This paper utilizes census records, multiple comprehensive surveys, and mortality records from England to trace out the effect of reaching retirement age on retirement status, health behaviors, health care utilization, and health outcomes. Applying a regression discontinuity design leveraging the pension age, I find that retirement substantially improves well-being and satisfaction with health, along with fewer individuals reporting having health issues. However, I find no immediate effect of retirement on behavioral outcomes such as smoking or exercising, and no evidence of changes to cognitive ability or healthcare utilization. Finally, death certificate data indicate that retirement does not affect mortality. While prior literature has considered the effects of retirement on specific outcomes, this paper is the first to systematically examine the full range of health-related outcomes with administrative and survey data in a unified context.
Among the elderly population that is hospitalized in a given year, about 20 percent are discharged to skilled nursing care facilities, at a cost of over $30 billion annually. Skilled nursing facilities (SNFs) provide high-level medical care in an outpatient setting with the intent of reducing individuals' time in the hospital and preventing readmissions. Despite the large role SNFs play in the provision of health care, there is limited evidence on their effectiveness. This is primarily due to the systematic differences between individuals that go to a SNF and those that are discharged to their homes. In this paper, I leverage a Medicare policy that induces a sharp discontinuity in the probability of being transferred to a SNF to estimate the effectiveness of SNF care. I find that SNF care reduces the probability of readmission to the hospital within 30 days by 33 percent, suggesting that SNF care substantially improves patient outcomes.
With Nicole Maestas and Carlos Dobkin
With Rob Fairlie, Jeremy West, and Bryan Pratt
Despite substantial use of prosocial persuasion to reduce negative externalities, there remains scant evidence on how moral suasion interacts with conventional pecuniary policies. We empirically examine this question in the context of water conservation. Methodologically, we use a randomized control trial that provided households with peer comparisons to their neighbors’ water consumption, coupled with a regression discontinuity design based on an arbitrary cutoff for pecuniary threats regarding irrigation violations. We find the pecuniary threat was effective only during targeted hours of the week, with evidence of significant intertemporal substitution. In contrast, the prosocial treatment yielded blanket reductions in water consumption.