Recently, Russ Roberts interviewed David Autor of MIT on the evolution of the Social Security Disability Insurance program. The conversation was extremely illuminating. We’ve discussed Autor’s work with Mark Duggan on reforming SSDI through the use of a layer of universal private disability insurance coverage in the past, but I recommend listening to the interview to get a sense of the historical context surrounding the rise of the disability rolls. For a more systematic treatment, you might turn to Autor’s 2011 paper on the same subject.
While the SSDI program is a central component of the U.S. social safety net, it suffers from two substantial ailments that limit its effectiveness and threaten its long-term viability. First, the program is ineffective in assisting the vast majority of workers with less severe disabilities to reach their employment potential or to earn their own way. In fact, the program provides strong incentives to applicants and beneﬁciaries to remain out of the labor force permanently, and it provides no incentive to employers to implement cost-effective accommodations that would enable disabled employees to remain on the job. Consequently, large numbers of work-capable individuals voluntarily exit the labor force, apply for, and ultimately receive SSDI annually. In 2010, fully 2 percent of the SSDI-insured population—2.9 million workers—applied for SSDI beneﬁts. Somewhere between 50 and 60 percent of applicants will eventually receive an award. But even those who are ultimately denied beneﬁts will spend substantial time—typically one to three years—out of the labor force before they have exhausted all appeals. At that point, their reemployment prospects may be substantially worse than they were at the time of initial application.
Second, the program’s expenditures are extremely high and growing rapidly (Figure 1). In 2010, SSDI cash transfer payments totaled $124 billion, while the cost of Medicare for SSDI beneﬁciaries was $59 billion. These outlays, exceeding $1,500 for every U.S. household, comprised 7.3 percent of federal non-defense spending last year—a sum that is larger than interest payments on the federal debt. In the last two decades, outlays grew at 5.6 percent in real terms, compared to just 2.2 percent for all other Social Security spending. As a consequence SSDI’s share of total Social Security outlays has risen from one in ten dollars in 1988 to almost one in ﬁve dollars at present (Figure 2). Perhaps most ominously, SSDI expenditures now exceed by 30 percent the payroll tax revenue dedicated to funding the program. The Trustees of the Social Security Administration project that the SSDI Trust Fund will be exhausted between 2015 and 2018, at least two decades ahead of the trust fund for Social Security retirement beneﬁts.
Most important, Congress directed the Social Security Administration to give additional weight to pain and related subjective factors in making its disability determination decisions, and to relax its strict screening of mental illness by placing less weight on diagnostic and medical factors and relatively more weight on the ability to function in a work setting. A key consequence was that applicants with difficult-to-verify disorders such as muscle pain and mental disorders could more easily qualify for beneﬁts.
At first, this shift didn’t seem to have much of an impact, in large part because the labor market improved dramatically. Yet during the 1990-91 recession, the disability rolls once again began to rise. Autors argues that the deterioration in the labor market prospects of less-skilled workers has been a key driver of this phenomenon. Living on SSDI benefits is hardly an attractive option, as the benefits are far from lavish; yet the benefits are very stable, which is why workers at the low end of the labor market might in some cases find it preferable to remain on the disability rolls rather than reenter the labor force, particularly as they age — rather than worry about holding on to your job, you have a guaranteed income stream. And if working part-time would jeopardize this guaranteed income stream, most will choose not to work part-time.
As highlighted by Autor and Duggan (2003), the effective replacement rate of labor earnings with SSDI benefits has also risen in recent decades due both to the rising value of in-kind Medicare beneﬁts and to a subtle interaction between the SSDI beneﬁts formula and rising income inequality in the U.S. This interaction causes SSDI’s effective generosity for low-wage workers to rise as wages in the lower deciles of the distribution fall relative to the mean.
One of the inevitable dangers of talking about SSDI reform is that it is easy to seem as though you’re accusing all SSDI beneficiaries of being dishonest layabouts. The truth is very different. A large number of beneficiaries really are profoundly disabled, and many more are emotionally distressed and socially isolated. Indeed, a lack of attachment to the working world can in some cases exacerbate these problems.
SSDI reform needs to be part of a comprehensive approach to the labor market that recognizes the challenges facing less-skilled workers. Part of the problem is that it is very hard to meaningfully alter the trajectory of beneficiaries who are now in their 40s and 50s. The best solution for these workers is to recreate the tight labor markets of the mid- to late 1990s. Over the longer term, the best solution is to improve U.S. human capital policies by reforming education and making safety net programs more sustainable. To that end, the SSDI reforms proposed by Autor and Duggan seem extremely sensible. But the larger SSDI story is a reminder that the inadequate (yet extremely expensive) human capital policies we’ve pursued in recent decades have left millions of prime-age workers with few attractive alternatives to going on the disability rolls.
Another part of the story, which Autor touches on, is the role of disabling mental illness. In the past, we’ve discussed the hypothesis advanced by Robert Whitaker, i.e., that our approach to the treatment of mental illness has actually contributed to its prevalence and severity.