Download presentation
Presentation is loading. Please wait.
Published byAmelia O’Brien’ Modified over 9 years ago
1
Two-stage least squares 1
2
D1 S1 2
3
P Q D1 D2D2 S1 S2 Increase in income Increase in costs 3
4
D1 S2 D2 D3 S3 S1 4
5
Births in 1997 Variable Smoked during pregnancyDid not smoke Birthweight of child (grams) 31283347 Mom Unmarried51.2%29.7% Mom a teen16.2%12.7% Received inadequate prenatal care 17.2%11.5% Black12.6%17.6% Mom is hispanic4.2%15.5% Mom < HS degree*47.9%30.1% Mom HS degree *24.1%10.6% 5
6
6 Z i =1 1 =0.57 Z i =0 0 =0.80
7
7 0 =3186 1 =3278 Sample Sizes
8
8 Data set used in previous 2 slides bweight.dta Three variables –treat (1=yes, 0=no) –smoke (1=yes, 0=no) –bweight (birth weight in grams)
9
9. * how many people are treated. tab treat dummy | variable, | =1 if | treated, =0 | otherwise | Freq. Percent Cum. ------------+----------------------------------- 0 | 438 50.52 50.52 1 | 429 49.48 100.00 ------------+----------------------------------- Total | 867 100.00
10
10 Two-sample t test with equal variances ------------------------------------------------------------------------------ Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval] ---------+-------------------------------------------------------------------- 0 | 438.7990868.0191673.4011415.7614152.8367583 1 | 429.5687646.0239388.4958271.5217124.6158167 ---------+-------------------------------------------------------------------- combined | 867.6851211.0157833.4647357.6541432.716099 ---------+-------------------------------------------------------------------- diff |.2303222.0306003.1702627.2903816 ------------------------------------------------------------------------------ diff = mean(0) - mean(1) t = 7.5268 Ho: diff = 0 degrees of freedom = 865
11
11. ttest bweight, by(treat) Two-sample t test with equal variances ------------------------------------------------------------------------------ Group | Obs Mean Std. Err. Std. Dev. [95% Conf. Interval] ---------+-------------------------------------------------------------------- 0 | 438 3178.131 26.1705 547.708 3126.695 3229.566 1 | 429 3278.05 25.859 535.5997 3227.224 3328.877 ---------+-------------------------------------------------------------------- combined | 867 3227.572 18.46639 543.7403 3191.328 3263.816 ---------+-------------------------------------------------------------------- diff | -99.91993 36.79962 -172.1469 -27.69294 ------------------------------------------------------------------------------ diff = mean(0) - mean(1) t = -2.7152 Ho: diff = 0 degrees of freedom = 865
12
12
13
13 * 1st stage. reg smoke treat Source | SS df MS Number of obs = 867 -------------+------------------------------ F( 1, 865) = 56.65 Model | 11.4969824 1 11.4969824 Prob > F = 0.0000 Residual | 175.54108 865.202937665 R-squared = 0.0615 -------------+------------------------------ Adj R-squared = 0.0604 Total | 187.038062 866.215979287 Root MSE =.45049 ------------------------------------------------------------------------------ smoke | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- treat | -.2303222.0306003 -7.53 0.000 -.2903816 -.1702627 _cons |.7990868.0215251 37.12 0.000.7568393.8413342 ------------------------------------------------------------------------------
14
14. * reduced form. reg bweight treat Source | SS df MS Number of obs = 867 -------------+------------------------------ F( 1, 865) = 7.37 Model | 2163797.34 1 2163797.34 Prob > F = 0.0068 Residual | 253872143 865 293493.807 R-squared = 0.0085 -------------+------------------------------ Adj R-squared = 0.0073 Total | 256035940 866 295653.51 Root MSE = 541.75 ------------------------------------------------------------------------------ bweight | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- treat | 99.91993 36.79962 2.72 0.007 27.69294 172.1469 _cons | 3178.131 25.88585 122.77 0.000 3127.324 3228.937 ------------------------------------------------------------------------------
15
15. reg bweight smoke (treat) Instrumental variables (2SLS) regression Source | SS df MS Number of obs = 867 -------------+------------------------------ F( 1, 865) = 6.65 Model | -25215738.4 1 -25215738.4 Prob > F = 0.0101 Residual | 281251678 865 325146.449 R-squared =. -------------+------------------------------ Adj R-squared =. Total | 256035940 866 295653.51 Root MSE = 570.22 ------------------------------------------------------------------------------ bweight | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- smoke | -433.8268 168.1696 -2.58 0.010 -763.895 -103.7585 _cons | 3524.796 116.8327 30.17 0.000 3295.487 3754.105 ------------------------------------------------------------------------------
16
Angrist: Vietnam Draft Lottery 16
17
17 Vietnam era service Defined as 1964-1975 Estimated 8.7 million served during era 3.4 million were in SE Asia 2.6 million served in Vietnam 1.6 million saw combat 203K wounded in action, 153K hospitalized 58,000 deaths http://www.history.navy.mil/library/online/america n%20war%20casualty.htm#t7
18
1980 Men, 1940-1952 Cohorts VariableNon-veteransVeterans In labor force93.2%95.9% Unemployed5.0%4.7% Labor earnings$15,155$15,875 Nonwhite16.8%12.3% < HS degree21.5%8.8% HS degree49.4%67.8% College degree28.9%23.3% Married72.1%75.5% 18
19
OLS Estimates Impact of Viet Vet Status Independent Variable Labor EarningsUnemployed Age510 (6.5)-0.0021 (0.0001) Non-white-3446 (67)0.029 (0.0014) < high school-9449 (74)0.078 (0.0015) High school-4800 (57)0.0032 (0.0011) Viet vet523 (53)-0.000 (0.0010) Mean of outcome $15,3724.9% R2R2 0.120.02 19
20
20 Vietnam Era Draft 1 st part of war, operated liked WWII and Korean War At age 18 men report to local draft boards Could receive deferment for variety of reasons (kids, attending school) If available for service, pre-induction physical and tests Military needs determined those drafted
21
21 Everyone drafted went to the Army Local draft boards filled army. Priorities –Delinquents, volunteers, non-vol. 19-25 –For non-vol., determined by age College enrollment powerful way to avoid service –Men w. college degree 1/3 less likely to serve
22
22 Draft Lottery Proposed by Nixon Passed in Nov 1969, 1 st lottery Dec 1, 1969 1st lottery for men age 19-26 on 1/1/70 –Men born 1944-1950. Randomly assigned number 1-365, Draft Lottery number (DLN) Military estimates needs, sets threshold T If DLN<=T, drafted
23
23 If volunteer, could get better assignment Thresholds for service DraftYear of BirthThreshold 19701946-50195 19711951125 19721952 95 Draft suspended in 1973
24
24
25
25 Model Sample, men from 1950-1953 birth cohorts Y i = earnings X i = Vietnam military service (1=yes, 0=no) Z i = draft eligible, that is DLN <=T (1=yes, 0=no)
26
26 Put this all together Model of interest Y i = β 0 + x i β 1 + ε i First stage x i = θ 0 + z i θ 1 + μ i θ 1 =(dx/dz)
27
27 1 st stage Because Z is dichotomous (1 and 0), this makes it easy 1 - 0 = θ 1 (change in military service from having a low DLN)
28
28 Reduced form y i = π o + z i π 1 + v i π 1 = dy/dz=(dy/dx)(dx/dz)
29
29 Intention to treat y i = γ o + z i γ 1 + v i 1 - 0 = π 1 (difference in earnings for those drafted and those not)
30
30 Divide reduced form by 1 st stage π 1 /θ 1 = (dy/dx)(dx/dz)/(dx/dz) = dy/dx Recall the equation of interest y i = β 0 + x i β 1 + ε i The units of measure are β 1 = dy/dx So the ratio π 1 /θ 1 is an estimate of β 1
31
31 β 1 = dy/dx β 1 = [ 1 - 0 ]/[ 1 - 0 ]
32
32 1 o
33
33 Graph of 1 - 0
34
34 1 - 0 in numbers
35
35 Β iv = ( 1 - 0 )/( 1 - 0 ) = -487.8/0.159 = $3067.9 CPI 78 = 65.2 CPI 81 =90.9 65.2/90.9 =.7173.717*3067.92 = $2199
36
36 Although DLN is random, what are some ways that a low DLN could DIRECTLY change wages
37
37
38
38
39
39 Angrist and Evans: The impact of children on labor supply
40
40 Introduction 2 key labor market trends in the past 40 years –Rising labor force participation of women –Falling fertility These two fact are intimately linked, but how? –Are women working more because they are having less children –Are women having less children because they are working more
41
41
42
42
43
43
44
44 Note that between 1970 and 1990 –Mean children ever born has fallen by 33%, from 1.78 to 1.18 –% worked last year increased by 32%, from 60 to 79% Hundreds have studies have attempted to address these questions Lots of persistent relationships, but what have we measured?
45
45 Women with children are not randomly assigned Who is most likely to have large families? –Lower educated –Those with lower wages –Certain minority groups –Certain religious groups –Those who want more children
46
46 Problem is, many of these same groups are also those most likely to be out of the labor force Of the lower labor supply women among women with young children, how much is due to the kids, how much is attributable to some of these other factors?
47
47 Preferences for sex mix Among married couples who desire 2+ kids –66% wives and 75% of husbands prefer mix Of women with 2 boys and desiring a 3 rd, 85% would prefer a girl Of women with 2 girls and desiring a 3 rd, 84% would prefer a boy
48
48
49
49
50
50
51
51 “The desire for a son is the father of many daughters”
52
52 The sex composition is only impacting 6 percent of women So the change in labor supply should be for this group only, So, if we divide -0.008 by 0.06, we get -0.008/0.06 = -0.133 Having a 3 rd child will reduce labor supply by 13.3 percentage points
53
53 Exactly identified model With 1 instrument Over-identified Model with 2 instruments
54
54
55
55
56
56
57
57
58
58 Exactly Identified Model
59
59
60
60
61
61 Over-identified Model
62
62 Test the coefficients on twoboys and twogirls are the same test the coefficient on twobots and twogirls are both equal to zero
63
63
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.