What it costs to wind down Social Security, and which choices move that cost

v1.1 · Created Sep 22, 2026 for v3.8.339 · Rebuilt Sep 22, 2026 for v3.8.340 · content last changed v3.8.457

Jason Robertson

Ohio · 2026

Sources Baseline. Every figure in this document is recomputed by tools/build_analysis_results.py, stored in 09_Meta_Tracking/transition_analysis_results.json, and written here by tools/build_transition_design_doc.py. The models run in today's dollars on the sealed 2026 Social Security Trustees Report. The results rest on tools/transition_lever_model.py reproducing 04_Combined_Reform_Model.xlsx to the cent, which the audit checks on every release (ANALYSIS-TOOL-DRIFT).

THE PLATFORM PUBLISHES THIS BOTH WAYS. Whether interest on the debt raised during the transition is charged against the platform, or refinanced as ordinary federal debt like any other borrowing, is an accounting judgment, and the author publishes both rather than choosing for the reader (ASSUM-transition-debt-treatment, decided v3.8.342). Charged to the platform, 972 of 972 combinations of the platform's own settings close the deficit. Refinanced, with the platform held to its steady state, 972 close, 243 of them keeping both the age-55 promise and the $9,500 healthcare target. Both sets are set out below.

Refinanced, three combinations keep everything the platform promises: the full ladder repayment, grandfathering at 55, the $9,500 healthcare target and no additional payroll contribution. They ask for one change, the fund's share of the retirement contribution rising from 20 to 20%, which takes individual accounts from 9.6 to 8.4 percent of pay.

Why this document exists

The platform replaces Social Security with a contribution-funded system. Winding down the old one is its largest fiscal commitment, larger than any pillar, and its size depends on design choices rather than on economic conditions: how much departing workers are repaid, when they leave, who keeps the old system, and how long the wind-down takes. This document sets out those choices, what each costs, and which have been decided.

The starting point: with the design as decided, phasing out Social Security requires about $85 trillion of cumulative borrowing in today's dollars, peaking around 2085. With no replacement system at all it would require about $73 trillion by 2085, in today's dollars, with no reform. Over the first 25 years the Social Security transition alone accounts for about $35 trillion over the first 25 years, for the Social Security transition alone.

What departing workers are repaid

A worker who leaves gets back the contributions they made before enactment, credited with interest, and paid before they retire. Workers leave oldest first, each age band inside its remaining working years, so nobody is repaid after retiring. The credited rate is the government's own real borrowing cost. The table shows why the rate is capped there: above it, a promise owed for decades grows faster than the debt that funds it.

This rule stands on its own terms: you get back what you paid, plus what it cost the government to hold it. It is deliberately not offset against the savings the platform delivers to households, which stand on their own terms in 05_Per_Citizen_Benefits_and_Costs.docx. Offsetting the two would count one benefit twice, settle specific debts with an average, and invite the reply that the new contributions are what buy the savings.

What moves the cost

Each figure below is the peak borrowing, in today's dollars, with one setting changed and everything else held.

The phase-out window and the repayment share are the strongest levers. Stretching the window asks nothing of any individual, at the price of the old system lingering longer; cutting the repayment pays people less for contributions they actually made.

Who leaves when

The order in which workers leave was not specified before v3.8.338. It matters: a worker still in the old system pays payroll tax into it, while one who has left is being repaid and is contributing to the platform instead.

Youngest first is cheaper and was rejected: its saving comes from leaving older workers in a system they will never draw from, paying in for decades and qualifying for nothing. Oldest first costs more and repays people while the money still matters to them.

Can the gap be closed?

A search over 972 combinations of these settings, together with pillar contribution rates and the healthcare target, finds 972 that close the platform's deficit including interest on the transition debt, and several hundred that close the steady state alone. The difference between those two counts is the open accounting decision described at the top of this document.

The least disruptive of the combinations that close the full picture asks for: a repayment at 100% of the ladder, a 50-year phase-out, grandfathering reduced to 13% of workers, 20% of the retirement contribution going to the fund, 0 additional payroll point(s), and a healthcare target of $9,500. It leaves peak borrowing at $85.2 trillion.

No combination closes the full picture while keeping both the age-55 promise and the $9,500 healthcare target: 243 of them do so. That is the platform's central design tension, stated plainly.

What the design rests on

Social Security is absorbed: the trust fund moves into the Sovereign Fund at year 1 and every payroll tax dollar collected during the wind-down goes to the fund, with benefits and repayments financed by borrowing. The fund then compounds on the whole payroll base rather than on a fifth of the new contribution, and every combination of the levers above closes the platform's deficit at its published return. THE LEVERS ARE THEREFORE NOT WHAT THIS DESIGN DEPENDS ON. It depends on the spread between what the fund earns and what the government pays to borrow, which is 1.5625 percent in today's dollars.

Below about 3 percent the design reverses and the platform adds to the deficit again, because it pays interest on a larger debt with a fund no longer outrunning it. Two cautions are recorded with the assumption: a fund of this size would be several times US GDP and a large share of global capital markets, while the return assumptions are anchored to funds a fraction of that size; and retirees' benefits become dependent on the fund outrunning the borrowing rate for sixty years. The stochastic analysis puts the fund's year-60 range at $27 trillion to $333 trillion at the conservative return, which is the same point stated as a distribution.

What remains open

Three questions are unresolved and recorded as such. How transition debt is treated, described above. Whether contributions made after enactment are also returned: if they are, the payroll tax collected during the wind-down is a loan rather than revenue, and the cost rises sharply. And the age mix of the workers grandfathered at enactment, which is assumed rather than sourced, along with the benefit level used for new retirees, which is the overall average and therefore understates what they would actually draw.

Robertson, J. (2026). Transition Design: Options and Costs. We The People Platform (Version 3.8.483). https://wethepeopleplatform.com/_web_html/05_Analytical_Framing/05_Transition_Design_Options_And_Costs.html

Robertson, Jason. 2026. "Transition Design: Options and Costs." We The People Platform v3.8.483. https://wethepeopleplatform.com/_web_html/05_Analytical_Framing/05_Transition_Design_Options_And_Costs.html.

@misc{wtpp_2026_160_transition_design_options_and_costs,
  author    = {Robertson, Jason},
  title     = {Transition Design: Options and Costs},
  year      = {2026},
  publisher = {We The People Platform},
  version   = {3.8.483},
  url       = {https://wethepeopleplatform.com/_web_html/05_Analytical_Framing/05_Transition_Design_Options_And_Costs.html},
  note      = {Document 160 of 149}
}