"Does Welfare Inhibit Success? The Long-Term Effects of Removing Low-Income Youth from Disability Insurance" (Job Market Paper) [UPDATED December 2014]
Abstract: I estimate the long-term effects of removing low-income youth with disabilities from Supplemental Security Income (SSI) on the level and variance of their earnings and income in adulthood. Using administrative data from the Social Security Administration, I implement a regression discontinuity design based on a change in the probability of SSI removal at age 18 created by the welfare reform law of 1996. I find that SSI youth who are removed earn on average $4,000 per year, an increase of just $2,600 relative to those who remain on SSI. This increase in earnings covers only one-third of the $7,700 they lose in annual SSI income, and they lose an additional 10% each year in other transfer income. As a result, removed SSI youth experience a present discounted income loss of $73,000, or 80% of the original SSI income loss, over the 16 years following removal. In addition to the large drop in income levels, the within-person variance of income quadruples as a result of the SSI loss. Based on back-of-the-envelope calculations assuming risk aversion and limited intertemporal consumption smoothing, I find that up to one-quarter of the recipient's welfare loss from SSI removal is attributable to the increase in income volatility rather than to the fall in income levels. This result suggests that ignoring the income stabilization effects of welfare and disability programs could substantially underestimate their value to recipients. (Online Appendix)
"The Effect of Disability Payments on Household Earnings and Income: Evidence from the Supplemental Security Income Children's Program"
Abstract: I estimate the effect of removing children with disabilities from the Supplemental Security Income (SSI) program on parental earnings and household income. Using administrative data from the Social Security Administration, I implement both a regression discontinuity design and a difference-in-differences design based on changes in the budget for medical reviews, which increase the likelihood of removal from SSI. I find that a loss of $1,000 in the child’s SSI payment increases parental earnings—exclusively on the intensive margin—by $700-$1,400, indicating that parents fully offset the SSI loss with increased earnings. The loss of the child’s SSI payment also discourages parents and siblings from applying for disability insurance. In addition, I find evidence that family members often apply for disability insurance at the same time, which suggests the importance of household-level shocks in the decision to apply. Using the unique institutional context of the SSI program, I provide suggestive evidence that the large response in parental earnings is driven mostly by an income effect rather than a substitution effect.
Research in Progress
"Testing the Coase Theorem in the Context of the Sulfur Dioxide Market: Does the Initial Allocation of Permits Affect Firms' Pollution Decisions?"
Abstract: I test whether the initial allocation of permits in the U.S. Acid Rain Program had an effect on the emissions decisions of covered firms. The initial allocation could affect subsequent sulfur dioxide emissions if firms face market uncertainty, transaction costs, or liquidity constraints, or if they exhibit behavioral biases. To address potential endogeneity of unobserved marginal abatement costs, I use variation in temperature across counties in 1985-1987 as an instrument for the number of permits allocated to a plant. The formula for the initial allocation was based on a plant’s heat input during this time period, and higher temperatures lead to greater electricity demand. I find that the initial allocation of permits has a large effect on firms’ pollution decisions. Increasing a plant’s permits by one ton raises the firm’s emissions on average by two tons. I propose and find evidence consistent with a lumpy abatement hypothesis in which firms that receive fewer permits are more likely to invest in high-fixed-cost technologies that reduce emissions in discrete amounts.
"Does the Timing of Removal from Disability Insurance Affect Children's Long-Term Outcomes? Evidence from the Supplemental Security Income Children's Program"
Abstract: The effect of removing children from disability insurance on their long-term outcomes may depend on whether families have time to adjust education and human capital investment decisions. I estimate the effect of removing low-income children with disabilities from the Supplemental Security Income (SSI) program during childhood, before education decisions are complete, on their long-term outcomes in adulthood, including earnings, income, health care utilization, and criminal activity. To identify these effects, I will use both a regression discontinuity design and a difference-in-differences design based on changes in the budget for medical reviews, which increase the likelihood of removal from SSI. I will use administrative data from the Social Security Administration linked to administrative data on health care utilization and criminal outcomes from states. I will compare the estimates from this paper to previous estimates of the effect of removing children at the age of 18, when education decisions for these low-income children are largely complete.