ABLE Accounts and the Household Economic Response to Asset-Test Relief

Asset tests in means-tested transfer programs impose implicit marginal taxes on saving that can exceed one million percent. The ABLE Act is the first federal statute since the 1996 welfare reforms to relax an asset test in a means-tested program, and the paper provides causal evidence that the policy raises employment, wages, and disposable income for adults with disabilities. The response concentrates among adults with baseline household net worth below the $2,000 cap, consistent with the theoretical mechanism. Take-up in the nine states with publicly reported account counts is currently far below the level at which the population-level gain would be realized.

The Constraint

In 1988, Congress set the Supplemental Security Income resource limit at $2,000 for an individual and $3,000 for a couple. Both numbers have remained frozen for thirty-seven years. Adjusted for inflation, the individual limit is now worth just over $1,000 in the currency of 1988. An adult with a disability whose countable resources rise above the $2,000 threshold forfeits cash income of approximately $11,000 per year and, in states whose Medicaid categorical eligibility rules operate through the SSI pathway, forfeits Medicaid coverage worth another $10,000 per year in fee-for-service equivalents. Expressed as an implicit marginal tax on the second-to-last dollar of savings, the cliff exceeds one million percent.

Standard economic theory has predicted for three decades that a marginal tax of this magnitude should compress precautionary saving among the households that face the constraint. The prediction is borne out in the data. Adults with disabilities in the 2020 FDIC survey are twice as likely as adults without disabilities to lack a bank account, and their conditional mean bank-account balance is approximately half of the corresponding population figure. Forty-three percent of adults with disabilities in the 2019 Federal Reserve Household Economics and Decisionmaking survey report that they could not cover a $400 emergency expense from savings, compared with 26 percent of adults without disabilities.

Research Design

The Stephen Beck, Jr. Achieving a Better Life Experience Act of 2014 relaxes an asset test in a means-tested program on the asset margin alone. Congress modeled the Act on the Section 529 college-savings program from 1996. The Act permits tax-advantaged saving in a state-administered account whose balances up to $100,000 are excluded from the SSI resource test, whose balances up to state-defined ceilings are excluded from Medicaid asset tests, and whose distributions are tax-free when applied to qualified disability expenses. Eligibility was originally restricted to individuals whose qualifying disability began before age 26. The Age Adjustment Act of 2022 raises the threshold to 46 effective January 2026.

Congress delegated program administration to states, creating staggered launch timing. Ohio, Tennessee, and Nebraska launched the first programs in June 2016. Nineteen states followed in 2017, eleven more in 2018, and additional states launched programs through 2023. Four states are treated as never-adopters in the main design. This staggered rollout provides the identifying variation for a difference-in-differences design across three state-year panels covering the 2000 to 2024 period, an individual-level American Community Survey panel, an individual-level SIPP household-wealth panel, and Social Security Administration OASDI state-year counts.

Preliminary headline estimates

Main Results

Main state-year TWFE estimates show that state ABLE program availability raises log disposable income for adults with disabilities by 1.49 percent (mean disposable income by $321 per person per year, aggregate by $1.6 billion). Transfer share of income falls 0.53 pp. Individual-level ACS estimates show employment rises 0.73 pp, labor force participation 0.82 pp, and log wage income 8.11 percent. Corresponding triple-differenced estimates, which sweep out any state-year shock common to adults with and without disabilities, are 1.73 pp for employment, 2.34 pp for participation, and 12.54 percent for log wage income (Table 2 columns 8-10).

Under linear scaling from the reduced-form estimate, every one percentage point of take-up increase corresponds to a $0.55 billion increase in the aggregate annual disposable-income gain. Take-up in the nine states with publicly reported account counts currently ranges from 0.19 percent to 3.04 percent of the state adult-with-disabilities population, with a median of 0.41 percent. Raising take-up in these nine states to Nebraska Enable Savings' 3.04 percent, roughly seven times the current median, would raise the aggregate gain in these states to about $11 billion per year.

Mechanism

The disposable-income gain operates through higher labor earnings and reduced reliance on cash transfers rather than through disability-insurance substitution. The baseline TWFE on state-year OASDI counts shows a 1.7 percent decline in log SSDI in launcher states, but a parallel decline in never-adopter states over the same window, together with pre-2016 differential trends in launcher states, means the effect on SSDI rolls is not identified once state-specific linear trends are absorbed. The point estimates in the state-linear-trend specification are 0.24 percent for SSDI, 0.09 percent for OASDI, and 0.42 percent for SSI, none statistically distinguishable from zero. This distinguishes ABLE from earlier asset-test reforms that operated through changes in cash-transfer eligibility.

Effects concentrate in states that did not expand Medicaid under the Affordable Care Act. In the ABLE times Medicaid-expansion interaction specification (Table 6), the base ABLE effect on log disposable income is 2.95 percent in non-expansion states. The interaction with Medicaid expansion is minus 2.36 percent. Medicaid expansion had already provided partial relief from the effective cliff, leaving less residual response for ABLE to capture. Transfer share shows the symmetric pattern: minus 0.81 pp base ABLE effect, plus 0.49 pp interaction with expansion.

Credibility

Policy Implications