Guide · Return metrics

Cap rate vs IRR vs equity multiple.

Three numbers claim to describe the same deal, and each one hides something the other two reveal. Cap rate prices the asset, IRR prices the time, and the equity multiple prices the total ride. Here is what each actually measures, where each misleads, and why serious underwriting quotes all three.

7-minute read · by the team at Nivora

Cap rate: the price of the income stream

The capitalization rate is the simplest of the three: net operating income divided by price. A property producing $1,100,000 of NOI at an $18,000,000 price trades at a 6.11% cap. No debt, no time, no projections — just this year’s unlevered income yield.

That simplicity is the point. Cap rate is the market’s common currency: it lets you compare a deal in Tampa against one in Phoenix in a single number, and it’s how your exit gets valued — the buyer in year five will pay your stabilized NOI divided by whatever cap rate the market demands then. What it hides: leverage, capital expenditures, growth, and every year after the first.

Direction matters more than magnitude: caps compressing (falling) means values rising for the same income; caps expanding means the reverse. Underwriting an exit cap below your going-in cap is a bet that the market will pay more per dollar of income in five years — a bet on timing, not on the asset.

IRR: the price of time

The internal rate of return is the annualized rate that discounts every cash flow — equity out on day one, distributions along the way, sale proceeds at exit — back to zero. It is the only metric of the three that knows when money moves, which is why institutional LPs anchor on it and why waterfall hurdles are written in IRR terms.

Its weakness is the same sensitivity: IRR can be flattered without making anyone richer. A cash-out refinance in year two that returns half the equity early can push IRR up several points while total profit barely moves. Two deals with identical 17% IRRs can return very different amounts of cash — which is exactly the blind spot the equity multiple covers.

Equity multiple: the price of the whole ride

The equity multiple (MOIC) is total cash returned divided by total cash invested. Put in $5,600,000, get back $11,700,000 across the hold — that’s a 2.09× multiple. It ignores timing completely, which makes it the honest counterweight to IRR: nobody engineers a multiple with a refi. A 2.0× over five years and a 2.0× over ten years are the same multiple but wildly different annual returns — which is why neither metric stands alone.

One deal, three lenses

The 120-unit example from our underwriting guide — $18.0M purchase, 65% leverage, five-year hold, sale at a 6.25% exit cap:

Same deal, three answers

Going-in cap rateNOI $1.10M ÷ $18.0M price6.11%
Levered IRRtime-weighted, after debt≈17%
Equity multipletotal cash back ÷ cash in≈2.0×

Read together, they answer the three questions an investment committee actually asks: is the price fair for the income (cap rate), is the return worth the hold (IRR), and is the absolute profit worth the effort (multiple)?

Where each one lies

  • Cap rate lies about growth. A 5-cap with rents 20% under market can be cheaper than a 6-cap at full market. It also says nothing about leverage — the same 6-cap deal produces very different equity returns at 50% vs 70% LTV.
  • IRR lies about scale and timing. Early distributions inflate it; a small, quick deal can post a dazzling IRR on trivial profit. Always ask what the multiple was.
  • Equity multiple lies about time. 2.0× sounds great until you learn it took eleven years. Always ask what the IRR was.
  • All three lie when the inputs do. Every metric is downstream of the rent roll, the T-12, and the expense normalization. Garbage in, confident-looking garbage out.

Quote all three, always

The discipline is simple: never present one without the others, and never compute them by hand twice. In Nivora, IRR (deal-level and LP-level), equity multiple, cash-on-cash, and the full cap-rate picture are computed from the same monthly cash-flow model and reconciled against each other on every change — so the three lenses can never quietly drift apart in separate spreadsheet tabs.

See it computed, not explained.

Every concept on this page is a number Nivora computes live — the real engine is running on a fictional 104-unit sample deal right now, no login required.