Wednesday, September 23, 2015

Disability in the United States

This is Joseph.

The New York times had a very nice article on social security disability insurance.  Of especial note were:
If people on disability were faking it, they wouldn’t have such high death rates. People on disability are three to six times more likely to die than people in their age group who are not on disability.
Disability claims are not skyrocketing. Rather, the population most likely to go on disability, those aged 50 to 64, is growing. The potential disability population is also larger now than in the past because today’s older women are more likely to have worked enough to qualify for disability than in earlier generations.
These points suggest that increases in disability claims are much less of a point of concern than one might think.  An elevated death rate is a pretty strong indicator of morbidity and it is impossible to fake.  Nor do I think an increased level of workforce participation (making it possible for more people to be eligible for benefits) is bad news. 

I don't want to minimize worries about fraud, but there are times when there is unlikely to be a lot of low hanging fruit. 


  1. Unless, of course, people on disability were engaging in some other dangerous activity.

    1. Unless the disability insurance, itself, is harmful, you would need an absolutely huge risk factor (unrelated to disability) that is also a risk of death. Since penetrance of the population is never 100%, you need to double or triple such a risk factor.

      It's not impossible that this could exist but it is implausible that epidemiologists have missed such a massive risk. If the risk of death was 25% higher then this would be a serious concern. But three to six times suggests that something directly involve with the participants in the program is happening (and they are being screened for having a disability).

      Disabled people dying more quickly is both plausible and a known source of mortality. Smart money suggests that you don't have healthy fraudsters who engage in other risky behaviors in large numbers (or else you wouldn't be able to sustain the increase in mortality, as the fraudsters would dilute the sample).