1=590 ∑ X 2= 14 ∑ yx1 =16900 ∑ yx 2
=590 ∑ x 1 x2 =78.0 ∑ y =127000 ∑ x 21=2280 ∑ x 22=2.8 ∑ e =666.6667
2 2

Answer
A. Estimate the parameter using OLS

^β = ∑ x1 y ∑ x 2 −∑ x1 x 2 ∑ x 2 y
2

1 2
∑ x 12 ∑ x 22− ( ∑ x 1 x 2 )
2.8× 16900 – 78 ×590
=
2280 ×2.8− ( 78.0 ) 2
1300
¿ = 4.3333
300
^β = ∑ x 2 y ∑ x 1 −∑ x1 x 2 ∑ x 1 y
2

2 2
∑ x 12 ∑ x22−(∑ x 1 x 2 )
2280× 590 – 78 ×16900
¿
300
27,000
= =90
300
^
β 0=Ȳ − ^β 1 X̄ 1− ^β2 X̄ 2
^
β 0 =840 - 4.3333×118 – 90×2.8
^
β 0 =840 - 511.3333 - 252 = 76.6667.
B. Y^ = 76.6667 + 4.3333X1 + 90 X2. This equation will be read as follow
 ^ β 0 = 76.6667 means if the price of the size & bedrooms is zero the price of the property will be
76.6667 units. But it is meaningless to interpret the constant term ( ^
β 0 ).
^
 β 1 is the coefficient of size of the property. The value 4.3333 signifies that holding bedrooms
constant, a one-unit increase in size (1 m²) is associated with an average increase in price of
about 4,333 Birr.
 ^β 2 is the coefficient of the bedrooms. Holding size constant, an additional bedroom raises price by
about 90,000 Birr on average.
C. Test the statistical significance of ( ^
β 0 , β^1∧ β^ 2 ¿.
Standard error test
To test this, we have to have variance & the standard error of the parameters ( ^
β 0 , β^1∧ β^ 2 ¿.The variance of
^
β 0 , β^1∧ β^ 2 will have a value of σ 2 (see equation number 4.16 – 4.18).

[ x̄ ∑ x + x̄ ∑ x −2 x̄ x̄ ∑ x x
]
2 2 2 2

1 1 2 2 1 1 2 1 2
(^
β0 ) =¿ n
+ 2
∑ x ∑ x −( ∑ x x )
2
2
σu
2

Var 1 2 1 2
. . .. . . . . . .. 4.16

[ ∑ x2
]
2

( ^β1 )=¿ σu ∑ ∑
2
2 2

Var x x −( ∑ x x 1 2 1 2)
2
. . . . . . . . . . . . . . . . . . . . . 4.17

8
 ∑e
[ ]
2
∑ x1
2
1
( ^β2 ) σu 2
2
σu
x 1 x 2−( x 1 x 2 ) . . . . . . . .. . . . . . . . . . .. . . . 4.18. Where
2 2

Var =
∑ ∑ ∑ 2
= n−k 1 , n is
number of samples, k is number of parameters which are estimated.

Where σ 2u= ∑
2
ei
from our example n=sample size of 5
n−k
K is number of estimated parameters which is 3 (α ^ β^ ¿
^ , β∧ 2

2 ∑
2
e i 666.6667
σ u= = =¿ 333.3333
N −k 5−3
Using equation number 4.16 we can calculate Var( ^ β0 )
1
Var( ^
β0 ) = +¿
5
Var( ^β0 ) = 5977.7778
Var( ^β ) again we can calculate using equation number 4.18
1

Var( ^β1 ) = 333.3333

2.8
300 [ ]
= 3.111111

Var( ^β2 ) = the Var of ^β 2can be calculated using equation number 4.19

Var( β^ 2 ) = 333.3333
2280
300 [ ]
= 2533.3333

From the above values we can calculate S.E ( ^ β0 ), S.E ( ^β1 ) & S.E ( ^β2 ) as follows
S.E ( β^0 ) = √ Var β^ =√ 5977.7778 ≈77.3187
0

√
S.E ( ^β1 ) = Var ( ^β 1) =√ 3.111111≈ 1.7638
√
S.E ( β^ 2 ) = Var ( ^β 2 )=√ 2533.3333 ≈ 50.3322
Having calculated S.E of the coefficients of the variables ( ^ ^ ^β ¿ we can undertake t-tests for individual
β 0 , β∧ 2
significance
We test H0: βj=0 against H1: βj≠0 at 5% significance level.
^
β0
t= 76.6667
Computed value of t = =0.992
S .E( ^ β0 ) 77.3187
^β
t= 1 4.3333
Computed t = =2.457
S . E ( ^β1 ) 1.7638
^β
t= 2 90
Computed t = =1.788
^
S . E ( β2 ) 50.3322

9
Compare computed with table value

Degrees of freedom for t-table: n−k=5-3=2. At 5% significance level, two-tailed, t-critical

value t0.025,2=4.303(from t-distribution table).

Since tB1=2.457<4.303, we fail to reject H0 for β₁.

tB2=1.788<4.303, fail to reject H0 for β₂.

tB0=0.992<4.303, fail to reject H0 for β₀.

Thus, none of the coefficients are statistically significant at the 5% level with only 5 observations. This is
due to the very low degrees of freedom.
Coefficient of determination /R2/ 126333
^β ∑ y x + β^ ∑ y x RRS 666.6667
1 1 2 2
R2= or R2= 1- =1- =1−0.005249=0.99475
∑y 2
TSS 127000
126333.3
R2 = = 0.99475
127000
This means 99.48% of price of the property is explained by size & bedrooms of the property.
Adjusted R̄2
2 2 n−1 5−1
R̄ =1-(1- R ) = 1−(1−0.99475) = 0.9895
n−k 5−3
F - test
The overall significance of the explanatory variables can be tested using F-test. Just like t- test in the case of
F test we will have computed & table value of F. Calculated value of F * can be obtained using the following
formula
H0:β1=β2=0 vs H1: at least one β ≠ 0.
2
R 0.99475 0.99475
k −1 3−1 2 0.99475
F* = 2= = =
1−R 1−0.99475 1−0.99475 0.00525
n−k 5−3 2
=189.5
Critical F-value: F0.05, k−1,n−k=F0.05,2,2. From F-table, F0.05,2,2=19.00. Since computed F=189.5>19.0, we
reject H0 and conclude that the model as a whole is statistically significant. This contradicts the individual
t-tests; it is a result of multicollinearity and low degrees of freedom.
 Significance: Despite the high R2, the individual coefficients are not statistically significant at the
5% level because of the very small sample size (n=5) and limited degrees of freedom. The model as
a whole is significant (F-test), but caution is needed.
 Hedonic model advantages: decomposes value into attribute contributions, handles heterogeneity,
uses market data, flexible for policy analysis.
Note: With only 5 observations, the regression results are unstable and should be treated as illustrative. A
larger sample is required for reliable inference.

Presentation of regression results: - Different books used different presentation methods but the most
commonly is the one which we write under here using our previous example.
^
Price=76.6667 + 4.3333 ¿ ¿ + 90 Bedrooms
SE (77.3187) (1.7638) (50.3322)
t (0.992) (2.457) (1.788)
2= 2=
R 0.99475 R̄ 0.9895 F = 189.5
N=5 ∑ e i2 = 666.6667
Valuing Residential Houses with Three Attributes 2. A researcher wants to model the market price of
residential houses (in 1,000 Birr) using three explanatory variables:
 X1 = Size (in 10 m² units)

10
  X2 = Age of the house (in years)
The model is: Y=β0+β1X1+β2X2 +ε
The following sample data (n = 12) have been collected:
House Price Size Age
(Y) (X₁) (X₂)
1 320 8 5
2 380 9 8
3 450 10 6
4 290 7 4
5 520 12 7
6 410 10 9
7 350 8 6
8 480 11 5
9 310 7 3
10 390 9 7
11 430 10 8
12 470 11 4
From these data, the following summary statistics have been calculated (in deviation form):

Ȳ = 400 X¯1=9.3333 X¯2 =6 ∑ Y =4800 ∑ X 1=112 ∑ X 2= 72 ∑ yx1= 1, 280 ∑ yx 2= 510

∑ x 1 x2= 13 ∑ y = 58,800
2
∑ x 21= 28.6667 ∑ x 22= 38 ∑ e = 1,491.85
2

Required:

A. Estimate the parameters using OLS.

B. Interpret the estimated coefficients.
C. Test the statistical significance of the individual coefficients at the 5% significance level using both the
standard error test and the t-test.
D. Compute the coefficient of determination (R2) and the adjusted R2.
E. Test the overall significance of the model (F-test) at the 5% level.
F. Present the regression results in the standard format.

Solution:
A. Estimate the parameter using OLS

^β = ∑ x1 y ∑ x 2 −∑ x1 x 2 ∑ x 2 y
2

1 2
∑ x 12 ∑ x 22− ( ∑ x 1 x 2 )
1, 280 ×3 8 – 13× 510
=
28.6667 × 3 8−( 13 ) 2
42,010
¿ = 45.647
920.33
^β = ∑ x 2 y ∑ x 1 −∑ x1 x 2 ∑ x 1 y
2

2 2
∑ x 12 ∑ x22−(∑ x 1 x 2 )
510× 28 .6667 – 13 ×1 28 0
¿
920.33
−2,019.97
= =−¿ 2.195
920.33
^
β 0=Ȳ − ^β 1 X̄ 1− ^β2 X̄ 2
^
β 0=400 – 45.647×9.3333 – 6×-2.195
^
β 0=400 – 426.037+13.17 = −¿12.867.

11
 B. Interpretation of the CoefficientsY^ = −¿12.867 + 45.647X1 – 2. 195X2. This equation will be read
as follow
 ^ β 0 = -12.867 When all explanatory variables are zero, the predicted price is -12,867Birr. This
is not economically meaningful because a house with zero size and zero age does not exist; it
merely serves as the intercept.
 ^β 1 =45.647: Holding age constant, an increase of one unit in X1 (i.e., 10 m² in size) raises the
price by about 45,647 Birr on average.
 ^β 2=−2. 195 : Holding size constant, each additional year of age reduces the price by about
2,195 Birr on average, which is plausible due to depreciation.
C. Test the statistical significance of ( ^ β 0 , β^1∧ β^ 2 ¿.
Standard error test
To test this, we have to have variance & the standard error of the parameters ( ^ β 0 , β^1∧ β^ 2 ¿.The variance of
^
β 0 , β^1∧ β^ 2 will have a value of σ 2 (see equation number 4.16 – 4.18).

[ x̄ ∑ x + x̄ ∑ x −2 x̄ x̄ ∑ x x
]
2 2 2 2

1 1 2 2 1 1 2 1 2
(^
β0 ) =¿ n
+
∑ x ∑ x −( ∑ x x )
2 2
2
2
σu
Var 1 2 1 2
. . .. . . . . . .. 4.16

[ ]
2
∑ x2
( ^β1 )=¿ σu ∑ ∑
2
2 2

Var x x −( ∑ x x ) 1 2 1 2
2
. . . . . . . . . . . . . . . . . . . . . 4.17

[ ]
2
∑ x1
( ^β2 ) σu 2 2 2
∑ x 1 ∑ x 2−( ∑ x 1 x 2 )2
Var = . . . . . . . .. . . . . . . . . . .. . . . 4.18.

∑e
2

2 1
σu
Where = n−k , n is number of samples, k is number of parameters which are estimated.
1

Where σ u2 = ∑
e i2
from our example n=sample size of 12
n−k
K is number of estimated parameters which is 3 ( ^
β 0 , β^1∧ β^ 2 ¿

σ 2u=
∑ e i2 = 1,491.85 =¿ 165.761
N −k 12−3
Using equation number 4.16 we can calculate Var( ^ β0 )
Var( ^
β ) =¿ 0

Var( ^
β0 ) = 533.650
Var( ^β1 ) again we can calculate using equation number 4.18
38
Var( ^β1 ) = 165.761
920.33
= 6.844 [ ]
Var( ^β2 ) = the Var of ^β 2can be calculated using equation number 4.19

Var( ^β2 ) = 165.761

28.6667
920.33
= 5.163 [ ]
From the above values we can calculate S.E ( ^ β0 ), S.E ( ^β1 ) & S.E ( ^β2 ) as follows
S.E ( ^β0 ) = √ Var β^ =√ 533.650 ≈23.101 0

√
S.E ( ^β1 ) = Var ( ^β 1) =√ 6.844 ≈ 2.616
√
S.E ( ^β2 ) = Var ( ^β )=√ 5.163 ≈ 2.343 2

Having calculated S.E of the coefficients of the variables ( ^

β 0 , β^1∧ β^ 2 ¿ we can undertake t-tests for individual
significance

12
We test H0: βj=0 against H1: βj≠0 at 5% significance level.
^
β0
t= 12.867
Computed t = =0.557
^
S . E ( β0 ) 23.101
^β
t= 1 45.647
Computed t = =17.449
S . E ( β^ 1 ) 2.616
^β
t= 2 2.195
Computed t = =0.937
S . E ( ^β2 ) 2.343

Compare computed with table value

Degrees of freedom for t-table: n−k=12-3=9. At 5% significance level, two-tailed, t-critical

value t0.025,9=2.262(from t-distribution table).

Since tB1=17.449>2.262, we reject H0 for β₁.

tB2=0.937<2.262, fail to reject H0 for β₂.

tB0=0.557<2.262, fail to reject H0 for β₀.

Thus, except ^
β 1none of the coefficients are statistically significant at the 5% level with only 12 observations.
Coefficient of determination /R2/
^β ∑ y x + β^ ∑ y x RRS 1491.85
1 1 2 2
R=2
or R2= 1- =1- =1−0.03=0. 97
∑y 2
TSS 58,800
57308.15
R2 = = 0.97
58,800
This means 99.48% of price of the property is explained by size & bedrooms of the property.
Adjusted R̄2
2 2 n−1 12−1
R̄ =1-(1- R ) = 1−(1−0.97) = 0.97
n−k 12−3
F - test
The overall significance of the explanatory variables can be tested using F-test. Just like t- test in the case of
F test we will compute & read table value of F. Calculated value of F * can be obtained using the following
formula
H0:β1=β2=0 vs H1: at least one β ≠ 0.
2
R 0.97 0.97
k −1 3−1 2 0.016
F* = 2= = = =¿4.5
1−R 1−0.97 1−0.97 0.0034
n−k 12−3 9
Critical F-value: F0.05, k−1, n−k=F0.05,2,2. From F-table, F0.05,2,2=19.00. Since computed F=4.5<19.0, we
accept H0 and conclude that none of the explanatory variable is statistically significant. This is inconsistent
with the individual t-tests; it is a result of multicollinearity and low degrees of freedom.
 Significance: Despite the high R2, the individual coefficients are not statistically significant, except
β₁, at the 5% level because of the very small sample size (n=12) and limited degrees of freedom. The
model as a whole is insignificant (F-test) as well.
Note: With only 12 observations, the regression results are unstable and should be treated as illustrative. A
larger sample is required for reliable inference.

Presentation of regression results: - Different books used different presentation methods but the most
commonly is the one which we write under here using our previous example.

Y^ = -12.867 + 45.647X1 – 2. 195X2.

SE (23.101) (2.616) (2.343)
13
 t (0.557) (17.449) (0.937)
2=
R 0. 97 R̄ 0.97 F = 4.5
2=

n = 12 ∑ ε2i = 1491.85

Exercise for chapter three

1) From the following data compute the regression of automobile expenditure on consumer expenditure &
other travel expenditure (Take the linear regression analysis)

Automobile 212 158 180 253 175 429 437 419 318 355
expenditure

Other travel expense 29 46 28 26 29 64 119 81 74 66

Consumer expenditure 2437 2476 213 2256 225 3566 4486 360 3446 3736
2 8 2
A. Calculate the coefficient of parameters using OLS regression equation of Consumer expenditure on
other travel expense & other travel expense?
B. Write the equation
C. Test the significance of the parameters using standard error & t-test?
D. Construct the 95% confidence interval for the parameters?
E. Calculate the unadjusted & adjusted R2. Why the difference is arising?
F. Test the overall significance of the regression?
G. If the relationship is non-linear, how would you interpret the results (coefficients)
H. Calculate the partial correlation coefficients of the parameters & interpret the results?
2: Commercial Property Valuation: A researcher wants to explain the market value of commercial plots (in
1,000 Birr) using the size of the plot (in 100 m²) and the distance to the central business district (CBD, in
km). The model is: Value=β0+β1Size+β2Distance+ε,

Plot Value (y) Size (x₁) Distance (x₂)

1 250 5 2
2 320 6 3
3 410 8 4
4 190 4 1
5 530 10 5
6 280 5 2.5
7 470 9 4
8 220 4 1.5
9 390 7 3
10 310 6 2
11 450 8 4.5
12 270 5 2
From the data, the following sums are given:
n=12 , Ȳ = 340.83 X¯1=6.42 X¯2 =2.88 ∑ Y =4,090 ∑ X 1=77 ∑ X 2= 34.5 ∑ yx1 = 2,315.83
∑ yx 2= 1,446.25 ∑ x 1 x2= 26.13 ∑ y 2= 127,291.67 ∑ x 21= 42.92 ∑ x 22= 17.56 ∑ e2= 1,523.03
Required:
1. Estimate the parameters using OLS.
2. Interpret the coefficients.
3. Test the statistical significance of each coefficient (5% level) using the t-test.
4. Compute R2 and adjusted R2.
5. Perform the F-test and interpret.
14
6. Present the regression output.

15