In 2034, American Social Security’s reserves will be depleted. Continued revenue would cover only approximately 83% of scheduled benefits in 2034 according to the 2026 report. The decisions Congress makes in the years ahead will determine who bears the adjustment: workers or retirees, affluent households or modest-income households. This choice will have lasting effects on the distribution of retirement income.
The Essentials
- According to administrators in 2026, combined reserves would be exhausted in 2034 and 83% of scheduled benefits would be payable.
- Closing the gap through immigration would require 3.9 million net entries per year, a threshold never reached since 1940.
- The share of people 65 and older is today close to 18% and is projected around 23% toward 2080, while the ratio of workers to retirees declines (Stanford SIEPR).
- Political divisions have limited negotiated reforms. The Social Security Fairness Act, enacted on January 5, 2025, eliminated the WEP and GPO, a major reform of the WEP and GPO, but not the first substantive modification to Social Security since 1983. Two scenarios remain open for overall financing: one gradual, one abrupt.
- Retirees most dependent on the program, those without assets or supplemental income, will absorb the sharpest blow if no decision is made before the deadline.
2034: A Date, A Mechanism, A Constraint
American Social Security operates on a pay-as-you-go basis. Worker contributions finance retiree benefits. For decades, the program accumulated reserves because contributions exceeded payments. This surplus is depleting. The program administrators’ projections place complete exhaustion in 2034.
On that date, no political decision is necessary for benefits to be reduced. Absent legislative action, payable benefits are limited by the revenues of legally distinct funds; exhaustion of the OASI is projected as early as 2032. At the combined exhaustion projected for 2034, approximately 83% of scheduled benefits would be payable; OASI alone would fall to 78% in 2032 without legislative change. No vote, no debate: an accounting mechanism.
This mechanism has been known for years. Administrators publish an annual report that specifies its deadline. The Congressional Research Service and Stanford SIEPR have documented the available adjustment margins. What is missing is political decision.
Demographics Cannot Close the Financing Gap
The reasoning appears simple at first glance: fewer workers to finance more retirees, so add more. Immigration is often presented as the natural adjustment variable.
The numbers contradict this intuition. The figure of 3.9 million concerns a scenario for eliminating the long-term actuarial deficit, not an objective of maintaining the worker-to-retiree ratio. The historical record since 1940 does not approach this threshold. Even a very expansive immigration policy, politically improbable in the current American context, would cover only a fraction of the gap.
Birth rates obey the same logic of delay. A child born today enters the labor market in 2043. Babies born in 2026 do not feed contribution revenues in 2034. The demographic transition is already underway and its effects are irreversible within the timeframe of the crisis.
Stanford SIEPR states it clearly: demographic levers are too slow and too limited to act before the deadline. The problem is structural, not cyclical. Choices must be made, not awaited.
Three Levers, Each With Its Political Cost
Economists and the Congressional Research Service identify three families of parametric reforms. They are not mutually exclusive, but each transfers the burden toward a different group.
Raise contributions. The current rate is 12.4% of wages, shared between employer and employee. A hike refills the program but weighs on low wages and small employers. It affects proportionally more workers without capital income, those who have only their wages to contribute.
Raise the full retirement age. The United States moved progressively from 65 to 67 years. A new increase would reduce total benefit payments but concentrate adjustment on workers in physical trades, whose healthy life expectancy is shorter. Employment policies play a critical role here: keeping senior workers employed requires a labor market that accepts them.
Broaden the contribution base. Today, income above $168,600 annually (2024 threshold) is no longer subject to Social Security contribution. Raising or eliminating the cap would increase contribution on high wage income, without subjecting capital income to OASDI contribution. This measure weighs almost exclusively on the wealthiest 10%, which makes it politically easier to defend, and earns it organized resistance of equivalent magnitude.
Each reform is feasible. Their combination is too. What is missing is the political coalition to adopt them.
The Trap of Institutional Deadlock
The United States has already reformed Social Security. In 1983, under Reagan, a bipartisan commission negotiated a contribution increase, a gradual increase in retirement age, and partial taxation of high-income benefits. This compromise pushed the deadline back by several decades.
The political context of 2026 is different. Congress is fragmented. Each camp has an electoral base, a significant part of which is directly concerned by the choices at stake: Republican retirees dependent on benefits, Democratic workers hostile to any contribution increase. Reform requires explicit losers in each camp. No party has an interest in naming them first.
The Swedish analogy is often cited by economists. Sweden created in the 1990s an automatic adjustment mechanism: when the ratio between contributions and commitments crosses a threshold defined by law, benefits and contributions adjust automatically, without a vote. The mechanism removes the decision from the electoral cycle. It requires in return a trust in institutions that Sweden possessed, and that the current American system struggles to build.
Two Trajectories Through 2034
Without reform, OASI would be limited to ongoing revenues starting in 2032; the combined 2034 scenario presupposes a modification allowing transfers between funds. Combined benefits would be payable at approximately 83% in 2034 under the assumption of combined fund treatment; OASI alone would be at 78% in 2032. A retiree receiving $1,500 per month would receive $1,215. Roughly one-quarter of elderly people live in households where Social Security represents at least 90% of family income. For affluent retirees with investment income, it is absorbable.
This scenario of partial default is neither catastrophe nor solution. Absent reform, payable benefits would be limited to available revenue after reserve depletion, without the distribution of the effect among beneficiaries being established. Data on intergenerational mobility suggest that retirees at the bottom of the income distribution had less access to asset accumulation during their working lives. They are the ones with no cushion.
The alternative scenario assumes a reform negotiated before reserve depletion. It can be early, from 2026-2028, when margin for maneuver is still broad, or late, around 2032, when the political pressure of the countdown finally forces compromise. An early reform allows gradual implementation: contributions rise over ten years, retirement age shifts gradually, high earners contribute more. A late reform imposes the same measures but urgently, with more abrupt effects.
Two signals make it possible to track which trajectory is underway. The first is the annual ratio between incoming revenue and scheduled benefits: according to 2025 intermediate projections, the gap between revenues and scheduled cost is approximately 18% of cost in 2033 and 17% in 2034; it alone does not allow establishing the probability of last-minute failure to reform. The second is the effective replacement rate perceived by the most modest retirees, the bottom 20% by income: its stability would indicate that adopted reforms protect the bottom of the distribution; its fall would indicate the opposite.
Poor Retirees Have No Reserves to Absorb a Cut
Debates on Social Security solvency are often conducted in aggregates: average replacement rate, overall actuarial deficit, total cost of adjustments. These aggregates mask dispersion.
A retiree who contributed on a median wage for forty years receives a benefit that covers a fraction of his or her last income. If he or she has savings, a 401(k) plan, or a spouse with a pension, a cut of approximately 17% is painful but absorbable. For someone who worked her entire career in part-time jobs or in sectors with high turnover, and who accumulated little or no private retirement savings, a profile overrepresented among women with discontinuous careers, minorities, and less-skilled workers, Social Security constitutes the sole source of income.
Raj Chetty’s work on intergenerational mobility documents that Americans born in lower income quintiles have significantly lower probability of accumulating assets over their lifetime. What this concretely means for retirement: choices about Social Security reform redistribute not only between generations but within each generation, along lines of class that preexist.
A reform that uniformly raises the retirement age ignores that total life expectancy varies by income level: an American study found a gap of 14.6 years in men and 10.1 years in women between the richest 1% and the poorest 1%. A reform that raises contributions without broadening the base weighs more heavily on average wages than on capital income. Each parameter has a social geography.
The Window Before the Lock
One of the most documented approaches combines two distinct mechanisms. An independent automatic adjustment committee, on the Swedish model, would trigger gradual corrections as soon as the ratio of workers to beneficiaries crosses a legal threshold, without awaiting a congressional vote. This mechanism removes the decision from the electoral calendar and makes adjustment predictable for workers planning retirement.
In parallel, mixed financing—wage-based contributions and general taxation on capital income—would allow broadening the base without concentrating the increase on low wages. This option requires a shift in fiscal doctrine, but it exists in several European systems. It amounts to explicitly posing the question of who pays for longer lives, and answering: those whose income has most benefited from decades of productivity growth.
These approaches are known. They have appeared in Congressional Research Service literature and academic institutions for at least two decades. Their implementation requires political legitimacy that neither Congress nor the executive has yet built. The window to act before the countdown imposes its own terms is closing. Before the projected exhaustion of OASI reserves in 2033, legislative reform would notably determine the scope and timeline of necessary changes.
Sources
- Stanford SIEPR, Strengthening Social Security’s Safety Net, https://siepr.stanford.edu/publications/policy-brief/strengthening-social-security-safety-net
- Congressional Research Service, reports on Social Security solvency (available at crsreports.congress.gov)
- Social Security Administration, The 2024 Annual Report of the Board of Trustees, https://www.ssa.gov/oact/TR/2024/
- Raj Chetty et al., work on intergenerational mobility and asset accumulation (Opportunity Insights, Harvard), https://opportunityinsights.org
