Reform Social Security

Key Points

  • Social security is facing a crisis. Without reform, the trust fund that pays a portion of the Old Age and Survivor Insurance (OASI) benefits will be depleted in 2033. From that point on, the Trustees estimate that payroll contributions from current workers will cover only 77% of earned benefits.

  • Congress has so far failed to unite around a plan, or even an approach, to fixing social security. Earlier this year, the Brookings Institute proposed a “blue print” for reform that meets the need without the poison pills that have defeated other attempts at reform.

  • My proposal below adopts most of the Brookings blue print, with a few changes:

    • Increase revenue from payroll taxes by increasing the maximum taxable ceiling to cover 90% of earnings and changing the rules for pass-through payroll tax.

    • Reduce benefits by increasing the number of working years in benefit calculations from 35 to 40, taxing all social security benefits of high earners, ending the dependent spouse benefit, eliminating child retiree benefits, restoring a reformed Windfall Elimination Provision and Government Pension Offset, and implementing one of the Congressional Budget Office options to slow growth of benefits by changing one or more bend points used in calculating them.

    • Increase social security coverage by increasing caps on legal immigration, including new employees in the roughly 6% of jobs that are not currently covered by social security, and repealing the Retirement Eligibility Test.

    • Change some of the current money policies affecting the OASI and Disability Insurance (DI) trust funds; specifically,

      • Devote all proceeds from taxes on social security benefits to those trust funds, and

      • Invest half of the OASDI Trust Fund in a total stock market index fund.

  • The challenge to social security is basically one of demographics: Americans are living longer now and having fewer babies. But with these changes, we can preserve social security for the long haul.

BackGround

  • Social security was established in 1935 to reduce poverty among the elderly. At that time, about 50% of elderly Americans lived in poverty. Social security brought that number down dramatically, and now about 6% to 8% of the elderly live in poverty.

  • According to the 2025 Trustees annual report for 2025, social security is providing benefit payments to about 68 million people, including:

    • 54 million retired workers and dependents of retired workers,

    • 6 million survivors of deceased workers, and

    • 8 million disabled workers and dependents of disabled workers

  • According to several studies, just over 40% of social security beneficiaries rely on the program for more than half their income, and about 14% rely on it for 90% of their income or more. In total, social security keeps about 20 million elderly Americans and about 1 million children out of poverty.

  • Social security has two parts with separate trust funds: Old Age and Survivors Insurance (OASI) and Disability Insurance (DI). These are usually referred to in combination as OASDI.

  • The program really is about insurance against poverty – it is not intended to be a retirement savings program. The program was built around two principles: adequacy (benefits need to be sufficient to provide a basic, reliable income in retirement to protect against poverty, even for low-wage workers) and equity (benefits need to be fair relative to contributions; e.g., given equal periods of retirement, no one who pays more in should get less out).

  • OASDI is collected from employers and employees, 6.2% each, for a total of 12.4% of paycheck for each working person. In the past, this was enough to pay benefits for current retirees and put some aside in a trust fund for future years.

  • Starting in 2021, worker and retiree demographics changed enough that social security had to start drawing down the OASI trust fund to pay part of the benefits owed to retirees. That situation continues today, and the Trustees project that OASI trust fund will be exhausted in 2033. At that point, the 12.4% contributed by and on behalf of current workers will still be used to pay benefits for retirees, but will only cover about 77% of current benefits.

  • The challenge to the OASI is primarily one of demographics. People are living longer and having fewer babies. We do not expect longevity to decrease in the future, and not even the most optimistic projections show the birthrate increasing up to the replacement level (2.1 babies per woman). This means that the problem is very unlikely to cure itself with time – if we want social security to endure, we need to reorient the program to meet our new reality.

Reform

  • Social security was reformed several times after 1935, most recently in 1983. Key elements of that reform were to gradually increase the full retirement age from 65 to 67 years of age, put all new federal employees in the program, and have federal income tax cover some social security benefits. These reforms secured the program at full benefits for 50 years.

  • Senate reconciliation rules bar provisions that recommend changes to social security, which means that neither party can push through changes on a party line vote with a simple majority. The parties have been split on how to reform the program. Broadly speaking, Democrats have generally proposed changes that increase revenue (for example, increasing taxes) to restore solvency and Republicans have generally proposed changes that reduce benefits (for example, increasing the age of full retirement).

  • In a February 11, 2025 proposal, the Brookings Institute put forward a “blue print” for reform that they expect will meet the needs of the program for the next 75 years. The proposal includes some revenue increases, some decreases to benefits, some increases to benefits, and some elements that they call “coverage and transfers,” including more people in the program and shifting some funds around.

  • My proposal below adopts almost all of what Brookings proposes, with a few adjustments.

My Proposal

  • Tax-Based Revenue Enhancements:

    • Increase the taxable maximum ceiling (Brookings). This proposal would increase the taxable maximum (currently $176,100) each year so that it eventually covers 90% of earnings, and then increase it as wages grow to keep it at 90%.

    • Change rules for pass-through payroll tax (Brookings). This proposal closes a loophole in which some types of business categorize income so as to not have to pay social security taxes on it.

    • Don’t increase overall payroll tax (me). Brookings recommends using a very small increase to the current 12.4% payroll tax as the final balancer to achieve solvency – an additional 0.2% split evenly between employer and employee contribution. I don’t want to increase payroll taxes at all. Social security is an insurance program – insurance against poverty. It is not a retirement savings program, and it mustn’t be allowed to crowd out the opportunity for people to save for retirement or build wealth in other ways, such as paying a mortgage. Once increased, payroll contributions will never go down.

  • Benefit reductions:

    • Increase the number of working years used to calculate social security’s average indexed monthly earnings (Brookings). Gradually increase the number of years used to calculate average indexed monthly earnings, so that by 2040 benefits would be calculated on the highest 40 years of earnings, rather than the current 35 years.

    • Tax all social security benefits of high earners (Brookings). Currently, for individuals earning between $25,000 and $34,000, and joint filers earning between $32,000 and $44,000, only 50% of their benefits are subject to federal income tax. Above those levels, more income is eligible for income tax, up to 85% of it. This proposal would change that so that all social security benefits of single individuals and joint filers with adjusted gross income above $100,000 and $125,000, respectively would be subject to income tax. These thresholds would be adjusted annually for wage growth.

    • End the dependent retiree spouse benefit (Brookings). This proposal would gradually eliminate the dependent spouse benefit for new retiree beneficiaries over time.

    • Eliminate child retiree benefits (Brookings). The proposal would end benefits to children of retirees and the associated caretaker (father or mother) benefit beginning in January 2027. Only newly retired beneficiaries would be affected, meaning everyone getting child retiree benefits in January 2027 would be protected. This benefit would continue for disabled children, adopted children, and children in the care of a grandparent or eligible relative.

    • Restore a reformed Windfall Elimination Provision (WEP) and Government Pension Offset (GPO), both of which were discontinued by the Social Security Fairness Act signed into law in 2025. These provisions were intended to address an inequity in which people who worked for some years in a job that contributed to OASDI and in other years in a job that did not were deprived of the social security benefits they should have earned. In some situations, though, discontinuing the WEP and the GPO could create an inequity the other way, with people receiving benefits not supported by what they paid into social security. I think we are capable of creating a reform to treat those groups fairly: give them the full benefits they earned and not more than that.

    • Adopt one of the Congressional Budget Office (CBO) options to either add another bend point for earners above the 70th percentile in determining benefit amount or to make the benefit structure more progressive (CBO). (This is in place of the Brookings proposal to increase the retirement age for high earners.) I think the Brookings proposal would carry the perception (and possibly the reality) that at some level of income, people will see their lifetime benefits actually reduced because they earned more. The CBO options also reduce the rate of growth of benefits as the individual’s income increases, but in a more even-handed way.

  • Benefit improvements.

    • Brookings proposes benefit improvements to targeted populations. All are well grounded and I don’t take issue with them.

    • However, Brookings is silent on some benefit improvements that I would like to see considered:

      • There is a good case for a minimum benefit for people who have contributed over a long career but whose distribution leaves them at or below the poverty line.

      • There is also a good case to remove the reduction of benefits that someone who left their career to be a caregiver for many years experiences. Any years short of 35 (40, if the proposal above is enacted) in which a person did not contribute payroll taxes are assigned a value of 0 in the average indexed monthly earning figure used to calculate benefits. People – predominantly women – who left their careers to provide caregiving services to children or elders are unfairly punished for that choice.

  • Increase program coverage:

    • Expand the labor force by changing legal immigration policies (Brookings). Raise the caps on legal migration by specified percentages in to get more people in the workforce.  Using Brookings’ stated percentage increases, my calculations imply that the specific changes they recommend would add about 1.1 million new temporary workers and about 1.2 million permanent residents to the US economy over 10 years, with continuing slow growth after that. This proposal directly addresses the ongoing challenges to the solvency of the OASI program posed by demographic realities of this century.

    • Achieve universal coverage in social security (Brookings). In 2024, about 6% of American workers held jobs considered “uncovered employment” in which they did not pay into social security. The largest share of uncovered workers is the 5.9 million state and local government employees who are insured by alternative public pension plans instead of social security. The proposal is that in 5 years (allowing the time to change existing retirement and pension plans) all newly hired employees of state and local governments would be covered by social security.

    • Repeal the Retirement Eligibility Test (RET). RET confuses people and may encourage them to not work once they become eligible for the full retirement amount.

  • Change money policies relating to the trust funds:

    • Devote all proceeds from taxes on social security benefits to OASDI trust funds (Brookings). The taxation of up to 85% of social security benefits was established in 1993 with the additional revenue from those taxes directed to Medicare’s Hospital Insurance fund. This proposal would move the proceeds from those taxes into the OASDI trust funds as soon as possible without harming the Medicare Hospital Insurance fund.

    • Invest half of the OASDI Trust Fund in a total stock market index fund (several sources). This proposal has been aired for several decades, but most have dropped it because the depletion of the trust fund is so close at hand. I think it is worth pursuing, even if just to extend full payments by a very short time; we will rebuild at least a small trust fund eventually. Social security was first established in 1935, during the Great Depression. The law as written directs that the trust fund can only be invested in Treasury securities. Our experience in the 90 years that followed, as well as smaller-scale initiatives like the National Railroad Retirement Investment Trust (NNRIT) that have been investing public retirement trust funds in the market for decades, gives us more confidence that the risk is bearable.