Introduction to Econometrics (4th Edition)

by

James H. Stock and Mark W. Watson

Solutions to Odd-Numbered End-of-Chapter Exercises:

Chapter 3

(This version September 14, 2018)

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 1
_____________________________________________________________________________________________________

3.1. The central limit theorem says that when the sample size ( n ) is large, the distribution
s Y2
of the sample average ( Y ) is approximately N æçè µY , s Y2 ö÷ø with s Y2 = n . Given a

population µY = 100, s Y2 = 43.0, we have

2
(a) n = 100, s Y2 = snY = 100
43
= 0.43, and

æ Y - 100 101 - 100 ö

Pr (Y < 101) = Pr ç < ÷ » F(1.525) = 0.9364.
è 0.43 0.43 ø

2
(b) n = 64, s Y2 = s64Y = 43
64 = 0.6719, and

æ 101 - 100 Y - 100 103 - 100 ö

Pr(101 < Y < 103) = Pr ç < < ÷
è 0.6719 0.6719 0.6719 ø
» F(3.6599) - F (1.2200) = 0.9999 - 0.8888 = 0.1111.

2
(c) n = 165, s Y2 = snY = 165
43
= 0.2606, and

æ Y - 100 98 - 100 ö
Pr (Y > 98) = 1 - Pr (Y £ 98) = 1 - Pr ç £ ÷
è 0.2606 0.2606 ø
» 1 - F(-3.9178) = F(3.9178) = 1.0000 (rounded to four decimal places).

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 2
_____________________________________________________________________________________________________

3.3. Denote each voter’s preference by Y, with Y = 1 if the voter prefers the incumbent
and Y = 0 if the voter prefers the challenger. Y is a Bernoulli random variable with
(Y = 1) = p (Y = 0) = 1 - p.
probability Pr and Pr From the solution to Exercise 3.2, Y
p p(1 - p).
has mean and variance

(a) pˆ = 215
400 = 0.5375.

!
(b) The estimated variance of p̂ is var( p̂) = p̂(1 − p̂)
= 0.5375 × (1 − 0.5375)
= 6.2148 × 10−4.
n 400
1
The standard error is SE ( pˆ ) = (var( pˆ )) 2 = 0.0249.

(c) The computed t-statistic is

pˆ - µ p,0 0.5375 - 0.5

t act = = = 1.506.
SE( pˆ ) 0.0249

Because of the large sample size (n = 400), we can use Equation (3.14) in the
text to compute the p-value for the test H 0 : p = 0.5 vs. H1 : p ¹ 0.5:

p-value = 2F(-|t act |) = 2F(-1.506) = 2 ´ 0.066 = 0.132.

(d) Using Equation (3.17) in the text, the p-value for the test H 0 : p = 0.5 vs.
H1 : p > 0.5 is

p-value = 1 - F(t act ) = 1 - F(1.506) = 1 - 0.934 = 0.066.

(e) Part (c) is a two-sided test and the p-value is the area in the tails of the standard
normal distribution outside ± (calculated t-statistic). Part (d) is a one-sided test
and the p-value is the area under the standard normal distribution to the right of
the calculated t-statistic.

(f) For the test H0: p = 0.5 versus H1: p > 0.5, we cannot reject the null hypothesis at
the 5% significance level. The p-value 0.066 is larger than 0.05. Equivalently the
calculated t-statistic 1.506 is less than the critical value 1.64 for a one-sided test
with a 5% significance level. The test suggests that the survey did not contain
statistically significant evidence that the incumbent was ahead of the challenger at
the time of the survey.

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 3
_____________________________________________________________________________________________________

3.5

(a) (i) The size is given by Pr(|pˆ - 0.5| > .02), where the probability is computed assuming
that p = 0.5.
Pr(|p̂ − 0.5| > .02) = 1− Pr(−0.02 ≤ p̂ − 0.5 ≤ .02)
⎛ −0.02 p̂ − 0.5 0.02 ⎞
= 1− Pr ⎜ ≤ ≤ ⎟
⎝ .5 × .5/1055 .5 × .5/1055 .5 × .5/1055 ⎠
⎛ p̂ − 0.5 ⎞
= 1− Pr ⎜ −1.30 ≤ ≤ 1.30⎟
⎝ .5 × .5/1055 ⎠
≈ 0.19

where the final equality using the central limit theorem approximation.

(ii) The power is given by Pr(|pˆ - 0.5| > .02), where the probability is computed
assuming that p = 0.53.

Pr(|p̂ − 0.5| > .02) = 1− Pr(−0.02 ≤ p̂ − 0.5 ≤ .02)

⎛ −0.02 p̂ − 0.5 0.02 ⎞
= 1− Pr ⎜ ≤ ≤ ⎟
⎝ .53× .47/1055 .53× .47/1055 .53× .47/1055 ⎠
⎛ −0.05 p̂ − 0.53 −0.01 ⎞
= 1− Pr ⎜ ≤ ≤ ⎟
⎝ .53× .47/1055 .53× .47/1055 .53× .47/1055 ⎠
⎛ p̂ − 0.53 ⎞
= 1− Pr ⎜ −3.25 ≤ ≤ −0.65⎟
⎝ .53× .47/1055 ⎠
≈ 0.74

where the final equality using the central limit theorem approximation.
0.54 - 0.5
(b) (i) t = 0.54 ´ 0.46/1055
= 2.61, Pr(|t| > 2.61) = .01, so that the null is rejected at the 5% level.

(ii) Pr(t > 2.61) = .004, so that the null is rejected at the 5% level.

(iii) 0.54 ± 1.96 0.54 ´ 0.46 /1055 = 0.54 ± 0.03, or 0.51 to 0.57.

(iv) 0.54 ± 2.58 0.54 ´ 0.46 /1055 = 0.54 ± 0.04, or 0.50 to 0.58.

(v) 0.54 ± 0.67 0.54 ´ 0.46 /1055 = 0.54 ± 0.01, or 0.53 to 0.55.

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 4
_____________________________________________________________________________________________________

3.5 (continued)

(c) (i) The probability is 0.95 is any single survey, there are 20 independent
surveys, so the probability if 0.9520 = 0.36

(ii) 95% of the 20 confidence intervals or 19.

(d) The sample size must be chosen so that 1.96×SE( p̂ ) < 0.01. Noting that SE( p̂ )
2
p(1− p)/n then n must be chosen so that n > .012 , and the value of n
1.96 p(1− p)
=

that satisties this depends on the value of p. The largest value that p(1 − p) can
take on is 0.25 (that is, p = 0.5 makes p(1 − p) as large as possible). Thus if
2
n > 1.96.01´20.25 = 9604, then the margin of error is less than 0.01 for all values of p.

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 5
_____________________________________________________________________________________________________

3.7. The null hypothesis is that the survey is a random draw from a population with p
t = pSE- (0.11
ˆ
pˆ ) ,
SE ( pˆ ) = pˆ (1 - pˆ )/n.
0.11. The t-statistic is where (An alternative
p̂ 0.11´ (1 - 0.11) / n,
formula for SE( ) is which is valid under the null hypothesis
p = 0.11).
that The value of the t-statistic is 2.71, which has a p-value that is less
p = 0.11
than 0.01. Thus the null hypothesis (the survey is unbiased) can be
rejected at the 1% level.

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 6
_____________________________________________________________________________________________________

3.9. Denote the life of a light bulb from the new process by Y . The mean of Y is µ and
Y s Y = 200 Y
the standard deviation of is hours. is the sample mean with a
n = 100. Y
sample size The standard deviation of the sampling distribution of is
sY = sn =
Y 200
= 20 H 0 : µ = 2000 H1 : µ > 2000 .
100 hours. The hypothesis test is vs.
Y > 2100
The manager will accept the alternative hypothesis if hours.

(a) The size of a test is the probability of erroneously rejecting a null hypothesis
when it is valid.

The size of the manager’s test is

size = Pr(Y > 2100|µ = 2000) = 1 - Pr(Y £ 2100|µ = 2000)

æ Y - 2000 2100 - 2000 ö
= 1 - Pr ç £ |µ = 2000 ÷
è 20 20 ø
-7
= 1 - F (5) = 1 - 0.999999713 = 2.87 ´10 .

Pr(Y > 2100|µ = 2000) means the probability that the sample mean is greater
than 2100 hours when the new process has a mean of 2000 hours.

(b) The power of a test is the probability of correctly rejecting a null hypothesis
when it is invalid. We calculate first the probability of the manager erroneously
accepting the null hypothesis when it is invalid:

æ Y - 2150 2100 - 2150 ö

b = Pr(Y £ 2100|µ = 2150) = Pr ç £ |µ = 2150 ÷
è 20 20 ø
= F(-2.5) = 1 - F(2.5) = 1 - 0.9938 = 0.0062.

The power of the manager’s testing is 1 - b = 1 - 0.0062 = 0.9938.

(c) For a test with size 5%, the rejection region for the null hypothesis contains
those values of the t-statistic exceeding 1.645.
act
- 2000
t act = Y > 1.645 Þ Y act > 2000 + 1.645 ´ 20 = 2032.9.
20

The manager should believe the inventor’s claim if the sample mean life of the
new product is greater than 2032.9 hours if she wants the size of the test to be
5%.

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 7
_____________________________________________________________________________________________________

3.11. Assume
1
that
n
n is an even number. Then Y! is constructed
3
by applying
n
a weight of
2 2 2 2
to the “odd” observations and a weight of to the remaining
observations.

1 ⎛⎜ 1 3 1 3 ⎞
E(Y! ) = ⎜ E(Y1 ) + E(Y2 ) +" E(Yn−1 ) + E(Yn ) ⎟⎟
n ⎜⎝ 2 2 2 2 ⎟⎠

1⎛ 1 n 3 n ⎞
= ⎜ ⋅ ⋅ µY + ⋅ ⋅ µY ⎟ = µY
n⎝ 2 2 2 2 ⎠
1 ⎛1 9 1 9 ⎞
var(Y! ) = 2 ⎜
var(Y1 ) + var(Y2 ) +" var(Yn−1 ) + var(Yn )⎟
n ⎝4 4 4 4 ⎠
1 ⎛ 1 n 2 9 n 2⎞ σ Y2
= ⋅ ⋅ σ + ⋅ ⋅ σ = 1.25 .
n2 ⎜⎝ 4 2 Y 4 2 Y ⎟⎠ n

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 8
_____________________________________________________________________________________________________

3.13 (a) Sample size n = 420, sample average Y = 646.2 sample standard deviation
sY = 19.5. Y (Y ) = sYn = 19420
.5
= 0.9515.
The standard error of is SE The 95%
confidence interval for the mean test score in the population is

µ = Y ±1.96SE(Y ) = 646.2 ±1.96 × 0.9515 = (644.3, 648.1).

(b) The data are: sample size for small classes n1 = 238, sample average Y 1 = 657.4,

sample standard deviation s1 = 19.4; sample size for large classes n2 = 182,

sample average Y 2 = 650.0, sample standard deviation s2 = 17.9. The standard

s12 2
error of Y1 - Y2 is SE(Y1 - Y2 ) = + ns22 = 19.42
+ 17182 = 1.8281. The hypothesis
2
.9
n1 238

tests for higher average scores in smaller classes is

H 0 : µ1 - µ2 = 0 vs. H1 : µ1 - µ2 > 0.

The t-statistic is

t act = Y 1 - Y 2 = 657.4 - 650.0 = 4.0479.

SE(Y 1 - Y 2) 1.8281

The p-value for the one-sided test is:

p-value = 1 - F(t act ) = 1 - F(4.0479) = 1 - 0.999974147 = 2.5853 ´10-5.

With the small p-value, the null hypothesis can be rejected with a high degree of
confidence. There is statistically significant evidence that the districts with
smaller classes have higher average test scores.

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 9
_____________________________________________________________________________________________________

3.15. From the textbook equation (2.45), we know that E( Y ) = µY and from (2.46) we
s Y2
Y
know that var( ) = n . In this problem, because Ya and Yb are Bernoulli random

pˆ a Ya pˆ b Yb s Ya2 s Yb2
variables, = , = , = pa(1–pa) and = pb(1–pb). The answers to (a)

follow from this.

For part (b), note that

var( pˆ a – pˆ b ) = var( pˆ a ) + var( pˆ b ) – 2cov( pˆ a , pˆ b ).

But, pˆ a and pˆ b are independent (they depend on data chosen from independent

samples), and this means that cov( pˆ a , pˆ b ) = 0. Thus var( pˆ a – pˆ b ) = var( pˆ a ) + var(

pˆ b ).

For part (c), equation (3.21) from the text (replacing Y with p̂ and using the result in
(b) to compute the SE).

For (d), apply the formula in (c) to obtain

0.859(1 - 0.859) 0.374(1 - 0.374)

95% CI is (.859 – .374) ± 1.96 + or 0.485 ±
5801 4249
0.017.

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 10
_____________________________________________________________________________________________________

3.17. (a) The 95% confidence interval is Ym, 2015 − Ym, 1996 ± 1.96 SE(Ym, 2015 − Ym, 1996 ) where
2 2
S m, S m,
SE(Ym, 2015 − Ym,1996 ) = + = 14.37 2
+ 11.44 = 0.45;
2015 1996 2

nm, 2015 nm, 1996 1917 1387

the 95% confidence interval is

(28.06 – 24.87) ± 0.88 or 3.19 ± 0.88 = ($2.31, 4.07).

(b) The 95% confidence interval is Yw, 2015 − Yw, 1996 ± 1.96 SE(Yw, 2015 − Yw, 1996 ) where

2 2
S w, S w,
SE(Yw, 2015 − Yw,1996 ) = + = 11.222
+ 8.93 = 0.37; the 95% confidence
2015 1996 2

nw, 2015 nw, 1996 1816 1232

interval is (23.04-20.99) ± 0.72 or 2.05 ± 0.72 = ($1.33, 2.77).

(c) The 95% confidence interval is

(Ym, 2015 − Ym, 1996 ) − (Yw, 2015 − Yw, 1996 ) ± 1.96 SE[(Ym, 2015 − Ym, 1996 ) − (Yw, 2015 − Yw, 1996 )],

where

(Ym, 2015 − Ym,1996 ) − (Yw, 2015 − Yw,1996 )

2 2 2 2
S m, S m, S w, S w,
= 2015
nm, 2015
+ 1996
nm, 1996
+ 2015
nw, 2015
+ 1996
nw, 1996

= 14.37 2
+ 11.44 + 11.22 + 8.93 = 0.58.
2 2 2

1917 1387 1816 1232

The 95% confidence interval is

(28.06 – 24.87) − (23.04-20.99) ± 1.96 ´ 0.58 or 1.14 ± 1.13 = ($0.01, 2.17)

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 11
_____________________________________________________________________________________________________

3.19. (a) No. E (Yi 2 ) = s Y2 + µY2 and E (YY

i j ) = µY for i ¹ j.Thus
2

2
æ1 n ö 1 n 1 n 1 2
E (Y ) = E ç å Yi ÷ = 2 å E (Yi 2 ) + 2 åå E (YY
2
i j ) = µY +
2
sY
è n i =1 ø n i =1 n i =1 j ¹i n

(b) Yes. If Y gets arbitrarily close to µY with probability approaching 1 as n gets

large, then Y 2 gets arbitrarily close to µY2 with probability approaching 1 as n

gets large. (As it turns out, this is an example of the “continuous mapping
theorem” discussed in Chapter 18.)

©2018 Pearson Education, Inc.

 Stock/Watson - Introduction to Econometrics 4th Edition - Answers to Exercises: Chapter 3 12
_____________________________________________________________________________________________________

3.21. Set nm = nw = n, and use equation (3.19) write the squared SE of Ym - Yw as

1 1
åin=1 (Ymi - Ym ) 2 åin=1 (Ywi - Yw ) 2
(n - 1) (n - 1)
[ SE (Ym - Yw )]2 = +
n n

åin=1 (Ymi - Ym ) 2 + åin=1 (Ywi - Yw ) 2

= .
n(n - 1)

Similary, using equation (3.23)

1
⎡⎣ ∑ i=1
n
(Ymi − Ym )2 + ∑ i=1
n
(Ywi − Yw )2 ⎤⎦
2(n − 1)
[SE pooled (Ym − Yw )]2 =
2n

∑ i=1
n
(Ymi − Ym )2 + ∑ i=1
n
(Ywi − Yw )2
= .
n(n − 1)

©2018 Pearson Education, Inc.