Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: calculating standard deviation of age at death from period life tables


From   Austin Nichols <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: calculating standard deviation of age at death from period life tables
Date   Tue, 9 Mar 2010 10:57:31 -0500

Johannes Schoder <johannes.schoder@soi.uzh.ch>:
I would bet that investments in health or technology changes do
immediately translate into longer life expectancy, but that life
expectancy is not measured for the affected cohorts but for a mix of
cohorts, only some of which enjoy the full effect of the
investment/change (life expectancy for a given cohort cannot actually
be measured until everyone in the cohort is dead).  But assuming you
want to assess the short-run effect of changes in X and the long-run
effect of changes in X, you might consider including the first
difference divided by two and the two-period moving average, i.e.
y_it = b ( X_it - X_it-1 )/2 + c  ( X_it + X_it-1 )/2 +e
where if b=c the model is identical to one with only contemporaneous
regressors.  The MA term captures long-run effects and the FD term
captures short-run effects.  For finer structure, you can apply some
kind of filtering; see e.g.
http://www.jstor.org/stable/2660649
on MA and FD and an extension to multiple terms and see
http://www.stata-journal.com/sjpdf.html?articlenum=st0116
on those multiple terms, and
http://www.stata.com/meeting/5nasug/TSFiltering_beamer.pdf
on further methods.

On Tue, Mar 9, 2010 at 9:46 AM, Johannes Schoder
<johannes.schoder@soi.uzh.ch> wrote:
> Dear Statalist users:
>
> I am trying to find the optimal lag structure of my health production
> function (lags are usually included because e.g. lifestyle effects do not
> immediately translate into longer life expectancy) which looks like that:
> LIFE EXPECTANCY_ct=b_o+b_1 HEALTH CARE EXPENDITURE_ct + b_2 GDP_ct + b_3
> ALCOHOL CONSUMPTION_ct
>
> where c=COUNTRY and t=YEAR
> Is there a test implemented in Stata which tells me e.g. whether to include
> GDP lagged 5 years, 10 years or 0 years?
>
> Thanks a lot for any hint!
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index