The Class B Multifamily That Lost LPs 47%
1. The deal as it was sold in 2021
The offering memorandum was 134 pages. The marketing deck was 28 slides. The numbers are reproduced here as they appeared in 2021.
| Metric | 2021 Underwriting |
|---|---|
| Acquisition price | $29.6M |
| Per-unit basis | $195,000 |
| Going-in cap rate | 4.2% |
| Senior debt | $22.2M (75% LTV bridge) |
| Common equity raise | $7.4M |
| Stabilization period | 2 years |
| Hold period | 5 years |
| Projected IRR | 18.0% |
| Equity multiple | 2.05x |
| Projected exit cap | 4.0% |
| Annual rent growth | 5.0% |
This was a typical 2021 Sun Belt multifamily underwriting. The cap rate, the rent growth assumption, the bridge debt structure, and the exit cap compression were all in line with what was being marketed across the asset class at the time. Hundreds of similar deals were sold in 2020 and 2021. Many of them are now in the same place this one ended up.
2. What the PPM said
Three clauses from the original 2021 offering memorandum, reproduced verbatim with the language that turned out to matter.
3. What actually happened
Four things that were not in the underwriting model but were in the world.
4. The capital call letter, December 2023
The capital call was the inflection point. The letter, two pages long, made three things clear and one thing unclear.
Clear: the amount required ($2.4M, or roughly 32 percent of original LP equity), the timeline (60 days), and the mechanism (pro rata participation with dilution for non-participants).
Unclear: what the capital call would actually accomplish. The letter described the funds as bridging the property to refinance. It did not describe what would happen if the refinance market did not improve, which was the realistic scenario. There was no projection showing LP recovery in the case where the additional capital did not produce the desired outcome.
LPs faced a binary choice (contribute or be diluted) without the information needed to evaluate which option produced a better expected recovery. The contributors ended up worse off than the non-contributors because the additional capital was deployed into a deal that continued to lose value.
5. The forced sale outcome
The economics of the disposition, from settlement statement, anonymized.
| Line | Amount |
|---|---|
| Sale price | $22,040,000 |
| Selling costs (3%) | ($661,200) |
| Net proceeds | $21,378,800 |
| Senior debt payoff (extended bridge plus accrued) | ($17,580,000) |
| Cash to equity | $3,798,800 |
| Original LP equity | $7,400,000 |
| Capital call equity (Dec 2023) | $2,400,000 |
| Total LP capital deployed | $9,800,000 |
| Recovery on original equity (non-contributors) | ~53c per dollar |
| Recovery on total deployed (contributors) | ~38c per dollar |
6. The PPM language that should have flagged this in 2021
Reading the original PPM today, four sections deserved more weight than they got at the time of subscription.
- The debt structure section. A 75 percent LTV bridge against a property requiring 24 months of stabilization is a structural bet that the rate environment will not move materially against the borrower during the stabilization window. In 2021, with rates at historic lows, the asymmetry was already obvious. The PPM did not stress-test the debt structure against a 200 bps move, let alone the 425 bps move that occurred.
- The exit cap section. An exit cap 20 bps tighter than going-in is an assumption, not a projection. PPM language describing it as the "General Partner's view" should have prompted the question: what happens at a flat exit cap, or at 50 bps of decompression. That sensitivity was not modeled.
- The rent growth section. 5.0 percent annual rent growth was reasonable in the trailing context but extraordinary in the historical context. The 30-year average for the submarket was 2.4 percent. The PPM did not model rent growth at the long-term mean.
- The capital call provision. Uncapped, discretionary, dilutive. This clause is standard in many PPMs. The right question at subscription is not whether it exists but what scenarios would trigger it and what LP economics look like in those scenarios. That question was not answered in the offering documents.
7. What the sponsor did right
This section matters because the failure was not primarily one of execution or character. It was structural.
- Quarterly communication throughout. The sponsor sent quarterly updates with actual NOI, occupancy, and rent growth against projections. The deterioration was visible to LPs in real time.
- A real workout process. The capital call was preceded by a detailed memo explaining the situation. The sale process was run by a competent broker. The sponsor did not delay the disposition to protect its promote at the expense of LP recovery.
- Did not blame the market without context. The disposition memo to LPs acknowledged the underwriting assumptions that did not hold and identified the structural choices that converted underperformance into capital risk. That is an honest post-mortem from the sponsor's side.
8. What the sponsor did wrong
The underwriting was aggressive but not unusual for the vintage. Many sponsors used 5 percent rent growth, 4 percent exit caps, and 75 percent bridge debt in 2021. The capital stack was the actual problem.
A sponsor who acquired the same asset in 2021 with 60 percent agency debt at a 10-year fixed rate would have ridden out the rate move, would not have faced a refinance event, and would have delivered a return profile in the range of 6 to 9 percent IRR rather than the marketed 18. That is a significant underperformance against the original projection but it is not a 47 percent loss of capital.
The 47 percent loss is what bridge debt converts moderate underperformance into when rates and cap rates move together against the borrower.
9. Lessons for evaluating 2026 deals
Every cycle has its set of assumptions that look reasonable in the moment and look obvious in hindsight. The 2021 vintage assumption that aged worst was the 4 percent exit cap. What is the 2026 equivalent?
Three candidates for the 2026 vintage's "4 percent exit cap" moment.
- The "rates will normalize" assumption. Many 2026 underwritings assume the 10-year Treasury settles back to 3.5 to 4.0 percent inside the hold period. If rates remain at or above the current 4.5 percent through 2030, the agency refinance environment will not provide the relief these deals are pricing in.
- The "AI-driven employment growth" thesis in secondary markets. Several 2025 and 2026 multifamily deals are using employment growth assumptions tied to AI-related job creation in markets that have not historically been employment hubs. The job-creation forecasts are real; the geographic distribution is the open question.
- The "office is bottoming" thesis. Office cap rates have decompressed materially. Some sponsors are now using current cap rates as exit caps, on the theory that the asset class has reset. That may prove correct. It also assumes that the structural shift in office demand has fully run, which is not yet established.
The 2021 deals that survived the cycle were the ones where the capital stack did not need a rate environment to cooperate, where the exit assumption did not need cap rate compression, and where the rent growth assumption was at or near the long-term mean. The same three filters apply to 2026 deals. The specific assumptions that age worst in this vintage will be different. The structural protection criteria will not.
Closing observation
The 152-unit deal did not fail because the sponsor was incompetent or dishonest. It failed because the structural choices made at acquisition (75 percent bridge debt, 4 percent exit cap, 5 percent rent growth) converted an unfavorable rate environment into a 47 percent loss of capital. Every one of those choices was inside the band of what the market was offering in 2021. None of them, individually, was a red flag at the time.
The lesson is not that 2021 underwriting was uniquely bad. It is that the structural choices LPs accept at acquisition determine what happens to their capital when assumptions miss. That is true in every vintage.