The Disability Ladder

The problem

Social Security Disability Insurance pays a benefit to a worker whose impairment prevents "substantial gainful activity," and the law defines that phrase as a number: in 2026, $1,690 a month. A beneficiary may test working for nine months at any wage, and for a period after that may earn up to the number without consequence. Then the rule is binary. Earn $1,689 and the cheque continues; earn $1,691 and it stops, entirely, along with, until № 004 takes effect, the Medicare coverage that many beneficiaries fear losing more than the cash. A person with a serious impairment who could work half-time at a good wage faces a tax on the marginal dollar, at the line, of more than 100 percent.

The results are what the design predicts. About 7 million disabled workers receive the benefit. In any year, about 11 to 13 percent of them do some work, nearly all of it under the line, and fewer than 10 percent ever leave the programme by returning to work or recovering. Most leave by reaching retirement age or by dying. The programme spends about $150 billion a year and was built in 1956 for a workforce of miners and steelworkers whose impairments were total; it now serves a population whose largest diagnoses are musculoskeletal and psychiatric, for whom capacity is partial and variable, and for whom the binary is a trap.

Every other rich country has a partial tier. The Netherlands, Sweden, Germany and most of the OECD pay a reduced benefit to a worker assessed as having reduced capacity, and let the worker earn alongside it. The United States has one tier and one line.

The honest complication is that the government tested the obvious fix. The Promoting Opportunity Demonstration, a randomised trial that replaced the cliff with a gradual offset, one dollar of benefit for every two dollars earned above a threshold, found that the offset did not significantly increase employment. Beneficiaries did not respond to the ramp the way the theory said they would, apparently because the rules were still too complex and the other barriers, health, employers, the fear of losing coverage, remained. This proposal is built on that finding rather than around it.

The proposal

The Disability Ladder: a benefit offset of one dollar for every two earned above the threshold, in place of the cliff, because a cliff is unjust whether or not it changes behaviour. A partial-disability tier, paying half the benefit to a worker assessed as able to work part-time, with earnings alongside it. Expedited reinstatement for ten years after leaving the rolls, without a new application. And an employer retention obligation, on the Dutch model, for the first year after a worker's injury or illness, so that fewer people reach the programme at all.

How it would work

1. The ramp. Above the substantial-gainful-activity threshold, the benefit falls by one dollar for every two dollars earned, and reaches zero at roughly three times the threshold, about $5,000 a month. No beneficiary is ever worse off for earning a dollar more. The demonstration found this did not raise employment much; it also found that the beneficiaries who did work kept more of what they earned, which is the point of a fair rule even when it is not an incentive.

1. The partial tier. A worker assessed as able to work part-time, on the medical evidence, receives half the full benefit and may earn without the offset up to the threshold, then on the ramp above it. The tier is the missing rung between full disability and none, and it is the rung that most of the OECD has and the United States does not. Eligibility is assessed on the same medical evidence as the full benefit, by the same examiners.

1. Ten years to come back. A beneficiary who leaves the rolls by working may return, without a new application and with benefits restarting the next month, at any time within ten years if the impairment reasserts itself. Today the window is five years and the process is slow. The fear that stops most beneficiaries from trying to work is the fear of not being able to get back; this is the answer to it.

1. Coverage that does not end. Under № 004 the question of Medicare disappears, because every citizen is covered from birth. Until then, Medicare continues for any beneficiary on the ramp or in the partial tier for as long as they work, without time limit. The evidence is that coverage, not cash, is the barrier that binds.

1. Retention before entry. The Netherlands cut the flow into its disability programme in half by requiring employers to keep paying a sick worker for up to two years and to make a documented effort at accommodation and return before the state takes over. The Ladder adopts a one-year version: an employer of 50 or more must hold the job, fund a return-to-work plan through its disability insurer, and may not terminate for the impairment during the year. The state's programme begins where the employer's obligation ends.

1. Simple enough to use. The demonstration's own evaluators concluded that complexity blunted the offset. The rules in items 1 to 4 are published as one page, the calculation is done by the Social Security Administration from wage reports rather than by the beneficiary, and a beneficiary is told each month, in advance, what working more would do to the cheque.

The numbers

The beneficiaries. About 7 million disabled workers, plus dependents; about $150 billion a year in benefits.

The cost of the ramp. A gradual offset pays partial benefits to people who today would lose the whole benefit; the demonstration found the offset raised programme costs modestly. If 5 percent of beneficiaries earn on the ramp at an average partial benefit of $500 a month, the cost is about $2 billion a year. The partial tier adds new beneficiaries at half the benefit; if it draws 200,000 workers who would otherwise have applied for the full benefit and been awarded it, it saves money, and if it draws 200,000 who would otherwise not have applied, it costs about $2 billion. The honest range is that the Ladder costs a few billion a year, and the proposal does not claim it pays for itself.

The retention obligation. The Dutch reform reduced disability inflow by about half over a decade. An American version at one year, applied to employers of 50 or more, could plausibly reduce new awards by a tenth to a fifth, on the order of 100,000 fewer awards a year, which at an average lifetime benefit cost of several hundred thousand dollars is a saving that outweighs the Ladder's cost several times over, if it materialises. It is the one item on the list with a large upside, and the one with the least American evidence.

The honest objections

"The government's own trial showed a ramp doesn't get people working. Why build one?" Because a cliff is wrong on its own terms: a rule under which earning one more dollar costs a person $1,500 is not a rule a country should keep because fixing it turns out not to be an incentive. The demonstration is the reason this proposal does not claim the ramp will empty the rolls, and the reason it adds the partial tier, the reinstatement window and the retention obligation, which the demonstration did not test.

"A partial tier will expand the rolls, not shrink them." It may, and item 2 says so. A worker who today applies for the full benefit and wins it because the system has no other answer would, under the tier, receive half; a worker who today does not apply and struggles on might apply for half. The first saves money, the second costs it, and the net depends on the mix. The OECD countries with partial tiers spend less on disability as a share of GDP than the United States, which is suggestive and not proof.

"Employers cannot carry a year of a sick worker's job." The obligation is on employers of 50 or more, and it is to hold the job and fund a plan through the insurer, not to pay full wages for the year. Most of these employers already carry short-term disability insurance; the change is that the insurer must try to bring the worker back before handing them to the state. The Dutch obligation is two years and applies to every employer, and Dutch employment is high.

"This is complex, and complexity is what killed the demonstration." It is, and item 6 is the attempt to answer that within a programme that will never be simple. The test of the Ladder is whether a beneficiary can be told, on one page and in advance, what working will do to their cheque. If they cannot, the design has failed.

"Fraud will rise with a partial tier." The partial tier uses the same medical evidence and examiners as the full benefit, and a worker in the tier is by definition working, reporting wages, and visible. The full benefit, paid to a person who must not work, is the harder one to police.

Sources

• Social Security Administration, The Red Book: A Summary Guide to Employment Supports, 2026 substantial gainful activity and trial work period amounts
• Social Security Administration, Office of Research, Evaluation and Statistics, Annual Statistical Report on the Social Security Disability Insurance Program; "Employment among Social Security Disability Program Beneficiaries," Social Security Bulletin 71:3
• Mathematica Policy Research, Promoting Opportunity Demonstration: Final Evaluation Report (2022), on the benefit offset trial
• OECD, Sickness, Disability and Work: Breaking the Barriers (2010), on partial disability tiers and the Dutch reforms
• Koning and Lindeboom, "The Rise and Fall of Disability Insurance Enrollment in the Netherlands," Journal of Economic Perspectives (2015)
• Proposals № 004 (CitizenCare); № 013 (The Portfolio Act); № 026 (The Landing); № 040 (The Retirement Compact)