Retired people in 2012 1-year ACS PUMS

Hi!

I do not think you can get this information from the ACS. The Medical Expenditure Panel Survey Household Component data (released by AHRQ) has a variable on ever retired. See meps.ahrq.gov/.../download_data_files_detail.jsp for recent year file. The downside of the MEPS is that you can only get regional and national data.

Tom Bell
tbell@mathematica-mpr.com
312-994-1010

Unfortunately there is no retired variable. One way you could try to measure retired is using ESR (employment status recode) =6 (not in labor force) and the an age combination agep

I will just add that the RETP variable contains the amount of retirement income, if any. So, one way to defined "retired people" is AGE>=65, ESR=6, and (optionally) RETP>0. However, this all depends on your definition of "retired". For example if the person is not in the labor force and is receiving retirement income, how relevant is the age cutoff?

Doug Hillmer

Good point!

I looked for something similar these days and I agree that the combination of "not in labor force" variable with retirement income would be the best... Though, I am not sure if this combination is already available in FactFinder...

Definition from ACS:
"Not in labor force category consists mainly of students, housewives, retired workers, seasonal workers interviewed in an off season who were not looking for work, institutionalized people, and people doing only accidental unpaid family work (less than 15 hours during the reference week)".

Thank you all for your answers! I’ve tried to compare the retired number resulting from the combination of variables ESR, RETP and AGE with the one published by Social Security Administration (www.socialsecurity.gov/.../OAbenies.html) - 36,719,288 retired workers in 2012. An overall incidence (among all indiv. 0+y.o.) would be around 12%.

-> If I filter by ESR=6 and AGEP>=62 and RETP>0 I obtain about 16,000,000 retired people, so about 5.3% within all indiv. 0+ y.o. I think this difference is too big.
-> It seems that only 8% of indiv (24,068,309 ) are receiving retirement incomes greater than 0 (RETP>0). So I have to exclude this variable from my filter in order to get closer to 12%= the overall incidence of retired.
-> If I take into consideration the variables: ESR=6 and AGEP>=62 I obtain 41,166,421 retired people, about 13.5% of all indiv. Here are also included, as Galatia noted, those housewives, seasonal workers and other groups besides retred. How can I exclude them?
Another question is: should I use different ages when targeting retired? For axemple women and men have the same standard retirement age? What about those who retire earlier than the standard age?

Thanks,
Adriana