Econometric Analysis, 4th Edition
ERRATA
Last updated 10/7/01
The reader will note a large number of corrections below dated May 7, 2000. These
were compiled by Professors Manabu Asai (Ritsumeikan University), Teruo Nakatsuma (Keio University),
Tsunemasa Shiba (Hitotsubashi University), and Toshiyuki Takahashi (Shizuoka Prefectural Government Office)
as part of their work during a translation of Econometric Analysis to Japanese.
I would like to thank them for their painstaking, diligent efforts in
collecting these items for us.
| Page Number | Date Posted | Errata | |
| All on front cover, left side | all 7/16/99 |
|
|
| All on front cover, right side | all 7/16/99 |
|
|
| xiii | 7/16/99 | In 6.6.8, e should be bold | |
| xiv | 7/16/99 | In 9.3.1, b should be bold | |
| xiv | 7/16/99 | In 9.3.1, b should be bold | |
| xv | 7/16/99 | In 10.3.1, Sqares should be Squares | |
| xvi | 7/16/99 | W should be bold in 11.4, 11.4.2, 12.4, 12.5. | |
| xvii | 7/16/99 | W should be bold in 13.6, 13.7, 14.4.2 | |
| xxiii | 7/16/99 | Line 11, "single linear equation" should be "single equation linear" | |
| xxiii | 7/16/99 | Last paragraph, "already run to over 3000 ... unnecessary." should read "and Volumes I (Macroeconomics) and II (Microeconomics) of theHandbook of Applied Econometrics already run to over 4000 pages, it is also unnecessary." | |
| xxiii | 7/16/99 | Second to last line, wisdon should be wisdom | |
| xxv | 7/16/99 | Before William Lott, should appear "Linan Liu, New York University;" | |
| 24 | 7/7/00 | At the beginning of section 2.4.6, [a b] should be [a1 a2] | |
| 30 | 8/22/01 | In the first equation at the top of the page, the subscript on a after the summation sign should be "i" not "j." | |
| 35 | 8/22/01 | Five lines up from the beginning of Section 2.6.5, "lower right of" should be "lower right corner of". | |
| 37 | 5/7/00 | In the second line, "lower right" should be "lower left". | |
| 50 and 51 | 5/7/00 | The approximately equals sign used in (2-121) should be used as well in (2-122), (2-123), and (2-127). | |
| 51 | 5/7/00 | In the expression for b2, there should be no leading minus sign, and the "+" sign should be "-". | |
| 58 | 5/7/00 | At the top of the page in the equations, three occurrences of J should not be italic. | |
| 60 | 5/7/00 | First line, the partial derivative signs should not be bold. | |
| 62 | 5/7/00 | In the first line of Section 3.2.2, two ocurrences of x should be X. | |
| 63 | 1/21/00 | In the definition of the median, in Section 3.3, Prob(X > m) < 1/2 should be Prob(X > m) > 1/2. | |
| 67 | 3/8/00 | Two lines up from the last bullet point, the reference to Table 4 should be to Table 3. | |
| 69 | 1/3/00 | In equation (3-39), x < 0 should be x > 0. | |
| 70 | 9/22/99 | In equation (3-40), "x/c" and "1-x/c" should both be enclosed in parentheses, and the exponent in each case should be outside the parentheses. | |
| 71 | 1/24/00 | In the last two equations on the page, both occurrences of "(yb)" should be "(y < b)" | |
| 73 | 5/7/00 | In (3-44) under the summations, Xx should be X < x and Yy should be Y < y | |
| 76 | 3/8/00 | In the second to last equation on the page, two occurrences of (2p)1 should be (2p)-1 | |
| 77 | 3/8/00 | In Example 3.5, 0 > x > 1 should be 0 < x < 1. | |
| 77 | 8/29/01 | In the equation before Section 3.7, in the integral, dy2 should be Jdy2 | |
| 78 | 5/7/00 | In (3-62), there should be a large left brace, "{" after the equals sign. | |
| 78 | 5/7/00 | In the first equation in Example 3.5, y ≤ 0 should be y ≥ 0. | |
| 85 | 5/7/00 | In the line above (3-87), there should be a prime after aj. | |
| 88 | 5/7/00 | In the first sentence after (3-103), "does not exist." should be modified to say "does not exist in the in the full dimensional space of x though it does exist in the subspace of dimension equal to the rank of S." | |
| 92 | 5/7/00 | In the denominator of the definition of t[n-1], there should be large braces around the entire expression not including the superscript 1/2. | |
| 96 | 8/18/01 | In part c. of Exercise 30, there should not be a minus sign in the exponent of the moment generating function for the normal distribution. The MGF is Mz(t)=exp(t2/2). | |
| 97 | 5/7/00 | In the first line of Section 4.2.1, there should not be a comma before "is a." | |
| 105 | 5/7/00 | In the last line of Example 4.4, the "10 percent" should be "1.2 percent." | |
| 110 | 8/16/01 | Eight lines up from the end of the page, In "Now to prove Theorem 4.4" 4.4 should be 4.3. | |
| 114 | 7/16/99 | After Definition 4.9, "The limiting..." should not begin a new paragraph | |
| 114 | 8/16/01 | In the first line of Definition 4.9, F(x) should be F(xn) and in the second line, cdf of x should be cdf of xn. | |
| 122 | 8/16/01 |
Definitions 4.13 and 4.14, should read
DEFINITION 4.13: Order nd . A sequence cn is of order nd, denoted O(nd), if and only if plim(1/nd) cn is a finite nonzero constant.. DEFINITION 4.13: Order less than nd . A sequence cn is of order less than nd, denoted o(nd), if and only if plim(1/nd) cn equals zero.. In the next sentence, O(1/n) should be O(n-1), O(1/n2) should be O(n-2), o(1/n) should be o(n-1), -2 should be +2, and -1 should be +1. Five lines later, O(1/n) should be O(n-1), O(1/n2) should be O(n-2), O(1/n3) should be O(n-3), O(1/n) should be O(n-1). |
|
| 134 | 1/21/00 | There is a typographical error in the first line of item 1. "l13ikelihood" should be "likelihood" | |
| 135 | 8/16/01 | In the ninth line of Example 4.23, "The models" should read "Otherwise, the models" | |
| 135 | 5/25/00 | In example 4.23, 5 lines up from the end of the page, "3.4.7" should be "3.4.8" and "3.6 through 3.8" should be "3.6 and 3.7." | |
| 136 | 8/18/01 | In the second and third expressions on the page, the three terms should be enclosed together in square brackets. | |
| 136 | 8/18/01 | In the second to last expression on the page, the y2i should be yi2. The subscripts are reversed. | |
| 138 | 8/16/01 | In the second line of Section 4.7.1, E[xk] should be E[x2k]. | |
| 139 | 8/18/01 | Eight lines up from the end of the page, "Theorem 4.3" should be "Theorem 4.6 and the corollary (4-20)" | |
| 141 | 9/8/01 | In the pairs of results given in the continuation of Example 4.26, the first pair in the third line should be changed from (2.607765,0.080441) to (2.60905,0.080475). In the pair given as th emaximum likelihood estimates at the end of the example, 2.4097 should be 2.4106. | |
| 167 | 8/25/01 | In Section 5.3.2, in the 4th line after Step 4, "The set 32 bits" should be "The set of 32 bits." | |
| 169 | 2/2/01 | In the first equation on the page, the expression given there (large fraction) should be T minus that same expression. In the first line of text below this equation, the expression T = ln(1/H2) the correct expression is the square root of this logarithm. | |
| 170 | 8/1/99 | Footnote 8, Chib and Greenberg (1995)should be Chib and Greenberg (1996) | |
| 170 | 8/1/99 | Footnote 9, Zellner and Min (1992) should be Zellner and Min (1995) | |
| 170 | 5/7/00 | Two lines up from the list of 3 items, f(y|x1) should be f(y|x0) | |
| 172 | 1/9/00 |
A computational problem appears to have distorted the values in Table 5.1. Since it is a simulation, there is no right answer, but the values there are implausible. The problem was analyzed further by Roger Koenker of the University of Illinois, who reports the following very plausible results for his run of the simulation with the parameters given in the margins: Degrees of Freedom
Observations 3 6 10
10 0.5152 0.8070 1.3388
25 0.5528 1.3426 1.6529
100 0.6133 1.5548 1.3190
In addition, the statement "As the sample size increases, the sample mean becomes the better estimator." is, in fact, not correct. Whether or not the mean is the better estimator is dependent only on the degrees of freedom parameter, not on the sample size. With 3 degrees of freedom, the ratio will converge to its true value of .62 as n grows to infinity. The sentence in quotes above should be deleted. The next sentence should begin with "As" and the "once again" should be deleted. |
|
| 175 | 5/7/00 | 14 lines up from the bottom of the page, "conversion" should be "convergence." | |
| 175 | 5/7/00 | 14 lines up from the bottom of the page, "conversion" should be "convergence." | |
| 177 | 9/20/01 | In the recursion in the center of the page, x after the equals sign should be -x and the next plus sign should be a minus sign. (See exercise 18 on page 94.) | |
| 179 | 8/29/01 | In the second equation on the page, after the integral sign, "(lnP)" should be "(lnt)" | |
| 179 | 5/7/00 | In the expression for F(x) at mid page, the limits of integration should be L(x) and U(x) and "f(x)dxshould be "f(t)dt | |
| 179 | 5/7/00 | In footnote 14, "int(P)" should be "INT(P)." | |
| 180 | 7/7/00 | In the last line of Example 5.4, "Hermite" should be "Laguerre." | |
| 182 | 5/7/00 | Third line, "(1973b)" should be "(1973)." | |
| 183 | 5/7/00 | Second to last line of Example 5.7, 1.648 should be 1.649. | |
| 194 | 8/25/01 | In the expression that appears just above (5-25), the "Prob[yi|" should be "Prob[yi = 1|". | |
| 196 | 7/16/99 | The line with bold subheader "The Lagrange Multiplier Statistic," the -H at the end of the line should be (-H)-1. | |
| 196 | 5/7/00 | In the numerator of the expression after (5-26), f(xi,q) should be f(yi|xi,q) | |
| 196 | 5/7/00 | In the second to last line of the section on Covariance Matrix Estimation, "positive" should be "nonnegative." | |
| 197 | 5/7/00 | First equation in Example 5.9, "log" should be "ln." | |
| 200 | 5/7/00 | First line of Section 5.6, "two" should be "several." | |
| 202 | 5/7/00 | 5 lines up from the bottom of the page, "(4,1)" should be followed by a comma. | |
| 203 | 5/7/00 | 6 lines up from the end of the page, the reference to Figure 5.5 should be to 5.6. | |
| 205 | 5/7/00 | In the first equation on the page, three occurrence of "log" should be "ln." | |
| 205 | 5/7/00 | In the three lines which precede the equation before the list of 4 steps at the end of the page, three occurrences of g should be gt. I.e., make bold and add subscript. | |
| 205 | 7/16/99 | 13 lines up from bottom, the -1/2(Z'Z) is mistypeset; the minus sign is at too low on the line. | |
| 206 | 5/7/00 | Fourth line, "log" should be "ln." | |
| 207 | 5/7/00 | Lines 3, 7, and 8, 5 occurrences of "log" should be "ln." | |
| 219 | 5/7/00 | In the third equation on the page, Ex should be EX. | |
| 219 | 5/7/00 | In the line after the third equation, three occurrences of e should not be bold. | |
| 229 | 5/7/00 | In the discussion before the normal equations, there should not be a comma between "time" and "trend." | |
| 248 | 5/7/00 | In the line after (6-42), Ex should beEX. | |
| 249 | 3/2/00 | In footnote 17, 3.10.3a should be 3.10.3. | |
| 257 | 5/7/00 | In the center line of (6-51), the subscript (x) should be (k). | |
| 261 | 7/16/99 | Third equation, there should be a prime before b inside the square brackets. | |
| 263 | 3/8/00 | In footnore 32, 8.4 should be 8.5. | |
| 273 | 5/7/00 | In the first line of Example 7.1, "chapter" shoulld be "section." | |
| 280 | 5/7/00 | In the 2nd, 3rd, and 4th equations, the first (Rb-q) in each quadratic form should be followed by a prime. | |
| 292 | 5/7/00 | In the paragraph that begins with "There are direct..." Professor Asai, et al have raised the possibility that this paragraph may be a bit misleading in that the size of the test described here may differ from what we suggest. This would depend on the alternative hypothesis, of course, but the point is well taken. | |
| 294 | 6/6/00 | In the second line after the second equation, 1/T should be T. (Multiply, don't divide.) Later in the same paragraph, "1.01 for K=1, 1.68 for K=5, 3.54 for K=15, and 4.52 for K=20." should be revised to "1.01 for K=2, 1.90 for K=6, 3.75 for K=15 and 4.52 for K=19." | |
| 297 | 6/6/00 | In Example 7.13, 62.0968 should be 1.7249, "is consistent" should be "is not consistent" "3.54" should be "1.90," and "is, once again, rejected" should be "is now not rejected." The second sentence of the next paragraph might now be replaced with "The results here are more or less consistent with the preceding results for the CUSUM of squares test but differ from those for the CUSUM test." | |
| 301 | 5/7/00 | In the first line of Section 7.10.1, H1 should be H0 and H0 should be H1. | |
| 303 | 10/15/01 | Line 6 after the second equation, "a constant Y" should be "a constant, Y". There is a comma missing. | |
| 303 | 10/3/00 | In Example 7.15, the values given, 2.544, 4.28, -19.24, -3.25, should be 2.58, 4.30, -21.25, -3.29, respectively. | |
| 305 | 10/3/00 | In the equation on the last line of the page, in the center term, (e'X.ZXeX.ZX/sZX2)2 should be (e'X.ZXeX.ZX)/sZX4. The numerator should not be included inside the parentheses. | |
| 306 | 5/7/00 | Lower half of the page, two occurrences of "Schwartz" should be "Schwarz." | |
| 307 | 5/7/00 | In the expression for the forecast interval, in "e0" the "e" should be italic. | |
| 309 | 5/7/00 | First line of the second bullet point, "T" should be "n." | |
| 314 | 5/7/00 | In Exercise 15, in the second of the three restrictions, dd should be dn | |
| 322 | 5/23/00 | In the first line of the page, qshould be q' (i.e., transposed). | |
| 322 | 5/23/00 | In the row 4, column 3 position of the matrix at the top of the page, b22 should be b32. I.e., the subscript is incorrect. | |
| 322 | 5/7/00 | After the second equation of Section 8.2.6, "element by element" should be "element by element (Hadamard, or direct)." | |
| 333 | 5/7/00 | Second to last equation, there should be a "1 - " immediately after the equals sign. | |
| 333 | 5/7/00 | In footnote 17, x should be X. | |
| 347 | 5/7/00 | First line, "1961b" should be "1961." | |
| 348 | 5/7/00 | 5 lines below equation (1), g < 0 should be g > 0 and g = 0 should be g < 0. | |
| 356 | 7/16/99 | Header for Section 3.3, "(b)" should be "[b]" | |
| 364 | 5/7/00 | Last sentence of Theorem 9.4, (X'X) should be (X'X)-1. | |
| 367 | 5/7/00 | In footnote 18, third line, ">=" should be ">." | |
| 373 | 3/27/00 | In Equation (9-24), s2 should be s2/n | |
| 382 | 5/7/00 | 2 lines after second equation, "inocuous" should be "innocuous." | |
| 394 | 10/20/00 | In the last equation on the page, the term "-log2/p" should be "+(1/2)log2/p" | |
| 395 | 3/15/01 | In the first equation on the page, the expression in large parentheses should be (-ei/sv - qsv). | |
| 396 | 7/16/99 | Line 5, "of the function" should be "of the production function" | |
| 400 | 1/6/00 | In the definition of the pdf for the Laplace distribution, the 2 in the exponent must be deleted. | |
| 400 | 1/9/00 | Roger Koenker (see the item for Page 172 above) questions the expression for the asymptotic covariance matrix of
bq given near the end of the page. For the case of independent, identically distributed
disturbances, he suggests Asy.Var.[ bq] = [q(1-q)/(f(F-1(q)))2(X'X)-1. The expression given in the text is the estimator, whereas Koenker gives the actual asymptotic result. The one given will converge to Koenker's in any event, but as such, it makes sense to use the asymptotic result straight away. The estimator given has no intrinsic virtue in its own right. Finally, there is a question as to what is meant by "easier to compute." With modern software and hardware, the LAD estimator is not appreciably more difficult, nor much more time consuming to compute than least squares, but in this instance, much depends on the programmer and how the computation is being done. Suffice to say that reasonable observers will differ on this question. |
|
| 407 | 5/7/00 | In the second equation on the page, the first x should be X. | |
| 408 | 5/7/00 | In (9-70) and the equation which follows it, three occurrences of S should be boldface. (The fourth is already.) In the equation for F which follows (9-70), there is a stray "+" in the first superscript which should be deleted. | |
| 408 | 5/7/00 | Mid page, the "This matrix is..." should be "This vector is..." | |
| 409 | 8/13/01 | Equation (9-77) should be "j=n+m." | |
| 410 | 5/7/00 | Last equation on the page, "+" should be "=." | |
| 411 | 5/7/00 | In the equation for B01, the H must be preceded by "X," in both numerator and denominator. (Condition on X.) | |
| 411 | 5/7/00 | Just before Example 9.12, "Bayes's" should be "Bayes." | |
| 412 | 5/7/00 | In the line after the second equation on the page, "Bayes's" should be "Bayes." | |
| 419 | 5/7/00 | In (10-9) and two lines up from (10-10), all occurrences of b should be b0. | |
| 419 | 8/29/01 | In the fifth line on the page, "same" should be "same one." | |
| 421 | 5/7/00 | In the first two lines of the paragraph which begins "As usual...", e should be e0 in both occurrences. | |
| 426 | 5/7/00 | In the third line and the line before (10-19), "(10-26)" should be "(10-16)." | |
| 427 | 5/7/00 | In the left margin headers of Table 10.2, "log" should be "ln." | |
| 428 | 5/7/00 |
The results given on pages 428 and 429 were based on an earlier set of regression results for this model,
obtained before a correction to the data set. The regression results shown on page 427 are correct. The
results on pages 428 and 429 must be changed as follows:
|
|
| 429 | 5/7/00 | In Figure 10.1, "KN" should be "KS." | |
| 432 | 7/16/99 | 7th line of Example 10.6, "and instrumental variables" should be "and instrumental variables estimates" | |
| 436 | 7/16/99 | Line 2, xi0 should be bold, not italic. | |
| 436 | 7/16/99 | Second line after the expression for V-hatc, the w in E[zi|wi] should be bold, not italic. | |
| 443 | 5/7/00 | In the second line of the second paragraph of Example 10.9, "10.4" should be "10.6." | |
| 443 | 10/15/01 | First line of last paragraph, "Regressions of M on a constant r and Y, and ln M on a constant ln r and ln Y," should be "Regressions of M on a constant, r and Y, and ln M on a constant, ln r and ln Y," There are commas missing before r and ln r. | |
| 448 | 5/7/00 | In (10-52), ths subscript on x should be ik not k. | |
| 450 | 5/7/00 | First line after second equation, "ln y" should be "y." In the line before the second to last equation on the page, "y" should be "lny" and "x" should be "lnx." | |
| 451 | 5/7/00 | Line after the first equation, "k" should be "K." | |
| 451 | 5/7/00 | Third equation, "log M" should be "ln M." | |
| 451 | 10/3/00 | In the second column of table 10.7, 0.0053504 should be 0.00636. | |
| 452 | 5/7/00 | Line before the last equation on the page, "(e)" should be "[e]." | |
| 452 | 5/7/00 | Last equation on the page, the numerator of the fraction should be "y-hatl-1." | |
| 464 | 4/3/00 | At mid page, in the definition of the Newey-West estimator, wl should equal 1-l/(L+1). | |
| 474 | 5/7/00 | Second equation, there should be a caret (hat) above the second s. | |
| 480 | 5/7/00 | Second to last equation, in the matrix, lower case "p" should be upper case. | |
| 480 | 9/8/01 | In the 4th line, 2.106 should be 2.4106. | |
| 483 | 5/7/00 | Second equation on the page, "[J-K]" should be "[L-K]" | |
| 488 | 3/2/00 | In the expression for bGMM at the top of the page, X'y should be Z'y. | |
| 489 | 7/16/99 | First line of last paragraph, b should be bold. | |
| 491 | 5/7/00 | Line 11 of the paragraph that begins "One might wonder..." f(v)dv should be f(vi)dvi. | |
| 493 | 8/22/01 | In the first equation on the page, there should be bars over the ms in the two occurrences of m(c0). | |
| 496 | 7/16/99 | Exercise 4, 3rd line, "of previous" should be "of the previous" | |
| 499 | 5/7/00 | Second equation, "E[" should be "E[" (italic E). | |
| 503 | 5/7/00 | In (12-2), the third equals sign should be preceded by "and plim k" | |
| 510 | 9/21/99 | Second line on the page, (e'e/n)-1 should be (e'e/n)-1. The -1 should not be a superscript | |
| 510 | 10/30/00 | Next to last line of the description of White's test, the value 21.03 should be 22.36. | |
| 511 | 5/7/00 | Third line of Section 12.3.4, "12.6" should be "11.2." | |
| 511 | 10/20/00 | At the end of the example, just before Section 12.3.4, the text "the F statistics are 3.2432, 6.9346, and 12.0677, respectively, which are all larger than the 95 percent critical value. The Wald statistic can be computed as 2F in each case." must be replaced with only "the Wald statistics are 3.5763, 13.3057, and 19.2226, respectively. The second and third of these are larger than the 95 percent critical value of 5.99 from the chi-squared table for two degrees of freedom." | |
| 515 | 7/16/99 | Second line inside box, "on log zi" should be "on (1,log zi)" | |
| 515 | 10/15/10 | Table 12.3, 4th row heading, the si2 after the equals sign should not have a subscript i. | |
| 515 | 5/7/00 | In the 5th line of Example 12.7, za should be zia. | |
| 515 | 10/15/01 | Table 12.3. 4th, 5th and 6th model headings, "logIncome" should be "Income." The models use the data in levels, not logs. | |
| 516 | 5/7/00 | 11 occurrences of "log" should be "ln." | |
| 517 | 5/7/00 | 7 occurrences of "log" in the equations should be "ln." | |
| 519 | 5/7/00 | In the partitioned matrix in the center of the page, the prime on 0 should be on the lower left element, not the upper right. | |
| 520 | 5/7/00 | In the discussion of the Likelihood Ratio Test, 4 occurrences of "log" in the equations should be "ln." | |
| 520 | 5/7/00 | In item 4 in the list, "If dt..." should be "If dt+1..." | |
| 520 | 7/16/99 | First line of LM test part of Section 12.5.3, 0 after equals sign should be bold. | |
| 524 | 5/7/00 | In exercise 9, in lines 3 and 6, q should be a. | |
| 529 | 5/7/00 | In the 1st line of footnote 4, "x" should be italic. | |
| 533 | 5/7/00 | In the line before (13-8), t > s should be t < s. | |
| 541 | 2/14/00 | In the list of standard errors at the bottom of the page, (0.342) should be (0.362). | |
| 544 | 8/13/01 | In equation (13-38), in the equation for z*2, in the numerator of the second part, inside the parentheses, the subscript on q should be 2, not 1. | |
| 545 | 5/7/00 | In the 1st equation, in the last term, "yT-1" should be "yT-1." | |
| 547 | 3/2/00 | In footnote 26, Park and Mitchell should be Parks and Mitchell. | |
| 551 | 5/7/00 | In the line after the second equation, "z's" should be "instrumental variables." | |
| 552 | 5/7/00 | In the last equation on the page, subscripts Gt, NCt, UCt should be G,t, NC,t, UC,t. | |
| 553 | 5/7/00 | In the first equation on the page, subscripts Gt, NCt, UCt,Gt-1, NCt-1, UCt-1 should be G,t, NC,t, UC,t, G,t-1, NC,t-1, UC,t-1. | |
| 553 | 7/16/99 | First equation on page, all compound subscripts such as Gt should have a comma before the "t" as G,t | |
| 559 | 3/23/00 | In the equation for log cost, the coefficient on load factor should be b4, not b3. | |
| 566 | 5/7/00 | In the second to last line of the example, [14,72] should be [14,67]. | |
| 568 | 1/19/00 | In equation (14-21), the matrices in the Kronecker product are reversed. The In should be at the left. | |
| 571 | 1/19/00 | In equation (14-31), the degrees of freedom in the denominator should be n-K-1, not n-K. | |
| 578 | 3/8/01 | In the second to last equation on the page, the degrees of freedom in the denominator should be n-K-1, not n-K. (Same change as on page 571.) | |
| 582 | 5/7/00 | In the 2nd line after the first equation, change "2 - T" to "two through T." | |
| 583 | 8/22/01 | In footnote 28, the Arellano and the Bover and Ahn and Schmidt references should both be 1995, not 1993. | |
| 583 | 8/22/01 | In the large three line equation in the center of the page, three occurrences of Md should be Mi0. | |
| 584 | 5/7/00 | In the first equation on the page, the comma after yis should be deleted. | |
| 586 | 5/7/00 | In the 3rd equation, in the fraction, the subscript in the numerator should be "u" and the subscript in the denominator should be "e." They are reversed. | |
| 587 | 5/7/00 | In the line after the 4th equation, (14-16) should be (14-17). | |
| 591 | 5/7/00 | The "4" at the beginning of item 4 should be bold. | |
| 599 | 5/7/00 | In the 1st equation, two occurrences of "log" should be "ln." | |
| 602 | 8/26/99 | The MLEs reported on page 602 are not quite correct, as the iterations that produced them were exited a bit too soon - the convergence criterion was too loose. If the iterations are forced to continue until the gradient rule gradient'Hessian-1gradient reaches a value of 1.D-8 instead of the 1.D-4 in place for the results shown, the new estimates will be "log-Likelihood = -515.42" with coefficients (standard errors) of [-2.22(1.96)], [0.02361(0.004291)], and [0.17089(0.01523)]. The second term in the LR statistic above Table 15.2 will be 98.23 instead of 88.53, so the sum changes to 219.15. (Thanks to the folks at Stata, Inc. for reporting this one.) | |
| 603 | 5/7/00 | In the 3rd line of the page, "odd" should be "off." | |
| 605 | 9/20/01 | In (15-20) in the second expression for r, the right hand summation in the denominator should run from 1 to T not from 2 to T. | |
| 608 | 8/13/01 | The "(12-39)" above the expression for Est. Var[b] should be "(12-18)." | |
| 609 | 8/22/01 | In the first set of equations, E[vi]=0 should be E[vi|Xi]=0 and E[vivi']=G should be E[vivi'|Xi ]= G. | |
| 614 | 5/7/00 | Two lines up from Section 15.4, "Levinson" should be "Levinsohn." | |
| 616 | 8/22/01 | In point 1 at the top of the page, "there is obviously no payoff to GLS. Indeed, GLS is OLS." should read "there is no payoff to GLS estimation of the full set of equations. Indeed, full GLS is equation by equation OLS." | |
| 622 | 5/7/00 | In the column headers of Table 15.7, "GE" and "US" are reversed. | |
| 629 | 5/7/00 | In the line before the first equation, "minimizes" must be "maximizes" and the "1/2" should be changed to "-1/2." | |
| 630 | 5/7/00 | In the first line, "because" should be "from the gradient because." | |
| 630 | 5/7/00 | The third equation on the page should be labeled "(15-58)." | |
| 633 | 5/7/00 | In the second equation, x2 should be x2t. | |
| 634 | 3/2/00 | In the last sentence, Park (1967) should be Parks (1967). | |
| 638 | 5/7/00 | In the set of share equations, the last one is not aligned vertically with the others. | |
| 642 | 5/7/00 | In (15-72), subscript mn should be ij and mm should be ii. | |
| 647 | 5/7/00 | In the equation after (15-78), the M should not be bold. | |
| 656 | 5/7/00 | 12 lines up from the bottom, "endogenous. This..." should be "exogenous. This..." | |
| 664 | 5/7/00 | In the third expression for plims, there is a prime after 1/T that should be deleted. | |
| 665 | 5/7/00 | In the 4th line after the example, "and S and" should be "and S as." | |
| 673 | 10/6/01 | In the first equation on the page, the subscript on q and q* should be s not d. | |
| 677 | 5/7/00 | In the line after the last equation, "(16-24)" should be "(16-28)." | |
| 677 | 5/7/00 | 7 lines up from the end of the page, at the beginning of the line, there should be a comma after wj. | |
| 682 | 10/7/01 | The claim about (16-20) is incorrect. In order for the estimator so constructed to be consistent, the prediction in (16-20) must include the included exogenous variables, Xj as well. | |
| 687 | 8/29/00 | In the 2SLS results in Table 16.4, the standard error for the P coefficient should be 0.118, not 1.118. Also, it should be noted, these standard errors do not include a degrees of freedom correction in the computation of the residual variance. If the correction is made, the four values would be 1.467, 0.131, 0.119, 0.045, respectively. | |
| 689 | 5/7/00 | In the third line of the second paragraph, change "temporally" to "contemporaneously." | |
| 689 | 8/13/01 | A reader points out that the observation in the first full paragraph, "neither the reduced form estimates at the first step nor rho-hat at the second will be consistent" must be qualified, since if a consistent estimate of rho is in hand, the method shown there, with iteration, does produce a consistent estimator of delta. (See Fair, Econometrica 1970, pp. 507-516 and Davidson and MacKinnon's text.) The point is correct, but moot. It is precisely because rho is estimated inconsistently at step 2 that step 3 fails produce a consistent estimator. Arguing that having a consistent estimator in hand already makes the estimator valid begs the question. If a consistent estimator of rho is in hand, the transformed data can be produced, and rho can be eliminated from consideration. | |
| 695 | 5/7/00 | In the second equation, in the large partitioned matrix, each Zi should be followed by a prime (transposed). | |
| 697 | 5/7/00 | In the second and last equations on the page, "M" appearing in subscripts should all be italic. (26 Ms altogether) | |
| 697 | 10/7/01 | In the equation for q at the top of the page, the large bracketed matrix at the left should be the same as the one on the right. In the expression given, the X that appears at the right of the parentheses in each term should be at the left. I.e., "(...)'X" should be "X'(...)" in each case. | |
| 701 | 5/7/00 | In the 4th line of Example 16.18, three are 20 observations, not 21. | |
| 701 | 5/7/00 | In the 5th line of Example 16.18, "Theoveridentifying" should be "The overidentifying." | |
| 702 | 5/7/00 | In the 2nd line of the first equation, "[YjXjx*]" should be "[Yj, Xj, x*]" | |
| 702 | 5/7/00 | In the first equation of Example 16.19, d should have a hat (carat) over it. | |
| 707 | 6/5/01 | In Example 16.22, in the matrix D-hat, the (2,2) element given, 0.0848, should be 0.848. | |
| 710 | 5/7/00 | In Exercise 1, in the second equation, e1 should be e2. | |
| 713 | 5/7/00 | In the first line, "(1998)" should be "(1998a, 1998b)." | |
| 717 | 5/7/00 | In (17-11) and the equation after it, "log" should be "ln" and in the second line, "T-1" should be "T-2." | |
| 717 | 5/7/00 | Second to last line of text, "The information criteria retain" should be "The information criterion retains." In the last line, before "They do avoid..." insert the sentence "In contrast, SC(q) has been seen to lead to underfitting in some finite sample cases." | |
| 718 | 9/6/99 | Equation (17-15), the matrix H should have another row at the top. It should be a (q+1)x(p+1) matrix. The first row should be (1,0,0,...,0). | |
| 719 | 5/7/00 | In the row headings in Table 17.1, "Log" should be "Ln" 10 times. | |
| 721 | 5/7/00 | 6 lines up from the end of the page, "of the estimates" should be "of the estimators of" | |
| 721 | 8/22/01 | In the line after (7-20) it should be noted that xt* is the expectation of the price in the next period formed in the current period. A notation sometimes used is xt+1|t* | |
| 723 | 5/7/00 | In the row headings in Table 17.2, "Log" should be "Ln" 16 times. | |
| 725 | 5/7/00 | In the first equation, bi should be bj. | |
| 726 | 8/22/01 | In the first line of the first equation in Section 17.3.2, wt should be dwt. In the second line of this equation, the same d should appear before the summation involving wt-l. | |
| 728 | 4/12/00 | In the first equation, the second t-1 should be t-2. | |
| 730 | 4/12/00 | In the third equation of Section 17.3.4, subscript T-p should be T-p+1. | |
| 731 | 5/7/00 | In the second equation, there should be a "}" after the "]." | |
| 731 | 4/12/00 | In the third equation on the page, subscript T-p+1 should be T-p+2. | |
| 731 | 5/7/00 | In the fourth equation on the page, subscript T-p should be T-p+1. and subscript T-p+1 should be T-p+2. | |
| 732 | 5/7/00 | In the first equation, second line, there should be a caret (hat) over the second m. | |
| 733 | 5/7/00 | In the second equation on the page, subscript T-2 should be T-p. | |
| 738 | 5/7/00 | In the 5th equation, the second subscript t-1 should be t-2. | |
| 739 | 4/13/00 | At the beginning of the line before the second equation, (1-r) should be (1-rL). | |
| 740 | 8/22/01 | In equation (7-29), vt should be et and add immediately after this line "where et is a nonautocorrelated vector of innovations." | |
| 741 | 5/7/00 | In the first line after the first equation, change "lmth" to "(l,m)." | |
| 742 | 5/7/00 | In the last equation on the page, subscripts 1 and 2 on e should be 1t and 2t. | |
| 743 | 5/7/00 | In the second equation two occurrences of "log" should be "ln." | |
| 744 | 5/7/00 | Add a sentence after "in the fit." "In contrast, the usual lag selection procedure in line with the general to simple principle discussed in Section 17.4.3 would be to set the maximum lag length and test down from it until the last estimated parameter matrix is significant." | |
| 745 | 5/7/00 | In the 4th equation, add parentheses around "m, l." | |
| 747 | 5/7/00 | Line before the exercises, change subscript "it" to "lt." | |
| 753 | 5/7/00 | In footnote 6, "(-1,2),(-1,-2), and (1,0)" should be "(2,-1), (-2,-1), (0,1)." | |
| 755 | 5/7/00 | The equation after (18-15), the correct expression for l0 is se2 [(1 - g2) / (1 + g2)] /(1 - g12 - g22) | |
| 756 | 5/7/00 | Second equation, z should be 1/z. | |
| 759 | 5/7/00 | In the first line, the second "t-1" should be "t-2." | |
| 772 | 8/17/99 | 2nd equation, subscript t-j should be t-k | |
| 772-773 | 8/19/99 | In four expressions for the spectrum, in hY(wj), there should be no subscript j on w. | |
| 774 | 7/16/99 | 3rd to last line on the page, "... in (18-35) or (18-36)" should be "... in (18-16) or (18-17)" | |
| 781 | 8/22/01 | In equation (18-22), in the first line, b(1-g) should be bg. In the second line, insert +ut before the period. | |
| 784 | 5/29/00 | Near the end of the page, the reference to Table 18.7 should be to Table 18.5. | |
| 784 | 6/30/00 | The t-statistic for the yt-1 term in the "simple AR(1) regression" in Example 18.6 should be -2.410 instead of -2.401. | |
| 787 | 5/7/00 | 8th line of Example 18.7, delete the left parenthesis before "This assumption..." | |
| 791 | 5/7/00 | Second to last line of the proof, y1 should be yt1. | |
| 792 | 5/7/00 | Second and third equations, 1, -g10, -g11, -g12, -g13 should be [1, -g10, -g11, -g12, -g13]'. 1, -g20, g21, g21, 0 should be [1, -g20, g21, g21, 0]' |
|
| 793 | 12/12/00 | In the last equation on the page, in the third term on the right hand side, (yt - qzt) should be (yt-1 - qzt-1). | |
| 795 | 5/7/00 | Last equation, the right hand side should be [Dyt-1, Dyt-2,.... Dyt-p+1]. |
|
| 797 | 5/7/00 | In the expression for "Var[et]" there is a "]" missing at the end of the line. | |
| 797 | 5/7/00 | Three lines up from the end of the page, delete "normally." In the last line, change "estimator" to "unbiased estimator." | |
| 816 | 5/7/00 | Change two occurrences of "Weibull" to "extreme value." | |
| 816 | 5/31/01 | In Table 19.1, 4th column, 0.499 should be 0.449. | |
| 822 | 5/7/00 | Third line, change 19-17 to 19-19. | |
| 830 | 4/3/00 | In the paragraph after the last equation on the page, 21.3 should be 21.94, 0.70412 should be 0.6789 and 0.534 should be 0.533. | |
| 830 | 5/31/01 | In the middle of the page, the Wald statistic should be (1.09337)2/(0.8806)2 = 1.5416. | |
| 831 | 5/7/00 | In the line before Section 19.4.5, 1978 should be 1979. | |
| 831 | 5/7/00 | In footnote 16, Cramer (1998) should be Cramer (1999). | |
| 832 | 5/7/00 | First line, delete b. In the next line, x should be xi. | |
| 835 | 9/21/99 | In the 4th equation on the page, after the equality, the 1/f(pi) should be 1/f(b'xi). In the next equation, fi is used to denote f(b'xi). | |
| 847 | 8/22/01 | Second to last and last equations on the page, in the definition of the Klein and Spady estimator. This needs numerous changes of subscripts. In the first equation, Gn(z) should be Gn(zi). (Add subscript "i.") On the right hand side of this equation, add subscript "i" to all three occurrences of z and to all three occurrences of y and likewise to z and y in the function in the next line. At the end of that line, change bn'x to zi = bn'xi. In the equation that follows, on the left hand side, add subscripts "i" to x and y. On the right hand side, change the index of the summation from "i" to "j." Change the subscript on y and x from "i" to "j." Add subscript "n" to b. Finally, add subscript "i" to z. | |
| 848 | 8/22/01 | In Table 19.5, in the third column of numbers, the first 0.44 should -0.44. (Missing minus sign.) | |
| 849 | 5/7/00 | 6 lines up from the end of the page, "national" should be "notational." | |
| 849 | 3/8/01 | In the last equation on the page, the subscript on r should be "i" not "t." | |
| 850 | 5/24/01 | At the bottom of the page, in the second derivative for b1 then r, immediately after the summation sign, qi2 should be qi1. | |
| 858 | 4/26/00 | In the last equation on the page, after the second equals sign, in both numerator and denominator of the fraction, in ai the subscript i should not be present. | |
| 863 | 5/17/00 | The values in the lower part of Table 19.11 are computed incorrectly. The correct values are:
C/D C/P Train Bus
+--------------------------------------------------------
C/D | -.04595 .001046 .008658 .031805
C/P | .35800 -.026485 .008658 .031805
Train| .35800 .001046 -.610191 .031805
Bus | .35800 .001046 .008658 -.482338
|
|
| 864 | 11/4/00 | Section header and fifth line of Section 19.7.3, "Independence of Irrelevant Alternatives" should be "Independence from Irrelevant Alternatives." | |
| 865 | 8/13/01 | In the header for Example 19.17, "Independence of Irrelevant Alternatives" should be "Independence from Irrelevant Alternatives." Above this example, two lines above the expression for the Hausman statistic, "choices are eliminated" should be "choices are included." | |
| 871 | 8/13/01 | In the last equation on the page, inside the square brackets, "(x1-xq)" should be "(xq-x1)" and "(xJ-xq)" should be "(xq-xJ)." | |
| 872 | 8/17/01 | In the first line on the page, the statement
"allows an unrestricted
(J-1) by (J-1) covariance matrix
" is not quite correct. One
of the diagonal elements of this matrix must be normalized to one for identification.
An appropriate, albeit inconvenient characterization might be
"allows an unrestricted
(J-1) by (J-1) correlation matrix and
J-2 free standard deviations on the diagonal"
Eighteen lines later, "(s1,...,sJ-1,1)" should be "(s1,...,sJ-2,1,1)." |
|
| 879 | 8/22/01 | At mid page, fifth line of text after the table, "F(0.8479)" should be "F(-0.8479)." Near the end of the same line, the 0.628 should be 0.629. | |
| 880 | 5/7/00 | In footnote 58, Greene (1996a) should be Greene (1997a). | |
| 883 | 5/25/00 | Line 3 on the page, "dropped from" should be "added to." | |
| 884 | 5/25/00 | In the first equation, bx should be bk. The subscript is wrong. | |
| 885 | 5/25/00 | In the 4th line after the first equation, 0.934 should be 0.922 | |
| 887 | 5/25/00 | At mid page, "hypothesis q = 0" should read "hypothesis a = 1/q = 0." | |
| 887 | 5/7/00 | In footnote 65, Terza (1997) should be Terza (1998). | |
| 892 | 8/26/99 | 3 lines below Table 19.19, +6.118 should be +1.9119. The number was computed incorrectly. This value is less than the 1.96 critical value, so the conclusion drawn in the text is reversed. A suitable revision might be "Vuong's diagnostic statistic is only +1.9119, which suggests that in spite of the preponderance of zeros, the Poisson model adequately describes the data. Indeed, if we examine the predictions of the model, then it appears that the Poisson model performs a bit better than the model with the regime split." (Thanks to David Drukker at Stata, Inc. for his assistance with this computation.) | |
| 904 | 2/16/00 |
In Example 20.6:
|
|
| 909 | 5/7/00 | In the first equation in the proof, at the end of the second line of the first equation, "|x" should be ",x." (Change vertical bar to comma.) | |
| 909 | 5/7/00 | In the last line of the page, the subscript i on yi* should be deleted. | |
| 915 and 971 | 9/9/99 | In several citations to Fin and Schmidt on page 915 and in the reference on page 971, Fin should be Lin. | |
| 920 | 5/7/00 | In the website, "work.ss" should be "worksd" | |
| 921 | 2/14/00 | In Table 20.4, in the third column of values, (0.184) should be (0.0184), in the fourth column, (0.484) should be (0.0484), in the fifth column, (0.102) should be (0.0102). | |
| 922 | 5/7/00 | In the 6th line of the paragraph that begins "Before leaving..." the "[" before "This" must be deleted. | |
| 923 | 5/7/00 | In the last line of the first paragraph, 0.107 should be 0.101. | |
| 924 | 5/7/00 | In the 4th line of the last paragraph, 19.25 should be 19.26. | |
| 928 | 5/7/00 | The first and second sentences of Example 20.15 should be combined to read "The model of migration analyzed by Nakosteen and Simmer (1980) fits precisely into the framework described earlier." | |
| 929 | 5/7/00 | In the first line of footnote 22, subscript M should be u. | |
| 930 | 5/7/00 | In the equation before Section 20.4.3, x and w should both have subscript i. | |
| 930 | 2/18/00 | In Section 20.4.3, in item 1, in the definition of di, inside the parentheses, the minus sign should be a plus sign. | |
| 932 | 5/7/00 | In the third line after (20-20), change "y on x*" to "y on X*." | |
| 933 | 5/7/00 | In (20-21), the vertical bars are too long. | |
| 942 | 5/7/00 | Two lines up from Section 20.5.2.d., change "normal" to "lognormal." | |
| 943 | 5/7/00 | In the second to last equation, delete the term (1/s). | |
| 943 and 944 | 9/28/99 | At the end of page 943, in the derivation of the density of wi, the Jacobian has been misplaced. In the second to last equation on the page, in the density for wi, the term (1/s) must be deleted. The first equation on page 944 remains correct as is. In order to reach the second equation on page 944, it is necessary to make another change of variable, to yi = lnti = swi + b'xi. With this change of variable, the Jacobian is (1/s), and the log likelihood function in the second equation on page 944 results, again, as is. However, it should be noted there that this is the log likelihood function for the observed data, (yi = lnti)| b,s. Thus, one might want to write the log likelihood in terms of vi = yi - b'xi instead of in terms of wi. | |
| 946 | 12/12/00 | In the last equation on the page, inside the large square brackets, ei should not have a subscript, both times. It should just be e, the base of the natural logs. | |
| 948 | 5/7/00 | In the expression for l(ti) in last equation on the page, delete the minus sign in the exponent. | |
| 949 | 5/7/00 | In the second equation, the summation term inside the square brackets should be the log of that sum. | |
| 954 | 5/7/00 | In the text for Table A9.1, change "capital" to "capital=capital input." | |
| 966 | 5/7/00 | In the entry for R. Brown, the first initial should be B. | |
| 967 | 8/1/99 | In Chib, S., and E. Greenberg... change"Manuscript...,1995" to EconometricTheory, 12, 1996, pp. 409-431. | |
| 967 | 8/1/99 | Cameron and Trivedi, 1998, the correct title is Regression Analysis of Count Data. | |
| 971 | 5/7/00 | Add entry to references, Garber, S. and S. Klepper, "Extending the Classical Normal Errors in Variables Model," Econometrica, 48, 1980, pp. 1541-1546. | |
| 977 | 5/7/00 | In the second Judge et al., entry, W. Griffiths should precede C. Hill. | |
| 980 | 5/7/00 | Manski's 1995 book should be entitled Identification Problems in the Social Sciences. Also in this entry, 1995a should be 1995. | |
| 980 | 5/7/00 | The McCullagh and Nelder and the McCullough entries should precede the first McCullough entry. | |
| 982 | 3/2/00 | In the bibliography, Park, R. should be Parks, R. in both references. | |
| 987 | 5/7/00 | The first White, 1982 entry should be 1982a; the second should be 1982b. | |
| 987 | 5/7/00 | Add an entry after Veall and Zimmerman: Vinod, H., "Bootstrap, Jackknife, Resampling and Simulation: Applications in Econometrics," in Maddala, G., C. Rao, and H. Vinod, eds., Handbook of Statistics: Econometrics, Vol II., Chapter 11, Amsterdam, North Holland, 1993. | |
| 988 | 8/1/99 | In Zellner, A., and C.-K. Min... change "Unpublished...,1992." to Journal of the American Statistical Association, 90, 1996, pp. 921-927. | |
| 989 | 7/16/99 | In Cramer, H., "e" should have an accent mark. | |
| 989 | 5/7/00 |
|
|
| 990 | 5/7/00 | In the author index, the "Hansen, C 375" entry should be included in the "Hansen, L., ..." There is no separate entry for a Hansen, C. | |
| 990 | 5/7/00 |
|
|
| 991 | 7/16/99 | First McKelvey entry with page 875 should be combined with the second one which is after McFadden | |
| 991 | 5/7/00 |
|
|
| 992 | 5/7/00 |
|
|
| 992 | 3/2/00 | In the author index, reference to Park, R. should be to Parks, R. | |
| 993 | 5/7/00 |
|
|
| 995 | 5/7/00 |
|
|
| 996 | 5/7/00 |
|
|
| 997 | 5/7/00 |
|
|
| 998 | 5/7/00 |
|
|
| 999 | 5/7/00 |
|
|
| 1000 | 5/7/00 |
|
|
| 1001 | 5/7/00 |
|
|
| 1003 | 5/7/00 |
|
|
| 1004 | 5/7/00 |
|