HealthRWM {momentfit} | R Documentation |
The dataset is used in Greene (2012) and is taken from Riphahn, Wambach, Million (2003).
data("HealthRWM")
A data frame with 27326 observations on the following 25 variables.
ID
Person-identification number
female
Female=1; male=0
year
Calendar year of the observation
age
Age in years
hsat
Health satisfaction, coded 0 (low) to 10 (high)
handdum
Handicapped=1; otherwise=0
handper
Degree of handicap in percent (0 to 100)
hhninc
Household nominal monthly net income in German marks/10,000
hhkids
Children under age 16 in the household=1; otherwise=0
educ
Years of schooling
married
Married=1; otherwise=0
haupts
Highest schooling degree is Hauptschul degree=1; otherwise=0
reals
Highest schooling degree is Realschul degree=1; otherwise=0
fachhs
Highest schooling degree is Polytechnical degree=1; otherwise=0
abitur
Highest schooling degree is Abitur=1; otherwise=0
univ
Highest schooling degree is university degree=1; otherwise=0
working
Employed=1; otherwise=0
bluec
Blue-collar employee=1; otherwise=0
whitec
White-collar employee=1; otherwise=0
self
Self-employed=1; otherwise=0
beamt
Civil servant=1; otherwise=0
docvis
Number of doctor visits in last three months,
hospvis
Number of hospital visits in last calendar year,
public
Insured in public health insurance=1; otherwise=0
addon
Insured by add-on insurance=1; otherwise=0
On Greene (2012) online resources, and on the Journal of Applied Econometrics website (http://qed.econ.queensu.ca/jae/2003-v18.4/riphahn-wambach-million/).
Riphahn, R.T. and Wambach, A. and Million, A. (2003), Incentive Effects in the Demand for Health Care: A Bivariate Panel Count Data Estimation, Journal of Applied Econometrics, Vol. 18, No. 4, 387–405.
Green, W.H.. (2012). Econometric Analysis, 7th edition, Prentice Hall.
###### Example 13.7 of Greene (2012) #################################### ## Selecting the same data point and scaling income ########## data(HealthRWM) dat88 <- subset(HealthRWM, year==1988 & hhninc>0) dat88$hhninc <- dat88$hhninc/10000 ### A guess start thet0 <- c(b0=log(mean(dat88$hhninc)),b1=0,b2=0,b3=0) ## Table 13.2 First column g <- hhninc~exp(b0+b1*age+b2*educ+b3*female) res0 <- nls(g, dat88, start=thet0, control=list(maxiter=100)) summary(res0)$coef ## Table 13.2 Second column ## Trying very hard to reproduce the results, ## Who is right? h1 <- ~age+educ+female model1 <- momentModel(g, h1, thet0, vcov="MDS", data=dat88) res1 <- gmmFit(model1, control=list(reltol=1e-10, abstol=1e-10)) summary(res1)@coef ## Table 13.2 third column (close enough) ## Here a sandwich vcov is required because it is not ## efficient GMM h2 <- ~age+educ+female+hsat+married model2 <- momentModel(g, h2, thet0, vcov="MDS", data=dat88) res2 <- gmmFit(model2, type="onestep") summary(res2, sandwich=TRUE)@coef ## Table 13.2 fourth column (Can't get closer than that) res3 <- gmmFit(model2) summary(res3)@coef # Lets see what happens if we start on Greene solution update(res3, theta0=c(b0=-1.61192, b1=.00092, b2=.04647, b3=-.01517)) ## No...