Performance tests aren’t pass/fail thresholds. They’re option contracts.
The performance test in the term sheet reads like a gate. Two prongs, tested each year, termination if the operator misses both. Read that way, it looks like the sharpest risk in the deal. Read as an option, it’s a different risk — and usually a smaller one.
The test exists for a legitimate reason: an owner’s exit right against genuine underperformance. This piece takes that as given. The question isn’t whether the owner should hold the right. It’s what the right is actually worth, and how often it gets used.
Start with a number. On a recent analysis of a convention headquarters hotel under a long-term management agreement, the GOP margin prong failed about 92% of the time across the Monte Carlo runs. The owner’s estimated termination rate was about 11%. The test failed nine years in ten. The owner acted on it one time in nine. That gap is the whole story.
The gap exists because a failed test doesn’t trigger termination. It hands the owner an option. The owner can act or not, and there are several ways to act, each with a different cost. Termination is only one of them — and it’s the one owners choose least. Once the option is visible, the cure mechanics stop reading as boilerplate. They become the terms that decide which path the owner takes.
What follows: what a performance test actually is, why the binary framing misprices the operator’s risk, and which cure terms an operator should fight for once the structure is clear.
The Standard Structure
Most HMAs run a two-pronged test. One prong is RevPAR Index — the property’s revenue per available room against a competitive set, usually 85 to 92 percent. The other is GOP margin — gross operating profit margin against an absolute floor, usually 28 to 34 percent depending on property type. Both prongs have to fail before the owner’s termination right opens.
Cure rights usually follow. The operator gets 12 to 24 months to fix the failure. Two ways to fix it: improve performance, or make a cure payment — a set dollar amount that satisfies the test without operational correction. Some contracts cap the payment. Some don’t. Some limit how many times the operator can cure. Some don’t. Those details look minor. They aren’t.
The industry reads this as binary. Clear the test, keep the contract. Fail it, face termination. Financial models often compress the whole thing into one number: the chance of losing the deal. But termination is only one path. And it’s the path owners take least often.
The Owner’s Option: Six Paths, Ranked by Cost
A failed test doesn’t force the owner’s hand. It gives the owner a choice among several paths. Owners are rational. They take the path that gets the most value for the least cost. Rank the paths by cost and the 11% termination rate stops being a surprise.
Do nothing. The cheapest path is to let it pass. The test failed, the contract permits a response, and the owner does nothing. The cost is zero. Operators often never learn the owner considered the failure material at all.
Threaten, without meaning it. The next-cheapest path is to invoke the failure as leverage. A letter, a phone call, a reference to the provision — used to gain ground in some other negotiation, like a fee discussion or a reserve dispute. The owner never intends to terminate. The operator carries the uncertainty anyway.
Neither path costs the owner capital, transition, or disruption. Both are attractive when the failure is mild or the relationship is otherwise working.
Demand a cure. A middle path is to demand operational improvement. This costs the owner some monitoring, some legal review, some relationship friction. In return, the owner keeps the incumbent and gets better performance. It works when the failure is fixable and the operator is likely to make the fix.
Take a cure payment. The owner can also take the payment. That costs some legal and accounting work, and it signals that future failures will be resolved by payment rather than by escalation. But the payoff is immediate: cash, test satisfied, contract intact. This becomes the preferred path when the failure is serious, operational cure looks unlikely, and termination would cost more than the payment. The cure payment cap is the strike price on this path. Set it right, and cure payment beats termination across a wide range of failures.
Renegotiate. A more expensive path is to reopen the whole agreement. Legal fees, advisor fees, management time, and the risk that talks fail and leave the relationship damaged. The payoff is fee concessions or better terms going forward. It makes sense when the failure is severe enough to give the owner leverage and the operator is likely to concede.
Terminate. Termination is the most expensive path by a wide margin. The direct costs are real — transition, search, legal, key money to the incoming operator. The indirect costs are larger. The property drops into a performance valley that runs 12 to 24 months. Corporate accounts scatter. Group pipelines take quarters to rebuild. Rankings reset under the new flag. Reviews track the dip. Comp-set position takes 18 to 36 months to recover.
That valley usually costs the owner more than any fee savings a new operator would bring. Termination isn’t inherently irrational — it’s the correct choice when the replacement operator’s improvement exceeds the transition cost. That calculation just doesn’t clear in most cases. Termination makes sense only in narrow cases: a genuinely incompetent operator, a broken relationship, or a replacement offering terms good enough to cover the transition.
The 11% termination rate is that cost structure showing up in the data. Owners terminate only when the cheaper paths won’t work.
Cure Mechanics Are the Real Terms
Once the test is an option, the cure terms stop being boilerplate. They decide which path the owner takes.
Cure payment caps work as strike prices. Higher caps keep the cure payment path economically rational for the owner. Lower caps push the owner toward termination when failures are severe. The instinct is to negotiate for lower caps. The math runs the other way.
Here’s why. Say underperformance costs the owner $800,000 a year and the cure cap is $200,000. The owner recovers 25 cents on the dollar, and the problem continues. Termination starts to look rational even with a big transition cost. Raise the cap — or remove it — and the payment can actually compensate the owner. Now cure payment beats termination across more of the failure range. The operator’s risk was never the size of the payment. It was the chance the owner walks because the payment is too small.
Cure windows work as exercise periods. A longer window — 18 to 24 months — gives the operator time to fix performance before paying. It raises the odds operational cure succeeds, which keeps the owner off the expensive paths. A short window compresses the operator’s response time and pushes the owner to escalate.
Cure rights work as renewals. Unlimited cure rights mean the contract never turns into sudden death. Every failure runs the same option, and the owner faces the same cost-benefit each time. Limited cure rights, common in older HMAs, mean later failures carry more weight, because the cure payment path may already be spent. The operator’s risk climbs sharply as the cures run out.
Back to the convention headquarters hotel. The GOP prong was failing 92% of the time. The operator’s risk-weighted NPV stayed positive anyway. The reason was in the cure terms: the payment was uncapped and the rights were unlimited. Those two terms did the work the headline failure rate hid. They steered the owner toward cure payment across nearly the whole failure range.
The terms interact. Unlimited rights, a 24-month window, and a high or uncapped payment steer the owner toward the cheap paths. One cure right, a 12-month window, and a low cap push the owner toward the expensive ones. The risk difference between those two contracts is large. Standard NPV models miss it.
Where the Standard Model Misprices
Most models treat the test as binary. Clear it and keep full value, or fail it and lose some share of value. Even Monte Carlo models, which run thousands of scenarios, usually assume a fixed chance of termination once the test fails. That misprices the risk two ways.
First, it ignores the paths. A 92% failure rate sounds fatal if failure means termination. It’s manageable once you know that 92% of failures produce an 11% termination rate, because the owner takes a cheaper path nine times out of ten. The operator’s real risk isn’t how often the test fails. It’s how often the test fails and the owner then chooses an expensive path. Collapse that into a single termination number and the structure is gone.
Second, it ignores the cure terms. Two contracts can have the same failure probability and completely different risk, because their cure terms steer the owner differently. Strong cure terms channel the owner toward cure payment. Weak ones leave termination as the least-bad option when failures get severe.
The better model runs the failures, then asks which path the owner takes each time. That answer depends on three things: how severe the failure is, what the cure terms allow, and how much termination would cost this owner. Severity decides whether “do nothing” holds. Cure terms decide whether cure payment beats termination. Termination cost — property type, market, replacement availability, lender consent — decides the point where termination finally makes sense.
Run it that way and a deal failing 92% of the time with strong cure terms and a high termination cost can be less risky than a deal failing 40% of the time with weak terms and a low termination cost. The failure rate is a poor proxy for the risk. The path the owner would take is the risk.
What the Operator Should Negotiate
Cure and remedy rights exist on both sides of most HMAs, on different terms for each party. This piece focuses on the operator’s cure rights against the owner’s termination option.
The priorities shift once the structure is clear. Thresholds still matter — a lower failure rate beats a higher one. But the cure terms often matter more.
Push for higher cure caps, not lower. This runs against instinct and it’s still correct. A high or uncapped payment keeps cure payment rational for the owner even when failures are severe. A low cap makes cure payment uneconomic and pushes the owner to terminate. The risk was never the payment size. It’s the chance the owner terminates instead.
Push for unlimited cure rights. Limited rights create sudden-death years, where one failure ends the operator’s options. Unlimited rights keep every failure running the same option, with the same owner incentives. The contract never turns brittle.
Push cure windows to 18 to 24 months. More time raises the odds operational cure works, which keeps the owner off the expensive paths. Short windows do the reverse.
These terms compound. Any one of them helps at the margin. All three together change the risk structure — they steer the owner toward the cheap paths across the scenarios that would otherwise trigger the expensive ones.
And model the paths, not just the failure rate. That means estimating what termination would cost this owner, on this property, in this market — and working out how the cure terms move the owner’s choice. The output isn’t one NPV. It’s a range, weighted by which path the owner is likely to take.
The larger point holds across deals. One that fails often but carries strong cure terms and a high termination cost can be safer than its headline suggests. The convention headquarters hotel failed 92% of the time and was still worth doing, because termination was too expensive for the owner and the cure terms kept steering the owner toward the manageable paths. The failure rate said dangerous. The structure said fine.
The Takeaway
A performance test is an option the owner holds. The failure isn’t the risk. What the owner does with the failure is the risk. And that comes down to option terms the operator can negotiate.
The binary framing — pass or terminate — misprices the risk and sends operators to fight the wrong battles. They spend leverage moving a GOP threshold from 30% to 28%, or a RevPAR Index threshold from 90% to 87%, when the cure terms often matter more. A deal with a 90% threshold, unlimited cures, a 24-month window, and no cap can be safer than a deal with an 85% threshold, one cure, a 12-month window, and a low cap.
Once the option structure is visible, it’s hard to read a performance test any other way. So it’s worth asking: on your current deals, is the performance test being priced on how often it fails — or on the path the owner would actually take?
Dash Decisions provides vendor-side financial deal desk services for hotel operators bidding on HMAs.