maximum benefit entitlement and retirement as competing risks

[marcel spijkerman]

Dear Stata users,

I am analyzing durations of unemployment benefit entitlements (not unemployment, people can remain unemployed after entitlement ends). The maximum duration of benefit entitlement depends on employment history. The spell of benefit entitlement can end for several reasons; finding a job, reaching the retirement age, reaching maximum duration of benefit entitlement. Someone with a maximum duration of benefit entitlement of, say, 30 months will never be at risk for this event before the 30^th month end their spell will end automatically if the spell not ended for other reasons. The same holds for retirement; many people in the data will not be at risk experiencing this event because they are to young. On the other hand, the probability of ending the spell because of retirement or reaching the maximum period of entitlement will be higher for those that are less likely to find a job.
My question is whether reaching the retirement age and maximum duration of benefit entitlement can be considered as competing risks or are there other methods to deal with this phenomena?

Kind regards, Marcel

[Stephen Jenkins]

For research addressing similar methodological issues, have a look at the following paper, especially section 4 (Econometric Model), and the references cited:

Wiji Arulampalam, Robin A. Naylor and Jeremy P. Smith, "A hazard model of the probability of medical school drop-out in the UK", Journal of the Royal Statistical Society, Series A, (2004) 167, Part 1, pp. 157–178

[marcel spijkerman]

Thanks for the reference!