------------------------------------------------------------------------------------------------------ log: c:\Imbook\bwebpage\Section2\mma04p4ivweak.txt log type: text opened on: 17 May 2005, 13:45:59 . . ********** OVERVIEW OF MMA04P4IVWEAK.DO ********** . . * STATA Program . * copyright C 2005 by A. Colin Cameron and Pravin K. Trivedi . * used for "Microeconometrics: Methods and Applications" . * by A. Colin Cameron and Pravin K. Trivedi (2005) . * Cambridge University Press . . * Chapter 4.9.5 pages 110-2 . * IV regression with potentially weak instruments . * (1) Compares OLS and IV estimation of log-wages on schooling regression . * where schooling, experience and experience-squared are endogenous . * and proximity to 4-year college, age and age-squared are instruments . * so model is just-identified. . * (2) Verifies that here can treat errors as homoskedastic . * (3) Looks at weak instruments . * (A) instrument relevance: Whether Shea's partial R-squared is low . * (B) finite sample bias: whether first-stage partial F is low . * (4) Provides Table 4.5 . * (5) Does more analysis than reported in the book . . * To run this program you need data and dictionary files . * DATA66.dat ASCII data set . * DATA66.dct Stata dictionary that labels variables . . ********** SETUP ********** . . set more off . version 8.0 . set memory 20m (20480k) . set linesize 150 /* Permits long inputline commands with delimit */ . . ********** ORIGINAL DATA SOURCE ********** . . * Program mma4p4ivweak.do based on Kling Analys66.d0 September 2003 . * written for Jeffrey R. Kling (2001) "Interpreting Instrumental Variables Estimates . * of the Return to Schooling", Journal of Business and Economic Statistics, . * July 2001, 19 (3), pp.358-364. . * This program focuses on Columns (1) and (2) of Kling's Table 1 on p.359 . * in turn based on . * David Card (1995), "Using Geographic Variation in College Proximity to . * Estimate the Returns to Schooling", in . * Aspects of Labor Market Behavior: Essays in Honor of John Vanderkamp, . * eds. L.N. Christofides et al., Toronto: University of Toronto Press, pp.201-221. . . ********** READ IN DATA and SUMMARIZE ********** . . infile using DATA66.dct, using(DATA66.dat) dictionary using DATA66.dat { _column(1) id %8f "ID CODE (r0000100) n= 5225 mean= 2613.000 min= 1 max= 5225 " _column(9) black %3f "Race (r0002300) n= 5225 mean= 1.296 min= 1 max=3 " _column(13) imigrnt %3f "Was r's brthpl in the US? (r0038000) n=4965 mean=0.98 mn=0 mx=1 " _column(17) hhead %8f "Person R lived w/ @ age 14 (r0039700) n= 5213 mean=1.92 mn=1 mx=9" _column(28) mag_14 %10f "Were magznes avail at age 14 (r0039900) n=5167 mean=0.69 mn=0 mx=1 " _column(40) news_14 %10f "Were nwspaprs avail at age 14 (r0040000) n=5195 mean=0.85 mn=0 mx=1" _column(52) lib_14 %10f "Were lib-card avail at age14 (r0040100) n=5204 mean=0.66 mn=0 mx=1 " _column(63) num_sib %8f "Tot # sibs r 66 (r0056900) n=5168 mean=3.408 min=0 max=18" _column(72) fgrade %8f "Hgc by father, 66 (r0063100) n=3930 mean=9.937 min=0 max=18" _column(81) mgrade %8f "Hgc by mother, 66 (r0063300) n=4573 mean=10.25 min=0 max=18" _column(90) iq %8f "Iq_score (r0171100) n= 3369 mean=101.582 min=50 max=158 " _column(99) bdate %8f "Birthdate - STATA formatted " _column(108) gfill76 %8f "'76 Grade level, some values filled from prevs reports" _column(117) wt76 %8f "'76 Weight " _column(126) grade76 %8f "'76 Grade level" _column(135) grade66 %8f "'66 Grade level" _column(144) age66 %8f "Age reported by screener (r0002200) " _column(153) smsa66 %8f "If lived in SMSA in 1966 (r0002455=1,2)" _column(162) region %8f "Census Region in 1966 (r0002900) " _column(171) smsa76 %8f "If lived in SMSA in 1976 (r0437515=1,2)" _column(180) col4 %8f "If any 4-year college nearby (r0004000!=4) " _column(189) mcol4 %8f "If male 4-year college nearby (r0004100=1,2) " _column(198) col4pub %8f "If public 4-year college nearby (r0004000=2,3)" _column(207) south76 %1f "If lived in South in 1976 (r0437511=1) " _column(209) wage76 %10f "'76 Wage" _column(219) exp76 %8f "'76 experience, (10 + age66) - grade76 - 6)" _column(230) expsq76 %10f "'76 experience, exp76 ^2/100 " _column(243) age76 %8f "'76 age (age66 +10) " _column(252) agesq76 %8f "'76 age squared (age76^2) " _column(261) reg1 %8f "region==NE" _column(270) reg2 %8f "If lived in Region 2 (region= MidAtl)" _column(279) reg3 %8f "If lived in Region 3 (region= ENC) " _column(288) reg4 %8f "If lived in Region 4 (region= WNC) " _column(297) reg5 %8f "If lived in Region 5 (region= SA ) " _column(306) reg6 %8f "If lived in Region 6 (region= ESC) " _column(315) reg7 %8f "If lived in Region 7 (region= WSC) " _column(324) reg8 %8f "If lived in Region 8 (region= M ) " _column(333) reg9 %8f "If lived in Region 9 (region= P ) " _column(342) momdad14 %8f "If lived with both parents at age 14 " _column(351) sinmom14 %8f "If lived with mother only at age 14 " _column(360) nodaded %1f "If father has no formal education " _column(362) nomomed %1f "If mother has no formal education " _column(365) daded %10f "Mean grade level of father " _column(377) momed %10f "Mean grade level of mother " _column(396) famed %8f "Father's and mother's education " _column(405) famed1 %8f "If mgrade> 12 & fgrade> 12 (famed=1) " _column(414) famed2 %8f "If mgrade>=12 & fgrade>=12 (famed=2) " _column(423) famed3 %8f "If mgrade==12 & fgrade==12 (famed=3) " _column(432) famed4 %8f "If mgrade>=12 & fgrade==-1 (famed=4) " _column(441) famed5 %8f "If fgrade>=12 (famed=5) " _column(450) famed6 %8f "If mgrade>=12 & fgrade> -1 (famed=6) " _column(459) famed7 %8f "If mgrade>=9 & fgrade>=9 (famed=7) " _column(468) famed8 %8f "If mgrade> -1 & fgrade> -1 (famed=8) " _column(477) famed9 %8f "If famed not in range (1-8)" _column(486) int76 %8f "If wt76 not missing " _column(495) age1415 %8f "If in age group =14-15" _column(504) age1617 %8f "If in age group =16-17" _column(513) age1819 %8f "If in age group =18-19" _column(522) age2021 %8f "If in age group =20-21" _column(531) age2224 %8f "If in age group =20-24" _column(540) cage1415 %8f "If in age group =14,15 and lived near college" _column(549) cage1617 %8f "If in age group =16,17 and lived near college" _column(558) cage1819 %8f "If in age group =18,19 and lived near college" _column(567) cage2021 %8f "If in age group =20,21 and lived near college" _column(576) cage2224 %8f "If in age group =20-24 and lived near college" _column(585) cage66 %8f "Age in 66 and whether lived near college " _column(594) a1 %8f "If age in 66 = 14 (age66= 14)" _column(603) a2 %8f "If age in 66 = 15 (age66= 15)" _column(612) a3 %8f "If age in 66 = 16 (age66= 16)" _column(621) a4 %8f "If age in 66 = 17 (age66= 17)" _column(630) a5 %8f "If age in 66 = 18 (age66= 18)" _column(639) a6 %8f "If age in 66 = 19 (age66= 19)" _column(648) a7 %8f "If age in 66 = 20 (age66= 20)" _column(657) a8 %8f "If age in 66 = 21 (age66= 21)" _column(666) a9 %8f "If age in 66 = 22 (age66= 22)" _column(675) a10 %8f "If age in 66 = 23 (age66= 23)" _column(684) a11 %8f "If age in 66 = 24 (age66= 24)" _column(693) ca1 %8f "Not lived near college in 66" _column(702) ca2 %8f "If age in 66 = 14 and lived near college" _column(711) ca3 %8f "If age in 66 = 15 and lived near college" _column(720) ca4 %8f "If age in 66 = 16 and lived near college" _column(729) ca5 %8f "If age in 66 = 17 and lived near college" _column(738) ca6 %8f "If age in 66 = 18 and lived near college" _column(747) ca7 %8f "If age in 66 = 19 and lived near college" _column(756) ca8 %8f "If age in 66 = 20 and lived near college" _column(765) ca9 %8f "If age in 66 = 21 and lived near college" _column(774) ca10 %2f "If age in 66 = 22 and lived near college" _column(777) ca11 %2f "If age in 66 = 23 and lived near college" _column(780) ca12 %8f "If age in 66 = 24 and lived near college" _column(782) g25 %12f "Grade level when 25 years old " _column(795) g25i %12f "If =g25 and intrvwed in year used for determining g25 " _column(819) intmo66 %8f "Intvw month in 1966, used to identify cases incl by CARD" _column(828) nlsflt %8f "Flag to identify if the case was used by CARD" _column(837) nsib %8f "Number of siblings " _column(846) ns1 %8f "If number of siblings = 0 (nsib= 0)" _column(855) ns2 %8f "If number of siblings = 2 (nsib= 2)" _column(864) ns3 %8f "If number of siblings = 3 (nsib= 3)" _column(873) ns4 %8f "If number of siblings = 4 (nsib= 4)" _column(882) ns5 %8f "If number of siblings = 6 (nsib= 6)" _column(891) ns6 %8f "If number of siblings = 9 (nsib= 9)" _column(900) ns7 %8f "If number of siblings =18 (nsib=18)" } (5226 observations read) . * save DATA66, replace . desc Contains data obs: 5,226 vars: 101 size: 2,132,208 (89.8% of memory free) ------------------------------------------------------------------------------- storage display value variable name type format label variable label ------------------------------------------------------------------------------- id float %9.0g ID CODE (r0000100) n= 5225 mean= 2613.000 min= 1 max= 5225 black float %9.0g Race (r0002300) n= 5225 mean= 1.296 min= 1 max=3 imigrnt float %9.0g Was r's brthpl in the US? (r0038000) n=4965 mean=0.98 mn=0 mx=1 hhead float %9.0g Person R lived w/ @ age 14 (r0039700) n= 5213 mean=1.92 mn=1 mx=9 mag_14 float %9.0g Were magznes avail at age 14 (r0039900) n=5167 mean=0.69 mn=0 mx=1 news_14 float %9.0g Were nwspaprs avail at age 14 (r0040000) n=5195 mean=0.85 mn=0 mx=1 lib_14 float %9.0g Were lib-card avail at age14 (r0040100) n=5204 mean=0.66 mn=0 mx=1 num_sib float %9.0g Tot # sibs r 66 (r0056900) n=5168 mean=3.408 min=0 max=18 fgrade float %9.0g Hgc by father, 66 (r0063100) n=3930 mean=9.937 min=0 max=18 mgrade float %9.0g Hgc by mother, 66 (r0063300) n=4573 mean=10.25 min=0 max=18 iq float %9.0g Iq_score (r0171100) n= 3369 mean=101.582 min=50 max=158 bdate float %9.0g Birthdate - STATA formatted gfill76 float %9.0g '76 Grade level, some values filled from prevs reports wt76 float %9.0g '76 Weight grade76 float %9.0g '76 Grade level grade66 float %9.0g '66 Grade level age66 float %9.0g Age reported by screener (r0002200) smsa66 float %9.0g If lived in SMSA in 1966 (r0002455=1,2) region float %9.0g Census Region in 1966 (r0002900) smsa76 float %9.0g If lived in SMSA in 1976 (r0437515=1,2) col4 float %9.0g If any 4-year college nearby (r0004000!=4) mcol4 float %9.0g If male 4-year college nearby (r0004100=1,2) col4pub float %9.0g If public 4-year college nearby (r0004000=2,3) south76 float %9.0g If lived in South in 1976 (r0437511=1) wage76 float %9.0g '76 Wage exp76 float %9.0g '76 experience, (10 + age66) - grade76 - 6) expsq76 float %9.0g '76 experience, exp76 ^2/100 age76 float %9.0g '76 age (age66 +10) agesq76 float %9.0g '76 age squared (age76^2) reg1 float %9.0g region==NE reg2 float %9.0g If lived in Region 2 (region= MidAtl) reg3 float %9.0g If lived in Region 3 (region= ENC) reg4 float %9.0g If lived in Region 4 (region= WNC) reg5 float %9.0g If lived in Region 5 (region= SA ) reg6 float %9.0g If lived in Region 6 (region= ESC) reg7 float %9.0g If lived in Region 7 (region= WSC) reg8 float %9.0g If lived in Region 8 (region= M ) reg9 float %9.0g If lived in Region 9 (region= P ) momdad14 float %9.0g If lived with both parents at age 14 sinmom14 float %9.0g If lived with mother only at age 14 nodaded float %9.0g If father has no formal education nomomed float %9.0g If mother has no formal education daded float %9.0g Mean grade level of father momed float %9.0g Mean grade level of mother famed float %9.0g Father's and mother's education famed1 float %9.0g If mgrade> 12 & fgrade> 12 (famed=1) famed2 float %9.0g If mgrade>=12 & fgrade>=12 (famed=2) famed3 float %9.0g If mgrade==12 & fgrade==12 (famed=3) famed4 float %9.0g If mgrade>=12 & fgrade==-1 (famed=4) famed5 float %9.0g If fgrade>=12 (famed=5) famed6 float %9.0g If mgrade>=12 & fgrade> -1 (famed=6) famed7 float %9.0g If mgrade>=9 & fgrade>=9 (famed=7) famed8 float %9.0g If mgrade> -1 & fgrade> -1 (famed=8) famed9 float %9.0g If famed not in range (1-8) int76 float %9.0g If wt76 not missing age1415 float %9.0g If in age group =14-15 age1617 float %9.0g If in age group =16-17 age1819 float %9.0g If in age group =18-19 age2021 float %9.0g If in age group =20-21 age2224 float %9.0g If in age group =20-24 cage1415 float %9.0g If in age group =14,15 and lived near college cage1617 float %9.0g If in age group =16,17 and lived near college cage1819 float %9.0g If in age group =18,19 and lived near college cage2021 float %9.0g If in age group =20,21 and lived near college cage2224 float %9.0g If in age group =20-24 and lived near college cage66 float %9.0g Age in 66 and whether lived near college a1 float %9.0g If age in 66 = 14 (age66= 14) a2 float %9.0g If age in 66 = 15 (age66= 15) a3 float %9.0g If age in 66 = 16 (age66= 16) a4 float %9.0g If age in 66 = 17 (age66= 17) a5 float %9.0g If age in 66 = 18 (age66= 18) a6 float %9.0g If age in 66 = 19 (age66= 19) a7 float %9.0g If age in 66 = 20 (age66= 20) a8 float %9.0g If age in 66 = 21 (age66= 21) a9 float %9.0g If age in 66 = 22 (age66= 22) a10 float %9.0g If age in 66 = 23 (age66= 23) a11 float %9.0g If age in 66 = 24 (age66= 24) ca1 float %9.0g Not lived near college in 66 ca2 float %9.0g If age in 66 = 14 and lived near college ca3 float %9.0g If age in 66 = 15 and lived near college ca4 float %9.0g If age in 66 = 16 and lived near college ca5 float %9.0g If age in 66 = 17 and lived near college ca6 float %9.0g If age in 66 = 18 and lived near college ca7 float %9.0g If age in 66 = 19 and lived near college ca8 float %9.0g If age in 66 = 20 and lived near college ca9 float %9.0g If age in 66 = 21 and lived near college ca10 float %9.0g If age in 66 = 22 and lived near college ca11 float %9.0g If age in 66 = 23 and lived near college ca12 float %9.0g If age in 66 = 24 and lived near college g25 float %9.0g Grade level when 25 years old g25i float %9.0g If =g25 and intrvwed in year used for determining g25 intmo66 float %9.0g Intvw month in 1966, used to identify cases incl by CARD nlsflt float %9.0g Flag to identify if the case was used by CARD nsib float %9.0g Number of siblings ns1 float %9.0g If number of siblings = 0 (nsib= 0) ns2 float %9.0g If number of siblings = 2 (nsib= 2) ns3 float %9.0g If number of siblings = 3 (nsib= 3) ns4 float %9.0g If number of siblings = 4 (nsib= 4) ns5 float %9.0g If number of siblings = 6 (nsib= 6) ns6 float %9.0g If number of siblings = 9 (nsib= 9) ns7 float %9.0g If number of siblings =18 (nsib=18) ------------------------------------------------------------------------------- Sorted by: Note: dataset has changed since last saved . sum Variable | Obs Mean Std. Dev. Min Max -------------+-------------------------------------------------------- id | 5225 2613 1508.472 1 5225 black | 5225 .2752153 .4466655 0 1 imigrnt | 5225 .0237321 .1522277 0 1 hhead | 5225 -.3783732 47.95128 -999 9 mag_14 | 5225 .6861566 .4616275 0 1 -------------+-------------------------------------------------------- news_14 | 5225 .8483024 .3577176 0 1 lib_14 | 5225 .658469 .4733619 0 1 num_sib | 5168 3.407701 2.586307 0 18 fgrade | 3930 9.93715 3.777654 0 18 mgrade | 4573 10.25104 3.17986 0 18 -------------+-------------------------------------------------------- iq | 3369 101.5818 15.93225 50 158 bdate | 5204 472926.6 31765.04 360823 521224 gfill76 | 5225 12.78718 2.802705 0 18 wt76 | 3695 475512.5 265188.5 98617 2582192 grade76 | 3671 13.23018 2.747627 0 18 -------------+-------------------------------------------------------- grade66 | 5225 10.58431 2.433696 0 18 age66 | 5225 18.09129 3.157657 14 24 smsa66 | 5225 .6599043 .4737864 0 1 region | 5225 4.721722 2.300767 1 9 smsa76 | 5225 .491866 .4999817 0 1 -------------+-------------------------------------------------------- col4 | 5225 .691866 .4617664 0 1 mcol4 | 5225 .6874641 .4635713 0 1 col4pub | 5225 .5129187 .4998809 0 1 south76 | 3695 .3964817 .4892328 0 1 wage76 | 3078 1.658013 .4430234 0 3.1797 -------------+-------------------------------------------------------- exp76 | 3671 8.933533 4.212664 0 25 expsq76 | 3671 .9754971 .8778352 0 6.25 age76 | 5225 28.09129 3.157657 24 34 agesq76 | 5225 799.0896 182.0539 576 1156 reg1 | 5225 .04 .1959779 0 1 -------------+-------------------------------------------------------- reg2 | 5225 .1617225 .3682313 0 1 reg3 | 5225 .1900478 .3923763 0 1 reg4 | 5225 .0639234 .2446399 0 1 reg5 | 5225 .2126316 .4092083 0 1 reg6 | 5225 .0895694 .2855912 0 1 -------------+-------------------------------------------------------- reg7 | 5225 .1083254 .3108206 0 1 reg8 | 5225 .0304306 .1717855 0 1 reg9 | 5225 .1033493 .3044437 0 1 momdad14 | 5225 .7680383 .4221251 0 1 sinmom14 | 5225 .1182775 .3229673 0 1 -------------+-------------------------------------------------------- nodaded | 5225 .2478469 .4318038 0 1 nomomed | 5225 .1247847 .3305062 0 1 daded | 5225 9.937162 3.276134 0 18 momed | 5225 10.25103 2.974812 0 18 famed | 5225 6.05933 2.643855 1 9 -------------+-------------------------------------------------------- famed1 | 5225 .0610526 .2394497 0 1 famed2 | 5225 .0742584 .262216 0 1 famed3 | 5225 .1144498 .3183872 0 1 famed4 | 5225 .0474641 .2126498 0 1 famed5 | 5225 .077512 .2674276 0 1 -------------+-------------------------------------------------------- famed6 | 5225 .1245933 .3302888 0 1 famed7 | 5225 .0486124 .215077 0 1 famed8 | 5225 .2273684 .4191726 0 1 famed9 | 5225 .224689 .4174173 0 1 int76 | 5225 .707177 .4551014 0 1 -------------+-------------------------------------------------------- age1415 | 5225 .2595215 .4384141 0 1 age1617 | 5225 .2482297 .4320271 0 1 age1819 | 5225 .1751196 .3801058 0 1 age2021 | 5225 .11311 .3167576 0 1 age2224 | 5225 .2040191 .4030216 0 1 -------------+-------------------------------------------------------- cage1415 | 5225 .1755024 .3804327 0 1 cage1617 | 5225 .1680383 .3739361 0 1 cage1819 | 5225 .1245933 .3302888 0 1 cage2021 | 5225 .0796172 .2707256 0 1 cage2224 | 5225 .1441148 .3512397 0 1 -------------+-------------------------------------------------------- cage66 | 5225 12.56115 8.785895 0 24 a1 | 5225 .1314833 .3379605 0 1 a2 | 5225 .1280383 .3341644 0 1 a3 | 5225 .1326316 .3392086 0 1 a4 | 5225 .1155981 .3197729 0 1 -------------+-------------------------------------------------------- a5 | 5225 .098756 .2983627 0 1 a6 | 5225 .0763636 .2656045 0 1 a7 | 5225 .0560766 .2300915 0 1 a8 | 5225 .0570335 .2319288 0 1 a9 | 5225 .0666029 .2493568 0 1 -------------+-------------------------------------------------------- a10 | 5225 .0683254 .2523275 0 1 a11 | 5225 .0690909 .2536329 0 1 ca1 | 5225 .308134 .4617664 0 1 ca2 | 5225 .0876555 .2828203 0 1 ca3 | 5225 .0878469 .2830992 0 1 -------------+-------------------------------------------------------- ca4 | 5225 .0870813 .2819812 0 1 ca5 | 5225 .0809569 .2727951 0 1 ca6 | 5225 .0708134 .2565374 0 1 ca7 | 5225 .0537799 .2256044 0 1 ca8 | 5225 .0390431 .193716 0 1 -------------+-------------------------------------------------------- ca9 | 5225 .0405742 .1973204 0 1 ca10 | 5225 .0465072 .2106009 0 1 ca11 | 5225 .0484211 .2146748 0 1 ca12 | 5225 12.52593 2.740455 0 18 g25 | 5225 12.53923 2.749407 0 18 -------------+-------------------------------------------------------- g25i | 4148 12.77929 2.740756 0 18 intmo66 | 5225 -5.790239 128.4984 -999 12 nlsflt | 5225 .9835407 .1272459 0 1 nsib | 5225 2.818565 2.473752 0 18 ns1 | 5225 .2547368 .4357549 0 1 -------------+-------------------------------------------------------- ns2 | 5225 .3534928 .4780998 0 1 ns3 | 5225 .0109091 .1038853 0 1 ns4 | 5225 .1892823 .3917702 0 1 ns5 | 5225 .135311 .3420882 0 1 ns6 | 5225 .0558852 .2297218 0 1 -------------+-------------------------------------------------------- ns7 | 5225 .0003828 .0195628 0 1 . . * Define the exogenous regressors using the global macro exogregressors . global exogregressors black south76 smsa76 reg2-reg9 /* > */ smsa66 momdad14 sinmom14 nodaded nomomed daded momed famed1-famed8 . . * Write data to a text (ascii) file so can use with programs other than stata . outfile wage76 grade76 exp76 expsq76 col4 age76 agesq76 black south76 smsa76 reg2-reg9 /* > */ smsa66 momdad14 sinmom14 nodaded nomomed daded momed famed1-famed8 /* > */ using mma04p4ivweak.asc, replace . . . ********** (1) OLS AND IV ESTIMATES: COLUMNS 1 AND 2 OF KLING TABLE 1 . . * RETAIN cases for the analysis . * Here drop if missing wages or missing schooling or not at first interview . keep if wage76!=. & grade76!=. & nlsflt==1 (2216 observations deleted) . . * DESCRIBE dependent variable, regressors and instruments . desc wage76 grade76 exp76 expsq76 col4 age76 agesq76 $exogregressors storage display value variable name type format label variable label ------------------------------------------------------------------------------- wage76 float %9.0g '76 Wage grade76 float %9.0g '76 Grade level exp76 float %9.0g '76 experience, (10 + age66) - grade76 - 6) expsq76 float %9.0g '76 experience, exp76 ^2/100 col4 float %9.0g If any 4-year college nearby (r0004000!=4) age76 float %9.0g '76 age (age66 +10) agesq76 float %9.0g '76 age squared (age76^2) black float %9.0g Race (r0002300) n= 5225 mean= 1.296 min= 1 max=3 south76 float %9.0g If lived in South in 1976 (r0437511=1) smsa76 float %9.0g If lived in SMSA in 1976 (r0437515=1,2) reg2 float %9.0g If lived in Region 2 (region= MidAtl) reg3 float %9.0g If lived in Region 3 (region= ENC) reg4 float %9.0g If lived in Region 4 (region= WNC) reg5 float %9.0g If lived in Region 5 (region= SA ) reg6 float %9.0g If lived in Region 6 (region= ESC) reg7 float %9.0g If lived in Region 7 (region= WSC) reg8 float %9.0g If lived in Region 8 (region= M ) reg9 float %9.0g If lived in Region 9 (region= P ) smsa66 float %9.0g If lived in SMSA in 1966 (r0002455=1,2) momdad14 float %9.0g If lived with both parents at age 14 sinmom14 float %9.0g If lived with mother only at age 14 nodaded float %9.0g If father has no formal education nomomed float %9.0g If mother has no formal education daded float %9.0g Mean grade level of father momed float %9.0g Mean grade level of mother famed1 float %9.0g If mgrade> 12 & fgrade> 12 (famed=1) famed2 float %9.0g If mgrade>=12 & fgrade>=12 (famed=2) famed3 float %9.0g If mgrade==12 & fgrade==12 (famed=3) famed4 float %9.0g If mgrade>=12 & fgrade==-1 (famed=4) famed5 float %9.0g If fgrade>=12 (famed=5) famed6 float %9.0g If mgrade>=12 & fgrade> -1 (famed=6) famed7 float %9.0g If mgrade>=9 & fgrade>=9 (famed=7) famed8 float %9.0g If mgrade> -1 & fgrade> -1 (famed=8) . . * SUMMARIZE dependent variable, regressors and instruments . sum wage76 grade76 exp76 expsq76 col4 age76 agesq76 $exogregressors Variable | Obs Mean Std. Dev. Min Max -------------+-------------------------------------------------------- wage76 | 3010 1.656664 .443798 0 3.1797 grade76 | 3010 13.26346 2.676913 1 18 exp76 | 3010 8.856146 4.141672 0 23 expsq76 | 3010 .9557907 .8461831 0 5.29 col4 | 3010 .6820598 .4657535 0 1 -------------+-------------------------------------------------------- age76 | 3010 28.1196 3.137004 24 34 agesq76 | 3010 800.5495 180.7484 576 1156 black | 3010 .2335548 .4231624 0 1 south76 | 3010 .4036545 .4907113 0 1 smsa76 | 3010 .7129568 .4524571 0 1 -------------+-------------------------------------------------------- reg2 | 3010 .1607973 .367405 0 1 reg3 | 3010 .1956811 .39679 0 1 reg4 | 3010 .0641196 .2450066 0 1 reg5 | 3010 .2083056 .406164 0 1 reg6 | 3010 .0960133 .2946584 0 1 -------------+-------------------------------------------------------- reg7 | 3010 .1099668 .3129003 0 1 reg8 | 3010 .0282392 .165683 0 1 reg9 | 3010 .0903654 .2867522 0 1 smsa66 | 3010 .6495017 .4772053 0 1 momdad14 | 3010 .7893688 .4078247 0 1 -------------+-------------------------------------------------------- sinmom14 | 3010 .1006645 .3009339 0 1 nodaded | 3010 .2292359 .4204111 0 1 nomomed | 3010 .1172757 .321802 0 1 daded | 3010 9.988262 3.266511 0 18 momed | 3010 10.33675 2.987507 0 18 -------------+-------------------------------------------------------- famed1 | 3010 .0614618 .2402153 0 1 famed2 | 3010 .0787375 .2693734 0 1 famed3 | 3010 .1249169 .3306796 0 1 famed4 | 3010 .0475083 .2127588 0 1 famed5 | 3010 .0790698 .2698925 0 1 -------------+-------------------------------------------------------- famed6 | 3010 .1328904 .3395126 0 1 famed7 | 3010 .0504983 .2190073 0 1 famed8 | 3010 .2202658 .4144947 0 1 . . * OLS estimates of return to schooling. . * This regression computes schooling coeff, se for Table1 col 1 p.359 . * based on all cases (age grp 14-24) reported highest grd cmpl 76 . . reg wage76 grade76 exp76 expsq76 $exogregressors Source | SS df MS Number of obs = 3010 -------------+------------------------------ F( 29, 2980) = 44.94 Model | 180.320527 29 6.21794919 Prob > F = 0.0000 Residual | 412.32209 2980 .138363117 R-squared = 0.3043 -------------+------------------------------ Adj R-squared = 0.2975 Total | 592.642616 3009 .196956669 Root MSE = .37197 ------------------------------------------------------------------------------ wage76 | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- grade76 | .072635 .0036984 19.64 0.000 .0653833 .0798868 exp76 | .0845293 .0066819 12.65 0.000 .0714277 .0976308 expsq76 | -.2289581 .0319499 -7.17 0.000 -.2916041 -.1663121 black | -.1894065 .0194462 -9.74 0.000 -.2275358 -.1512773 south76 | -.1464841 .0260345 -5.63 0.000 -.1975314 -.0954368 smsa76 | .1377121 .0201334 6.84 0.000 .0982353 .1771889 reg2 | .1023805 .0360137 2.84 0.005 .0317662 .1729947 reg3 | .1488958 .0352521 4.22 0.000 .0797748 .2180168 reg4 | .0601267 .0417556 1.44 0.150 -.021746 .1419994 reg5 | .1348504 .0419098 3.22 0.001 .0526752 .2170255 reg6 | .1452831 .0453155 3.21 0.001 .0564302 .2341359 reg7 | .1301968 .044965 2.90 0.004 .0420312 .2183624 reg8 | -.0444289 .0513937 -0.86 0.387 -.1451997 .0563419 reg9 | .1285658 .0389959 3.30 0.001 .0521042 .2050274 smsa66 | .0233775 .019544 1.20 0.232 -.0149436 .0616987 momdad14 | .0693317 .0263402 2.63 0.009 .017685 .1209785 sinmom14 | .0335387 .0354168 0.95 0.344 -.0359052 .1029825 nodaded | -.0390477 .0531089 -0.74 0.462 -.1431815 .0650862 nomomed | .0168143 .0348295 0.48 0.629 -.051478 .0851066 daded | -.0017839 .0043977 -0.41 0.685 -.0104068 .0068389 momed | .0081443 .0041513 1.96 0.050 4.64e-06 .0162839 famed1 | -.1166029 .0788125 -1.48 0.139 -.2711354 .0379296 famed2 | -.052544 .0712753 -0.74 0.461 -.1922977 .0872097 famed3 | -.0719675 .0654608 -1.10 0.272 -.2003205 .0563856 famed4 | -.0197095 .0437058 -0.45 0.652 -.1054062 .0659872 famed5 | -.0252185 .0643526 -0.39 0.695 -.1513985 .1009615 famed6 | -.0733887 .0621076 -1.18 0.237 -.1951667 .0483894 famed7 | -.059927 .0656929 -0.91 0.362 -.188735 .068881 famed8 | -.0738951 .0572428 -1.29 0.197 -.1861345 .0383444 _cons | -.0278815 .1005974 -0.28 0.782 -.2251288 .1693659 ------------------------------------------------------------------------------ . estimates store ols . . * IV Instrumental variables estimates of return to schooling. . * This regression computes schooling coeff and se for Table 1. col 2 p.359 . * Endogenous variables: schooling, experience, experience squared . * Excl instruments: college in cnty, age age^2 . * based on all cases (age grp 14-24) reported highest grd cmpl 76 ***/ . . ivreg wage76 $exogregressors /* > */ (grade76 exp76 expsq76 = col4 age76 agesq76 $exogregressors) Instrumental variables (2SLS) regression Source | SS df MS Number of obs = 3010 -------------+------------------------------ F( 29, 2980) = 34.56 Model | 122.395448 29 4.22053269 Prob > F = 0.0000 Residual | 470.247169 2980 .157801063 R-squared = 0.2065 -------------+------------------------------ Adj R-squared = 0.1988 Total | 592.642616 3009 .196956669 Root MSE = .39724 ------------------------------------------------------------------------------ wage76 | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- grade76 | .1324485 .0493419 2.68 0.007 .0357009 .2291961 exp76 | .0632411 .0241061 2.62 0.009 .0159748 .1105074 expsq76 | -.1266694 .1184765 -1.07 0.285 -.3589735 .1056347 black | -.1643766 .0292248 -5.62 0.000 -.2216795 -.1070737 south76 | -.1400178 .0283887 -4.93 0.000 -.1956812 -.0843545 smsa76 | .0909867 .0441338 2.06 0.039 .0044509 .1775224 reg2 | .0753178 .0444167 1.70 0.090 -.0117726 .1624083 reg3 | .1231473 .0431763 2.85 0.004 .038489 .2078057 reg4 | .0241968 .0534911 0.45 0.651 -.0806865 .1290801 reg5 | .1247819 .0455148 2.74 0.006 .0355383 .2140255 reg6 | .135761 .0490304 2.77 0.006 .039624 .2318979 reg7 | .1063645 .0519274 2.05 0.041 .0045472 .2081817 reg8 | -.0850609 .064327 -1.32 0.186 -.2111907 .0410688 reg9 | .0916464 .0515551 1.78 0.076 -.0094409 .1927337 smsa66 | .0379821 .0241116 1.58 0.115 -.0092951 .0852592 momdad14 | .043168 .0354056 1.22 0.223 -.0262539 .11259 sinmom14 | .025849 .0383465 0.67 0.500 -.0493392 .1010373 nodaded | -.0462392 .0570684 -0.81 0.418 -.1581366 .0656583 nomomed | .0266252 .0383434 0.69 0.487 -.048557 .1018074 daded | -.0110565 .0089768 -1.23 0.218 -.0286579 .0065449 momed | -.0017539 .0093223 -0.19 0.851 -.0200326 .0165249 famed1 | -.213271 .1160049 -1.84 0.066 -.4407287 .0141867 famed2 | -.1567074 .1145696 -1.37 0.171 -.3813508 .0679361 famed3 | -.1354685 .0872725 -1.55 0.121 -.3065889 .035652 famed4 | -.0707323 .0627189 -1.13 0.260 -.193709 .0522444 famed5 | -.0699675 .077928 -0.90 0.369 -.2227656 .0828306 famed6 | -.1171712 .0754408 -1.55 0.120 -.2650926 .0307502 famed7 | -.0921498 .0749801 -1.23 0.219 -.2391679 .0548683 famed8 | -.1184618 .0713021 -1.66 0.097 -.2582681 .0213445 _cons | -.4311125 .3567904 -1.21 0.227 -1.130693 .2684678 ------------------------------------------------------------------------------ Instrumented: grade76 exp76 expsq76 Instruments: black south76 smsa76 reg2 reg3 reg4 reg5 reg6 reg7 reg8 reg9 smsa66 momdad14 sinmom14 nodaded nomomed daded momed famed1 famed2 famed3 famed4 famed5 famed6 famed7 famed8 col4 age76 agesq76 ------------------------------------------------------------------------------ . estimates store iv . . ********** (2) NEW ANALYSIS: HETEROSKEDASTIC ROBUST STANDARD ERRORS ********** . . * Heteroskedastic errors makes little difference here. . . quietly reg wage76 grade76 exp76 expsq76 $exogregressors . hettest /* Shows that here there is no heteroskeadsticity for OLS */ Breusch-Pagan / Cook-Weisberg test for heteroskedasticity Ho: Constant variance Variables: fitted values of wage76 chi2(1) = 0.42 Prob > chi2 = 0.5191 . quietly reg wage76 grade76 exp76 expsq76 $exogregressors, robust . estimates store olshet . . quietly ivreg wage76 $exogregressors /* > */ (grade76 exp76 expsq76 = col4 age76 agesq76 $exogregressors), robust . estimates store ivhet . . **** DISPLAY RESULTS IN TABLE 4.5 p.111 . . * Table 4.5 p.111: OLS and IV estimates, s.e.'s and R^2 in Table 4.5 . . * Table reports only the coefficient and standard erros for grade76 . estimates table ols olshet iv ivhet, /* > */ se stats(N ll r2 rss mss rmse df_r) b(%10.4f) ------------------------------------------------------------------ Variable | ols olshet iv ivhet -------------+---------------------------------------------------- grade76 | 0.0726 0.0726 0.1324 0.1324 | 0.0037 0.0039 0.0493 0.0488 exp76 | 0.0845 0.0845 0.0632 0.0632 | 0.0067 0.0068 0.0241 0.0241 expsq76 | -0.2290 -0.2290 -0.1267 -0.1267 | 0.0319 0.0322 0.1185 0.1182 black | -0.1894 -0.1894 -0.1644 -0.1644 | 0.0194 0.0198 0.0292 0.0285 south76 | -0.1465 -0.1465 -0.1400 -0.1400 | 0.0260 0.0280 0.0284 0.0292 smsa76 | 0.1377 0.1377 0.0910 0.0910 | 0.0201 0.0193 0.0441 0.0440 reg2 | 0.1024 0.1024 0.0753 0.0753 | 0.0360 0.0350 0.0444 0.0432 reg3 | 0.1489 0.1489 0.1231 0.1231 | 0.0353 0.0338 0.0432 0.0418 reg4 | 0.0601 0.0601 0.0242 0.0242 | 0.0418 0.0412 0.0535 0.0531 reg5 | 0.1349 0.1349 0.1248 0.1248 | 0.0419 0.0428 0.0455 0.0459 reg6 | 0.1453 0.1453 0.1358 0.1358 | 0.0453 0.0452 0.0490 0.0483 reg7 | 0.1302 0.1302 0.1064 0.1064 | 0.0450 0.0457 0.0519 0.0516 reg8 | -0.0444 -0.0444 -0.0851 -0.0851 | 0.0514 0.0509 0.0643 0.0619 reg9 | 0.1286 0.1286 0.0916 0.0916 | 0.0390 0.0388 0.0516 0.0504 smsa66 | 0.0234 0.0234 0.0380 0.0380 | 0.0195 0.0187 0.0241 0.0231 momdad14 | 0.0693 0.0693 0.0432 0.0432 | 0.0263 0.0257 0.0354 0.0352 sinmom14 | 0.0335 0.0335 0.0258 0.0258 | 0.0354 0.0359 0.0383 0.0384 nodaded | -0.0390 -0.0390 -0.0462 -0.0462 | 0.0531 0.0511 0.0571 0.0550 nomomed | 0.0168 0.0168 0.0266 0.0266 | 0.0348 0.0344 0.0383 0.0375 daded | -0.0018 -0.0018 -0.0111 -0.0111 | 0.0044 0.0044 0.0090 0.0089 momed | 0.0081 0.0081 -0.0018 -0.0018 | 0.0042 0.0042 0.0093 0.0093 famed1 | -0.1166 -0.1166 -0.2133 -0.2133 | 0.0788 0.0792 0.1160 0.1160 famed2 | -0.0525 -0.0525 -0.1567 -0.1567 | 0.0713 0.0698 0.1146 0.1132 famed3 | -0.0720 -0.0720 -0.1355 -0.1355 | 0.0655 0.0644 0.0873 0.0865 famed4 | -0.0197 -0.0197 -0.0707 -0.0707 | 0.0437 0.0416 0.0627 0.0601 famed5 | -0.0252 -0.0252 -0.0700 -0.0700 | 0.0644 0.0625 0.0779 0.0763 famed6 | -0.0734 -0.0734 -0.1172 -0.1172 | 0.0621 0.0601 0.0754 0.0735 famed7 | -0.0599 -0.0599 -0.0921 -0.0921 | 0.0657 0.0640 0.0750 0.0730 famed8 | -0.0739 -0.0739 -0.1185 -0.1185 | 0.0572 0.0545 0.0713 0.0682 _cons | -0.0279 -0.0279 -0.4311 -0.4311 | 0.1006 0.0997 0.3568 0.3528 -------------+---------------------------------------------------- N | 3010.0000 3010.0000 3010.0000 3010.0000 ll | -1279.2297 -1279.2297 r2 | 0.3043 0.3043 0.2065 0.2065 rss | 412.3221 412.3221 470.2472 470.2472 mss | 180.3205 180.3205 122.3954 122.3954 rmse | 0.3720 0.3720 0.3972 0.3972 df_r | 2980.0000 2980.0000 2980.0000 2980.0000 ------------------------------------------------------------------ legend: b/se . . ********** (3) NEW ANALYSIS: CHECK FOR WEAK INSTRUMENTS ********** . . * Model is y = b1*x1 + x2'b2 + u . * where x1 is scalar endogenous (grade76) . * where x2 is vector of regressors that includes . * exp76 and exp76 which are also endogenous . * and $exogregressors which are exogenous . * and the instruments Z are grade76 col4 age76 agesq76 $exogregressors . . * Check for weak instruments . * Focus on grade76 but can also do this for the other two endogenous regressors. . * In this example no problems for the other two: . * as age and age-squared are good instruments for exp and exp-squared. . . **** (A) Simple analysis R-squared and F-test [Given in Table 4.5] . . * R2 from regress endogenous regressor on instruments . * This is same as correlation between x1 and projection of x1 on Z . quietly reg grade76 col4 age76 agesq76 $exogregressors . di e(r2) " r2 of x1 on Z" .29677588 r2 of x1 on Z . . * Do the partial F-test on the three instruments . * This is the standard first-stage regression F-test . . **** DISPLAY RESULT IN TABLE 4.5 page 111 . . * First-stage F statistic given in Table 4.5 . test col4 age76 agesq76 ( 1) col4 = 0 ( 2) age76 = 0 ( 3) agesq76 = 0 F( 3, 2980) = 8.07 Prob > F = 0.0000 . . * Compare this to R-squared when only regress on instruments without Z . quietly reg grade76 $exogregressors . di e(r2) " r2 of x1 on Z with the three additional instruments dropped" .29106483 r2 of x1 on Z with the three additional instruments dropped . . * Obtain first-stge F for the other two endogenous . quietly reg exp76 col4 age76 agesq76 $exogregressors . test col4 age76 agesq76 ( 1) col4 = 0 ( 2) age76 = 0 ( 3) agesq76 = 0 F( 3, 2980) = 1772.03 Prob > F = 0.0000 . quietly reg expsq76 col4 age76 agesq76 $exogregressors . test col4 age76 agesq76 ( 1) col4 = 0 ( 2) age76 = 0 ( 3) agesq76 = 0 F( 3, 2980) = 1542.36 Prob > F = 0.0000 . . **** (B) Minimum eigenvalue of matrix analog of the first-stage F statistic . * proposed by Stock et al (2002) and tables in Stock and Yogo (2003) . * This test is not done here. . . **** (C) Bound et al (1995) partial R-squared . . * Not relevant here as more than one endogenous regressor . * If only one endogenous regressor x1 Bound et al purge the effect of x2 . * by (1) get residual from regress x1 on x2 . * (2) get the residuals from regress z on x2 . * and then get the R-squared from regress (1) on (2). . . **** (D) Shea (1997) partial R-squared [Given in Table 4.5] . . * Here we have three endogenous regressors. . * Focus on the endogenous schooling regressor. . * For the other two just need to replace the first line of (1) . * e.g. quietly reg exp76 grade76 expsq76 $exogregressors . * and replace the first line of (2B) . * e.g. quietly reg exp76hat grade76hat expsq76hat $exogregressors . . * (1) Form x1 - x1tilda: residual from regress x1 on other regressors . quietly reg grade76 exp76 expsq76 $exogregressors . predict x1minusx1tilda, resid . . * (2) Form x1hat - x1hattilda: residual from regress x1hat on fitted values of other regressors . * (2A) First get the fitted values from regress endogenous on instruments . quietly reg grade76 col4 age76 agesq76 $exogregressors . predict grade76hat, xb . di e(r2) " r2 from regress x1 on Z" .29677588 r2 from regress x1 on Z . quietly reg exp76 col4 age76 agesq76 $exogregressors . predict exp76hat, xb . di e(r2) " r2 from regress second endog regressor on Z" .70622765 r2 from regress second endog regressor on Z . quietly reg expsq76 col4 age76 agesq76 $exogregressors . predict expsq76hat, xb . di e(r2) " r2 from regress third endog regressor on Z" .67573235 r2 from regress third endog regressor on Z . * Fitted values for the exogenous from regress exogenous on instruments are the exogenous . * (2B) Run the regression of x1hat on fitted values of other regressors . quietly reg grade76hat exp76hat expsq76hat $exogregressors . di e(r2) " r2 from regress prediction of x1 on predictions of x2 .98987117 r2 from regress prediction of x1 on predictions of x2 . predict x1hatminusx1hattilda, resid . . * (3) Form the correlation between (1) and (2) . corr x1minusx1tilda x1hatminusx1hattilda (obs=3010) | x1minu~a x1hatm~a -------------+------------------ x1minusx1t~a | 1.0000 x1hatminus~a | 0.0800 1.0000 . . **** DISPLAY RESULT IN TABLE 4.5 page 111 . . * Shea's Partial R^2 in Table 4.5 . di r(rho)^2 " Shea's partial R-squared measure" .00640757 Shea's partial R-squared measure . . sum grade76 grade76hat exp76 exp76hat expsq76 expsq76hat grade76 x1minusx1tilda x1hatminusx1hattilda grade76hat Variable | Obs Mean Std. Dev. Min Max -------------+-------------------------------------------------------- grade76 | 3010 13.26346 2.676913 1 18 grade76hat | 3010 13.26346 1.458306 8.919074 17.42063 exp76 | 3010 8.856146 4.141672 0 23 exp76hat | 3010 8.856146 3.480551 1.329216 17.68953 expsq76 | 3010 .9557907 .8461831 0 5.29 -------------+-------------------------------------------------------- expsq76hat | 3010 .9557907 .6955874 -.3913698 2.917523 grade76 | 3010 13.26346 2.676913 1 18 x1minusx1t~a | 3010 -8.71e-10 1.833502 -6.948598 5.661138 x1hatminus~a | 3010 -6.86e-11 .1467669 -.3732457 .3033035 grade76hat | 3010 13.26346 1.458306 8.919074 17.42063 . . **** (E) Poskitt-Skeels (2002) partial R-squared . * Not done here . . **** (F) If model was over-identified then do test of over-identifying restrictions . * Not done here as model is just-identified . . ********** CLOSE OUTPUT . log close log: c:\Imbook\bwebpage\Section2\mma04p4ivweak.txt log type: text closed on: 17 May 2005, 13:46:03 ------------------------------------------------------------------------------------------------------------------------------------------------------