Arnab Mukherji, Satrajit Roychoudhury, Pulak Ghosh, and Sarah Brown,
"Estimating Healthcare Demand for an Aging Population: A Flexible and
Robust Bayesian Joint Model", Journal of Applied Econometrics,
Vol. 31, No. 6, 2016, pp. 1140-1158.

The data used in this paper come from the Health and Retirement Survey
(HRS) and are supported by the National Institute on Aging (NIA
U01AG009740) and the Social Security Administration. Health and
Retirement Study data products are available without cost to
registered users; certain "Conditions of Use" apply.

Once you are registered, the "Getting Started" page will help you
learn more about the data products produced by this complex study. The
HRS helpdesk advises users not to upload their data to journal
archives directly. Instead users are asked to place instructions on
creating their analytical sample for future users so that once they
are registered, they may be able to reconstructe the sample. This
readme (and the associated code file) attempts to provide such
instuctions.

After registering with the HRS website, the raw data may be downloaded
from

  http://hrsonline.isr.umich.edu/index.php?p=data

where we used the public data maintained by RAND Corporation (HRS
Public Data).

In this study, we focus on the HRS cohort, i.e. those born between
1931 and 1941. We select a random sample of 1,929 individual cases
where we require that each individual be observed in wave 1 and in
any two other subsequent waves. The hhids for the cases we use are
listed below the stata code. 

outcome variables we use are: rhsptim and roopmd
control variables: rbmi, female, gedplus, rdress, rwork, rliv, rage,
  rshlth, rshlthc, rinsured, covered, mostlycovered, partlycovered,
  rhibp, rdiab, rcancr, rlung, rheart, rstrok, rpsych, rarthr, n_cond,
  rchronic, anydisability, zerohosp, nobs, R, rshlt, rcovr, rcovs,
  rhigov, rhiothp

Stata code used to generate the final analytic dataset is available in
the file cleanhrs.do which is zipped in the file cleanhrs.zip. Since
cleanhrs.do is an ASCII file in DOS format, Unix/Linux users should
use "unzip -a".

Identifiers of the cases used in the analytical sample:

10394010
10451010
10465020
10533010
10652010
10818010
10893010
11477010
11522010
11590020
11591020
11612010
11810020
11830010
11835020
11930010
12010010
12068020
12162010
12196020
12481010
12512010
12792010
13635010
13637010
13894010
13902010
14036010
14115020
14247010
14253010
14272010
14359010
14359020
14397010
14568010
14624010
14648010
14754010
15015020
15339010
15516010
15982010
16086010
16211010
16221010
16245010
16400010
16407010
16412010
16412020
16420020
16648020
16672010
16700020
16710010
16789010
16789020
17042020
17235010
17374010
17374020
17520010
17922010
17924010
17938010
18092010
18276010
18352010
18451010
18508010
18776010
18896010
18973010
19007010
19280020
19423020
19454010
19496020
19585010
19620010
19625020
19629020
19694010
20168020
20221010
20255010
20363010
20383010
20384010
20438010
20501010
20554020
20588020
20719010
20860020
21169010
21337020
21350010
21354010
21422010
21442010
21479010
21671020
21672010
21742020
21863010
21863020
21928010
21928020
21946010
22125010
22377010
22454010
22499020
22516020
22563010
22572010
22738020
22757010
22811020
22893010
22999010
23112010
23387020
23677010
23748020
23760010
23919010
24109010
24132010
24187010
24187020
24366010
24411010
24539010
24956020
25281010
25411010
30254010
30972010
31326010
31354010
31460010
31592020
31714010
31821020
31929010
32264010
32266020

Arnab Mukherji
arnab [AT] iimb.ernet.in