Research
Health Affairs 42(6), 2023, with Caroline Hanson, Allison Percy, Emily Vreeland, Alexandra Minicozzi, and the CBO Coverage Team
The Congressional Budget Office estimates that in 2023, 248 million people in the US who are younger than age sixty-five have health insurance coverage (mostly through employment-based plans), and twenty-three million people, or 8.3 percent of that age group, are uninsured—with significant variations in coverage by income and, to a lesser extent, by race and ethnicity. The unprecedented low uninsurance rate is largely attributable to temporary policies that kept beneficiaries enrolled in Medicaid and enhanced the subsidies available through the health insurance Marketplaces during the COVID-19 pandemic. As the continuous eligibility provisions unwind in 2023 and 2024, an estimated 9.3 million people in that age group will transition to other forms of coverage, and 6.2 million will become uninsured. If the enhanced subsidies expire after 2025, 4.9 million fewer people are estimated to enroll in Marketplace coverage, instead enrolling in unsubsidized nongroup or employment-based coverage or becoming uninsured. By 2033 the uninsurance rate is projected to be 10.1 percent, which is still below the 2019 rate of about 12 percent.
The Opioid Crisis and Foster Care Admissions
Under Review
A growing literature suggests that the opioid crisis has negatively impacted children. This paper builds on prior work by separately identifying the causal effects of parental illicit opioid use and prescription opioid (PO) misuse on foster care entry and caseload after 2010. Two baseline opioid supply measures from 2000, at the onset of the opioid crisis, instrument for post-2010 changes in opioid overdose death rates among adults of prime parenting age. Using county-level mortality data and administrative foster care records, I find that moving from the 10th to the 90th percentile of change in the illicit opioid death rate leads to a 1.1 (2.4) standard deviations larger percent growth in foster care entry (caseload). In contrast, a 10th to 90th percentile difference in the change in PO misuse does not significantly affect total entry or caseload. Female illicit opioid use has larger effects than male illicit opioid use. Effects are similar across child age at the time of removal. Illicit opioid use has similar effects across all placement settings, whereas PO misuse only increases entry to kinship care.
Presented at: U.S. Census Bureau, Center for Economic Studies Interdisciplinary Seminar Fall 2021; Association for Public Policy Analysis & Management Fall 2021 and Fall 2022 Research Conferences
Reconciling Survey and Administrative Measures of Self-Employment
Journal of Labor Economics 39(4), 2021, with Katharine Abraham, John Haltiwanger, Kristin Sandusky, and James Spletzer
Good information on self-employment is needed to inform the ongoing discussion of the rise of the gig economy and its implications for workers. Tax data show significant growth in self-employment not captured in the Annual Social and Economic Supplement to the Current Population Survey (CPS-ASEC). The growing gap reflects both self-employment in tax data missing from the CPS-ASEC and self-employment misreported as wage and salary work. We document consistent patterns in the discrepancies between the tax and survey data but are able to explain only a modest share of the growing disagreement between them.
Presented at: AEA 2020 Annual Meeting (coauthor presented); Society of Labor Economists 2019 Annual Meeting (coauthor presented)
The Effect of Expanding Medicaid Eligibility on Self-Employment Dynamics
Working Paper (Draft available upon request)
How does an expansion of public health insurance affect self-employment dynamics? I study this question in the context of the Affordable Care Act (ACA) Medicaid expansion, which made Medicaid available to all non-elderly adults with family income up to 138% of the federal poverty level. By increasing access to health insurance outside of wage employment, expanding Medicaid eligibility could (i) induce some wage workers to become self-employed, and (ii) increase persistence in self-employment for the currently self-employed. Using the 2003-2016 CPS-ASEC linked to respondents' earnings histories from tax returns, I estimate the effect of expanding Medicaid eligibility on the probability of self-employment, wage employment to self-employment transition, and self-employment persistence among childless adults, the group that saw the largest increase in Medicaid eligibility as a result of the expansion. Using difference-in-differences that takes into account pre-ACA cross-county variation in population shares of “at-risk” groups and triple differences, I find suggestive evidence that the expansion of Medicaid eligibility increased persistence in self-employment, but no effects on wage employment to self-employment transitions or the probability of self-employment.
Self-Employment in the Transition to Retirement
Working Paper (early version), with Katharine Abraham, John Haltiwanger, Kristin Sandusky, and James Spletzer
Self-employment rates rise with age, especially past the age of 50. Using unique integrated survey and administrative data, we find that self-employment as a share of employment more than doubles from 47-52 to 65-70 – rising from under 10% to more than 20%. This growth reflects the differential patterns by age of all of the transitions among wage and salary employment, self-employment and non-employment. There is a sharp decline in the likelihood that workers switch from self-employment to wage and salary employment with age, but not the reverse. The share of wage and salary workers who transition to non-employment each year rises more rapidly with age between 53-58 and 65-70 than is the case for the self-employed. Just as important, there is a much sharper decline with age in the pace of transitions from non-employment to wage and salary employment than in the pace of transitions from non-employment to self-employment. The interaction of these changing transition rates, as opposed to simply their individual effects, plays a large role in accounting for the increase in the self-employment rate with age. We investigate how education, cumulative earnings over the prior 20 years, and earnings volatility over the prior 20 years affect these changing transition dynamics by age. We find, for example, that wage and salary workers who are more educated and have higher cumulative earnings are more likely to move to self-employment and less likely to move to non-employment, with both of these effects larger at older ages.
Presented at: Russell Sage Foundation, 2019 (coauthor presented); Stanford Institute for Economic Policy Research Working Longer and Retirement Conference, 2018 (coauthor presented)