---
name: cap-rate-decomposition
source: https://app.decimal.ai/s/cap-rate-decomposition@1/SKILL.md
source_sha256: 9a6f0ab7d30a
---

# Cap-rate decomposition

A capitalization rate is the first-year unlevered yield on a property:
**cap rate = NOI ÷ value (or price)**. Everything below hangs off that identity —
but a usable analysis does three things the raw formula does not: it separates the
*entry* yield from the *exit* yield, it explains *why* a given cap rate is what it is,
and it puts comps on a common footing before treating them as evidence of a market rate.

Use net operating income that is **truly net**: after a capital-reserve deduction and
stripped of one-time items. A cap rate computed on an NOI that still carries a
non-recurring insurance recovery or omits a replacement reserve is not comparable to one
that is clean.

## 1. Going-in vs. terminal cap

- **Going-in (entry) cap** = Year-1 forward NOI ÷ purchase price. It is the yield the
  buyer locks in at acquisition.
- **Terminal (exit / reversion) cap** = the NOI of the year *after* sale
  (Year N+1, the first year of the next owner's hold) ÷ projected sale price. You apply
  it to forward NOI, not the year you sell in.

Set the **terminal cap at or above the going-in cap**. The asset is older and further
from its warranty/capex cycle at exit, and the future is less certain, so a spread is
added — typically **+25 to +75 bps** over the going-in cap for a stabilized hold, wider
for shorter-lived or more cyclical assets. Holding the terminal equal to going-in (or
below it) assumes the asset gets *better* with age and should be flagged as aggressive.

## 2. Build up the cap rate from its parts

A cap rate is not an opaque market number — it is a required return net of growth.
Decompose it so you can judge whether the quoted rate compensates for the actual risk:

```
required return (discount rate) = risk-free rate
                                + general real-estate risk premium
                                + property-specific risk premium
cap rate ≈ required return − expected long-run NOI growth (g)
```

Typical / illustrative bands — calibrate to the current market, do not treat as fixed:

| Component | What it pays for | Typical band |
|-----------|------------------|--------------|
| Risk-free rate | Time value; use the current 10-yr Treasury | the live 10-yr yield |
| General RE risk premium | Illiquidity + real-asset / sector risk over bonds | ~150–300 bps |
| Property-specific premium | The add-ons below, summed | ~0–400 bps |
| less: expected NOI growth (g) | Rent-growth outlook over the hold | ~100–300 bps |

The **property-specific premium** is itself a sum of add-ons. Each of these *widens* the
cap (raises required yield):

- **Asset class** — Class A ≈ +0; Class B ≈ +25–75 bps; Class C ≈ +75–150 bps.
- **Market tier** — gateway ≈ +0; secondary ≈ +25–75 bps; tertiary ≈ +75–150 bps.
- **Tenant credit & rollover** — weak credit, short WALT (weighted-average lease term),
  or heavy near-term rollover ≈ +0–150 bps.
- **Physical / deferred capex** — meaningful deferred maintenance ≈ +0–100 bps.

So two assets can share a headline cap rate for very different reasons: a Class C
tertiary-market building at 8% is priced for its risk stack, while a Class A gateway asset
at 8% would signal distress or expected NOI decline. Reconcile the quoted cap against this
build-up; a large gap is the thing to explain.

## 3. Normalize comps before you compare

Observed sale cap rates are only evidence of a market rate if the NOIs behind them are
defined the same way as the subject's. Before averaging comps or reading a market cap,
put every comp on the subject's basis:

**Occupancy.** A cap rate computed on in-place NOI at 78% occupancy is not comparable to
one at stabilized 94%. Restate each comp's NOI to a **stabilized economic occupancy** —
mark the vacant space to market rent, then subtract a stabilized vacancy and credit-loss
factor — and compare on that basis. Comparing an in-place-occupancy cap to a stabilized
cap silently mixes a lease-up discount into the "market rate."

**Lease structure (NNN vs. gross).** NOI is defined differently under different leases, so
their cap rates are not directly comparable:

- Under a **triple-net (NNN)** lease the tenant pays taxes, insurance, and CAM, so the
  landlord's NOI ≈ **base rent** (few landlord-borne operating expenses).
- Under a **gross / full-service** lease the landlord pays operating expenses, so
  NOI = **gross rent − landlord operating expenses**, a margin often 25–40% below the
  gross rent.

Never compare an NNN cap rate to a gross cap rate as-is. Restate all comps to the **same
expense-responsibility basis** as the subject — e.g., gross up the NNN comp to a full-
service equivalent (add the reimbursed expenses to both rent and opex) or net the gross
comp down — so the NOIs are on one footing before you draw a market cap.

Also normalize for **below/above-market in-place rents** (mark toward market so the cap
reflects economic, not contractual, income) and strip **one-time items** from every comp's
NOI. Only after occupancy, lease structure, mark-to-market, reserves, and non-recurring
items are aligned is the spread of comp cap rates meaningful.