CHAPTER 6

Exercise Solutions

112
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 113

EXERCISE 6.1

(a) To compute R2 , we need SSE and SST. We are given SSE. We can find SST from the
equation

̂ y 
( yi  y ) 2 
SST
 13.45222
N 1 N 1
Solving this equation for SST yields
SST  ̂2y  (N 1)  (13.45222) 2  39  7057.5267

Thus,
SSE 979.830
R2  1   1   0.8612
SST 7057.5267

(b) The F-statistic for testing H 0 :  2  3  0 is defined as

(SST  SSE) (K 1) (7057.5267  979.830) / 2
F    114.75
SSE (N  K ) 979.830 /(40  3)

At  0.05 , the critical value is F(0.95, 2, 37)  3.25 . Since the calculated F is greater than
the critical F, we reject H0 . There is evidence from the data to suggest that 2  0 and/or
3  0 .
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 114

EXERCISE 6.2

The model from Exercise 6.1 is yi  1  2 xi  3 zi  ei . The SSE from estimating this
model is 979.830. The model after augmenting with the squares and the cubes of
predictions ŷ 2 and ŷ 3 is y     x   z   ŷ 2   ŷ 3  e . The SSE from estimating
i i i 1 2i 3 i 1 i 2 i i

this model is 696.5375. To use the RESET test, we set the null hypothesis
H0 : 1  2  0 . The F-value for testing this hypothesis is
(SSER  SSEU ) J (979.830  696.5375) 2
F    7.1175
SSEU (N  K ) 696.5373 (40  5)

The critical value for significance level   0.05 is F(0.95,2,35)  3.267 . Since the calculated
F is greater than the critical F we reject H0 and conclude that the model is misspecified.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 115

EXERCISE 6.3

(a) Let the total variation, unexplained variation and explained variation be denoted by SST,
SSE and SSR, respectively. Then, we have
SSE   êi2   N  K   ̂ 2  (20  3)  2.5193  42.8281
Also,
SSE
R2  1   0.9466
SST
and hence the total variation is
SSE 42.8281
SST    802.0243
1 R2
1  0.9466
and the explained variation is
SSR  SST  SSE  802.0243  42.8281  759.1962

(b) A 95% confidence interval for 2 is

b2  t(0.975,17)se(b2 )  0.69914  2.110  0.048526  (0.2343, 1.1639)

A 95% confidence interval for 3 is

b2  t(0.975,17)se(b3 )  1.7769  2.110  0.037120  (1.3704, 2.1834)

(c) To test H0: 2  1 against the alternative H1: 2 < 1, we calculate

b   0.69914 1
t  se2 b 2 
 2  0.048526 1.3658
At a 5% significance level, we reject H0 if t  t(0.05,17)  1.740 . Since 1.3658  1.740 ,
we fail to reject H 0 . There is insufficient evidence to conclude 2  1.

(d) To test H 0 : 2  3  0 against the alternative H1 : 2  0 and/or 3  0 , we calculate

explained variation  K 1 759.1962 / 2
F   151
unexplained variation  N  K  42.8281/17

The critical value for a 5% level of significance is F(0.95, 2,17)  3.59 . Since 151  3.59 , we
reject H0 and conclude that the hypothesis 2 = 3 = 0 is not compatible with the data.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 116

Exercise 6.3 (continued)

(e) The t-statistic for testing H 0 : 22  3 against the alternative H1 : 22  3 is

t
 2b2  b3 
se  2b2  b3 

For a 5% significance level we reject H0 if t  t(0.025,17)  2.11 or t  t(0.975,17)  2.11 . The

standard error is given by
se 2b  b   22  vˆ
ar(b2 )  vˆ
ar(b 3)  2  2  cˆ
ov(b ,b )
2 3 23

 4  0.048526  0.03712  2  2  0.031223

 0.59675

The numerator of the t-statistic is

2b2  b3  2  0.69914 1.7769  0.37862
leading to a t-value of
0.37862
t  0.634
0.59675
Since 2.11  0.634  2.11 , we do not reject H0 . There is no evidence to suggest that
22  3 .
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 117

EXERCISE 6.4

In each case we use a two-tail test with a 5% significance level. The critical values are
given by t(0.025,60)  2.000 and t(0.975,60)  2.000 . The rejection region is t  2 or t  2 .

(a) The value of the t statistic for testing the null hypothesis H0 : 2  0 against the alternative
H1 : 2  0 is
b2 3
t   1.5
se(b2 ) 4
Since 2  1.5  2 , we fail to reject H0 and conclude that there is no sample evidence to
suggest that 2  0.

(b) For testing H0: 1 + 22 = 5 against the alternative H1: 1 + 22  5, we use the statistic

t
 b1  2b2   5
se  b1  2b2 

For the numerator of the t-value, we have

b1  2b2  5  2  2  3  5  3
The denominator is given by

se(b1  b2 )  vˆ
ar(b 1 2b )2  vˆ
ar(b1 )  4  vˆ
ar(b 2)  4  cˆ
ov(b ,b
12
)
 
 3  4  4  4  2  11  3.3166
 

Therefore,
 
3
t   0.9045
3.3166
Since 2  0.9045  2 , we fail to reject H0. There is no sample evidence to suggest that
1  22  5 .
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 118

Exercise 6.4 (continued)

(c) For testing H0 : 1  2  3  4 against the alternative H1 : 1  2  3  4 , we use the

statistic
(b1  b2  b3 )  4
t
se(b1  b2  b3 )

Now,
(b1  b2  b3 )  4  2  3 1  4  6

and
se(b  b  b )  ˆ
v
ar(b  b 2 b )3

1 2 3 1

  vˆ
ar(b1 )  vˆ
ar(b 2 )  vˆ
ar(b 3)  2cˆ
ov(b ,b
12
)  2cˆ
ov(b ,b1 3)  2cˆ
ov(b ,b2 )3

  3  4  3  2  2  2 1  0
 4

Thus,
6
t  1.5
4
Since 2  1.5  2 , we fail to reject H0 and conclude that there is insufficient sample
evidence to suggest that 1  2 + 3 = 4 is incorrect.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 119

EXERCISE 6.5

Consider, for example, the model

yi  1  2 xi  3 zi  ei
If we augment the model with the predictions ŷi the model becomes

yi  1   2 xi  3 zi   ŷi  ei
However, ŷi  b1  b2 xi  b3 zi is perfectly collinear with xi and wi . This perfect
collinearity means that least-squares estimation of the augmented model will fail.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 120

EXERCISE 6.6

(a) Least squares estimation of yi  1  2 xi  3 wi  ei gives b3  0.4979 , se(b3 )  0.1174

and t  0.4979 0.1174  4.24 . This result suggests that b3 is significantly different from
zero and therefore wi should be included in the model. Additionally, the RESET test
based on the equation yi  1  2 xi  ei gives F-values of 17.98 and 8.72 which are much
higher than the 5% critical values of F(0.95,1,32)  4.15 and F(0.95, 2,31)  3.30 , respectively.
Thus, the model omitting wi is inadequate.

(b) Let b2 be the least squares estimator for 2 in the model that omits wi . The omitted-
variable bias is given by
cˆ
ov(x, w)
2 ) 2  3
*
E(b
vˆ
ar(x)
Now, cˆ ov(x, w)  0 because r xw  0 . Thus, the omitted variable bias will be positive. This
result is consistent with what we observe. The estimated coefficient for  2 changes from
0.9985 to 4.1072 when wi is omitted from the equation.

(c) The high correlation between xi and wi suggests the existence of collinearity. The
observed outcomes that are likely to be a consequence of the collinearity are the sensitivity
of the estimates to omitting wi (the large omitted variable bias) and the insignificance of
b2 when both variables are included in the equation.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 121

EXERCISE 6.7

(a) The coefficients of ln(Y), ln(K) and ln(PF) are 0.6792, 0.3503 and 0.3219, respectively.
Since the model is in log-log form the coefficients are elasticities. The estimate 0.6792 is
the percentage change in VC when Y changes by 1%, with the other variables held
constant. Similarly, 0.3503 is the percentage change in VC when K changes by 1%, and
0.3219 is the percentage change in VC when PF changes by 1%, keeping the other
variables constant in each case.

(b) An increase in any one of the explanatory variables should lead to an increase in variable
cost, with the exception of ln(STAGE). For a given level of output (passenger-miles) and a
given level of capital stock, longer flights should be cheaper than shorter ones. Thus,
positive signs are expected for all variables except ln(STAGE), whose coefficient should
be negative. All coefficients have the expected signs with the exception of ln(PM ) .

(c) The coefficient of ln(PM ) has a p-value of 0.4966 which is higher than 0.05, indicating
that this coefficient is not significantly different from zero. The p-values of the other
coefficients are all 0.0000, indicating that they are significant.

(d) Augmenting the equation with the squares of the predictions, and squares and cubes of the
predictions, yields the RESET test F-values of 3.3803 and 1.8601 with corresponding p-
values of 0.0671 and 0.1577, respectively. These two p-values are higher than the
conventional 0.05 level of significance indicating that the model is adequate.

(e) From the middle panel of Table 6.6 the F-value for testing H0 : 2  3  1 is 6.1048 with
a p-value of 0.014. This p-value is less than the significance level of 0.05. We reject H 0
and conclude that constant returns to scale do not exist.

(f) The F-value and the p-value for testing H0 : 4  5  6  1 can be read from the bottom
panel of Table 6.6. The F value is very large and the corresponding p-value of 0.00000 is
below the significance level of 0.05. We reject H0 and conclude that there is no evidence
to suggest that if all input prices increase by the same proportion, variable cost will
increase by the same proportion.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 122

Exercise 6.7 (continued)

(g) To test H 0 : 2  3  1, the value of the t statistic is

b2  b3 1 0.6792  0.3503 1
t   2.48
se(b2  b3 ) 0.01187
where the standard error is calculated from
se(b  b )  vˆ
ar(b 2 b )3
2 3

 vˆ
ar(b2 )  vˆ
ar(b 3)  2cˆ
ov(b ,b
23
)

 0.002851  0.002796  2(0.002753)

 0.011874
We reject H0 because 2.48  t(0.975,261)  1.969 . Note t 2  (2.48)2  6.15  F  6.10 . The
difference between t 2 and F is due to rounding error.

To test H0 : 4  5  6  1 , the value of the t -statistic is

b  b  b 1 0.2754  0.3219  0.0683 1
t 4 5 6    8.69
se(b4  b5  b6 ) 0.0542
where
se(b  b  b )  ˆ
var(b 4 b 5 b )6  0.002938  0.0542
4 5 6

with
ˆ ˆ
ar(b  b  b )  v
v ar(b )  vˆ
ar(b )  vˆ
ar(b )  2cˆ
ov(b ,b )  2cˆ
ov(b ,b )  2cˆ
ov(b ,b )
4 5 6 4 5 6 4 5 4 6 5 6

 0.001919  0.001303  0.010068  2  0.000088

 2  0.002159  2  0.002929
 0.002938
We reject H0 because 8.69  t(0.025,261)  1.969 . Note that t 2  (8.69)2  75.52 which
is approximately equal to F  75.43 .
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 123

EXERCISE 6.8

There are a number of ways in which the restrictions can be substituted into the model,
with each one resulting in a different restricted model. We have chosen to substitute out
1 and 3 . With this in mind, we rewrite the restrictions as
3  1  3.84
1  80  62 1.93  3.614
Substituting the first restriction into the second yields
1  80  62 1.9(1  3.84 )  3.614
Substituting this restriction and the first one 3  1  3.84 into the equation
S    P  A  A2  e
i 1 2 i 3i 4 i i

yields
S  80  6 1.9(1  3.8 )  3.61    P  1  3.8  A   A2  e
i 2 4 4 2 i 4 i 4 i i

Rearranging this equation into a form suitable for estimation yields

 Si  Ai  78.1  2  Pi 6  4 3.61  3.8A  A   e

i i
2
i
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 124

EXERCISE 6.9

The results of the tests in parts (a) to (e) appear in the following table. Note that, in all
cases, there is insufficient evidence to reject the null hypothesis at the 5% level of
significance.

Part H0 F-value df Fc (5%) p-value

(a) 2 = 0 0.047 (1,20) 4.35 0.831

(b) 2 = 3 = 0 0.150 (2,20) 3.49 0.862
(c) 2 = 4 = 0 0.127 (2,20) 3.49 0.881
(d) 2 = 3 = 4 = 0 0.181 (3,20) 3.10 0.908
(e) 2 + 3 + 4 + 5 = 1 0.001 (1,20) 4.35 0.980

(f) The auxiliary R2 s and the explanatory-variable correlations that are exhibited in the
following table suggest a high degree of collinearity in the model.

Correlation with Variables

Variable Auxiliary R2 ln(L) ln(E) ln(M)

ln(K) 0.969 0.947 0.984 0.959
ln(L) 0.973 0.972 0.986
ln(E) 0.987 0.983
ln(M) 0.984

To examine the effect of collinearity on the reliability of estimation, we examine the

estimated equation, with t values in parentheses,

lˆ
n Y   0.035  0.056 ln  K   0.226 ln  L   0.044 ln  E   0.670 ln  M 
(t) 0.800 0.216 0.511 0.112 1.855
R2  0.952
The very small t-values for all variables except ln(M ) , our inability to reject any of the
null hypotheses in parts (a) through (e), and the high R2 , are indicative of high
collinearity. Collectively, all the variables produce a model with a high level of
explanation and a good predictive ability. Furthermore, our economic theory tells us that
all the variables are important ones in a production function. However, we have not been
able to estimate the effects of the individual explanatory variables with any reasonable
degree of precision.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 125

EXERCISE 6.10

(a) The restricted and unrestricted least squares estimates and their standard errors appear in
the following table. The two sets of estimates are similar except for the noticeable
difference in sign for ln(PL). The positive restricted estimate 0.187 is more in line with
our a priori views about the cross-price elasticity with respect to liquor than the negative
estimate 0.583. Most standard errors for the restricted estimates are less than their
counterparts for the unrestricted estimates, supporting the theoretical result that restricted
least squares estimates have lower variances.

CONST ln(PB) ln(PL) ln(PR) ln(I )

Unrestricted 3.243 1.020 0.583 0.210 0.923

(3.743) (0.239) (0.560) (0.080) (0.416)
Restricted 4.798 1.299 0.187 0.167 0.946
(3.714) (0.166) (0.284) (0.077) (0.427)

(b) The high auxiliary R2s and sample correlations between the explanatory variables that
appear in the following table suggest that collinearity could be a problem. The relatively
large standard error and the wrong sign for ln(PL) are a likely consequence of this
correlation.

Sample Correlation With

Variable Auxiliary R2 ln(PL) ln(PR) ln(I)
ln(PB) 0.955 0.967 0.774 0.971
ln(PL) 0.955 0.809 0.971
ln(PR) 0.694 0.821
ln(I) 0.964

(c) We use the F-test to test the restriction H0 : 2  3  4  5  0 against the alternative
hypothesis H1 : 2  3  4  5  0 . The value of the test statistic is F = 2.50, with a p-
value of 0.127. The critical value is F(0.95,1, 25)  4.24 . Since 2.50  4.24 , we do not reject
H 0 . The evidence from the data is consistent with the notion that if prices and income go
up in the same proportion, demand will not change. This idea is consistent with economic
theory.
The F-value can be calculated from restricted and unrestricted sums of squared errors as
follows
(SSER  SSEU ) J (0.098901  0.08992) 1
F   2.50
SSEU (N  K ) 0.08992 25
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 126

Exercise 6.10 (continued)

(d)(e) The results for parts (d) and (e) appear in the following table. The t-values used to
construct the interval estimates are t(0.975, 25)  2.060 for the unrestricted model and
t(0.975, 26)  2.056 for the restricted model. The two 95% prediction intervals are (70.6,
127.9) and (59.6, 116.7). The effect of the nonsample restriction has been to increase both
endpoints of the interval by approximately 10 litres.

ln(Q) Q
lˆ
n(Q) se( f ) tc lower upper lower upper

(d) Restricted 4.5541 0.14446 2.056 4.257 4.851 70.6 127.9

(e) Unrestricted 4.4239 0.16285 2.060 4.088 4.759 59.6 116.7
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 127

EXERCISE 6.11

(a) The estimated Cobb-Douglas production function with standard errors in parentheses is

lˆ
n  Q   0.129  0.559 ln  L   0.488ln  K  R2  0.688
(se) 0.546 0.816 0.704
The magnitudes of the elasticities of production (coefficients of ln(L) and ln(K)) seem
reasonable, but their standard errors are very large, implying the estimates are unreliable.
The sample correlation between ln(L) and ln(K) is 0.986. It seems that labor and capital
are used in a relatively fixed proportion, leading to a collinearity problem which has
produced the unreliable estimates.

(b) After imposing constant returns to scale the estimated function is

lˆ
n  Q   0.020  0.398ln  L   0.602 ln  K 
(se) 0.053 0.559 0.559
We note that the relative magnitude of the elasticities of production with respect to capital
and labor has changed, and the standard errors have declined. However, the standard
errors are still relatively large, implying that estimation is still imprecise.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 128

EXERCISE 6.12

The RESET test results for the log-log and the linear demand function are reported in the
table below.

Test F-value df 5% Critical F p-value

Log-log 1 term 0.0075 (1,24) 4.260 0.9319
2 terms 0.3581 (2,23) 3.422 0.7028
Linear 1 term 8.8377 (1,24) 4.260 0.0066
2 terms 4.7618 (2,23) 3.422 0.0186

Because the RESET test returns p-values less than 0.05 (0.0066 and 0.0186 for one and
two terms respectively), at a 5% level of significance we conclude that the linear model is
not an adequate functional form for the beer data. On the other hand, the log-log model
appears to suit the data well with relatively high p-values of 0.9319 and 0.7028 for one
and two terms respectively. Thus, based on the RESET test we conclude that the log-log
model better reflects the demand for beer.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 129

EXERCISE 6.13

(a) The estimated model is

Yˆ  0.6254  0.0302t  0.0794 RG  0.0005 RD  0.3387 RF R2  0.6889
t t t t

(se) (0.2582) (0.0034) (0.0817) (0.0918) (0.1654)

(t) (2.422) (8.785) (0.972) (0.005) (2.047)

We expect the signs for 2 ,3 ,4 and 5 to be all positive. We expect the wheat yield to
increase as technology improves and additional rainfall in each period should increase
yield. The signs of b2 and b5 are as expected, but those for b3 and b4 are not. However,
the t -statistics for testing significance of b3 and b4 are very small, indicating that both of
them are not significantly different from zero. Interval estimates for 3 and 4 would
include positive ranges. Thus, although b3 and b4 are negative, positive values of 3 and
4 are not in conflict with the data.

(b) We want to test H0 : 3  4 , 3  5 against the alternative H1 : 3 , 4 and 5 are not all
equal. The value of the F test statistic is
(SSER  SSEU ) J (4.863664  4.303504) 1
F    2.7985
SSEU (T  K ) 4.303504 (48  5)
The corresponding p-value is 0.072. Also, the critical value for a 5% significance level is
F(0.95, 2, 43)  3.214 . Since the F-value is less than the critical value (and the p-value is
greater than 0.05), we do not reject H 0 . The data do not reject the notion that the response
of yield is the same irrespective of whether the rain falls during germination, development
or flowering.

(c) The estimated model under the restriction is

Ytˆ  0.6515  0.0314t  0.0138 RGt  0.0138 RDt  0.0138RFt

(se) (0.2679) (0.0035) (0.0567) (0.0567) (0.0567)
(t) (2.432) (8.89) (0.2443) (0.2443) (0.2443)

With the restrictions imposed the signs of all the estimates are as expected. However, the
response estimates for rainfall in all periods are not significantly different from zero. One
possibility for improving the model is the inclusion of quadratic effects of rainfall in each
period. That is, the squared terms RGt2 , RDt2 and RF t2 could be included in the model.
These terms could capture a declining marginal effect of rainfall. See Chapter 7.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 130

EXERCISE 6.14

(a) The estimated model is

ˆ
H W  8.1236  2.1933HE  0.1997HA R 2  0.1655
(se) 4.1583 0.1801 0.0675
(t) 1.954 12.182 2.958
An increase of one year of a husband’s education leads to a $2.19 increase in wages. Also,
older husbands earn 20 cents more on average per year of age, other things equal.

(b) A RESET test with one term yields F  9.528 with p-value = 0.0021, and with two terms
F  4.788 and p-value = 0.0086. Both p-values are smaller than a significance level of
0.05 , leading us to conclude that the linear model suggested in part (a) is not adequate.

(c) The estimated equation is:

ˆ
H W  45.5675 1.4580HE  0.1511HE 2  2.8895HA  0.0301HA2 R2  0.1918
(se) 17.5436 1.1228 0.0458 0.7329 0.0081
(t) 2.597 1.298 3.298 3.943 3.703
Wages are now quadratic functions of age and education. The effects of changes in
education and in age on wages are given by the partial derivatives
Hˆ W HˆW
 1.4580  0.3022HE  2.8895  0.0602HA
HE HA
The first of these two derivatives suggests that the wage rate declines with education up to
an education level of HEmin  1.458 0.30522  4.8 years, and then increases at an
increasing rate. A negative value of HW HE for low values of HE is not realistic. Only
7 of the 753 observations have education levels less than 4.8, so the estimated relationship
might not be reliable in this region. The derivative with respect to age suggests the wage
rate increases with age, but at a decreasing rate, reaching a maximum at the age
HAmax  2.8895 0.06022  48 years.

(d) A RESET test with one term yields F  0.326 with p-value = 0.568, and with two terms
F  0.882 and p-value = 0.414. Both p-values are much larger than a significance level of
0.05. Thus, there is no evidence from the RESET test to suggest the model in part (c) is
inadequate.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 131

Exercise 6.14 (continued)

(e) The estimated model is:

ˆ
H W  37.0540  2.2076HE  0.1688HE 2  2.6213HA
(se) 17.0160 1.0914 0.0444 0.7101
(t) 2.178 2.023 3.800 3.691
 0.0278HA2  7.9379CIT R2  0.2443
0.0079 1.1012
3.525 7.208
The wage rate in large cities is, on average, $7.94 higher than it is outside those cities.

(f) The p-value for b6 , the coefficient associated with CIT, is 0.0000. This suggests that b6 is
significantly different from zero and CIT should be included in the equation. Note that
when CIT was excluded from the equation in part (c), its omission was not picked up by
RESET. The RESET test does not always pick up misspecifications.

(g) From part (c), we have

Hˆ W ˆ
H W
 1.4580  0.3022HE  2.8895  0.0602HA
HE HA
and from part (f)
HˆW ˆ
H W
 2.2076  0.3376HE  2.6213  0.0556HA
HE HA
Evaluating these expressions for HE  6 , HE  15 , HA  35 and HA  50 leads to the
following results.

HW HE HW HA

HE  6 HE  15 HA  35 HA  50
Part (c) 0.356 3.076 0.781 0.123
Part (e) 0.182 2.855 0.678 0.156

The omitted variable bias from omission of CIT does not appear to be severe. The
remaining coefficients have similar signs and magnitudes for both parts (c) and (e), and
the marginal effects presented in the above table are similar for both parts with the
exception of HW HE for HE  6 where the sign has changed. The likely reason for
the absence of strong omitted variable bias is the low correlations between CIT and the
included variables HE and HA. These correlations are given by corr CIT , HE   0.2333
and corr(CIT , HA)  0.0676 .
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 132

EXERCISE 6.15

(a) The average price of a 40-year old house of size 3600 square feet is
PRICE(40,3600)  1  36002  403

The average price of a 5-year old house of size 1800 square feet is
PRICE(5,1800)  1  18002  53

The conjecture that we set up as the alternative hypothesis is

1  36002  403  2(1  18002  53 )
Thus, after simplifying this inequality, the null and alternative hypotheses are
H0 : 1  303  0 H1 : 1  303  0
The test statistic for testing H0 is
b1  30b3
t
se(b1  30b3 )
where

se(b1  30b3 )  vˆ
ar(b1 )  900vˆ
ar(b3)  60cˆ
ov(b ,b
13
)

The values for these quantities and the test results for each house category are as follows.

All houses Town houses French style

b1  30b3 19296 169063 291863
vˆ
ar(b )
1 48855007 130798354 1088235489
vˆar(b3 ) 19851 140902 12063976
ˆ
cov(b ,b )
13 497267 3248879 15538629
se(b1  30b3 ) 9826 21273 113482
t-value 1.964 7.947 2.572
df 1077 67 94
5% critical value 1.646 1.668 1.661
p-value 0.0249 1.0000 0.0058
Decision Reject H0 Accept H0 Reject H 0

For the all-house and French style categories, the data support the conjecture stated in the
alternative hypothesis, namely, that PRICE(40,3600)  2  PRICE(5,1800) . In the case of town
houses, whose estimated equation suggests that they quickly depreciate, the alternative
hypothesis is not supported.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 133

Exercise 6.15 (continued)

(b) The average prices for the three houses are as follows.
(i) PRICE(0,2000)  1  20002

(ii) PRICE(20,2200)  1  22002  203

(iii) PRICE(40,2400)  1  24002  403

Setting PRICE(0,2000)  PRICE(20,2200) , gives

1  20002  1  22002  203

which can be simplified to
102  3  0
Setting PRICE(0,2000)  PRICE(40,2400) , gives

1  20002  1  24002  40 3

which can be simplified to
102  3  0
Thus, all three houses will be equally priced if H0 :102  3  0 holds. The F-value for
testing this null hypothesis against the alternative H1 :10 2  3  0 is F  1.12 . The
corresponding p-value is 0.29. Thus H0 is not rejected. There is no evidence to suggest
the houses are not equally priced.

Remark: In the first printing of POE, the third house was given as 40 years old with 2300
(not 2400) square feet. In this case, the null and alternative hypotheses are H0 : 2  3  0
and H1 : 2  0 and/or 3  0 . The test values are F  773.6 and p-value = 0.0000. The
null hypothesis is rejected.

(c) The application of RESET tests to all houses, town houses and French style homes leads
to rejection of the adequacy of the model PRICE  1  2 SQFT  3 AGE  e in all cases.
The model might be improved by the inclusion of more variables such as type of
neighborhood, and whether the house has particular attributes such as a view, a pool and a
fireplace. Also, the functional form might be inadequate. Log-log or log-linear forms or
the inclusion of quadratic terms might improve the model
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 134

EXERCISE 6.16

(a) A 1% increase in GROWTH leads to a change in VOTE of 2 . A 1% increase in

INFLATION leads to a change in VOTE of 3 . Thus the change in VOTE from increasing
both by 1% is 2  3 . Thus, the null and alternative hypotheses are
H0 : 2  3  0 and H1 : 2  3  0
The rejection region for a 5% significance level is t  t(0.95,28)  1.701. The calculated value
of the test statistic is
b2  b3 0.64876  0.18622
t   0.95
se(b2  b3 ) 0.48685
where the standard error is computed from
se(b  b )  vˆ
ar(b2 )  vˆ
ar(b 2 )  2cˆ
ov(b ,b )
2 3 23

 0.028043  0.186606  2  0.011186

 0.48685

Because 0.95  1.701 , H0 is not rejected. Alternatively, H0 is not rejected because its p-
value of 0.175 is greater than 0.05. There is not enough evidence to suggest increasing
GROWTH and INFLATION by 1% will improve Willie’s VOTE.

(b) Willie will get reelected if E(VOTE)  1  42  53  50 . Thus, the null and alternative
hypotheses are
H0 : 1  42  53  50 and H1 : 1  42  53  50
The rejection region for a 5% significance level is t  t(0.95,28)  1.701. The calculated value
of the test statistic is
b1  4b2  5b3  50 52.4436  4  0.64876  5  0.18622  50 4.1075
t    2.68
se(b1  4b2  5b3 ) 1.5340 1.5340
where the standard error is computed from
se(b1  4b2  5b3 )
ˆ)  42 var(b
ˆ ˆ ˆ)  10  cov(b
ˆ ˆ)
 var(b 2)  5  var(b )3  8  cov(b 1,b 3 )  40  cov(b2 ,b
2
1 2 1 ,b 3

 2.21523  16  0.02804  25  0.18661  8  0.04551 10  0.50594  40  0.01119

 1.5340
Because 2.68  1.701, H0 is rejected. Alternatively, H0 is rejected because its p-value of
0.06 is less than 0.05. The evidence suggests Willie will get reelected when
GROWTH  4 and INFLATION  5 .
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 135

EXERCISE 6.17

(a) The delay from a train is 4 and the delay from a red light is 3 . Thus, the null and
alternative hypotheses are
H0 :33  4 and H1 :33  4
The test can be performed with an F or a t statistic, with the critical value for the F-test
being F(0.95,1, 227)  3.883 , and those for the t-test, t(0.025, 227)  1.970 and t(0.925, 227)  1.970 .
The rejection regions are F  3.883 for the F-test, and t  1.970 or t  1.970 for the t-
test. The calculated value of the t-test statistic is
3b3  b4 31.3353  2.7548
t   2.404
se(3b3  b4 ) 0.5205
where the standard error is computed from
se(3b  b )  9  vˆ
ar(b3 )  vˆ
ar(b 4 )  2  3 cˆ
ov(b ,b )
3 4 23

 9  0.019311  0.092298  6  0.00081

 0.5205

The calculated value of the F-test statistic is

(SSER  SSEU ) J (3824.793  3729.870) 1
F   5.78
SSEU (N  K ) 3729.870 227

Note that F  5.78  t 2  2.4042 . The null hypothesis is rejected. Using the t-distribution
rejection occurs because 2.404  1.970 . Using the F-distribution rejection occurs because
5.78  3.883 . In both cases the p-value is 0.017. The delay from a train is not equal to
three times the delay from a red light.

(b) This test is similar to that in part (a), but it is a one-tail test rather than a two-tail test. The
hypotheses are
H0 : 4  33 and H1 : 4  33
The rejection region for the t-test is t  t(0.05, 227)  1.652 , where the t-value is calculated
as
b4  3b3 2.7548  31.3353
t   2.404
se(b4  3b3 ) 0.5205
Since 2.404  1.652 , we reject H0 . The delay from a train is less than three times the
delay from a red light.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 136

Exercise 6.17 (continued)

(c) The delay from 3 trains is 34 . The extra time gained by leaving 5 minutes earlier is
5  52 . Thus, the hypotheses are
H0 :34  5  52 and H0 :34  5  52
The rejection region for the t-test is t  t(0.95, 227)  1.652 , where the t-value is calculated as

3b4  5b2  5 3 2.7548  5  0.36923  5

t   1.546
se(3b4  5b2 ) 0.9174
and the standard error is computed from
se(3b  5b )  9  vˆ
ar(b4 )  25  vˆ
ar(b2 )  30  cˆ
ov(b ,b )
4 2 24

 9  0.092298  25  0.000241  30  0.000165

 0.9174
Since 1.546  1.652 , we do not reject H0 at a 5% significance level. Alternatively, we do
not reject H0 because the p-value = 0.0617, which is greater than 0.05. There is
insufficient evidence to conclude that leaving 5 minutes earlier is not enough time.

(d) The expected time taken when the departure time is 7:15AM, and no red lights or trains
are encountered, is 1  45 2 . Thus, the null and alternative hypotheses are
H0 :1  452  45 and H1 : 1  452  45
The rejection region for the t-test is t  t(0.95, 227)  1.652 , where the t-value is calculated as

b1  45b2  45 19.9166  45  0.36923  45

t   7.44
se(b1  45b2 ) 1.1377
and the standard error is computed from

se(b1  45b2 )  vˆ
ar(b1 )  452  vˆ
ar(b2 )  90  cˆ
ov(b ,b
12
)

 1.574617  2025  0.00024121  90  0.00854061

 1.1377
Since 7.44  1.652 , we do not reject H0 at a 5% significance level. Alternatively, we do
not reject H0 because the p-value = 1.000, which is greater than 0.05. There is insufficient
evidence to conclude that Bill will not get to the University before 8:00AM.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 137

Exercise 6.17 (continued)

(e) The predicted time it takes Bill to reach the University is

Tˆ
IME  b  b  30  b  6  b 1  41.76
1 2 3 4

Using suitable computer software, the standard error of the forecast error can be calculated
as se( f )  4.0704 . Thus, a 95% interval estimate for the travel time is
Tˆ
IME  t se( f )  41.76  1.97  4.0704  (33.74, 49.78)
(0.975, 227)

Rounding this interval to 34 – 50 minutes, a 95% interval estimate for Bill’s arrival time is
from 7:34AM to 7:50AM.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 138

EXERCISE 6.18

(a) We are testing the null hypothesis H 0 : 2  3 against the alternative H1 : 2  3 . The test
can be performed with an F or a t statistic. Using an F-test, we reject H0 when
F  F(0.95,1,348) , where F(0.95,1,348)  3.868 . The calculated F-value is 0.342. Thus we do not
reject H0 because 0.342  3.868 . Also, the p-value of the test is 0.559, confirming non-
rejection of H0 . The hypothesis that the land and labor elasticities are equal cannot be
rejected at a 5% significance level.

Using a t-test, we reject H0 when t  t(0.975,348) or t  t(0.025,348) where t(0.975,348)  1.967 and
t(0.025,348)  1.967 . The calculated t-value is

b2  b3 0.36174  0.43285
t   0.585
se(b2  b3 ) 0.12165
In this case H0 is not rejected because 1.967  0.585  1.967 . The p-value of the test is
0.559. The hypothesis that the land and labor elasticities are equal cannot be rejected at a
5% significance level.

(b) We are testing the null hypothesis H0 : 2  3  4  1 against the alternative

H1 : 2  3  4  1 , using a 10% significance level. The test can be performed with an F
or a t statistic. Using an F-test, we reject H 0 when F  F(0.90,1,348) , where F(0.90,1,348)  2.72 .
The calculated F-value is 0.0295. Thus, we do not reject H 0 because 0.0295  2.72 . Also,
the p-value of the test is 0.864, confirming non-rejection of H0 . The hypothesis of
constant returns to scale cannot be rejected at a 10% significance level.

Using a t-test, we reject H0 when t  t(0.95,348) or t  t(0.05,348) where t(0.95,348)  1.649 and
t(0.05,348)  1.649 . The calculated t-value is

b  b  b 1 0.36174  0.43285  0.209502 1

t 2 3 4   0.172
se(b2  b3  b4 ) 0.023797
In this case H0 is not rejected because 1.649  0.172  1.649 . The p-value of the test is
0.864. The hypothesis of constant returns to scale cannot be rejected at a 10% significance
level.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 139

Exercise 6.18 (continued)

(c) In this case the null and alternative hypotheses are

2  3  0  2  3  0 and/or
H0 :      1 H1 :       1
 2 3 4  2 3 4

We reject H0 when F  F(0.95,2,348) , where F(0.95,2,348)  3.02 . The calculated F-value is

0.183. Thus, we do not reject H0 because 0.183  3.02 . Also, the p-value of the test is
0.833, confirming non-rejection of H0 . The joint null hypothesis of constant returns to
scale and equality of land and labor elasticities cannot be rejected at a 5% significance
level.

(d) The mean of log output when AREA  2 , LABOR  100 and FERT  175 is
E ln(PROD)  1  2  ln(2)  3  ln(100)  4  ln(175)
 1  0.693152  4.605173  5.164794
Thus, the null and alternative hypotheses are
H 0 : 1  0.693152  4.605173  5.164794  1.5
H1 : 1  0.693152  4.605173  5.164794  1.5
We reject H0 when F  F(0.95,1,348) , where F(0.95,1,348)  3.868 . The calculated F-value is
208. Thus, we reject H0 because 208  3.868 . Also, the p-value of the test is less than
0.0001, confirming rejection of H0 . The hypothesis that the mean of log output is equal to
1.5 when the inputs are set at the specified levels is rejected.
 Chapter 6, Exercise Solutions, Principles of Econometrics, 3e 140

EXERCISE 6.19

The results are summarized in the following table.

Full FERT LABOR AREA

model omitted omitted omitted
b2 ( AREA) 0.3617 0.4567 0.6633
b3 (LABOR) 0.4328 0.5689 0.7084
b4 (FERT ) 0.2095 0.3015 0.2682
RESET(1) p-value 0.5688 0.8771 0.4281 0.1140
RESET(2) p-value 0.2761 0.4598 0.5721 0.0083

(i) With FERT omitted the elasticity for AREA changes from 0.3617 to 0.4567, and the
elasticity for LABOR changes from 0.4328 to 0.5689. The RESET F-values (p-values) for
1 and 2 extra terms are 0.024 (0.877) and 0.779 (0.460), respectively. Omitting FERT
appears to bias the other elasticities upwards, but the omitted variable is not picked up by
the RESET test.

(ii) With LABOR omitted the elasticity for AREA changes from 0.3617 to 0.6633, and the
elasticity for FERT changes from 0.2095 to 0.3015. The RESET F-values (p-values) for 1
and 2 extra terms are 0.629 (0.428) and 0.559 (0.572), respectively. Omitting LABOR also
appears to bias the other elasticities upwards, but again the omitted variable is not picked
up by the RESET test.

(iii) With AREA omitted the elasticity for FERT changes from 0.2095 to 0.2682, and the
elasticity for LABOR changes from 0.4328 to 0.7084. The RESET F-values (p-values) for
1 and 2 extra terms are 2.511 (0.114) and 4.863 (0.008), respectively. Omitting AREA
appears to bias the other elasticities upwards, particularly that for LABOR. In this case the
omitted variable misspecification has been picked up by the RESET test with two extra
terms.