A public-interest project. Not a company, and it sells nothing. What that means
autonomist

The engine: limited-equity co-ops that people with options would choose

Jason Edelman and Claude · working paper · 2026-11-20

"What I'm really trying to do is solve the bootstrap and exit problems for limited-equity cooperatives in a way that is competitive with other investment options, for people that have investment options. I believe, that with a coordinated strategy, we can beat back private equity and vulture capitalism - but everything rides on that coordination." (Jason Edelman)

This is a working design, not a finding. The numbers come from the engine model (model/lec-core.cjs) on one sound 8-unit apartment building, using its listed price ($2.4 million) and cap rate (6.27%). Every rate and split is a scenario. The model's full results are in Results: the engine model.

1. Why limited-equity co-ops lose people with options

A limited-equity co-op has two exit problems.

There is also a bootstrap problem. A co-op raises equity slowly, usually from residents who have little, against buyers who close fast.

2. The design: split a building's equity growth in two

A building gains equity in two ways: its loan is paid down, and its value rises.

How the split is set is each co-op's own decision. Paydown, appreciation and indexation are governance choices, and the federation is built to let co-ops choose differently (section 5).

3. Three kinds of capital, each paid what it needs

partwho might supply itshare of costreturn (scenario)
senior notesallies with savings, credit unions, institutions70%5%, paid down over 30 years, secured by the building
solidarity moneyfoundations, congregations, patient allies20%2%, interest only
member equityresidents10% ($30,500 a home here)indexed 3% a year, plus the paydown credited

Blended, that is 4.2% a year. On the worked building (price plus repairs $2.44 million, net operating income $150,480):

4. The member's side

Members of the worked building pay charges below market rent and build an account. In the model's default run (which also funds reserves and a contribution to the federation pool), a member who joins at the start holds about $68,000 after 7 years, on a $31,400 buy-in. Against renting at market, the co-op member comes out ahead in every case the model ran. Against a comparable condo, the member comes out ahead when values grow about 3% a year or less, and behind when they grow about 6%, because the condo owner holds the appreciation with leverage. The member's exit is guaranteed at book value; a condo seller pays to sell and takes the market's timing. The tables are in Results, section 3.

5. A federation that firewalls risk

"If coops are financially isolated from each other but have access to a common pool of resources we can hopefully prevent failures from cascading - because they will happen. It's a balance between autonomous risk and the protection that comes with collectivization - that should be fluid." (Jason Edelman)

The model compares three designs under random shocks: co-ops on their own, co-ops that are separate entities sharing a pool, and co-ops merged onto one balance sheet. Merging fails least often, but when it fails, it fails whole. The firewalled pool fails more often, in pieces. Neither protects against a shock that hits every co-op at once. Spreading the federation across independent markets does: across four markets, the firewalled pool's worst run loses 60% of homes, while the merged federation's still loses all of them. The first quarter of savings pooled does most of the protecting. Leverage at purchase matters more than any other choice, which makes the patient solidarity money the resilience lever, not only a cheap one. Details: Results, section 4.

One result cuts against the design. Co-ops that learn by copying whichever peer looks richest drift toward saving less and crediting more appreciation, and that raises losses under severe stress. A public ledger of failures barely helped at 24 co-ops. For the federation's nervous system to teach the right lessons, it has to record why a co-op failed, not only that it did.

6. Risks and unknowns