Terminal Value and Reversion: Exit Cap, Hold Period, and the Year 11 Forward NOI Convention

Key Takeaways

Why Reversion Drives Returns

On a 5-year institutional multifamily hold, the terminal value (the projected sale proceeds at exit, also called the reversion) is the dominant component of present value. In the worked example below, the reversion is75% of total present value. The current cash flow stream from operations, the part most underwriters spend the most time modeling, contributes the remaining 25%. The exit assumption isn't a tail input; it's the dominant driver of the deal's IRR.

This ratio is the first thing a junior analyst tends to under-appreciate and the first thing a senior investor tests when sanity-checking an underwrite. Hold a 5.85% cap deal for five years at 3% NOI growth and sell at a 6.00% exit, and the reversion line is doing about three-quarters of the work. Move the exit cap 50 bps wider and the 5-year unlevered IRR falls by roughly 140 bps (about 250 bps at a 3-year hold, about 60 bps at 10 years). Lower the NOI growth assumption by 50 bps and the IRR moves by about 50 bps. At a 5-year hold the exit cap carries roughly three times the leverage of the growth input, and the institutional discipline reflects that ordering: the exit cap, hold period, and disposition cost assumptions get the most attention at IC because they have the most leverage on the answer.

The terminal value question comes up in two settings. Corporate-finance analysts pricing public companies or private equity buyouts use a perpetuity growth or exit multiple method. The CRE practitioner asks the same question but needs a different answer: CRE doesn't use perpetuity growth; CRE uses the exit cap method. The two are mathematically equivalent, but the convention, the inputs, and the institutional discipline are different. This article covers both, with the bridge that makes them the same math, and the CRE-specific details that equity-finance treatments don't cover: the forward NOI convention, the jurisdictional sale-cost detail, and the hold-period sensitivity behind the typical 5–7 year hold.

The Year 11 Forward NOI Convention

The single most overlooked detail in CRE terminal value, and one that equity-finance terminal value treatments rarely address because it's specific to the way real estate transacts, is which year's NOI gets capitalized at the exit cap. The institutional convention is forward NOI, the NOI of the year after the sale: Year 11 NOI for a 10-year hold (and Year 6 for a 5-year hold), not Year 10 trailing. The difference is one full year of NOI growth, and it propagates straight into terminal value at the exit cap rate.

The economic reasoning is structural. When an institutional buyer underwrites a property for purchase at the end of Year 10, they are acquiring the right to next year's income going forward. They run their own pro forma starting from Year 11 NOI (their Year 1 of operation is the seller's Year 11). The price they pay is determined by their underwriting of that forward income stream, which they then capitalize at the going-in cap rate appropriate for the asset at the time of sale. From the seller's side, that means the relevant NOI for computing reversion proceeds is Year 11 forward, not Year 10 trailing, because Year 10 trailing is the income the seller already collected, not the income the buyer is acquiring.

State it cleanly for a 10-year hold: Terminal Value = NOI_{Year 11} ÷ Exit Cap − Disposition Costs. The Year 11 NOI is grown one year out from the Year 10 stabilized NOI, typically at the same long-run growth rate used in the cash flow projection. If your model is projecting 3% annual NOI growth and your Year 10 NOI is $1,800,000, your Year 11 forward NOI is $1,854,000, the number that capitalizes at the exit cap. Using $1,800,000 (Year 10 trailing) instead would understate gross sale proceeds by 2.9% (1 − 1 ÷ 1.03). In the 5-year worked example below, that is a net reversion shortfall of about $985,000 and an unlevered IRR understatement of about 53 bps (about 90 bps at a 3-year hold and about 20 bps at 10 years). That is not a rounding error.

Models often default to Year 10 trailing because it's the last cash flow line in the projection table and the modeler reaches for the nearest number. The institutional discipline is to extend the projection one year further, either by computing Year 11 explicitly in a hidden helper cell, or by multiplying Year 10 NOI by (1 + g) in the terminal value formula directly. The audit trail should show which approach was used, because the choice changes the answer by a full year of growth.

The Exit Cap Choice

The exit cap is the single most consequential input in the terminal value formula and the input that gets debated longest at IC. The institutional discipline is to build the exit cap from the same Gordon Growth components as the going-in cap, with an explicit forecast of where each component sits at the exit date. The dedicated treatment of this build is in the sibling article ongoing-in vs. exit cap and the spread mechanics; the short version follows.

As illustrative modeling assumptions (not survey data; replace them with a current broker or research survey for the market and asset in question), exit cap conventions for institutional core sit roughly:

• Multifamily core: 5.50–6.25%, typically priced 15–50 bps wide of going-in.
• Retail STNL (investment-grade): 6.80% area, priced flat to 25 bps wide of going-in.
• Industrial logistics core: 6.50% area, priced flat to 25 bps wide of going-in.
• Office (Class A trophy): 7.50–9.00%, priced 50–150 bps wide of going-in given structural risk.

The conventional practitioner spread, "use going-in + 25 bps as exit cap", is a heuristic that happens to work for stabilized multifamily and industrial in flat-curve environments. It does not work when the 10-year Treasury is moving against the underwrite (as in 2022–2024), when sector growth views are shifting (suburban office post-2020, anchored retail mid-2010s), or when sponsor and structural premiums are in flux. The discipline is to build the exit cap from forward components: forward 10-year (the spot today, plus or minus an explicit forecast), forward risk premium (typically the same as going-in unless the asset-class risk profile is expected to shift), and forward growth view (typically the same as going-in for stabilized assets, materially different for assets in cycle transitions). Market references such as the 10-year Treasury series on FRED and Green Street's Commercial Property Price Index help anchor the direction of the forward components.

The structural-conservatism convention, adding 25–50 bps to the build-up implied exit cap as a buffer, is a common practice. The buffer protects against the forecast uncertainty in the build-up itself, not against any specific component. It is the line that most often distinguishes a defensible underwrite from a price-taker. (The worked example below uses a narrower 15 bp spread to the going-in cap as a simplifying assumption.)

Disposition Costs: The Sale-Cost Haircut

Disposition costs are the all-in transaction costs the seller pays at closing. In a mid-market jurisdiction they are commonly modeled at 1.75–3.5% of gross sale price (illustrative assumptions, not published schedules), decomposed into three line items: brokerage, transfer taxes, and legal/escrow. A model that skips the line treats the gross sale price (forward NOI ÷ exit cap) as the reversion proceeds. The institutional discipline is to net the costs out explicitly, because they are not optional and they are not small.

Brokerage: 1.0–1.5%

Sales brokerage commissions are negotiated deal by deal and are commonly modeled at 1.0–1.5% of gross sale price, paid by the seller. For modeling purposes, 1.25% is a reasonable midpoint. These are illustrative modeling assumptions; replace them with the actual engagement terms for a live sale.

Transfer taxes: 0.5–1.5% in a mid-market jurisdiction

Transfer taxes are state and local imposts on the recording of real property transactions. The rate varies materially by jurisdiction. The table below lists rates from the statutes and agency pages cited in Sources; 0.5–1.5% is a mid-market planning range, with some jurisdictions at zero and others far higher.

Table 1: Selected transfer tax rates by jurisdiction, from the statutes and agency pages listed in Sources. Rates change periodically and counties and cities impose additional levies, so check the current tax codes for any specific transaction. The "who pays" allocation can be negotiated in the purchase agreement; only the seller's share belongs in the disposition-cost line.

The spread between jurisdictions is wide. With the midpoints used in this article, the all-in disposition cost is about 1.25 + 0 + 0.30 = 1.55% where no transfer tax applies, versus about 1.25 + 3.275 + 0.30 = 4.83% for a New York City sale of non-residential property above $2M. For multi-market portfolios, modeling the jurisdiction-specific transfer tax in each line item rather than applying a flat assumption produces materially different reversion proceeds across the portfolio.

Legal and escrow: 0.25–0.50%

Legal fees, title insurance, escrow, and recording costs are commonly modeled together at 0.25–0.50% of gross sale price (an illustrative assumption). Legal fees scale with deal complexity, and who pays for title insurance varies by state and by contract. For modeling purposes, 0.30% all-in is a reasonable midpoint.

The total

Pulling the three lines together, a sale in Florida (0.70% transfer tax) at a 1.25% brokerage produces an all-in disposition cost of roughly 1.25 + 0.70 + 0.30 = 2.25%. The same transaction in New York City would be roughly 1.25 + 3.275 + 0.30 = 4.83%, and in Washington, DC about 1.25 + 1.45 + 0.30 = 3.00% on the seller's transfer tax alone (4.45% if the seller also bears the 1.45% recordation tax). The institutional convention is to model the actual jurisdiction-specific number rather than apply a flat 2.5% or 3% assumption, because the difference between 2.25% and 4.83% on a $30M sale is about $772,500: material at the deal level, and worth about 50 bps of 5-year unlevered IRR in the worked example below.

Hold Period Sensitivity

The hold period choice is an underwriting decision, not a default. Different strategies tend to use different holds: value-add 3–5 years (renovate, lease up, exit at stabilized cap); core 5–10 years (clip coupons through the cycle); opportunistic 3–5 years (develop, lease up, exit). These are typical ranges rather than rules. A 5–7 year hold for core/core-plus is a common balance point between two pressures: short-hold disposition cost drag and long-hold exit-cap forecast risk.

Run the same hypothetical 100-unit multifamily property (5.85% going-in cap, $1,800K Year-1 NOI, 3% NOI growth, 6.00% exit cap, 2.75% disposition costs, 8.5% unlevered discount, $30.77M purchase price) at four different hold periods:

Table 2: Hold-period sensitivity for the 100-unit MF core deal. NOI in year t is $1,800K × 1.03^t−1 with end-of-year cash flows; the sale at the end of year N is priced at year N+1 NOI ÷ 6.00%, less 2.75%, and discounted at 8.5%. The IRR is measured on the $30.77M purchase price. The cash flow PV rises with longer holds (more years of operating income); the reversion PV falls because it is discounted over more years, even though the sale price grows. The reversion-as-percent-of-total drops from 84% at 3 years to 57% at 10 years. The 5–7 year band is a common core convention.

Three things to read out of this grid. First, the unlevered IRR rises with the hold, from 7.1% at 3 years to 8.4% at 10 years, a spread of about 130 bps. Two effects drive it in roughly equal parts: the 2.75% disposition cost, which is spread over more years of operating return as the hold lengthens, and the 15 bp widening from the 5.85% going-in cap to the 6.00% exit cap, whose annualized drag also fades with a longer hold. (With no sale costs and an exit cap equal to the going-in cap, the IRR in this model would be 8.85%, the going-in cap plus growth, at every hold.) All four IRRs sit below the 8.5% discount rate, so each hold's total PV is below the $30.77M purchase price. Second, the reversion's share of total PV moves from 84% to 57% as the hold lengthens, so the exit cap assumption has materially different leverage on the IRR at different hold periods. Third, the difference in total PV between a 3-year and 10-year hold is only about $0.9M (3.0% of the 10-year total), but the cash flow vs reversion mix shifts dramatically. A sponsor that is more confident in its operating projections than in its exit cap forecast may prefer the longer hold, which produces a more cash-flow-dependent return profile.

The typical 5–7 year hold reflects a balance: long enough that the disposition cost drag is amortized to a manageable share, short enough that the exit cap forecast is still anchored to the current market environment rather than to a speculative view of a full cycle out. A 5-year hold lets the underwriter use the current 10-year Treasury and risk premium for the exit-cap build with modest forward adjustment; a 10-year hold forces an explicit view on where the 10-year Treasury and risk premium will be in 2036. That view is rarely defensible at IC with the same confidence as the 5-year version.

The Equity-Finance Bridge

Equity-finance practitioners approaching terminal value from a corporate-finance background expect to see one of two formulas: the perpetuity growth method (Gordon Growth applied at the terminal year) or the exit multiple method (EV/EBITDA times terminal year EBITDA). Both are mathematically equivalent to the CRE exit cap method, and the conversion is mechanical.

Perpetuity growth method = CRE exit cap method

The perpetuity growth method computes terminal value as TV = FCF_N+1 ÷ (r − g), where r is the discount rate and g is the long-run growth rate. Apply this to CRE: FCF_N+1 is forward NOI (Year 11 NOI for a 10-year hold), r is the unlevered required return, and g is the long-run NOI growth term. The denominator (r − g) is, by the Gordon Growth identity, the cap rate when the discount rate and growth rate are set consistently with the market cap rate. So the equity-finance formula reduces to:

TV = NOI_{Year 11} ÷ cap rate = NOI_{Year 11} ÷ exit cap

They are the same math. CRE uses the exit cap because cap rates are directly observable in transaction markets: brokers quote them, transaction databases such as MSCI RCA track the pricing behind them, and comparable sales fix them. Equity finance uses (r − g) because public-company discount rates and growth rates are inferred separately from CAPM construction and analyst growth forecasts. The CRE convention is more direct because the market data is more direct. In a working model the three inputs are usually set separately. In the worked example below, the 8.5% discount rate is an assumed hurdle and the 6.00% exit cap is a market assumption, so the identity implies a perpetual growth rate of 8.5% − 6.00% = 2.5%, below the 3% explicit-period NOI growth. Benchmarks such as the NCREIF Property Index record property-level returns historically and are a common cross-check on those assumptions.

Exit multiple method = inverse cap rate

The exit multiple method computes terminal value as TV = EBITDA_N × Multiple, where the multiple is an observed market multiple (e.g., 12x EBITDA). For CRE, the analogous construction isTV = NOI × (1 ÷ cap rate), which is the same as NOI ÷ cap rate rearranged. The "multiple" in CRE is the reciprocal of the cap rate: a 6.00% cap is a 16.7x NOI multiple, a 7.00% cap is a 14.3x NOI multiple. Practitioners occasionally quote the multiple instead of the cap rate, but the convention in institutional CRE remains the cap rate.

Discounting the terminal value

Whichever method produces the terminal value, the institutional discipline is to discount it back to present at the unlevered required return. PV of reversion = Net reversion ÷ (1 + r)^N, where N is the hold period. At an 8.5% unlevered discount and a 5-year hold, the discount factor is 1 ÷ (1.085)⁵ = 0.665, so the terminal value gets cut by one-third on the way back to present. At a 10-year hold, the discount factor is 0.442, so the terminal value is cut by more than half. That math is what produces the reversion-as-percent-of-total ratio shifting from 75% at 5 years to 57% at 10 years in the table above.

Worked Example: 100-Unit Multifamily Core, 5-Year Hold

A hypothetical stabilized 100-unit multifamily property, priced for institutional sale at a 5.85% going-in cap. All inputs below are illustrative assumptions. The acquisition team's underwrite:

• Purchase price: $30,769,000 (implied by $1,800K Year-1 NOI ÷ 5.85%)
• Year-1 NOI: $1,800,000 (institutional method, 5% vacancy, 45% expense ratio)
• Long-run NOI growth: 3.0% annual
• Hold period: 5 years
• Exit cap: 6.00% (15 bps wide of going-in; a simplifying assumption, narrower than the 25–50 bp buffer discussed above)
• Disposition costs: 2.75% (1.25% brokerage + 1.20% transfer tax + 0.30% legal/escrow, a blended illustrative assumption)
• Unlevered required return: 8.50% (an assumed unlevered hurdle rate, not derived from the Gordon identity; with the 6.00% exit cap it implies 2.5% perpetual growth. Seecap rate decompositionfor the cap rate build-up.)

For a 10-year hold, project Year 10 NOI: $1,800,000 × (1.03)⁹ = $1,800,000 × 1.30477 =$2,348,592. (Note: nine years of compounding because Year 1 is the first projected year and Year 10 is nine years out.) Project Year 11 forward NOI: $2,348,592 × 1.03 = $2,419,049.

The rest of this example uses the 5-year hold. The full 5-year math runs:

Table 3: The 5-year hold worked example. Year 6 forward NOI ($2,086,693) is the institutional input to the exit cap, not Year 5 ending. (For a 5-year hold, the "Year N+1 forward" convention means Year 6 forward, not Year 11; the "Year 11 forward" terminology is shorthand for the most common 10-year hold case.)

Sum the cash flow PVs over the 5-year operating period (NOIs of $1,800,000, $1,854,000, $1,909,620, $1,966,909, and $2,025,916, each discounted at 8.5% from year end): $7,495,530. Total deal PV: $22,493,047 + $7,495,530 = $29,988,577, about $780K (2.5%) below the $30,769K purchase price. The shortfall is expected: at the purchase price the deal's unlevered IRR is 7.88%, below the 8.5% hurdle, because the 15 bp exit-cap widening and the 2.75% sale costs pull the IRR down from 8.85% (the going-in cap plus growth).

The reversion's contribution to total deal PV: $22,493 / $29,989 = 75%. The reversion is doing three-quarters of the work in this stabilized core underwrite. A 50-bp widening of the exit cap to 6.50% would reduce gross sale price to $32,103K, net reversion to $31,220K, PV of reversion to $20,763K, and total PV to $28,258K, taking the unlevered IRR from 7.88% to 6.45%, a drag of roughly 140 bps on the same operating projection. That sensitivity is what makes the exit cap the most-debated input at IC.

How to Model It

The terminal value calculation fits on a single tab of any DCF model. Eight rows, in this order, each with an input cell and a source comment:

• 1. Year N ending NOI. The last year of the operating projection (Year 5 for a 5-year hold, Year 10 for a 10-year hold). Pull from the cash flow projection, not a separate input.

• 2. Year N+1 forward NOI. Year N NOI multiplied by (1 + long-run NOI growth rate). For a 10-year hold this is Year 11; for a 5-year hold this is Year 6. The naming convention defaults to "Year 11 forward" in institutional language because the 10-year hold is the canonical case, but the math is the same at any hold period.

• 3. Exit cap rate. Build from forward components, ideally using the Gordon Growth framework with a 25–50 bp structural conservatism buffer. Illustrative ranges: multifamily core 5.50–6.25%; retail STNL 6.50–7.00%; industrial 6.25–6.75%; office Class A 7.50–9.00%. Replace them with a current survey for the market in question.

• 4. Gross sale price. Year N+1 forward NOI ÷ exit cap. This is the gross top-line sale number; it is not the reversion.

• 5. Disposition costs. Sum three sub-lines: brokerage (1.0–1.5%), transfer tax (jurisdiction-specific, see Table 1), legal/escrow (0.25–0.50%). All as a percentage of gross sale price.

• 6. Net reversion. Gross sale price minus total disposition costs.

• 7. Discount factor. 1 ÷ (1 + unlevered required return)^{hold period}. For a 5-year hold at 8.5%, the factor is 0.665. For a 10-year hold at 8.5%, the factor is 0.442.

• 8. Present value of reversion. Net reversion × discount factor. This is the input to the deal-level NPV and the headline reversion contribution to total return.

Two audit checks. First, compare the reversion as a percent of total PV against a reference deal with similar yield and growth at the same hold period. For the 5.85% cap, 3% growth example, the ratio runs 84%, 75%, 67%, and 57% at 3, 5, 7, and 10 years. A ratio far from the reference means either the exit cap is mispriced or the operating projection is unusually strong or weak relative to the cap. Second, sensitize the exit cap by ±50 bps and observe the IRR drift. For a deal like the worked example, 50 bps of exit cap moves the unlevered IRR by about 140 bps at a 5-year hold and about 100 bps at 7 years (about 250 bps at 3 years and 60 bps at 10). A move far outside the range for the hold period points to an inconsistent terminal calculation or an IRR that is overexposed to a single exit-day assumption.

Common Mistakes

Seven errors that consistently distort terminal value calculations in practice, with IRR effects measured on the 5-year worked example where a figure is given:

• Using Year N trailing NOI instead of forward NOI. A common error that produces a terminal value understatement of one full year of growth (2.9% at 3% growth). In the 5-year worked example that is about $985K of net reversion and roughly 53 bps of unlevered IRR (about 90 bps at a 3-year hold and about 20 bps at 10 years). The institutional convention is forward NOI because the buyer at exit is acquiring next-year income, not income already collected.

• Forgetting disposition costs entirely. The disposition cost haircut on gross sale price is not optional. Models that compute terminal value as forward NOI ÷ exit cap and treat that as the reversion overstate proceeds by the full disposition cost percentage. On a $30M sale at 1.75–3.5%, that is $525K–$1.05M of phantom equity proceeds; at the 2.75% used in the worked example it overstates the 5-year unlevered IRR by about 50 bps.

• Applying a flat disposition cost percentage across jurisdictions. Using 2.5% as a portfolio-wide assumption when individual properties sit in New York City (4.83% all-in), DC (3.00% on the seller's transfer tax alone), a jurisdiction with no transfer tax (1.55% all-in), and Florida (2.25% all-in) understates costs by roughly 50–230 bps for the high-tax jurisdictions and overstates them for the low-tax ones. For multi-market portfolios, model the jurisdiction-specific number.

• Using the going-in cap as the exit cap by default. The "going-in plus 25 bps" heuristic works in flat-curve, stable-sector environments and produces material errors in any other context. The exit cap should be built from forward components, with structural conservatism added explicitly. See thegoing-in vs. exit cap sibling for the full spread mechanics.

• Treating hold period as a default rather than an underwriting choice. A 5-year hold, 7-year hold, and 10-year hold produce materially different reversion weights, exit-cap forecast risk, and cash flow vs reversion mix. The hold should be chosen to match the strategy (typically core 5–10, value-add 3–5, opportunistic 3–5) and the sponsor's confidence in operating vs exit-cap forecasts, not defaulted to 10 years because the cash flow projection has 10 columns.

• Failing to discount the reversion at the unlevered required return. The reversion sits at year N and has to come back to present. Some models discount the operating NOIs to present at the unlevered cost of capital but treat the reversion as a nominal year-N number without discounting, producing a wildly inflated NPV. The discount step is not optional; it is the bridge between the Year-N terminal value and the deal-level decision.

• Conflating perpetuity growth and exit cap inputs. Practitioners crossing from equity finance sometimes import a 2.5% perpetuity growth assumption into a CRE model and apply it to the cap rate computation, double-counting the growth already embedded in the exit cap. The CRE exit cap is the (r − g) of the Gordon Growth Model, so the growth term is already in the cap rate. Adding a separate growth subtraction on top understates the exit cap by the full growth assumption.

Related Articles

Terminal value spokes off the cap rate pillar and connects to every other piece of the valuation cluster:

• Cap Rate Calculator and Formula: the pillar. NOI ÷ Value, the formula's three rearrangements, asset-class benchmarks.
• Cap Rate Decomposition: the Gordon Growth build-up that connects exit cap to discount rate via r = cap + g.
• Going-In Cap vs. Exit Cap: Modeling the Spread: the exit cap input to terminal value, with full spread mechanics.
• DCF vs. Direct Capitalization: When to Use Which: the methodological choice between cap-rate-based and discount-rate-based valuation.
• NOI: Institutional vs. Broker Calculation: what counts as NOI for the Year 11 forward input.
• Replacement Cost Analysis: Development and Insurance Frameworks: the cost-side cousin for ground-up and casualty-loss contexts.

FAQ

Sources

• Federal Reserve Economic Data, 10-Year Treasury Constant Maturity Rate (DGS10)(daily 10-year Treasury yield series).
• NCREIF Property Index (NPI)(a historical measurement of property-level returns for institutional real estate).
• Green Street Commercial Property Price Index (CPPI)(monthly commercial property price index built on REIT-owned asset valuations).
• MSCI Real Capital Analytics(commercial real estate transaction pricing, volume, and liquidity data).
• Gordon, M.J. and Shapiro, E. (1956), "Capital Equipment Analysis: The Required Rate of Profit," Management Science, 3(1), 102–110(original publication of the Gordon Growth Model underpinning both perpetuity growth and exit cap terminal value methods).
• New York City Department of Finance, Real Property Transfer Tax (RPTT)(city rates for non-residential transfers).
• New York State Department of Taxation and Finance, Real Estate Transfer Tax(state base tax and the New York City additional base tax on non-residential conveyances of $2 million or more).
• D.C. Code § 47-903(deed transfer tax rate and additional tax).
• D.C. Code § 42-1103(deed recordation tax rate and liability of the parties).
• City of Chicago Department of Finance, Real Property Transfer Tax(city and CTA portions and the buyer/seller allocation).
• Florida Statutes § 201.02(documentary stamp tax on deeds).