Multifamily Exit Cap Rate: Where Most of the IRR Lives

Published October 25, 2026

In the five-year multifamily worked example below, sale proceeds account for roughly 84% to 90% of the dollars the equity investor ever receives. The exit cap rate that sets those proceeds is usually one cell, filled by convention — going-in cap plus 25 basis points — and defended less rigorously than the expense line it sits a thousand rows below.

That convention was cheap to hold when the long end of the curve was anchored. It is not cheap now. The 10-year Treasury was 5.21% on October 13, 2026, per the Federal Reserve's H.15 release, with the effective federal funds rate at 3.88% on the same day; it had eased to about 4.96% by October 22, per FRED. A going-in cap rate in the low-5% range against a long bond near 5% is approximately zero spread, which makes "entry plus 25 bps" a forecast of spread compression rather than a conservative placeholder.

Why the Multifamily Exit Cap Rate Decides More Than the Entry

The entry cap rate is observable. It is the price divided by an income figure both sides can argue about, and at closing it is a fact. The multifamily exit cap rate is an estimate of what a different buyer, with different debt, in a different rate environment, could pay five years out — and it is applied to the largest single cash flow in the model.

The asymmetry is arithmetic. A 50-basis-point error on a year-one expense line moves Net Operating Income (NOI) by a fraction of a percent. A 50-basis-point error on the exit cap rate moves the residual by roughly 9% of gross sale value, and because the loan balance beneath it barely moves over the hold, it lands almost entirely on equity. Leverage does not amplify the entry assumption the way it amplifies the exit.

Three Ways to Set an Exit Cap, and What Each One Assumes

Most models use the first method below and stop. Running all three is generally more informative, because they disagree, and the size of the disagreement is itself the finding.

The third test is the one most often skipped and the one that currently bites. On illustrative financing at a 6.0% rate and 30-year amortization, the debt constant is approximately 7.2%. A buyer at a 5.25% cap is paying well above where amortizing debt breaks even, which means the exit depends on equity willing to accept negative leverage — the same condition the seller is counting on, extended five years forward. If that condition does not hold at sale, the clearing price is set by a lower cap rate buyer who may not exist, or by a higher cap rate. Terms and pricing vary by lender and vintage; verify current quotes directly with a lender.

What the Cap Rate Surveys Actually Say Right Now

Separating the observable from the inferred matters here, so the sourced facts come first.

• CBRE's H1 2026 U.S. Cap Rate Survey, reflecting views as of late June 2026, reported that roughly 60% of respondents expected no change in cap rates over the following six months, and identified infill multifamily as the most bearish subtype — the one with the most respondents expecting cap rates to rise.
• In that same survey, stabilized multifamily cap rates in New York City widened from a 4.5%–5% range in the second half of 2025 to 5%–5.5% in the first half of 2026.
• Yardi Matrix data covering the 30 largest multifamily markets, as reported by Bisnow in October 2025, put the average transaction cap rate near 5.04% across roughly 929 deals through September 2025, with a spread of more than 200 bps between the lowest- and highest-priced markets.
• The 10-year Treasury traded in a roughly 5.1%–5.3% range over the first half of October 2026, per the Federal Reserve's H.15 daily series, and was 4.96% on October 22 per FRED.

What this could mean, read as an interpretation rather than a forecast: the survey consensus for flat pricing sits alongside a long bond above 5%, so a flat exit cap assumption is consistent with the consensus but offers no margin if the consensus is wrong in the direction the same survey's multifamily respondents lean. Cap rates are not a mechanical function of the 10-year, and these are national and metro averages — individual submarkets and vintages vary substantially.

A Worked Example: What 50 Basis Points Costs

Illustrative figures throughout. A 100-unit Class B property in a secondary market, purchased for $20M at a 5.25% going-in cap on $1.05M of year-one NOI. Financing is $12M at a 6.0% fixed rate on a 30-year amortization schedule, leaving $8M of equity. Annual debt service is approximately $863K, so year-one Debt Service Coverage Ratio (DSCR) is about 1.22x. NOI grows 3% a year, and the asset is sold at the end of year five on forward NOI of roughly $1.22M, net of a 1.5% cost of sale.

Two things fall out. The first is that 50 bps of exit cap movement cost roughly 370 bps of Internal Rate of Return (IRR) — roughly what halving the NOI growth assumption from 3% to 1.5% does in the same model, which lowers IRR by about 330 bps, to roughly 7.2%. The second is the line that tends to get missed: five years of 3% NOI growth created about $3.2M of value at a flat cap, and 50 bps of exit cap widening erased roughly 63% of it. The business plan and the residual assumption are not comparable in size, and the one that is harder to control is the larger.

Net sale proceeds were between 84% and 90% of total dollars returned across these scenarios. Cumulative cash flow over the hold was approximately $1.26M on $8M of equity — real, but not what the deal is actually about.

The Exit Cap Is Not Independent of the Rest of the Model

The most common failure is not an aggressive exit cap. It is an exit cap that contradicts the assumptions sitting above it. A model carrying above-trend rent growth, a refinance at a lower rate, and an exit cap at or below entry is making the same macro bet three times and counting it three times. Rent growth that strong generally implies an economy in which the long end is not falling, and cap rate compression generally implies a rate environment in which rent growth is harder to come by.

The denominator deserves the same scrutiny. An exit cap applied to a seller-style NOI — before the assessed-value reset a sale can trigger, with reserves below the agency minimum and a management fee under market — produces a residual that no buyer would underwrite. The reset alone can move the tax line materially in the second year, as covered in the year-two property tax reset. Applying a defensible cap rate to an indefensible NOI is the more expensive of the two errors, because it is invisible in the sensitivity table.

Where This Analysis Breaks

The strongest argument against all of this is that the exit cap rate is unknowable, and that disciplined spread analysis has been wrong for long stretches. A spread rule is also a blunt instrument. An underwriter who refused to buy whenever multifamily cap rates sat inside 150 bps of the 10-year would have kept buying through 2012–2021, when the 10-year mostly ran between roughly 1.5% and 3% (per FRED) and, per Yardi Matrix data reported by Bisnow, cap rates bottomed near 4.1% in 2021 — and would have bought the 2021 vintage at the tightest absolute pricing of the cycle. The rule only started to bite in 2022–2023, as the 10-year averaged about 4% in 2023 against cap rates near 5.2% (same sources), which is to say after the repricing was already underway. Spread discipline is a risk framework, not a timing signal, and treating it as the latter can miss both the top and the turn.

Cap rates are also not a clean derivative of Treasury yields. Replacement cost, capital flows, supply pipelines, and the depth of the agency bid all push on pricing independently, and the 2021–2023 repricing was not a tidy function of the curve. The current curve shape complicates it further: with the effective federal funds rate at 3.88% against a 5.21% 10-year, per the Federal Reserve, policy easing at the front end would not mechanically pull the long end down, and a long bond near 5% may not persist in either direction.

There is a reasonable case that precision here is theater. A point estimate carried to the basis point implies knowledge nobody has, and some underwriters would argue the better practice is to stop forecasting the exit cap entirely, present a band, and size the equity so the deal survives the wide end. That position has merit. It is also not an argument for leaving the cell at entry plus 25 bps — it is an argument for widening the band and saying so out loud.

The Underwriting Takeaway

Three things follow for a multifamily exit cap rate assumption. It belongs in a sensitivity table rather than a single cell, because the plausible error on it is larger than the plausible error on anything else in the model. It deserves a stated method — entry spread, Treasury spread, or financeability — rather than an unstated convention. And it needs to be consistent with the rent growth and refinance assumptions above it, since pairing optimistic operations with optimistic pricing counts one bet twice.

A deal that only works at a flat exit cap is a deal that requires a specific rate environment five years from now. That may happen. It is worth knowing it is the position being taken, rather than discovering it at sale, and the same year-by-year discipline that underwriting the 2026 turn and an occupancy-first revenue model demand everywhere else in the model applies here too. For the mechanics behind the return math, see the Internal Rate of Return reference.

UWmatic is an AI-powered underwriting platform built for multifamily and mobile home park investors. Run the exit cap rate as a band rather than a cell and see what each scenario does to IRR, equity multiple, and the share of return riding on the sale. Try the free underwriting calculator →

This analysis reflects current market interpretations as of the publication date and may evolve as new data becomes available. Figures cited are drawn from public sources including the Federal Reserve's H.15 Selected Interest Rates release, FRED (St. Louis Fed) Treasury yield data, CBRE's H1 2026 U.S. Cap Rate Survey, and Yardi Matrix transaction data as reported by Bisnow, and are subject to revision. Nothing in this post is investment advice; readers should conduct their own diligence and consult qualified professionals before making investment decisions.