My dissertation examines the effects that US Social Security policy changes have on labor supply and application behavior. In my first chapter, I use survey data matched to administrative records to measure the effect of a natural experiment in the provision of information - the introduction of the Social Security Statement - on Social Security Disability Insurance (DI) application behavior. I find that receipt of the Statement, a document gradually introduced in the 1990s which contained personalized information on all Social Security benefits, had a positive, substantial, and statistically significant effect on DI application, amounting to a 62% increase over the base rate. The overall effect was entirely driven by individuals reporting a work-limiting condition who were previously not employed. Furthermore, my analysis shows no evidence of applicants "shifting forward" their DI application. In the absence of these new applicants, the 32% growth rate of the per-capita DI rolls from 1995-2004 would have been approximately 25%, a 20% drop. My second chapter examines the impact of the Statement on individuals approaching retirement who are still working. I find that the effects are mixed: individuals working few hours per week prior to Statement receipt markedly increase their hours worked, while those working full time or more decrease their hours. However, it appears that individuals misunderstood that the Statement provided a retirement benefit projection based on constant earnings going forward. After second Statement receipt, those who previously decreased their labor supply then increased it. My third chapter uses program interactions between the two federal disability programs - DI and Supplemental Security Income - to estimate labor supply elasticities of DI beneficiaries. Given wide variation in SSI state supplements temporally and geographically and temporal variation in DI parameters, otherwise similar DI beneficiaries may face a 50% or 0% disability benefit reduction rate. This variation in marginal rate allows for a direct calculation of labor supply responsiveness to such a rate change. My elasticity estimates - the first for US disability beneficiaries based on differences in benefit reduction rates - are inelastic and low, ranging from -0.14 to -0.36 for the participation elasticity.