Guides/Financial Modelling

Real Estate Financial Modelling: Why Corporate Finance Mechanics Break Here

The Excel mechanics are the same. The economic logic is not — and that is what catches people.

By Surojit Chakraverti — ex-Citi, Rothschild, Morgan Stanley & hedge fundsUpdated August 23, 202613 min read

Analysts moving from corporate finance into real estate rarely fail on the arithmetic. They fail by applying a correct methodology in the wrong context — building a three-statement model around EBITDA multiples for an asset class where neither EBITDA nor multiples are how value is determined. In real estate the asset is the business: it produces cash directly through rent, and its value is that income divided by the market’s required return. Almost everything follows from that one difference.

What actually changes

In a corporate model, value derives from earnings applied at enterprise level, and the assets that generate those earnings sit on the balance sheet without being the direct object of valuation. In real estate, the property is what you are valuing, and its income is measured at property level before any financing.

Three structural consequences. There is no revenue line built from volume and price in the corporate sense — there is net operating income, built from a rent roll. There is no EV/EBITDA multiple applied to normalised earnings — there is asset value derived by dividing stabilised NOI by a market capitalisation rate. And there is no conventional working capital cycle: no inventory, no receivables cycle of the corporate kind, just timing differences in rent collection.

The Excel skills transfer completely. The mental model does not, and trying to force the corporate frame onto the asset is the error that produces confidently wrong numbers.

Net operating income is not EBITDA

NOI is gross potential rent, less vacancy and credit loss, plus other income, less property operating expenses. It excludes debt service, depreciation, capital expenditure and income taxes.

It resembles EBITDA in that both are pre-financing and pre-tax, which is exactly what makes the conflation tempting and wrong. NOI is measured at a single property, excludes capex entirely, and is the direct input into valuation, debt sizing and return analysis simultaneously. If NOI is wrong, every output is wrong, and unlike corporate EBITDA there is no second metric to triangulate against.

That is why NOI must be built bottom-up from the rent roll rather than estimated top-down. The model is only ever as credible as the lease data underneath it.

The cap rate is where the value actually sits

Capitalisation rate is the market’s required return on a stabilised income-producing property: cap rate equals NOI divided by asset value, or rearranged, asset value equals NOI divided by cap rate.

A property generating $5m of stabilised NOI at a 5.0% cap rate is worth $100m. The same property at 6.0% is worth $83.3m. A 100 basis point move on a $100m asset is a $16.7m swing — which makes cap rate sensitivity, not revenue growth, the most material analysis in any real estate model.

Critically, the cap rate is an input, not an output. It is derived from comparable transaction evidence, and the analyst’s job is to defend the assumption rather than to calculate it. A model that presents a cap-rate-derived value without a sensitivity grid across cap rate has hidden its own largest uncertainty.

Debt is sized on coverage, not on a multiple

Corporate analysts size debt as a multiple of EBITDA. Real estate lenders apply three constraints at once and lend to the tightest.

Loan-to-value caps the loan as a percentage of appraised value, commonly 65% to 75% for stabilised commercial property. Debt service coverage requires NOI to exceed annual debt service by a margin, typically 1.20x to 1.35x. Debt yield — NOI divided by the loan amount, commonly a 7% to 9% floor — is the lender’s appraisal-independent check: if values fall and cap rates expand, a low debt yield flags an over-levered loan regardless of what the valuation says.

The sequence is therefore inverted from a corporate LBO. NOI determines debt capacity, debt capacity determines the equity cheque, and the equity cheque determines the required return. You cannot start from a target leverage ratio.

  • Maximum LTV against appraised value.
  • Minimum DSCR: NOI divided by annual debt service.
  • Minimum debt yield: NOI divided by loan amount, independent of appraisal.
  • The loan is the minimum of the three. Model all three explicitly.

The rent roll and the lumpiness of everything

The revenue build is lease by lease: each tenant, current rent, lease expiry, renewal probability, and market rent at expiry. For a multi-tenant office asset that runs to dozens of lines; for a single-tenant net lease it may be three.

This lease-level structure is what makes vacancy a dynamic rather than a static assumption. Applying a flat 5% vacancy across ten years ignores that expiries create lumpy vacancy events, and that re-letting carries downtime, rent-free periods and tenant improvement allowances that hit cash flow in the quarters right after an expiry. A model that smooths this is understating volatility.

Capital expenditure is lumpy for the same reason. It is tied to lease events and physical condition rather than being a percentage of revenue, and splitting maintenance capex from value-add capex is essential because only the second is discretionary.

The equity waterfall — and the provision everyone forgets

Cash flow after debt service is split between the sponsor (general partner) and the investors (limited partners) through a waterfall: a preferred return to the LP, return of capital, then a promoted interest split to the GP above a hurdle IRR.

The step first-time models omit is the GP catch-up. A waterfall that pays the LP its preferred return and then splits everything 80/20 without a catch-up materially understates the sponsor’s promote, and therefore misstates the return distribution in every scenario above the hurdle. If your waterfall has no catch-up row, it is almost certainly wrong.

Model it period by period across the hold, not as a single terminal calculation, because the timing of distributions drives the IRR on both sides. And read the actual legal structure before building — promote percentages, multiple hurdles, clawback provisions and whether the preferred return compounds all change the arithmetic.

REITs: FFO, AFFO and NAV

At the public market level, GAAP earnings are close to meaningless for property companies because real estate depreciation is a large non-cash charge against assets that frequently hold or gain value.

Funds from operations adds real estate depreciation and amortisation back to net income and removes gains on property sales. Adjusted FFO goes further, deducting maintenance capex and straight-line rent adjustments, and is the closest proxy for distributable cash. REITs are valued on price to FFO and price to AFFO — applying an EV/EBITDA multiple to a REIT will systematically misvalue it.

Net asset value is the intrinsic framework: capitalise each property’s NOI at an appropriate cap rate, sum, add non-property assets, deduct debt, divide by shares. It is also a capital allocation lens — a REIT trading above NAV has cheap equity and can make accretive acquisitions; one trading below NAV may create more value buying back its own stock than buying assets.

Frequently asked questions

What is the difference between NOI and EBITDA?

NOI is a property-level figure — gross rent less vacancy and credit loss, plus other income, less property operating expenses — and it excludes debt service, depreciation, capital expenditure and taxes entirely. EBITDA is an enterprise-level measure that includes items NOI excludes. NOI feeds directly into valuation via the cap rate, debt sizing via DSCR, and returns, so an error in it propagates everywhere with no second metric to catch it.

How is debt sized in a real estate model?

By satisfying three constraints simultaneously and taking the tightest: maximum loan-to-value (commonly 65-75% for stabilised commercial property), minimum debt service coverage (typically 1.20-1.35x), and minimum debt yield, being NOI divided by the loan amount (commonly a 7-9% floor). Unlike a corporate LBO, you cannot start from a target leverage multiple — NOI determines capacity.

How does a cap rate determine property value?

Asset value equals stabilised NOI divided by the cap rate. A property with $5m of NOI at a 5.0% cap rate is worth $100m; at 6.0% it is worth $83.3m. The cap rate is a market-derived input taken from comparable transactions, not something the model calculates, and because a 100 basis point move swings value by roughly 17%, cap rate sensitivity is the most material analysis in the model.

What is a GP catch-up in a real estate equity waterfall?

The provision that pays the general partner until it has received its full agreed share of profits, after the limited partner has received its preferred return and capital back. Omitting it is the most common structural error in first-time waterfall models: without a catch-up the sponsor’s promote is materially understated and the return split is wrong in every scenario above the hurdle.

Why are REITs valued on FFO rather than earnings?

Because GAAP depreciation on real estate is a large non-cash charge against assets that often hold or increase in value, making net income a poor measure of a property company’s performance. FFO adds that depreciation back and removes property sale gains; AFFO further deducts maintenance capex and straight-line rent adjustments, giving the closest proxy for distributable cash.

Related guides

Get the foundation right first.

Every specialist model sits on three-statement mechanics. Become a Financial Modelling Pro is a 2-day sprint on exactly that — architecture, driver-based builds, working capital, the roll-forward, statement linking and error-proofing.