Papers Recently Released by the NBER Retirement and Disability Research Center
NB23-21: Barriers Faced by People Experiencing Homelessness in Los Angeles when Filing Social Security Disability Appeals: A Qualitative and Community-Engaged Study, Philip Armour, Alex Sizemore, and Jordy D. Coutin
Abstract: People experiencing homelessness (PEH) face many barriers to accessing public benefits. Over 75,500 people are unhoused in Los Angeles County and many live with severe disabilities. In this article, we examine barriers faced by PEH in Los Angeles County when filing a Social Security Administration (SSA) disability program appeal after receiving an initial denial, familiarity with and perceptions of the appeals process among PEH, and support resources available to PEH filing appeals. This qualitative and community-engaged study includes interviews with service providers (n=8) and appellants who are currently or formerly unhoused or at imminent risk of losing their housing (n=13). Snowball sampling was used to reach service-connected appellants, and field interviews were conducted in Skid Row to reach non-service-connected interviewees. The study was conducted in partnership with a Community Advisory Board (CAB) of community experts who deliver supportive services to PEH filing SSA disability program appeals in Los Angeles. Findings suggest that PEH face heightened barriers remaining in contact with SSA and service providers, receive lower quality healthcare (both exacerbating their chronic health conditions and leading to worse documentation of these conditions for disability determination), and experience significant confusion and frustration when navigating the appeals process. Delays in receiving benefits have a negative effect on the health and well-being of PEH and, in some cases, lead to significant declines in health.
NB23-20: Do Automatic Savings Policies Actually Increase Savings?, James J. Choi, David Laibson, Jordan Cammarota, and John Beshears
Abstract: Automatic enrollment and default contribution rate auto-escalation have become widely adopted in retirement savings plans on the belief that these nudges increase savings. We find that previous estimates of their savings effects are overstated. Our new estimates at eight companies incorporate the facts that employee turnover is high, a large percentage of 401(k) balances leave the retirement savings system upon employment separation, and employees may opt out of the auto-escalation default. The net savings rate increase generated by automatic enrollment, default auto-escalation introduced on top of pre-existing automatic enrollment, and the simultaneous introduction of automatic enrollment and default auto-escalation is only 0.5 percent, 0.3 percent, and 0.7 percent of income per year, respectively, on average. Employees with positive balances under automatic policies withdraw a higher proportion of these balances upon separation, and only 37 percent of those with an auto-escalation default accept it at their first auto-escalation date.
NB23-18: Public Housing Assistance and Social Security Disability Program Participation: Do Vouchers Crowd Out Disability Applications and Receipt?, Philip Armour, Jason Ward, and Javier Rojas Aguilera
Abstract: The Department of Housing and Urban Development (HUD) and the Social Security Administration (SSA) both administer programs providing economic support to non-elderly individuals with disabilities, but we know little about interactions between these independently funded and administered programs and the resulting implications for the economic security of individuals with disabilities. We address this knowledge gap by leveraging ZIP code-level variation in HUD funding to local housing authorities arising from policy changes to rental ceilings for the Housing Choice Voucher program. We use this variation to estimate the causal impact of voucher receipt on Supplemental Security Income and Social Security Disability Insurance program participation at the ZIP code level. Across numerous specifications, we consistently find a strongly statistically significant negative relationship, where additional voucher utilization leads to fewer SSI and SSDI applications and lower SSI rolls. Our estimates range from a reduction of one SSI applicant for between 10 and 40 additional vouchers utilized, smaller effect size for SSDI applicants, and a larger effect size for SSI recipients. These results point to significant substitutability between income support and housing support programs, potentially mediated by local housing authorities themselves, which has important implications for interactions between federal programs supporting individuals with disabilities as well as the net federal costs of funding additional housing vouchers.
NB23-14: The power of inclusive labor force participation for mitigating population aging closing gaps at the intersection between race/ethnicity and gender in the United States, René Böheim, Thomas Horvath, Thomas Leoni, and Martin Spielauer
Abstract: We develop a dynamic microsimulation model to project the labor force and economic dependency ratios in the United States from 2022 to 2060 taking population projections and the large inequalities between population groups of different race/ethnicity and gender into account. We contrast policy scenarios and show the potential impact that closing the gaps in ed- ucation, health and participation rates between population subgroups can have on increasing the US labor force. Our baseline projections indicate an increase of the labor force of about 27 million people by 2060 which is mainly caused by population growth. The downstream effects of removing disparities in population health and educational attainment on labor force participation can add about 10 percent (+2.6 million people) to our baseline projections. The potential effects of closing gaps between genders and between minority groups and the non-Hispanic White population however, are much larger if we assume the equalization of participation rates for individuals with similar characteristics. Closing gender gaps within ethnoracial groups, for instance, can add 9.9 to 14.3 million people to the labor force depending on the assumptions. Overall, reducing disparities in labor force participation rates has the potential to more than compensate for the effects of demographic aging on the economic dependency ratio.
NB23-13: Social Security Disability Reform and Implications for Employment, Hilary W. Hoynes, Nicole Maestas, and Alexander Strand
Abstract: Beginning in 2010, the Social Security Administration began implementing a series of data driven policy initiatives to improve the quality and consistency of disability case reviews performed by Administrative Law Judges (ALJs). The policy initiatives included revised training curricula for ALJs, new decision-support tools, and direct feedback about common decision errors. Around this time, the hearing-level allowance rate dropped by 22 percentage points and employment among people with disabilities rose for the first time in decades. However, it is not known what, if any, role the policy initiatives played in these changes. This paper investigates the impacts of the policy initiatives on the hearing-level allowance rate and the effects of these changes in allowance rates on the employment and earnings of applicants. We conclude that the policy initiatives accounted for 28-36percent of the decline in the hearing-level allowance rate and, as the margin of allowance gradually tightened, led to increased work activity among many denied applicants.
NB23-08: Examining the Impact of Inflation on the Economic Security of Disability Program Beneficiaries, Zachary Morris and Stephanie L. Rennane
Abstract: Disability program beneficiaries, including those who receive Social Security Disability Insurance (DI) benefits or Supplemental Security Income (SSI), often incur substantial out-of-pocket (OOP) expenses for disability-related goods and services. The market basket of goods and services purchased by disability program beneficiaries is thus likely to differ from the goods and services purchased by the average urban consumer, thereby creating different actual rates of inflation. Our study explores and documents these kinds of disability-related spending patterns and the potential impact of inflation on these items, providing new evidence to better understand the economic well-being of disability program beneficiaries. We present data collected from the Survey of Used and Needed Disability-related Goods and Services (SUNDiGS), a novel survey drawn from the Understanding America Study, a nationally representative panel with a large sample of disability program beneficiaries.
NB23-07: Understanding the Disparate Impacts of the Social Security Disability Insurance Family Maximum Rules, Timothy J. Moore
Abstract: Many dependents of DI beneficiaries – including spouses caring for children, and minor and disabled children – are eligible for additional DI payments. The aim of this project is to understand how the DI family maximum rules affect the economic security of households with DI beneficiaries, and whether the application of these rules increases or decreases socioeconomic disparities. To do this analysis, I use the Consumer Expenditure Survey and the National Beneficiary Survey as two complementary data sources. I present information on differences by “potential dependents” – children aged under 18 and spouses where such children are present. I find that there are large gaps in income and expenditure between DI and working households that generally increase with the number of potential dependents. More surprisingly, larger DI households generally have lower incomes than smaller DI households. Having the lowest absolute income is particularly concerning when one considers that no adjustments are made here for the number of people in a household. Given that DI is generally providing similarly sized dependent payments for one dependent as for multiple dependents, and lower dependent benefits for DI beneficiaries with poor earnings histories than for DI beneficiaries with better earnings histories, these results suggest that DI is doing better at insuring smaller households with dependents than larger households. The Social Security Administration (SSA) could consider doing more in general for DI beneficiaries with dependents, especially those with many dependents.
NB23-04: Racial and Ethnic Disparities in SSDI Entry and Health, Colleen Carey, Nolan H. Miller, and David Molitor
Abstract: Racial disparities in the Social Security Disability Insurance (DI) program have long been a concern, yet little is known about how the health and entry patterns of DI recipients vary by race and ethnicity. In this paper, we examine trends in the racial/ethnic composition of DI recipients and show how the health of DI entrants and the responsiveness of DI entry to economic conditions and program rules differ across race and ethnicity. Our analysis uses the racial/ethnic categorization in Medicare administrative data, which we first validate against U.S. Census self-reports. We then document the race and ethnicity of all DI recipients since 1992. In examining entry patterns, we find that per capita DI entry is highest among Blacks and lowest among Asians, while illness burden, as measured by medical expenditure and mortality, is lowest among Asians and Hispanics and highest among Blacks and Natives. Additionally, we analyze the effects of poor economic conditions on DI entry for different racial and ethnic subgroups. Finally, we show racial/ethnic variation in the effect of an age-based change in the program’s eligibility rules, finding that the impact of relaxing the eligibility rules at ages 50 and 55 is largest among Natives and smallest among Asians.
NB23-03: COVID-19 Pandemic Effects on the Composition of the Population with Disabilities, Otavio A. C. Bartalotti, Steven Dieterle, and Brent Kreider
Abstract: We consider the evolution of employment rates among the population identified as persons with disabilities (PWD) in the United States during the onset of the COVID-19 pandemic. Using data from the Survey of Income and Program Participation (SIPP), we analyze flows into and out of the population with reported disabilities and document compositional changes during the pandemic’s first year. We find an important composition shift among the PWD population, with a higher share reporting that they suffer from disabilities in surveys conducted after March 2020. Importantly, those reporting PWD status during 2020 were more likely to have stronger employment histories than observed for years before COVID. Furthermore, the cohort reporting disability in 2020 differs from previous cohorts in terms of the types of impairments affecting them, with a sizable increase in the share reporting cognitive/mental health difficulties. The increase in disability reported in 2020 came disproportionately from those employed and working from home at least some days a week, with smaller increases among those working away from home, or not working at all. We also find that while the share of workers in telework-amenable occupations rose among non-PWD workers during the early stages of the pandemic, it declined among PWD workers – signaling that potential structural changes in the labor market favoring these types of occupations and remote work might not have been advantageous for promoting employment among PWD. In a preliminary analysis of the SSA’s Disability Analysis File Public Use File (DAF PUF), we document increases in mortality for SSDI recipients in 2020 and 2021 relative to the previous eight years. This increase in mortality explains a nontrivial portion of the reduction in active claims in 2020 and 2021.
NB22-17: Variation in a Mid-Atlantic State’s Opioid Response Services by Race, Age, Gender, and Place-Based Economic Indicators, Tibebe Abebe Assefa, LaTanya Brown-Robertson, Nega Lakew, Ellen Meara, George C. Onoh, and Azene Zenebe
Abstract: As opioid overdose mortality and morbidity continue to rise in recent years, many states have pursued State Opioid Response (SOR) programs to facilitate access to opioid use disorder treatment. This study characterizes access to care and variation in federally funded SOR programs operating in a state in the Mid-Atlantic region since 2014. This state has experienced high and rising opioid-involved overdose deaths, especially among its Black residents. Using data from 53 jurisdiction-level service providers combined with information on demographic, economic, and social characteristics within a 10-mile radius of the SOR service providers, our research explores the equity of opioid treatment, referral services, and discharge in this Mid-Atlantic State’s SOR programs. From October 2020 through May 2022, our study area’s jurisdiction-level SOR providers served 8,659 adult clients. Among the rate of service received (per 100,000 population) adult men received more service, at 260, compared with adult women, at 110. The rate of service receipt of Black adults (240) was higher than White adults (186). Black SOR service users were prevalent at ages when Social Security Disability Insurance benefits are most common; 49 percent of Black adults served were aged 45 to 64, in contrast with 21 percent of White adults served. Furthermore, the results from the analysis of the services to clients at the SOR communities’ level (within 10 miles of a SOR service provider) show no significant difference in the association between the number of referrals, clients, or MOUD (medication for opioid use disorder) starts, and the poverty level and education level. In addition, the clients from communities with more Black residents were more likely to take up services. The findings also reveal many clients referred, enrolled, or starting MOUD in the SOR program are from the lowest Economic Vulnerability Index (EVI) communities with fewer clients coming from the high EVI communities. In conclusion, access to treatment and recovery services among our study area is relatively equitable across the White and Black race groups; SOR service providers were an important point of access for MOUD with OUD (opioid use disorder) treatment needs living in majority Black communities. There are still great opportunities for outreach in those more economically vulnerable communities. Since rates of opioid-involved overdose deaths continue to grow fastest among Black residents in our study area, future research should examine whether opioid treatment correlates with a decline in opioid-involved deaths and if there is any difference in the quality of SOR provider services delivered by race as well as by type of program.
NB22-16: Financial Inclusion, Inequality, and Retirement Trends among Older Workers, Isaac Marcelin and Wei Sun
Abstract: The research created a financial inclusion index with three dimensions: usage, barriers, and access to financial resources. A Principal Component Analysis was used to determine the weights for each dimension. This index aids in evaluating how financial inclusion affects various factors, such as ethnic groups, minorities, human capital, retirement, wealth outcomes, and mental well-being. Our findings have uncovered new psychological and sociological effects of having access to financial products. Households with higher financial inclusion scores are likelier to have higher income, own homes, and possess real estate wealth. They also have a greater chance of creating intergenerational wealth and escaping poverty. Financial inclusion leads to long-term improvements in wealth and retirement outcomes, benefiting minority groups and genders. It also improves family and work resilience, reduces stress, and lowers drug-related problems. Results have several policy implications, including bridging the wealth gap, enhancing retirement security, and bettering socioeconomic outcomes.
NB20-15: Medicaid vs Medicare: Evidence from Medicaid to Medicare Transitions at 65, Timothy Layton, Nicole Maestas, Daniel Prinz, Mark Shepard, and Boris Vabson
Abstract: The US has two predominant government health insurance programs, Medicaid and Medicare, which collectively cover over 100 million Americans. Given differences between Medicaid and Medicare in program design and costliness, there has been ongoing policy debate on how much of the population should be covered through one program versus the other, as well as whether the design of one program should more closely mimic the other. Unfortunately, little is known about how these programs actually compare on important outcomes, such as government spending and beneficiary well-being. We investigate these questions by leveraging mandatory age-based transitions into Medicare at 65, among those previously in Medicaid. We find that the government spends 13 percent more to cover the same beneficiary under Medicare compared to Medicaid, with most of this difference coming from higher payment rates to providers rather than through increased healthcare utilization. We find significantly higher rates of outpatient care usage under Medicare, alongside lower levels of acute care usage. These results may reflect improved primary care access under Medicare, which could arise through the program’s more generous physician reimbursement rates.
NB20-07: Retirement Benefits and Incentives for Work: New Evidence from Eligibility Criteria, Elira Kuka and Glenn Springstead
Abstract: Social Security is the largest federal program in the US, yet it is not a universal program. To qualify for benefits, individuals must have accumulated at least 40 quarters of coverage (QCs) during their work history. This requirement creates an upward notch in individuals’ lifetime (earnings and retirement) income, providing a very large incentive for individuals to work at least 40 QCs. Using standard bunching methods, I show that individuals do indeed bunch at this threshold by working a few extra quarters, and that bunching increases as individuals get closer to the retirement age. The size of the response is however very small when compared to the size of the benefit, suggesting very small elasticities of work with respect to retirement benefits for this population.
NB19-21: The (Lack of) Anticipatory Effects of the Social Safety Net on Human Capital Investment, Manasi Deshpande and Rebecca Dizon-Ross
Abstract: How does the expectation that a child will receive government benefits in adulthood affect parental investments in the child's human capital? Most parents whose children receive Supplemental Security Income (SSI) benefits overestimate the likelihood that their child will receive SSI benefits in adulthood. We present randomly selected families with the predicted likelihood that their child will receive SSI benefits in adulthood. Reducing parents' expectations that children will receive benefits in adulthood does not increase investments in children's human capital. This zero effect is precisely estimated. Likely explanations include parents working more themselves, non-financial goals influencing investment, and families facing investment constraints.
