How to Study the Moreno-Warren Social Security Reform

A student trying to write a Bernie Moreno Social Security reform study runs into the problem almost immediately: the numbers do not line up. One table suggests that removing the payroll tax cap could push trust fund depletion decades into the future. Another score books trillions in new revenue over 10 years. A dynamic model cuts that revenue estimate by more than half and predicts a smaller economy. A generational model can make some reform bundles look growth-enhancing over the long run. Distributional tables then move the question away from whether the trust fund lasts longer and toward who pays, who gains, and which cohorts absorb the transition.

That disagreement is not a nuisance to clean up before the real analysis begins. It is the analysis. Social Security reform is a good test of whether a policy memo can hold several correct but incomplete answers at once without turning one of them into the whole conclusion.

Start With the Measurement Problem

The 2026 Trustees baseline supplies the pressure behind the exercise. The Old-Age and Survivors Insurance trust fund is projected to be depleted in the fourth quarter of 2032, which would trigger an automatic 22% benefit cut for retirees if lawmakers made no changes. The combined Old-Age, Survivors, and Disability Insurance trust funds are projected to be depleted in 2034. The 75-year shortfall is 4.42% of taxable payroll, or about $31 trillion, and that gap is 16% larger than the prior year’s estimate.[1]

Those dates make delay costly, but they do not tell an analyst which reform is adequate. The same baseline also tells the analyst what must be measured: trust fund solvency, annual cash flow, long-range actuarial balance, revenue, economic response, and distributional burden. A reform can look strong on one of those dimensions and weak on another.

The Moreno-Warren proposal is useful precisely because it is not yet a fully specified bill. Senators Bernie Moreno and Elizabeth Warren described a bipartisan plan to lift the payroll tax cap in a June 23, 2026, New York Times op-ed, but the proposal has not been introduced as formal legislation. Details such as phase-in, self-employment treatment, effective date, and whether additional earnings would earn additional benefits could still change.[7] Treating “lift the cap” as a fixed policy object before those design choices are known is where many weak memos begin to go wrong.

Lens One: Solvency Is the Natural First Stop, Not the Last

For Social Security, the first lens is usually actuarial solvency. The question is narrow and important: if taxable earnings above the current cap were brought into the payroll tax base, how much longer would the trust funds last and how much of the long-range shortfall would close?

The Social Security Administration’s Office of the Chief Actuary has a provision table for eliminating the taxable maximum beginning in 2025 with no benefit credit for newly taxed earnings. Under that specification, OASDI depletion moves from 2034 to 2059, and the long-range actuarial deficit falls from -3.82% of taxable payroll to -1.28%. In plain terms, that provision closes 67% of the long-range shortfall. It does not create permanent balance: annual surpluses last only through 2028, and deficits resume in 2029.[2]

That is a strong result, but it is not a portable verdict on the Moreno-Warren plan. The SSA provision is a specific design: eliminate the taxable maximum, begin in 2025, and provide no benefit credit on earnings above the old cap. If a proposal credits those additional earnings toward future benefits, the solvency gain is smaller because the reform raises future obligations as well as current revenue.

There is also a baseline mismatch that deserves more than a footnote. The SSA provision table uses 2025 Trustees assumptions, while the current policy problem is now being discussed against the worse 2026 Trustees baseline. The 2026 long-range gap is larger than the prior year’s estimate, so a re-run of the same provision against 2026 assumptions would likely close a smaller share of the shortfall than the published 67% figure. The number remains useful; it is just not entitled to travel without its assumptions.

Lens Two: Conventional Scoring Turns Solvency Into Budget Arithmetic

The second lens asks a different question: how much revenue would the federal government book over a budget window under conventional assumptions? This is where the memo moves from “what happens to the trust fund date?” to “how much money is counted, over what period, under what scoring convention?”

The Peterson Foundation estimate cited by ABC News puts the 10-year revenue from lifting the cap at about $3.4 trillion.[3] Tax Foundation’s conventional estimate is close, at $3.2 trillion over 10 years.[4] Those figures are not the same as the SSA solvency figures. They are measured over a shorter window, they emphasize revenue rather than actuarial balance, and they do not by themselves say whether the system reaches sustainable long-range balance.

This is the point where a clean table is more useful than an argument. A 10-year conventional score can make a proposal look large in budget terms while still leaving the 75-year financing problem partly unresolved. Conversely, a provision that meaningfully improves actuarial balance may still be judged differently if lawmakers care about near-term cash flow, the unified budget, or effects outside the Social Security trust funds.

The benefit-credit design choice is especially important. The Committee for a Responsible Federal Budget’s Social Security Reformer tool shows that uncapping payroll taxes without benefit credit closes 61% of the 75-year shortfall, while uncapping with benefit credit closes 44%.[5] That is not a minor technicality. It changes the policy from a more redistributive financing increase into a tax-and-benefit expansion for high earners, with a smaller net solvency effect.

This is also why criticism of the proposal can be partly right and still incomplete. Forbes argued that the Warren-Moreno approach would create a “massive welfare entitlement for the rich” and would not save Social Security.[8] If newly taxed earnings receive benefit credit, the critique points to a real design issue. If the design provides no benefit credit, the critique is aimed at a different policy. A student memo should not treat that as a semantic dispute; it is the difference between scoring one proposal and scoring another.

Lens Three: Dynamic Models Ask What People and the Economy Do Next

Conventional scoring is deliberately disciplined. It counts the policy under standard assumptions rather than trying to make every household and firm react inside the score. Dynamic macroeconomic scoring opens that box. It asks whether high earners change labor supply, whether compensation shifts into untaxed forms, whether saving and investment change, and whether output is different after the reform.

Tax Foundation’s General Equilibrium Model gives a sharply different view of uncapping the payroll tax. Its conventional $3.2 trillion estimate falls to $1.5 trillion after behavioral responses, a 53% reduction. The model also estimates that the policy would reduce GDP by 1.5% and reduce employment by about 1.8 million jobs. Tax Foundation further argues that the top marginal income-plus-payroll tax rate could reach about 60% in high-tax states such as New York and California, above its estimated revenue-maximizing rate of about 52%.[4]

A weak response would be to declare that the dynamic model has “debunked” the conventional score. A better response is to identify exactly what has changed. The conventional score asks how much revenue is booked before macroeconomic feedback. The dynamic score asks how the tax base itself changes when households and firms respond. The second question is important, but it does not erase the first; it answers a different question under a more behaviorally ambitious model.

Penn Wharton Budget Model shows why this lens cannot be reduced to one anti-tax result. In its analysis of stylized Social Security reform bundles, all five options increase long-run GDP relative to a debt-financed baseline. Its revenue-focused Option A raises GDP 2.4% by 2060, while its benefit-focused Option E, which relies on benefit cuts rather than tax increases, raises GDP 6.1% by 2060.[6]

Those results do not contradict Tax Foundation in the simple sense of one model being right and the other wrong. They use different models, baselines, and policy specifications. Tax Foundation is analyzing an uncap proposal and emphasizing marginal tax effects, compensation shifting, and labor supply responses. Penn Wharton is analyzing stylized packages that restore financial balance relative to a debt-financed baseline in an overlapping-generations framework. A policy memo that writes “dynamic scoring says GDP falls” or “dynamic scoring says GDP rises” has already lost the plot.

The Hardest Comparison Is Not Static Versus Dynamic

The hardest comparison is between adequacy and efficiency. A reform can improve solvency and still impose efficiency costs. A reform can increase long-run output relative to a debt baseline and still place large losses on particular current households. A model can be well built for one judgment and poorly suited for another.

The Moreno-Warren case is tidy enough to teach and messy enough to be honest. Solvency scoring says that removing the cap without benefit credit can do a large share of the financing work. Conventional scoring says the proposal raises trillions over a 10-year window. Tax Foundation’s dynamic model says behavioral responses substantially reduce the take and shrink the economy. Penn Wharton’s OLG work shows that the long-run growth result depends heavily on the baseline and reform bundle being modeled. None of those sentences is a complete policy conclusion.

The analyst's job is to make the denominator visible. Percent of taxable payroll is not dollars over 10 years. Trust fund depletion is not GDP. GDP is not household welfare. Household welfare is not political feasibility. Once the denominator changes, the conclusion has moved.

Lens Four: Distribution Changes the Question Again

Distributional analysis begins with a blunt fact: only about 7% of workers earn above the $184,500 payroll tax cap, so the direct tax increase from lifting the cap falls on a narrow group of high earners.[4] That narrow base is part of the political appeal. It is also part of the economic concern, because concentrating the tax increase at the top raises the marginal-rate questions that dynamic models emphasize.

Public opinion belongs here, not at the beginning as proof that the design is sufficient. A 2025 Bipartisan Policy Center poll cited in the Warren-Moreno op-ed found that 65% of Democrats and 62% of Republicans support lifting the cap, including most households earning over $200,000 who would pay more.[7] ABC News also reported the proposal in the context of broad support for taxing higher earnings.[3] That tells an analyst something about political feasibility and perceived fairness. It does not settle the actuarial, macroeconomic, or generational questions.

Generational distribution is a separate issue from income distribution. Penn Wharton’s equivalent-variation tables show that revenue-focused reforms impose larger costs on current older cohorts but deliver smaller gains to future cohorts. In the reported table, the age-60 middle-quintile group faces a -$30,745 equivalent-variation change under the revenue-focused option, while those born in 2051 receive a +$42,025 gain. Benefit-focused reforms show the opposite pattern, with larger gains for future cohorts.[6]

That table should be handled carefully. Penn Wharton notes that the bottom quintile of Social Security beneficiaries is not the same as the bottom quintile of current income at age 60.[6] In other words, the analyst cannot casually translate a beneficiary quintile into a current-income class. This is exactly the kind of warning that belongs in a graduate policy memo because it prevents a precise-looking table from being used for a claim it does not support.

How to Write the Memo Without Overclaiming

A defensible analysis of the Moreno-Warren proposal should not begin by choosing the most convenient estimate. It should begin by specifying the policy variant. At minimum, the memo should state whether the cap is eliminated or merely lifted, whether earnings above the old cap receive benefit credit, when the change begins, whether it phases in, and how self-employment income is treated. If the proposal text does not answer those questions, the memo should say so and analyze explicit scenarios rather than pretending the details are settled.

• Use Trustees data to define the financing target: depletion dates, automatic benefit-cut stakes, and the 75-year shortfall.
• Use SSA-style provision scoring to ask how much a specific design improves actuarial balance.
• Use conventional scoring to show the booked revenue over a budget window, while keeping the window and assumptions visible.
• Use dynamic scoring to test whether behavioral and macroeconomic responses shrink the tax base or change output.
• Use distributional and generational tables to identify who bears costs, who receives gains, and which cohorts are protected or exposed.

The order matters because each lens disciplines the next one. Solvency scoring prevents the analyst from mistaking popularity for adequacy. Conventional scoring prevents a vague “tax the rich” claim from floating free of a dollar estimate. Dynamic scoring prevents the revenue estimate from being treated as if people and firms never respond. Distributional analysis prevents aggregate improvement from hiding concentrated losses.

A strong memo can also let credible estimates disagree in public. It can say that under one SSA provision, eliminating the cap without benefit credit closes 67% of the long-range gap using 2025 assumptions; that the current 2026 baseline is worse; that conventional 10-year estimates cluster around the low-$3-trillion range; that Tax Foundation’s dynamic model sharply reduces the expected revenue; and that Penn Wharton’s OLG results show long-run GDP effects depend on the reform bundle and baseline. That is not indecision. It is the minimum standard for not confusing the output of a model with the answer to a policy question.

What the Moreno-Warren Case Teaches

The Moreno-Warren proposal should not be reduced to “saves Social Security” or “does not save Social Security.” With no benefit credit and under a favorable provision design, lifting the cap can materially improve solvency. With benefit credit, the improvement is smaller. Under conventional scoring, the revenue is large. Under dynamic scoring, the revenue and GDP effects can look much less favorable. Under generational analysis, the fairness question depends on whether the analyst is looking at high earners today, older cohorts near retirement, future workers, or beneficiaries grouped by lifetime rather than current income.

The proposal is a useful case study because credible modelers can produce divergent verdicts without anyone needing to be dismissed as biased. The divergence comes from model purpose, baseline, behavioral assumptions, and policy specification. A rigorous policy analysis memo has to identify the exact design, choose the relevant metric, state the model assumptions, compare results across lenses, and separate adequacy, efficiency, distributional fairness, and political feasibility. The numbers matter, but the method determines what the numbers are allowed to mean.

References

1. Analysis of the 2026 Social Security Trustees’ Report, Committee for a Responsible Federal Budget.
2. E2.1: Eliminate the Taxable Maximum in Years 2025 and Later, Social Security Administration Office of the Chief Actuary.
3. Social Security fix proposed by Sens. Elizabeth Warren, Bernie Moreno, ABC News.
4. Save Social Security Payroll Tax Cap Proposal, Tax Foundation.
5. Social Security Reformer, Committee for a Responsible Federal Budget.
6. Six Options to Restore Social Security’s Financial Balance, Penn Wharton Budget Model, March 9, 2026.
7. Warren, Moreno Pen NYT Op-Ed: Our Bipartisan Plan to Save Social Security, Office of Senator Elizabeth Warren, June 23, 2026.
8. Elizabeth Warren And Bernie Moreno’s Plan Won’t Save Social Security, Forbes, June 26, 2026.