Every institutional acquisition eventually passes through a DCF. Cap rates do the screening, comps do the sanity check, and pro formas do the narrative, but when a committee needs to defend paying $150M for an asset, someone is building a 10-year model with a terminal reversion. The DCF is where assumptions stop being narrative and start being numbers — and where value-add deals either make sense or don’t.
Most CRE DCFs are decent. A minority are excellent. A worrying number are wrong in ways that don’t show up until year three, when the model’s Year 1 rent growth of 4% turns out to be 1.2% and the terminal cap rate that looked conservative at 6.5% now looks aggressive at 5.75%. This guide is about building the kind that holds up — the kind that survives diligence, lender review, and the IC memo revisions.
Why DCF in Real Estate Matters
A direct capitalization valuation divides NOI by a cap rate. It’s instant and works well when the cash flow is stable. But most CRE deals aren’t stable. They’re:
- Value-add acquisitions where rents are below market and NOI will grow materially in years 1–3
- Lease-up plays where occupancy climbs from 70% to 95% over 18–24 months
- Repositioning deals where capital is spent in years 1–2 before the revenue thesis materializes
- Development where year 1 is negative cash flow and the first stabilized NOI shows up in year 3
For any of these, a direct-cap valuation using Year 1 NOI understates value — and using stabilized NOI overstates it, because the buyer has to fund the gap between in-place and stabilized over time.
Discounted cash flow in real estate handles this explicitly. It models each year’s cash flow based on the actual trajectory: lease rollovers, rent growth, capital spend, operating expense escalation, and a specific exit assumption. Then it applies a discount rate that reflects the risk of those cash flows actually showing up.
The result is a valuation that captures the full investment thesis, not just Year 1. That’s why institutional acquisitions teams lead with DCF on anything that isn’t already stabilized.
The 5 Inputs That Drive the Entire Model
A 10-year CRE DCF has hundreds of cells, but five inputs drive roughly 80% of the output. Get these right and the model is defensible. Get any one of them materially wrong and the DCF becomes a sophisticated way to reach the answer you wanted.
1. Market Rent and Rent Growth
The single most important assumption. Market rent drives rollover renewals, new leases, and — at the terminal — the stabilized NOI that sets reversion value. Small changes compound: a 50 bps difference in annual rent growth applied over 10 years produces a 5% swing in terminal NOI, which translates to 3–4% on the DCF value.
How to get it right: base market rent on recent leasing, not asking rates. Use comps from the last 18 months with adjustments for size, floor, amenities, and lease terms. For growth, reconcile to submarket historical averages and employment projections, not national REIT consensus.
2. Discount Rate
The discount rate is the investor’s required return — the hurdle rate the model has to clear. For unlevered DCFs on stabilized CRE, it typically runs 7%–11%. For a levered DCF discounting equity cash flows, it’s the required return on equity (often 12%–20%).
How to get it right: anchor to a risk-free rate (the 10-year Treasury) and layer on an asset-class premium (100–400 bps) and a property-specific risk adjustment (+/- 100 bps for location, tenant credit, deferred maintenance, and execution risk). Don’t pick a round number because it feels right — document the buildup.
3. Terminal (Exit) Cap Rate
The terminal cap rate is applied to Year 11 NOI to calculate reversion value. On a 10-year hold, the reversion commonly drives 50%+ of the total DCF value. A 25 bps swing in terminal cap rate changes DCF value by 4–5%.
How to get it right: the conservative default is the going-in cap rate plus 25–50 bps to reflect the property being 10 years older at exit. For value-add deals where the property will be materially better at exit (newly renovated, fully leased, longer WALT), a flat or tighter terminal cap can be defended — but the IC memo should explain why.
4. Operating Expense Growth
Expense growth historically tracks roughly 2.5%–3.5% annually, but in inflation-sensitive categories (insurance, utilities, payroll for hotels/seniors housing) it can run 5%+. Expenses compounding faster than revenue is what turns a 4% rent-growth pro forma into a 2% NOI-growth reality.
How to get it right: break expenses into controllable (payroll, marketing, administrative) and non-controllable (insurance, taxes, utilities) categories and grow each appropriately. Insurance deserves special scrutiny — in hurricane, wildfire, and earthquake zones, insurance has grown 15%+ annually and is materially under-reserved in most pro formas.
5. Capital Expenditure Schedule
CapEx eats DCF value directly — every dollar spent on roof replacement, HVAC, or tenant improvements is a dollar subtracted from cash flow. For older assets, CapEx can easily exceed 15% of cumulative NOI over a 10-year hold.
How to get it right: pull the property condition assessment and schedule specific capital items by year, not a generic ”$/SF” assumption. A roof replacement in year 4 lands differently in the DCF than the same replacement amortized across 10 years.
Building a 10-Year Cash Flow from Rent Roll and T-12
A proper CRE DCF starts with two documents: the current rent roll and the trailing-twelve financial statements (T-12). Everything else builds from these.
Step 1: Normalize the rent roll. Consolidate amendments, reconcile stacked step rents, and produce a clean schedule of each tenant’s current rent, lease end date, renewal options, and recovery treatment (NNN, gross, modified gross). Every row should tell you: who the tenant is, what they pay today, what they pay at each scheduled step, when their lease ends, and what they’re likely to do at rollover.
Step 2: Project rollover. For each tenant, model the renewal scenario (renewal probability, new rent at market, free rent, TI allowance) and the non-renewal scenario (downtime in months, commission, TI, lease-up to market). Apply a probability weighting — for office, 60–75% renewal probability is common; for multifamily, 50–60% turnover; for industrial, 70–85% renewal depending on tenant size and credit.
Step 3: Layer in market rent growth. Apply annual rent growth to market rates (not to in-place contracts, which grow by their contractual terms). New leases post-rollover come in at the then-current market rent.
Step 4: Build the expense side. Start with T-12 operating expenses, segment by category, and grow each at a category-appropriate inflation rate. Reconcile to the rent roll’s expense recoveries — for NNN leases, recoverable expenses flow back as CAM income.
Step 5: Subtract capital expenditures. Pull the capital plan from the property condition assessment and schedule items by year. For tenant improvements and leasing commissions, tie them to the rollover schedule from Step 2.
Step 6: Calculate the reversion. Year 11 NOI divided by the terminal cap rate, minus selling costs (typically 1.5%–2.5% for broker commission, legal, and transfer taxes). Add net sale proceeds to Year 10 cash flow.
Step 7: Discount everything to present value. Apply the discount rate to each year’s cash flow and the reversion. The sum is the unlevered property value.
For levered DCFs, insert a debt schedule (loan proceeds in Year 0, debt service through Year 10, loan payoff at exit) and discount the resulting equity cash flows at the required equity return.
The Three Leverage Points in a CRE DCF
Every DCF has three numbers that dominate the output. Stress-testing these is more valuable than stress-testing anything else.
Discount rate — moves the entire valuation. A 50 bps change in discount rate produces a 4–6% swing in DCF value on a typical 10-year hold. Sensitivity tables should always include discount rate on one axis.
Terminal cap rate — drives reversion, which drives 50%+ of total value. A 25 bps change in terminal cap rate produces a 4–5% swing. This is where most pro forma manipulation hides: analysts inherit a slightly optimistic terminal cap from a template and don’t question it.
Market rent growth — compounds through the 10-year projection. A 100 bps change in annual rent growth produces an 8–10% swing in DCF value on value-add deals. For stabilized assets with long WALTs, the effect is muted because only a fraction of the rent roll turns during the hold.
The single most useful output from any DCF is a sensitivity matrix of these three variables. If the model’s value swings by 30% across plausible ranges, the precision of the model is false — the uncertainty is real.
DCF Pitfalls (Where Most Models Quietly Fail)
Discounted cash flow models fail in predictable ways. Catching these during diligence is worth more than any amount of formula sophistication.
Stale rent roll. The model is only as good as the rent roll. A rent roll from 8 months ago, missing two amendments and one expansion option, produces a model that’s confidently wrong. Every rent roll should be reconciled against the lease abstracts during diligence — amendment by amendment, option by option.
Wrong reversion cap rate. Pro forma models frequently use going-in cap rates flat to terminal, or (worse) a terminal cap tighter than going-in based on speculative market improvement. Default to terminal = going-in + 25 bps unless there’s a specific reason to argue otherwise. If the model uses a tighter terminal, the IC memo should explain why and the sensitivity analysis should show the downside.
Optimism bias in rent growth. Every pro forma ever produced has optimistic rent growth. Tie market rent growth to a specific, defensible source: submarket historical data, employment projections, or a comparable rent-growth benchmark. Discount it 50–100 bps in a downside case.
Missing capital expenditures. The two most commonly under-modeled items are roof replacement and HVAC systems. Both have typical useful lives of 15–25 years, both cost $10–$40/SF depending on asset, and both tend to land in the middle of the hold period where they haircut NOI materially. If the property condition assessment flags either as due within the hold period, the CapEx line needs to reflect it.
Expense inflation understated. Applying 2.5% expense growth to all categories — including insurance, utilities, and payroll — in a 5%+ inflation environment understates expenses materially. Segment and grow appropriately; check against recent T-12 trends, not 10-year averages.
Interest rate assumptions baked into DCF without mention. For levered DCFs, the exit refinance or loan payoff assumptions can quietly determine the outcome. If the model assumes a 6% refinance rate in year 5 but market rates by year 5 are 7.5%, the equity return collapses. Make the loan assumptions explicit and stress them.
Where AI Accelerates the Model
DCF construction has historically been hours of work per property — reconciling rent rolls, building rollover schedules, pulling expense benchmarks, and recalculating when assumptions change. A meaningful portion of this is now automated.
Rent roll normalization — AI-powered lease abstraction extracts and normalizes rent schedules, stacked step rents, and amendments in minutes rather than hours. For portfolios, the time savings run into multiple weeks.
Lease rollover modeling — renewal probabilities, market rent assumptions, and TI/LC schedules can be pre-populated based on asset class, market, and tenant credit, with the modeler overriding where specific diligence justifies.
Expense benchmarking — T-12 expenses can be automatically segmented and compared against market benchmarks, flagging categories that look materially out of line (insurance under-reserved, payroll above market, utilities inconsistent with square footage).
Sensitivity analysis — instead of building three-variable sensitivity tables manually, the model can produce them on demand with any set of variables the modeler specifies.
What AI doesn’t do is make the judgment calls. Whether the terminal cap should be 6.25% or 6.75%, whether rent growth should be 3% or 4%, whether the deferred maintenance warrants a 200 bps risk premium in the discount rate — these remain the analyst’s call, informed by diligence and market context. AI makes the mechanics faster; the thesis is still human.
The Output That Matters
A CRE DCF exists to answer one question: At what price does this deal produce the required return? Everything else — the rollover schedule, the expense buildup, the capital plan — is in service of that number.
The best DCFs produce three outputs, not one:
- Base case valuation — the number in the IC memo
- Sensitivity matrix — how the valuation moves across plausible ranges of discount rate, terminal cap, and rent growth
- Downside case — the valuation under conservative assumptions, showing what the deal is worth if the thesis only half-delivers
Teams that show all three in their underwriting pass IC reviews on the first pass. Teams that show only the base case get sent back for more work. The market rewards the ones who build models that hold up under stress — not the ones who build the tightest base cases.
Discounted cash flow in real estate is, in the end, less about the math than about the discipline of making every assumption explicit and defensible. The mechanical part — 10 years, a reversion, a discount rate — is straightforward. The hard part is admitting where the numbers are uncertain and showing the range of outcomes honestly. Deals that die because the DCF was optimistic cost far more than deals that died because the DCF was conservative.