Bivariate Probit and Logit Models SAS Program and Output.pdf ...

0 downloads 178 Views 53KB Size Report
Variable N Mean Std Dev Minimum Maximum. hlthe. dmdu. age. linc. ndisease. 5574. 5574. 5574. 5574. 5574. 0.5412630. 0.67
* Bivariate Probit Model in SAS; * Copyright 2013 by Ani Katchova; proc import out= work.data datafile= "C:\Econometrics\Data\bivariate_health.csv" dbms=csv replace; getnames=yes; datarow=2; run; %let ylist1=hlthe; %let ylist2=dmdu; %let xlist= age linc ndisease; proc means data=data; var &ylist1 &ylist2 &xlist; run; proc freq data=data; tables &ylist1 &ylist2; run; *Probit models; proc qlim data=data; model &ylist1 = &xlist / discrete (dist=normal); run; proc qlim data=data; model &ylist2 = &xlist / discrete (dist=normal); run; *Bivariate probit; proc qlim data=data; model &ylist1 &ylist2 = &xlist; endogenous &ylist1 &ylist2 ~ discrete; output out=mfx marginal; run; *Bivariate probit allowing different set of regressors; proc qlim data=data method=qn; model &ylist1 = &xlist; model &ylist2 = &xlist; endogenous &ylist2 &ylist1 ~ discrete; output out=mfx marginal; run; /*Marginal effects; proc means data=mfx n mean std; var Meff_P2_hlthe_age Meff_P2_hlthe_linc Meff_P2_hlthe_ndisease Meff_P2_dmdu_age Meff_P2_dmdu_linc Meff_P2_dmdu_ndisease; run;

The SAS System The MEANS Procedure Variable

N

Mean

Std Dev

Minimum

Maximum

hlthe

5574

0.5412630

0.4983392

0

1.0000000

dmdu

5574

0.6713312

0.4697715

0

1.0000000

age

5574

25.5761339

16.7301105

0.0253251

63.2751500

linc

5574

8.6969290

1.2205920

0

10.2832400

ndisease

5574

11.2052651

6.7889585

0

58.6000000

The SAS System The FREQ Procedure hlthe Frequency Percent Cumulative Cumulative Frequency Percent 0

2557

45.87

2557

45.87

1

3017

54.13

5574

100.00

dmdu Frequency Percent Cumulative Cumulative Frequency Percent 0

1832

32.87

1832

32.87

1

3742

67.13

5574

100.00

The SAS System The QLIM Procedure Discrete Response Profile of hlthe Total Frequency

Index

Value

1

0

2557

2

1

3017

Model Fit Summary Number of Endogenous Variables

1

Endogenous Variable

hlthe

Number of Observations

5574

Log Likelihood

-3554

Maximum Absolute Gradient

0.00358

Number of Iterations Optimization Method

9 Quasi-Newton

AIC

7116

Schwarz Criterion

7143

Goodness-of-Fit Measures Measure

Value Formula

Likelihood Ratio (R) 580.81 2 * (LogL - LogL0) Upper Bound of R (U) 7689.2 - 2 * LogL0 Aldrich-Nelson Cragg-Uhler 1

0.0944 R / (R+N) 0.099 1 - exp(-R/N)

Cragg-Uhler 2

0.1322 (1-exp(-R/N)) / (1-exp(-U/N))

Estrella

0.1027 1 - (1-R/U)^(U/N)

Adjusted Estrella

0.1013 1 - ((LogL-K)/LogL0)^(-2/N*LogL0)

McFadden's LRI

0.0755 R / U

Veall-Zimmermann

0.1628 (R * (U+N)) / (U * (R+N))

McKelvey-Zavoina

0.1567

N = # of observations, K = # of regressors Algorithm converged. Parameter Estimates Parameter DF

Estimate Standard Error t Value

Approx Pr > |t|

Parameter Estimates Parameter DF

Estimate Standard Error t Value

Approx Pr > |t|

Intercept

1 -0.230438

0.133559

-1.73

0.0845

age

1 -0.017819

0.001083

-16.46

Suggest Documents