SECTION II: PROBABILITY MODELS 1
SECTION II: Aggregate Data. Fraction of births with low birth weight per province. Model A: OLS, using observations 1 260 Heteroskedasticity-robust standard errors, variant HC1 const 1.91797 0.679055 2.8245 0.0051 m age 0.00116767 0.00176503 0.6616 0.5089 married 0.0301099 0.0113383 2.6556 0.0084 momage 0.126591 0.0439073 2.8832 0.0043 momage2 0.00210753 0.000719053 2.9310 0.0037 inlf 0.0332909 0.0102804 3.2383 0.0014 Sum squared resid 0.014364 S.E. of regression 0.007520 R 2 0.276533 Adjusted R 2 0.262291 F (5, 254) 18.55076 P-value(F ) 1.06e 15 Log-likelihood 905.5628 Akaike criterion 1799.126 Schwarz criterion 1777.761 Hannan Quinn 1790.537 3
Model B: Pooled OLS, using 260 observations const 1.89247 0.585101 3.2344 0.0014 m age 0.00173297 0.00156068 1.1104 0.2679 married 0.0292426 0.0128107 2.2827 0.0233 momage 0.124555 0.0381506 3.2648 0.0012 momage2 0.00208411 0.000620996 3.3561 0.0009 inlf 0.0323011 0.00917413 3.5209 0.0005 Dyear 2 0.00142028 0.00147778 0.9611 0.3374 Dyear 3 0.00295910 0.00150650 1.9642 0.0506 Dyear 4 0.00279514 0.00154592 1.8081 0.0718 Dyear 5 0.00102573 0.00160204 0.6403 0.5226 Sum squared resid 0.014054 S.E. of regression 0.007498 R 2 0.292124 Adjusted R 2 0.266640 F (9, 250) 11.46325 P-value(F ) 5.06e 15 Log-likelihood 908.3950 Akaike criterion 1796.790 Schwarz criterion 1761.183 Hannan Quinn 1782.476 ˆρ 0.715466 Durbin Watson 0.509472
Model C: Fixed-effects, using 260 observations Robust (HAC) standard errors const 0.789962 1.10661 0.7139 0.4761 momage 0.0502255 0.0736061 0.6824 0.4958 momage2 0.000844071 0.00122440 0.6894 0.4914 inlf 0.0413644 0.0171788 2.4079 0.0169 Sum squared resid 0.003921 S.E. of regression 0.004373 LSDV R 2 0.802520 Within R 2 0.095151 F (54, 205) 15.42737 P-value(F ) 2.41e 48 Log-likelihood 1074.357 Akaike criterion 2038.714 Schwarz criterion 1842.876 Hannan Quinn 1959.984 ˆρ 0.194313 Durbin Watson 1.938624 Joint test on named regressors Test statistic: F (3, 205) = 7.18573 with p-value = P (F (3, 205) > 7.18573) = 0.000130733 Test for differing group intercepts Null hypothesis: The groups have a common intercept Test statistic: F (51, 205) = 10.7062 with p-value = P (F (51, 205) > 10.7062) = 7.07532e-36
Model D: Fixed-effects, using 260 observations Robust (HAC) standard errors const 1.15502 1.11549 1.0354 0.3017 momage 0.0737986 0.0742726 0.9936 0.3216 momage2 0.00122654 0.00123696 0.9916 0.3226 inlf 0.0339295 0.0194595 1.7436 0.0828 dt 2 0.00127786 0.000893603 1.4300 0.1543 dt 3 0.00271773 0.00104817 2.5928 0.0102 dt 4 0.00268211 0.00110089 2.4363 0.0157 dt 5 0.000498857 0.00127686 0.3907 0.6964 Sum squared resid 0.003624 S.E. of regression 0.004246 LSDV R 2 0.817490 Within R 2 0.163743 F (58, 201) 15.52254 P-value(F ) 3.32e 49 Log-likelihood 1084.605 Akaike criterion 2051.210 Schwarz criterion 1841.130 Hannan Quinn 1966.755 ˆρ 0.212082 Durbin Watson 1.976061 Joint test on named regressors Test statistic: F (7, 201) = 5.6224 with p-value = P (F (7, 201) > 5.6224) = 6.14119e-06 Test for differing group intercepts Null hypothesis: The groups have a common intercept Test statistic: F (51, 201) = 11.3449 with p-value = P (F (51, 201) > 11.3449) = 3.2927e-37 Wald test for joint significance of time dummies Asymptotic test statistic: χ 2 (4) = 22.6249 with p-value = 0.000150459
Model E: Fixed-effects, using 260 observations Robust (HAC) standard errors const 1.15502 1.11549 1.0354 0.3017 momage 0.0737986 0.0742726 0.9936 0.3216 momage2 0.00122654 0.00123696 0.9916 0.3226 inlf 0.0339295 0.0194595 1.7436 0.0828 Dyear 2 0.00127786 0.000893603 1.4300 0.1543 Dyear 3 0.00271773 0.00104817 2.5928 0.0102 Dyear 4 0.00268211 0.00110089 2.4363 0.0157 Dyear 5 0.000498857 0.00127686 0.3907 0.6964 Sum squared resid 0.003624 S.E. of regression 0.004246 LSDV R 2 0.817490 Within R 2 0.163743 F (58, 201) 15.52254 P-value(F ) 3.32e 49 Log-likelihood 1084.605 Akaike criterion 2051.210 Schwarz criterion 1841.130 Hannan Quinn 1966.755 ˆρ 0.212082 Durbin Watson 1.976061 Joint test on named regressors Test statistic: F (7, 201) = 5.6224 with p-value = P (F (7, 201) > 5.6224) = 6.14119e-06 Test for differing group intercepts Null hypothesis: The groups have a common intercept Test statistic: F (51, 201) = 11.3449 with p-value = P (F (51, 201) > 11.3449) = 3.2927e-37
Model F: Random-effects (GLS), using 260 observations const 0.536011 0.805177 0.6657 0.5062 m age 0.00101057 0.00191789 0.5269 0.5987 married 0.0249506 0.0205637 1.2133 0.2261 momage 0.0356566 0.0526929 0.6767 0.4992 momage2 0.000605116 0.000870041 0.6955 0.4874 inlf 0.0369104 0.0128159 2.8801 0.0043 Dyear 2 0.00131835 0.000857269 1.5378 0.1254 Dyear 3 0.00277439 0.000934201 2.9698 0.0033 Dyear 4 0.00261512 0.00104517 2.5021 0.0130 Dyear 5 0.000622466 0.00118445 0.5255 0.5997 Sum squared resid 0.014416 S.E. of regression 0.007578 Log-likelihood 905.0915 Akaike criterion 1790.183 Schwarz criterion 1754.576 Hannan Quinn 1775.869 ˆσ 2 ε = 1.80274e 05 ˆσ 2 u = 4.47315e 05 θ = 0.716093 Breusch-Pagan test Null hypothesis: Variance of the unit-specific error = 0 Asymptotic test statistic: χ 2 (1) = 230.658 with p-value = 4.28436e-52 Hausman test Null hypothesis: GLS estimates are consistent Asymptotic test statistic: χ 2 (7) = 4.58549 with p-value = 0.7104
SECTION II. INSTRUMENTAL VARIABLES: Microdata. work status Birth weight and mother s Table 1: Birth weight and mother s work status: Models (1) (2) (3) (4) (5) (6) (7) (8) OLS OLS OLS OLS OLS 2SLS 2SLS OLS VARIABLES bw inlf inlf bw bw bw bw e1 inlf -11.27-230.78-255.02-230.78-255.02 (0.344) (24.362) (14.615) (24.938) (15.028) momage 1.26 0.02 0.02 6.67 7.27 6.67 7.27 0.00 (0.036) (0.000) (0.000) (0.601) (0.362) (0.615) (0.372) (0.036) married 46.40-0.02-0.02 41.50 40.96 41.50 40.96-0.00 (0.420) (0.000) (0.000) (0.687) (0.531) (0.704) (0.548) (0.433) qbirth1m -0.02-0.02-0.59 (0.001) (0.001) (0.601) qbirth1p -0.03 0.43 (0.001) (0.593) v1 219.55 (24.364) v2 243.89 (14.619) Constant 3,192.21-0.16-0.16 3,155.87 3,151.86 3,155.87 3,151.86 0.00 (1.054) (0.001) (0.001) (4.169) (2.640) (4.268) (2.715) (1.087) F excluded 1623.87 2268.03 instruments p-value= p-value= 0.00 0.01 Ex. Instr. qbirth1m qbirth1m and qbirth1p Observations 8244952 8244952 8244952 8244952 8244952 8244952 8244952 8244952 R-squared 0.0024 0.057 0.0572 0.0024 0.002404 0.0012 0.0013 7.4*10ˆ(-7) v1=residuals model column 2 v2=residuals model column 3 e1=residuals model column 7 The dependent variable and the estimator used are indicated at the top of each column. 9