Operators price key-money recovery as a simple probability. It’s a barrier option — and worth a fraction of what the simple math says.
Key money is the operator’s upfront investment — typically $2M to $8M on mid-market deals, amortized over 15 to 25 years. The operator recovers it by staying in the deal until amortization runs out, or, on early termination, through a clawback of the unamortized balance. Most operators price that recovery as a simple contingent claim: the unamortized balance times the chance of early termination. A $3.2M balance at an 18% chance of early exit — call it $576K. That pricing is wrong by a factor of three to five.
Key-money recovery isn’t a simple claim. It’s a barrier option. The operator’s right to recover knocks in or out depending on whether two barriers are crossed — the type of termination (for-cause versus convenience or sale) and the asset’s value (worth more or less than its debt). And both barriers move against the operator in exactly the situation that ends the deal. Underperformance drives early termination. Underperformance also knocks out both barriers. The recovery has real value only in the benign states — a convenience or sale termination of a healthy asset, where the deal ends without operator fault and the hotel is worth more than its debt. Those states are real. They just aren’t the states a struggling operator is exposed to.
The forfeit-on-for-cause, refund-on-convenience split isn’t arbitrary. It gives the owner a rational allocation — keep the operator’s money when the operator fails, return it when the owner chooses to exit through no fault of the operator. Seen that way, key money is partly a performance bond, and that’s a defensible structure. The operator’s problem isn’t the split. It’s that the states where the recovery is supposed to pay are the ones a struggling operator won’t reach.
What follows: the two barriers, the correlation that ties them to the termination event, a worked example contrasting naive and barrier-adjusted valuation, and the remedies the structure points to.
The Two Barriers
A barrier option’s payoff depends on whether some underlying variable crosses a set barrier during the option’s life. Cross it, and the option knocks in (becomes active) or knocks out (becomes worthless). Key-money recovery is built exactly this way. The structure was introduced by Merton (1973) for down-and-out options and generalized by Reiner and Rubinstein (1991) — though here it supplies the framework and intuition, not a closed-form price. The recovery is a real option, not a traded one, so the number comes from a transparent expected-value walk on the property’s own parameters, as the worked example shows.
The operator’s right to recover the unamortized balance turns on two barriers — one contractual, one economic.
Barrier 1: Termination Type
Recovery depends on why the deal ends. The common drafting convention: a for-cause termination (performance failure, material breach, bankruptcy) forfeits the unamortized key money — the recovery knocks out. A convenience or sale termination (the owner’s choice, no operator fault) refunds it — the recovery knocks in.
The contract draws a line. On the owner-initiated, no-fault side, the operator gets the balance back. On the operator-fault side, it forfeits. The recovery exists only if the termination lands on the right side of the line.
This is the common structure, not a universal rule. Some agreements refund regardless of cause; some forfeit in all early exits; some have no clawback at all. The barrier framing holds however the contract draws the line. Whether a specific clause forfeits on for-cause, refunds on convenience, or is enforceable is counsel’s determination.
For an operator under performance pressure, the type barrier is the first knock-out. If underperformance drives a for-cause termination, the recovery is gone before the operator even reaches the second barrier.
Barrier 2: Asset Value
Even when recovery knocks in on type — a convenience or sale exit, no operator fault — it’s still a claim against the SPE. If the hotel is worth less than its mortgage debt, the operator is an unsecured claimant behind the lender, against an entity with no spare cash. The recovery knocks out on collectability even though it knocked in on type.
The SPE collectability mechanics are developed in “Non-Recourse: The Clawback You Can’t Collect.” The short version: the owner entity is usually a single-purpose vehicle whose only asset is the hotel, the lender holds a first-priority lien, and if the hotel is underwater there’s nothing left to collect.
The asset-value barrier is binary. Worth more than its debt, the SPE has equity and the claim is collectible. Worth less, the SPE has no equity and the claim is worthless. And for an operator under pressure, this is the second knock-out: underperformance depresses the hotel’s value, and if it’s severe enough to trigger termination, it’s often severe enough to push the asset underwater.
The Adverse Correlation
The two barriers aren’t independent of the termination event. They move against it. Underperformance is the common driver, and it knocks out both barriers at once.
When the operator underperforms — repeated test failures, material breaches, losses that threaten the owner’s debt service — the owner terminates for cause, and Barrier 1 knocks the recovery out. The same underperformance depresses the hotel’s value. Buyers underwrite on trailing NOI; sustained underperformance drives NOI down, which drives valuation down, and if the decline is severe enough the hotel falls below its debt. The SPE is underwater, and Barrier 2 knocks the recovery out on collectability.
So the early terminations that actually happen to a struggling operator cluster in exactly the states where both barriers extinguish the recovery. The operator underperforms, the owner terminates for cause (Barrier 1 out), the hotel is worth less than its debt (Barrier 2 out). The recovery is worthless twice over.
The recovery has real value only in the benign states — a convenience or sale termination of a healthy asset. The owner sells into a portfolio transaction and the buyer wants its own operator; the hotel is performing, worth $80M against a $50M mortgage; the operator is terminated for convenience and collects from an SPE with $30M of equity. Or the owner’s thesis changes — it exits hospitality or consolidates brands — and again the hotel is performing and the operator collects.
These happen. Convenience and sale terminations are real. But they aren’t the states a struggling operator is exposed to. An operator bidding a marginal deal and pricing key-money recovery into the NPV isn’t pricing the benign-state recovery. It’s pricing the recovery in the underperformance states — where it might actually need it — and in those states, both barriers knock it out. Priced naively, the recovery looks valuable, because the probability includes every early-exit scenario. Priced correctly, most of that probability lands in the knocked-out states. The recoverable value is a small fraction of the naive figure, and the gap is entirely the correlation.
Worked Example: Barrier-Adjusted Valuation
Take one property and value the recovery three ways — face, naive, and barrier-adjusted. The numbers are illustrative and varied; the point is the correlation, not a market norm.
The property. A 180-room select-service hotel, 15-year HMA, $4.0M of key money invested upfront, amortizing straight-line at $267K a year. The mortgage is $22M at origination (65% LTV on a $34M purchase). Evaluate at the midpoint, year 7-8, where the unamortized balance is about $3.2M. Take an illustrative early-termination probability of 18% at this point.
Face value: $3.2M. The unamortized balance — the contractual refund if the deal ends early and both barriers knock in. That’s the ceiling, not the expected value.
Naive value: $576K. The unamortized balance times the chance of early termination — $3.2M at 18%. This is what most operators use in a bid NPV. It’s wrong: it assumes the recovery is collectible in every early-termination scenario, treats the 18% as uniform across termination types and asset values, and ignores both barriers and the correlation. It overstates recoverable value by a factor of three to five.
Barrier-adjusted value: $96K. Split the 18% into barrier-defined states, weight by the correlation between underperformance, termination type, and asset value, and value the recovery in each.
- For-cause, asset underwater (both barriers out) — 10% of the 18%. The operator underperforms, tests fail, the owner terminates for cause. NOI has fallen 30-40% from underwriting; market value is $28M or $20M against the $22M mortgage. The operator forfeits (Barrier 1 out), and even if it didn’t, the SPE has near-zero or negative equity (Barrier 2 out). Recovery: $0.
- For-cause, asset healthy (Barrier 1 out, 2 in) — 3%. The operator breaches a covenant unrelated to performance — an unauthorized transfer attempt, a lapsed insurance requirement. The hotel is performing, but the forfeiture clause applies. Recovery: $0.
- Convenience/sale, asset underwater (Barrier 1 in, 2 out) — 2%. The owner sells in distress or a portfolio fire sale; the buyer wants its own operator; the hotel is underwater on market conditions unrelated to the operator (a demand shock, new supply, cap-rate expansion). The operator is terminated for convenience but the SPE has no equity. Recovery: $0 (uncollectible).
- Convenience/sale, asset healthy (both barriers in) — 3%. The owner sells into a portfolio deal or exits hospitality; the hotel is performing, worth $38M against the $22M mortgage; the operator is terminated for convenience and collects from an SPE with $16M of equity. Recovery: the full $3.2M.
Only the last state pays. Three percent of the probability, at $3.2M, is about $96K. The other three states pay nothing. The total is $96K.
The gap: naive $576K, barrier-adjusted $96K — a difference of $480K, or 83% of the naive figure. All of it is the correlation. Fifteen of the eighteen percentage points of early-termination probability land in states where one or both barriers knock the recovery out.
What the Operator Does With It
The barrier structure isn’t just an observation. It points to the remedies. The operator’s job is to see the barriers, price the correlation-adjusted recoverable value, and negotiate either to move the barriers or to shrink the notional they apply to.
Underwrite at barrier-adjusted value. Not face, not naive — the expected recovery weighted by the probability the deal ends in a state where both barriers knock in. For this property that’s $96K, not $576K. Underwrite the bid at $576K and the NPV is overstated by $480K. Discipline starts with pricing the recovery at its actual expected value, not its contractual ceiling.
Move Barrier 1 (termination type). Make key money refundable regardless of cause, or pro-rata on for-cause. That recovers the for-cause-but-healthy state — the 3% where the asset is fine but a covenant breach triggers forfeiture. It doesn’t recover the 10% underwater state, which Barrier 2 knocks out independently. For this property, moving Barrier 1 lifts expected recovery from $96K to about $192K — the two healthy states together, 6% of the probability at $3.2M — a $96K gain. It’s the highest-value move, because it recovers a real slice of probability the standard forfeiture clause extinguishes. Owners resist, because forfeiture is their penalty for underperformance. The operator’s counter: key money is its capital investment, not a performance bond, and blanket forfeiture is a windfall, not damages calibrated to the harm.
Move Barrier 2 (asset value / SPE collectability). Tie recovery to something other than the SPE’s asset value — a parent guaranty, a letter of credit, a funded escrow. These are developed in “Non-Recourse: The Clawback You Can’t Collect“: a parent guaranty makes the owner’s parent liable, a letter of credit is a bank’s commitment to pay on demand, a funded escrow is cash set aside at closing. Each removes Barrier 2 by making recovery independent of the hotel’s market value. For this property, moving Barrier 2 recovers the convenience-but-underwater state — the 2% — lifting expected recovery from $96K to about $160K (the two convenience states, 5% at $3.2M): a $64K gain, real but smaller than moving Barrier 1.
Shrink the notional (faster amortization). Ten years instead of fifteen, or seven instead of ten. This doesn’t move the barriers, but it shrinks the balance they apply to. For this property, amortizing over ten years leaves $1.6M unamortized at year 7-8 instead of $3.2M, so barrier-adjusted recovery is about $48K instead of $96K — same 3%, smaller base. It’s the remedy when the operator can’t move the barriers.
The Recovery You Won’t Reach
Key-money recovery isn’t a simple claim. It’s a barrier option with two knock-out barriers — termination type and asset value — and both move against the operator in the situation that ends the deal. Underperformance drives early termination, and it knocks out both barriers. The recovery has real value only in the benign states, and those aren’t the states a struggling operator is exposed to.
Priced naively, it looks valuable. Priced correctly, most of the probability lands in the knocked-out states and the recoverable value is a small fraction of the headline. The gap is the correlation. The highest-value fix is making key money refundable regardless of cause — it recovers the for-cause-but-healthy state the standard forfeiture clause throws away.
So on the next deal with real key money on the table: is the operator underwriting the recovery at its contractual face value — or at what it can actually collect in the states where it would ever need it?
Dash Decisions provides vendor-side financial deal desk services for hotel operators bidding on HMAs.