Debate: How to Address America’s Retirement Savings Gap

Last updated 2 weeks ago. Our resources are updated regularly but please keep in mind that links, programs, policies, and contact information do change.

Kitty Ruderman had just turned 75 when she sat down before the House Ways and Means Subcommittee on Social Security in June 2021 to explain her budget. A retired legal secretary, she said Social Security was far and away the bulk of her income and she needed every penny of it. Then came the line that settled it: “My rent alone exceeds my monthly Social Security payments.” She skimped so much, she said, watching where the pennies go.

Her question is everybody’s question, asked later than most of us ask it. Will the check be there, and will it be enough?

The answer comes in two parts, because America has two retirement problems that get argued as one. Your check will almost certainly not stop, the Social Security gap is arithmetic that several unpopular measures each close on their own, and the deeper hole in most people’s retirement is the one their job never offered to fill.

What people your age actually have

Among households headed by someone aged 55 to 64, 57.0 percent had money in a workplace retirement account or an individual retirement account (IRA), and among those that did, the median balance was $185,000. For households 65 and older, 47.1 percent held an account, median $170,000.

That median describes only the households that saved something.

Against those balances sits the floor. The average retired worker collected $2,013 a month in November 2025, on a formula progressive by design: it replaces 83 percent of career-average earnings for a very low earner and 37 percent for a high earner.

What separates savers from everyone else is less discipline than the job. Among private-industry workers in the lowest-paid quarter in March 2025, 49 percent had access to a plan at work and 23 percent were in one. In the best-paid quarter: 91 percent access, 80 percent taking part.

You cannot contribute to a plan nobody offered you.

The month the reserves run out

The 2026 Trustees Report projects the Old-Age and Survivors Insurance Trust Fund reserves depleted in 2032, after which continuing tax income covers 78 percent of scheduled benefits for the rest of that year, declining to 62 percent for 2100.

Depletion is not bankruptcy. Payroll taxes keep arriving every payday; what runs out is the reserve stacked on top of them.

The law then ties the agency’s hands. Benefit payments “shall be made only from the Federal Old-Age and Survivors Insurance Trust Fund,” and nothing in that section lets the Managing Trustee borrow to cover a shortfall.

The Congressional Research Service (CRS) says it is unclear what the Social Security Administration (SSA) would do, and that the two workable options are full benefits on a delayed schedule or timely but reduced payments. Beneficiaries, it adds, “would remain legally entitled to full, timely benefits” and could sue for the balance.

So the check does not stop. It comes late, or it comes short, and the missing piece stays a legal debt.

Whether the benefit is yours: Ephram Nestor’s answer

Ephram Nestor came from Bulgaria in 1913 and worked in covered employment from December 1936 to January 1955, nineteen years of payroll taxes. He became eligible in November 1955 and was awarded $55.60 a month. In July 1956 he was deported for belonging to the Communist Party from 1933 to 1939, and a law enacted on September 1, 1954, cut his benefits off.

He sued. In 1960 the Supreme Court ruled against him, holding that a covered worker’s interest is noncontractual and cannot be soundly analogized to an annuity holder’s, whose right rests on contractual premiums. To engraft “accrued property rights” onto the system, the Court said, would strip it of the flexibility Congress had in mind when it reserved the right to alter, amend or repeal any provision of the Act.

That reservation is a single line of the statute: “The right to alter, amend, or repeal any provision of this chapter is hereby reserved to the Congress.”

You paid a tax, not a premium. Congress writes the schedule and can rewrite it, which is the unsettling half of the answer and also why it can raise your benefit as easily as cut it.

Fewer workers, and more pay above the cap

Two things moved. There were about 2.6 covered workers for every beneficiary in 2025, a ratio that held between 3.2 and 3.4 from 1974 through 2008, and the Trustees project it falling to about 1.9 by 2075.

The other shift is where the earnings went. The share of covered earnings falling below the taxable maximum dropped from 90 percent in 1982 to 83 percent in 2020, which CRS attributes to rising earnings inequality: top pay outran average wages, and the cap climbs only with the average.

The Trustees state the hole plainly. Over 75 years the shortfall is 4.42 percent of taxable payroll; closing it from today takes an immediate, permanent payroll tax increase of 4.25 percentage points, or an immediate, permanent cut of 25.2 percent in scheduled benefits.

Every proposal below chooses between those two sentences.

Scrap the cap

In 2026 the Social Security payroll tax stops at $184,500 of wages, a base that rises with average wage growth and by statute can never fall.

Sen. Bernie Sanders put the consequence on the Senate floor in February 2025, asking unanimous consent to pass his Social Security Expansion Act: “If you make $1 billion a year, you pay nothing more into the Social Security fund than someone making $176,000.”

Nancy Altman, president of Social Security Works, makes the affirmative case to House Ways and Means: benefits are “vital but inadequately low” and should be increased, with the cost of the expansion and the shortfall alike met by “requiring the wealthiest to pay.”

The scoring is less tidy than the slogan. SSA’s actuaries found that taxing all earnings immediately, with no benefit credit for the newly taxed wages, eliminates 73 percent of the shortfall, leaving about 0.96 percent of taxable payroll uncovered. Credit those earnings toward benefits as the formula normally would, and the fund still empties.

Mark Warshawsky of the American Enterprise Institute puts the case against. Social Security “is decidedly not a welfare program,” he writes, and scrapping the cap “would further reduce the link between earnings and benefits,” landing a 12.4 percentage point increase on about 10 million mostly middle-class workers. The Tax Foundation puts the combined top marginal rate on a New York City business owner as high as 60 percent, above the roughly 52 percent revenue-maximizing rate Treasury and Joint Committee on Taxation economists estimate.

Lifting the cap is the largest single lever anyone has scored. It is still not enough.

Work longer, and who can

On March 9, 1983, Rep. J.J. Pickle of Texas offered the floor amendment raising the full retirement age to 66 by 2009 and 67 by 2027. It passed 228 to 202. Rep. Claude Pepper of Florida offered a substitute raising the payroll tax rate from 6.20 to 6.73 percent in 2010 instead; it was rejected 132 to 296.

The tax alternative lost that afternoon. The age increase that beat it is still phasing in.

The Committee for a Responsible Federal Budget, which campaigns for deficit reduction, makes the modern case. Citing the program’s chief actuary, it argues that indexing the retirement age to longevity, so workers keep a constant ratio of years in work to years in retirement, would close one-fifth of the solvency gap and two-fifths of the structural gap, reaching 69 around 2075.

The objection is that longevity did not rise for everyone. The National Academies found that for men born in 1930 the highest-earning fifth outlived the lowest-earning fifth at age 50 by 5.1 years; for the 1960 cohort the projected gap widens to 12.7 years.

Nor is the work the same work. The Center for Economic and Policy Research, a progressive research group, found in 2009 data that 65.3 percent of workers aged 58 to 61 without a high school degree held physically demanding jobs, against 25.6 percent of college graduates.

A later retirement age is a benefit cut delivered through the calendar. Whose body does the waiting is the whole question.

Own it outright: the counties that left

In April 2005, hours after visiting the Bureau of the Public Debt in Parkersburg, West Virginia, President George W. Bush told an audience at West Virginia University there is no “trust fund,” “just IOUs that I saw firsthand.” The Parkersburg office stores them: “They’re stacked in a filing cabinet.” He was arguing for personal accounts.

Three Texas counties had already run the experiment. Galveston County told SSA in 1981 it intended to withdraw, its employees voted two to one for it, and Brazoria and Matagorda followed, the three collecting the same money to fund “Alternate Plans” covering about 3,000 workers.

The Government Accountability Office (GAO) compared the systems for hypothetical full-career workers. Over a 1964 to 1998 career, Social Security paid a low earner $750 a month against $617 from the Alternate Plan annuity; at high earnings the Alternate Plan won, $1,848 against $1,283. GAO’s caveat decides a long retirement: Social Security benefits are indexed for inflation and would eventually exceed the Alternate Plan annuities, which stay constant.

Then there is the bill for switching. Most payroll tax revenue is spent at once on current beneficiaries, so diverting part of it into accounts worsens the imbalance at least temporarily. Over 75 years, CRS reports, individual accounts could improve the system’s unfunded liability by $0.8 trillion or worsen it by up to $4.6 trillion, depending on design.

You can pre-fund the future. Somebody still has to fund the present.

Enrollment by default, and what it still misses

The private-side fix is already law and barely made the news. Newly established 401(k) and 403(b) plans must enroll workers automatically at a default of “not less than 3 percent and not more than 10 percent” of pay, rising a point each year. In plain terms: at a new plan, saving starts unless you stop it.

Defaults work, within limits. GAO reports studies finding participation can reach as high as 95 percent under automatic enrollment, while cautioning that participation in opt-in plans also climbs the longer people stay.

States moved first. As of July 31, 2026, state automatic IRA programs held $3.27 billion of the $3.35-plus billion administered across all state programs, in more than 1.36 million funded accounts. Marketplace, reporting on OregonSaves, said four economists who studied it found the plan “meaningfully increased” employee savings, and that opting out may have been rational for low-wage workers meeting current needs.

The leak is at the other end. The Employee Benefit Research Institute (EBRI) says “no effective solution has been implemented to address the problem of 401(k) cashout leakage“: each year roughly 40 percent of terminated participants cash out 15 percent of plan assets, an estimated $92.4 billion in 2015 alone.

A default decides whether you are in. It does not decide whether you stay in.

Is the shortfall as big as the number says?

The serious argument that the savings half is overstated rests on measurement, not optimism. Andrew Biggs of the American Enterprise Institute told the Senate Budget Committee that “Census Bureau research utilizing IRS data finds that retiree incomes are at record levels,” at the 25th, 50th and 75th percentiles alike.

That research, and a later Census working paper, matched survey respondents to their own records.

Adam Bee and Joshua Mitchell, the Census economists who ran it, traced the gap mainly to underreported pension income and retirement account withdrawals, not to survey design.

Alicia Munnell, whose index the crisis case rests on, counts a household as at risk when its projected replacement rate falls more than 10 percent short of what it needs to hold its living standard, assuming the household works to 65 and annuitizes every asset, reverse mortgage included. On the 2022 Survey of Consumer Finances the share at risk fell from 47 percent to 39 percent, the lowest since the index began in 2004, mostly because home values rose.

Both findings hold. Retirees have more income than the surveys credited them with, and 39 percent of today’s households are still projected to fall short on a measure whose own author calls it conservative.

What Congress already changed, and what forces the next move

The most recent change was not a solvency fix. On September 19, 2024, a discharge petition hit 218 signatures, forcing a House vote on the Social Security Fairness Act, a law that works by deletion: it strikes the government pension offset from section 202(k) and the windfall elimination provisions from section 215.

It also moved the deadline the wrong way. The Congressional Budget Office (CBO) scored it at about $198 billion in added direct spending over 2024 to 2034 and estimated it advances the combined funds’ exhaustion date by roughly six months.

Nothing now pending would close the gap.

Georgetown’s tracker lists the PROMISE Act of 2026, introduced that July by Sen. Richard Durbin as S. 4979, which would force a vote on a plan to keep the program solvent for 50 years, plus the Social Security 2100 Act.

Meanwhile the measured shortfall grew, from 3.82 percent of taxable payroll in last year’s report to 4.42 percent in this one.

The reserves have a date. Congress does not.