Cap Rate vs. IRR: Which Metric Answers Which Question

Two Numbers, Two Different Questions

The OM lands with a cap rate on the cover page and an IRR buried in the proforma tab. An LP calls asking which one to trust. A broker quotes one number over the phone and a lender's term sheet references the other entirely. Both are legitimate. Both are frequently misapplied — used to answer a question they were never built to answer.

Cap rate and IRR aren't competing measures of the same thing. They're answers to two different questions asked at two different points in a deal's life. Cap rate prices what the property is earning right now, in a single year. IRR prices what an investor's capital earns across an entire holding period, with every dollar in and out weighted for when it moved. Confusing the two — or worse, treating a strong number on one as proof of strength on the other — is how underwriting goes wrong quietly, months before anyone notices.

By the end of this piece, you should be able to state cleanly which question each metric answers, name the specific failure mode each one is prone to, and know exactly what to ask when a projected IRR looks too good.

What Cap Rate Prices: One Year, In Place

Cap rate is a snapshot. It divides one year of net operating income by price or value:

Cap Rate = NOI ÷ Price

That's the entire calculation. No debt, no time, no exit. It says: given this property's current income, at this price, an all-cash buyer earns this yield in year one, before financing and before anything changes. It's a pricing tool, not a return forecast — which is exactly why brokers put it on the cover page. It lets a buyer compare two properties, or one property against a market convention, in a single number.

That simplicity is the point, and it's also the limit. Cap rate says nothing about what happens in year three when a lease rolls, or year five when the roof needs replacing, or whether the buyer intends to hold five years or fifteen. It prices the income statement, not the investment. For the mechanics of building this number correctly — trailing income versus forward income, gross versus effective, what belongs in operating expenses — see our guide on how to calculate cap rate.

What IRR Prices: A Whole Hold, Time-Weighted

IRR asks a different question: given every dollar an investor put in, and every dollar that came back, at exactly what points in time, what single annualized rate makes the present value of those cash flows equal zero? It's solved by iteration, not by a simple division, because it has to account for the timing of each flow — a distribution in year two is worth more than the same dollar amount in year eight.

That time-weighting is what cap rate structurally cannot do. IRR captures acquisition price, every year of operating cash flow (including the effect of debt service, capital expenditures, and any refinance proceeds), and the sale price at exit — all folded into one number that reflects the entire strategy, not one year of it. A value-add plan with negative cash flow in year one, heavy renovation spend in year two, and a sale in year five can only be judged on IRR. Cap rate has nothing to say about a plan that unfolds over time; it can only describe income as it exists at a single moment. For the calculation walkthrough, see our IRR calculator for real estate.

Both metrics belong in the same underwriting model, alongside cash-on-cash and equity multiple — each answering a piece of the return question that the others don't. That fuller picture is covered in real estate investment metrics.

Where Cap Rate Misleads

The most common cap rate failure mode is a clean number sitting on top of a property that needs real capital. A seller quotes a cap rate calculated on trailing twelve-month NOI, before disclosing that half the roofs need replacement and a third of the units are below market because of deferred maintenance the current owner never addressed. The cap rate is accurate. It's also irrelevant to the actual return a buyer will earn, because it says nothing about the capital that has to go in before that income is stable or before it can grow.

A second failure mode: comparing cap rates across properties with structurally different income durability. A cap rate built on a single-tenant lease expiring in fourteen months is not comparable to the same cap rate built on a diversified multifamily rent roll, even at an identical number, because the risk to that income stream in year two is completely different. Cap rate treats both cases the same, because it was never designed to price risk to future income — only to price income that exists today. For the spreadsheet mechanics that keep this comparison honest, see our cap rate calculator in Excel.

Cap Rate vs. IRR: Where IRR Misleads — The Exit Cap Bet

IRR's failure mode is subtler and, in practice, more expensive, because it hides inside an assumption most models don't flag prominently: the exit cap rate.

Every multi-year proforma has to assume a sale price in the final year, and that sale price is almost always modeled as forward NOI divided by an assumed exit cap rate. Change that one assumption and the projected IRR moves substantially — often more than any single operating assumption in the entire model. This means a headline IRR is never purely a bet on the property's operating performance. It's partly a bet on where cap rates sit in the market at the exact year the model assumes a sale.

Consider a simple illustrative example with internally consistent, made-up numbers used purely to demonstrate the mechanic — not a market forecast. A sponsor buys a property for $10,000,000 at a 6.0% entry cap rate on $600,000 of in-place NOI. The business plan grows NOI to $720,000 by year five through renovation and lease-up. If the model assumes the exit cap rate stays at 6.0%, the year-five sale price is $720,000 ÷ 0.06 = $12,000,000. If the model instead assumes the exit cap rate compresses to 5.5% — a bet that the market gets more aggressive, not that the property performs any differently — the same $720,000 of NOI sells for $720,000 ÷ 0.055 ≈ $13,090,909, over a million dollars more, with zero change to the operating plan. Run that same $720,000 through a 6.5% exit cap instead and the sale price drops to $11,076,923. The operating story didn't change at all across these three cases. The IRR did, meaningfully, because the exit cap did.

This is why sensitivity testing an exit cap assumption — a simple grid across a range of plausible exit caps, holding the operating plan constant — belongs in every model before a projected IRR gets circulated to investors. A short holding period compounds the same risk: the shorter the hold, the more the total return depends on the sale price relative to the purchase price, and the less time there is for operating performance to offset an exit cap that moves against the sponsor.

When Cap Rate and IRR Disagree

The two metrics disagree most visibly on value-add and opportunistic deals, and that disagreement is informative rather than a contradiction to resolve.

A stabilized, fully-leased asset priced at a tight cap rate can carry a modest IRR, because there's limited room for NOI growth and the return is mostly the in-place yield plus whatever appreciation the market delivers on exit. A distressed or under-managed asset can trade at a wider cap rate — reflecting real, current risk and lower in-place income — while modeling a much higher IRR, because the business plan assumes substantial NOI growth from lease-up, repositioning, or expense reduction over the hold. Neither number is wrong. They're measuring different things: one is pricing today's income against today's risk; the other is pricing a multi-year plan to change that income.

The disagreement becomes a genuine warning sign only when the IRR is high but the underlying plan to get there is thin — heavy reliance on rent growth assumptions with no comparable evidence, or an exit cap meaningfully below the entry cap with no stated reason the market would compress. In that case, the gap between the two metrics isn't showing two valid perspectives on the same deal. It's showing a return that depends on the market doing the sponsor a favor.

The LP's Question: Sixteen Percent of What

When a projected IRR comes back at 16% — again, purely an illustrative figure for this discussion, not a benchmark — the correct LP response isn't to accept or reject the number. It's to ask what it's built from.

Specifically: what entry cap rate and exit cap rate does the model assume, and is the exit cap the same as, above, or below the entry cap? What holding period produces this IRR, and how much of the total return comes from the sale versus from operating cash flow along the way? How much of the projected NOI growth is contractual (in-place rent steps, signed leases) versus assumed (market rent growth, lease-up at a market rent that hasn't yet been achieved)? What happens to the same IRR if the exit cap rate is held flat, or moved 50 to 100 basis points against the sponsor, with everything else unchanged?

Those four questions decompose a single headline number into its actual sources: current income, projected growth, and a market-timing bet on the exit. Underwriting conventions around exit-cap spreads and hold-period assumptions vary by lender and by sponsor's own investment committee standards — there's no universal rule for how much cushion to build in, so confirm current practice with your own lender or investment committee rather than treating any specific spread as fixed.

Put Both to Work

Cap rate and IRR aren't rivals for the title of "the real return." They're a division of labor — one prices the income statement today, the other prices the whole strategy across time, exit assumption included. A model that only shows one is showing half the deal.

The Cap Rate & Property Valuation Toolkit builds both sides of that picture in one unlocked, documented workbook: clean cap rate valuation alongside multi-year IRR with an exit-cap sensitivity grid built in, so the assumption doing the most work in your projected return is visible rather than buried. See the Cap Rate & Property Valuation Toolkit and underwrite both questions at once, not just the one that fits on the cover page.