You are on page 1of 6

Tuesday, January 01, 2013, at 10:13 AM

Errata for Principles of Econometrics, 4e


First Printing
Page

Date

Correction

68

10-feb-11

Below equation (2.25), second line: the sentence should read In the
linear specification the expected price per additional square foot is
constant.

69

31-mar-11

Below equation (2.26), first line: x = SQFT 2. Thanks to


Michelle Savolainen, Louisiana State University.

70

18-apr-11

In footnote 4: See Appendix A.3.1, Derivative Rule 7. Thanks to


Genevieve Briand, Washington State University.

74

10-feb-11

Figure 2.18, top panel: The legend should read House Prices
($1,000) in Golden Oaks

79

17-feb-11

86

1-jan-13

Exercise 2.10: In the original capm data files, in all formats, the risk
free rate was based on 30-day T-bill, but it was an annual yield. It
has been, effective February 17, 2011, has been changed to the
correct monthly rate. The definition file capm.def also explains
this. Thanks to Jeffrey Parker, and his students, at Reed College for
pointing this out.
In the derivation of the variance of b2 the first four lines should be

var(b2 ) E 2 wi ei 2
E wi ei

E wi2 ei2 wi w j ei e j
(square of bracketed term)
i j

= wi2 E ( ei2 ) wi w j E ( ei e j ) (because wi is not random)


i j

Note that in the text there is a redundant 2 in the 3rd and 4th lines.
116

4-mar-11

In line 1, 1 2 should be 1/2.

116

4-mar-11

In line 4, a 100(1)%...

118

4-mar-11

In line 2, or H0: 1 2 20 250 0

120

4-jan-11

Exercise 3.6, line 10. H 0 : 2 0 should be H1 : 2 0

121

17-feb-11

Exercise 3.7: In the original capm data files, in all formats, the risk
free rate was based on 30-day T-bill, but it was an annual yield. It
has been, effective February 17, 2011, has been changed to the
correct monthly rate. The definition file capm.def also explains
this. Thanks to Jeffrey Parker, and his students, at Reed College for
pointing this out.

123

4-jan-11

Exercise 3.13, line 3. The model is said to be log-linear because the


right-hand side is linear in the parameters, even though it has a
quadratic term on the right-hand side. We should have perhaps said
more simply that it is a logarithmic model

130

16-june-11

In line 9 of first paragraph. the difference between the predicted


value of y and the actual value of y.. would be better worded
the difference between the actual value of y and the predicted
value of y.. Thanks to Michelle Savolainen, Louisiana State
University

141

4-mar-11

In line 5, above Figure 4.4: Using just these two algebraic .

147

4-mar-11

In line 4 of 2nd paragraph: dependence in the regression errors.

151

4-mar-11

In line 5: in addition to TIME3.

160

24-oct-12

4.10 part (e): total income (180) should be total expenditure


(180). Thanks to Lee Adkins, Oklahoma State University.

163

29-sep-12

In line 5: such as adding additional explanatory

163

29-sep-12

Line below the first equation: Subtract the term and add the
same term

182

2-may-11
13-oct-11

In line 6: The interval estimate is = (10.093, 5.723), and in the


upper bound calculation the point estimate should be 7.9079.
Thanks to Sandamali Kankanamge, Louisiana State University and
Diana Whistler, University of British Columbia.

184

18-feb-11

b1 , b2 ) should be cov(
b2 , b3 )
In expression for standard error cov(

189

5-may-11

Under point 4, in the second line of the equation, under the square

b , b ) should be cov(
b , b ) . Thanks to Joe Goss,
root sign, cov(
1

Louisiana State University.


197

28-nov-11

In line 4. the approximate increase in wages from an extra year


of experience is 100 Thanks to Alex James, Louisiana State
University

203

18-feb-11

Exercise 5.8(c): Change (iv), (v), (vi) to (i), (ii), (iii)

217

18-feb-11

First equation: delete last parenthesis

225

18-feb-11

Equation (6.5): In this equation the terms of the regression are


written as xkk rather than as kxk as in Equation (5.3). The two
terms are equivalent

228

18-feb-11

In the equation: Change ADVER to ADVERT.

234

2-may-11

In equation (6.21): the sign of the intercept is positive. That is


FAMINC 26191 5.155 HEDU

239

18-feb-11

First line in next-to-last paragraph: Change (6.26) and (6.27) to


(6.29) and (6.30)

250

29-jan-11

Exercise 6.15(c), line 4. price will be more than $20,000.

251

29-jan-11

Exercise 6.17(g), line 2: "increased the price of the house by more


than $20,000."

251

31-oct-11

Exercise 6.18 (a) and (c): Replace Report the estimated relationship
between ln(SPRICE), LIVAREA and AGE for two- three- and fourbedroom houses with: Using the estimates from this model, find
the regression functions for two- three- and four-bedroom
houses.

251

18-feb-11

In the equation in Exercise 6.19 DEPARTS should be DEPART.

251

18-feb-11

Below equation in Exercise 6.19: Find 95% interval estimates


Delete a

252

18-feb-11

In Exercise 6.22(c) and (e): Replace INC with INCOME

284

10-feb-11

Equation (7.20): The alternative hypothesis is H1 : 0 . The < 0


is missing in the text.

295

29-jan-11

Exercise 7.15, line 2. Using the equation estimated in part (c),


predict

297

29-jan-11

The running page head should read Appendix 7B The Differences


Estimator

297

29-jan-11

The heading for Appendix 7B should be Derivation of the


Differences Estimator

297

29-jan-11

line 5. To verify the expression for the differences estimator

315

18-feb-11

Label second equation as (8.36)

320

13-oct-11

In Table 8.1, the last digit in three of the four LS-robust standard
errors should be increased by 1. The correct ones are: (0.0653) for C;
(0.0604) for PRATIO; (0.0344) for DISP_PEPSI. Thanks to Diana
Whistler, University of British Columbia.

325

16-jul-11

In Exercise 8.9(a) should read as the true parameter values,

328

13-mar-11

In each of the last 3 equations, the coefficient of DISP _ PEPSI


not 3 .
should be

346

13-oct-11

In the upper part of Table 9.2, the t-value for Gt should be 6.120.
Thanks to Diana Whistler, University of British Columbia.

349

13-oct-11

Towards the bottom, the calculation for Z 2 should use 0.411, not
0.414, and the result should be 4.07 not 4.10.
Thus, the correct version is Z 2 98 0.411 4.07 . Thanks to
Diana Whistler, University of British Columbia.

383

2-aug-11

Exercise 9.2(b). The first line on this page should read Using a twotail test and

383

19-jul-11

In Exercise 9.3, after the first sentence add the sentence: The
amounts spent on advertising in the last two weeks were
ADV104 1.313 and ADV105 1.358 .

383

19-jul-11

In Exercise 9.3(b), after the first sentence add the sentence: The
estimated error variance is 2 2.3891 .

384

2-aug-11

In Exercise 9.6, line 7, Delete: (stored in the file homes.dat).

385

22-aug-11

In Exercise 9.8, the first line should read: In Section 9.6, not 9.1.

385

2-apr-11

In Exercise 9.9(a) 1 00 should be 1 01 .

387

9-mar-11

Exercise 9.13(e) Change by at least $6 million to by more than $6


million.

397

10-feb-12

Above equation (9C.1), the equation reference should be (9.43) not


(9.44). Thanks to Erdogan Cevher.

398

10-feb-12

Below equation (9C.4), b1 and b1 should be b1 and b2 . Thanks to


Erdogan Cevher.

423

18-oct-12

In Exercise 10.1 the variable name MDHOUSE in the table of results


should be MEDHOUSE. And in part (b), the regions of the country
should be REG2, REG3 and REG4. Thanks to Ethan Luedecke and
Max O'Krepki, Louisiana State University.

428

18-oct-12

In Exercise 10.9 there should be a parameter 5 before


ln QPRODt 1

435

4-jan-11

Equation (10E.1) should have a multiplication sign in the middle.


That is, it should read
Cragg Donald F N G B L rB2 1 rB2 . Thanks to
Deokrye Baek, Louisiana State Univ.

480

16-dec-12

Two lines below equation (12.2c): described by this model


with 0, 1, and 0.01 the delta=0.01 in the text is
replaced by lambda=0.01. Thanks to Erdogan Cevher, The
Ministry of Science, Industry and Technology of Turkey

487

16-jul-11

In the Dickey-Fuller test equation for B, change the coefficient of


Bt 1 from 0.237 to 0.290. Thanks to Genevieve Briand,
Washington State Univ.

503

12-Sep-12

In equation (13.9) a minus sign was omitted from the first t-value. It
should read 2.077 . Thanks to Erdogan Cevher.

507

29-sep-12

Line 2: Change subscript on v from t to 1 so that it reads x1 v1x x

520

4-jan-11

The Y axis of Figure 14.1(a) is labeled as starting at 12 instead of


30. Thanks to Genevieve Briand, Washington State Univ.

521

16-Aug-11

The Y axis in Figure 14.2(d) should be from 0.00 to 0.07 and not
from 0.00 to 0.09. Thanks to Randall Campbell, Mississippi State
University.

550

24-may-11

The following changes should be made in Table 15.7 under Cluster


Robust Standard Errors. For C: Change 0.06153 to 0.06149, and the
t-value to 23.58. For SOUTH: Change 0.06539 to 0.06535. For
UNION: Change 0.01885 to 0.01884.

550

4-jan-11

In footnote 9, the first sentence should read The cluster robust


standard errors are computed with the degrees of freedom correction
in EViews 7.0.

557

3-dec-12

In the first line, TENURE2 should be EXPER2. Thanks to Brad


Lewis, Louisiana State University

560

1-apr-11

Equation (15.38). The parameters should be


symbol appears twice

. Currently the

561

1-apr-11

Equation (15.39). The parameters should be


symbol appears twice

. Currently the

569

1-apr-11

Lines 3 & 4: In two places the value 0.729 should be replaced by


0.5314

569

1-apr-11

Line 4: The value 0.854 should be replaced by 0.729

574

13-oct-11

Exercise 15.8, first line: The data are discussed in Section 15.1.

575

28-aug-11

Exercise 15.8(h) should read: Using the differenced data in part (f),
estimate the wage equation in part (g), but

575

13-oct-11

Exercise 15.9, first line: The data are discussed in Section 15.1.

578

13-oct-11

Line -4: The value for 2 should be 6328.437.

620

7-oct-11

653

19-jan-11

In the middle of the page. The scale factor and marginal effect were
calculated at AGE = 42.5 rather than the correct AGE = 42.54. The
0.3630 .
revised text should read, The calculated scale factor is
Thus the marginal effect on observed hours of work of another year
of education is
E HOURS
2 73.29 0.3630 26.61
EDUC
Thanks to Diana Whistler, University of British Columbia.
Exercise A.1(b): "x =10 and at x = 50" should be "P = 10 and at P =
50"

654

2-feb-12

Exercise A.4(c) should read Using the results in parts (a) and (b),
express the . Thanks to Micah West and Deborah Williams,
Louisiana State University.

671

15-dec-12

In line 9. The second part of the statement is incorrect. It should


read The conditional variance in this example is smaller than the
unconditional variance. In general the conditional variance does
not have to be smaller than the unconditional one. In the example let
X 9 10 then the conditional variance is 0.07979 which is larger
than the unconditional variance.
What is true is if X and Y are jointly normal, then
var Y | X var Y 1 2 , where is the correlation between X
and Y. If 0 then the normal variables are also independent so the
original statement stands.
A more general statement is based on what William Greene (2012, p
1036) calls a Decomposition of Variance. This rule says that
var Y E X var Y | X varX E Y | X
The subscript X on the expectation and variance operators
indicates that those are the expectation and variance of the
expression treating X as random. For an elegant explanation of this
result see A thing of beauty, by Tony O'Hagan in Significance,
Volume 9, Issue 3, June 12, 2012, pp. 26-28.

709

5-may11

Inside rear
cover: left
side

31-mar-11

Inside rear
cover: left
side

31-mar-11

Thanks to Sergio Pastorello, Dipartimento di Scienze Economiche,


Universit di Bologna, ITALY
In line 2, nought should be naught, though they have equivalent
meanings. Thanks to Julie Leiby, Louisiana State University.
Under Prediction: the forecast error should be f y0 y0 . Thanks
to Michelle Savolainen, Louisiana State University
Under Log-Linear Model: the generalized goodness of fit measure
can be computed either as Rg2 corr y, y n or
2

Rg2 corr y, y c . These values are equal since the natural and
corrected predictors differ only by a scalar. Noted by Michelle
Savolainen, Louisiana State University
2

You might also like