What Does It Cost You To Leave?
Contracts either let you leave clean or clean you out
Executive Brief
A plan sponsor’s leverage comes down to three things in sequence. You need your data. You need to be able to verify it. And if what you find is bad enough, you need to be able to leave.
Break the last link and the first two stop mattering. An audit that proves you’re being overcharged is a document, not a remedy, if leaving costs more than the overcharge.
So we looked at what typical PBM contracts actually say about leaving. Across the 67 in our database, the median score on termination and clean exit is 49. The median is misleading, though, and the reason is the finding.
Eighteen contracts score Good or better on exit rights. Twenty-nine score Red Flag. Only 20 sit anywhere in between.
Most provisions have a middle. This one doesn’t. A contract either lets you go or it builds a trap, and the negotiated middle ground where somebody argued and got partway is largely empty.
The Contracts That Prove It
Six contracts in our database score Good or better on rebate and manufacturer revenue while scoring Red Flag on exit terms.
Three of those six carry rebate provisions between 95 and 100, against exit provisions in the low 40s.
Read that again. Someone negotiated the rebate pass-through language to near-perfect. Nobody touched the clause governing whether the plan can ever leave.
That’s not a contract with a weak spot. It’s a contract where the most-negotiated provision and the least-negotiated provision sit twenty pages apart, and only one of them determines whether the other can be enforced.
What the Standard Actually Requires
Before we get to the bad clauses, here’s the test we apply, because it’s more specific than most sponsors expect and it’s not hostile to PBMs.
A termination charge passes if it recovers stated up-front investment under a documented amortization method, or compensates for stipulated wind-down work.
It fails if it recovers lost profits, if it forfeits plan assets including accrued rebates and retained credits, or if it functions to deter exit rather than to recover cost.
A reasonable termination fee isn’t a problem. A fee tied to genuine up-front investment makes the vendor whole for money it actually spent, declines as the investment amortizes, and gives the sponsor a known exit cost on day one. That’s a business term, and any plan sponsor should accept one.
The structural failure is a fee recovering something the PBM was never entitled to recover. Accrued rebates aren’t the PBM’s. Retained administrative credits aren’t the PBM’s. Future profits aren’t the PBM’s. A contract that takes them on the way out isn’t recovering an investment. It’s redirecting plan assets.
That distinction is the whole issue.
Three Ways to Build the Trap
Stacked forfeiture. One PBM’s services agreement combines full forfeiture of pending and future rebates on early termination, a three-year lock, and repayment of a prorated administrative credit. Each element would be arguable alone. Together, exit isn’t priced. It’s foreclosed. That provision scored 15, the lowest exit score in our database.
Lost profits, relabeled. A bundled TPA arrangement calculates its early termination fee as current administrative rates times covered employees times months remaining. The contract calls this compensation for anticipated probable harm and concedes actual damages might be more or less. Runout is priced at 100% of three months plus another 50% of three months. Claims records are released only after payment of all fees. Three additional exit costs are named without amounts, so the total cost of leaving can’t be calculated at signing. That provision scored 19.
No fee at all. A family of government contracts from one regional PBM charges nothing to terminate, and withholds rebates at the PBM’s discretion on early termination or insufficient notice. Rebates are plan assets. A discretionary transfer of plan assets on exit fails the reasonableness test regardless of the dollar amount, which is why a contract with a zero-dollar exit fee still scores Red Flag.
Your Own Money as the Penalty
The third example points at something worth understanding on its own, because it’s the mechanism most sponsors have never had explained.
Rebates are earned on claims already adjudicated. Your plan already paid the pharmacy. The manufacturer money coming back is a receivable, and it belongs to the plan.
It arrives late. Standard remittance runs months behind the claims that generated it, which means at any given moment your PBM is holding a substantial sum of your money.
Then the forfeiture clause converts that balance into the price of leaving.
Nobody set out to post a bond with their PBM. But that’s the effect. A rolling deposit, held by the counterparty, forfeited if you go.
And unlike a security deposit, nobody quoted you the amount at signing, because the amount is whatever happens to be in flight on the day you decide.
One coalition arrangement we assessed this year states it plainly. The contract holds rebates until the end of the incurred quarter plus ninety days and expressly denies the plan sponsor any right to interest on money held during that period. There is a for-any-reason termination right, conditioned on the PBM retaining earned but unpaid rebates, and would relieve the PBM of providing a final financial reconciliation.
So the plan has the right to leave, forfeit the money in flight, and give up the accounting that would let it calculate what it surrendered. That provision scored 30.
Which is one more reason to reduce what flows through rebates. Not only because list prices inflate to fund them and patients pay coinsurance on the inflated number. Because the money sitting in transit is leverage held against you by the party you might need to leave.
29 Points, Same Dollars
Now the contract that shows what the test actually measures.
A PBM moved its termination and clean exit provision from 55 to 84 in a single review cycle.
The fee didn’t change. It was, and remains, 75% of average monthly administrative fees times the months remaining in the initial year. Not one dollar moved.
What changed is that the contract now states what the fee is for. That it represents unrecovered up-front implementation investment, amortized straight-line. That it excludes lost profits. That the calculation is documentable on request.
Three sentences.
Before, the fee failed the reasonableness test because nobody could evaluate it, and a failed test caps the provision no matter how strong the rest of it is. After, it passes, and the improvements the PBM had already made to rebate continuation and data return could finally register.
What Leaving Can Cost You
A termination fee is typically only one part of the cost to exit.
The rebates in flight. Covered above, and usually the largest number nobody calculates.
The data. Claims records released only after all fees are paid, in the incumbent’s format. You pay to leave, then pay again to receive your own history in a form your next vendor has to reconstruct.
The runout. Priced separately, often at a premium, sometimes twice for overlapping periods.
The costs with no amount. Several contracts name additional exit charges without stating a figure. An exit cost that can’t be modeled at signing isn’t a term. It’s an exposure.
What to Do First Thing Monday
- Write down what leaving costs, as one number. Termination fee, runout pricing, data extraction, and rebates in flight. If you can’t produce a total, that’s the finding, and it belongs in your fiduciary file today.
- Ask what happens to rebates earned during the term but paid after it ends. Get it in writing. Silence means the PBM keeps them, and the amount is larger than you think.
- Check whether your data release is conditioned on payment. If claims records arrive only after all fees are settled, your exit is sequenced by the party you’re exiting.
- Ask your PBM to state the basis for the termination fee. Not the amount. What it recovers, how it amortizes, and what documentation supports it.
In Closing
This issue closes a four-part series: Omission takes money by leaving a term out. Redefinition takes it by changing what a term means. Restriction makes the taking undiscoverable. Lock-in decides whether discovering it changes anything.
The binary result is what stays with me. Eighteen contracts let a plan leave clean. Twenty-nine make leaving cost more than staying wrong. Relatively little in between.
Which means this isn’t a provision that erodes through inattention. Somebody built each of those traps deliberately, and somebody could have built the more reasonable provision instead.
One PBM proved that in a single review cycle. Same fee, three sentences, twenty-nine points. Ten PBMs now sit in the Excellent tier, and all but one started in Red Flag. None of them gave up anything they should have wanted to keep.
Signing Season Is Open
The contract you sign this fall governs through 2028 at least. CAA 2026 readiness needs to be at the heart of your renewal plan.
Tell your advisor and PBM, “We will only sign a CAA 2026 Ready contract scoring 90+ on Contract X-Ray.”
You have more power than you think.
Once you see it, you can change it.
Here’s to clearer thinking, stronger plans, and better outcomes for the people who rely on us.
All the best,
P.S. Special edition this Wednesday. Mark Cuban has been building a model PBM contract in public, revising it daily, and we scored it. More than ten PBMs have now moved their standard contracts into the Excellent tier, all but one starting in Red Flag, average improvement 39 points. Wednesday covers what a 90+ contract actually contains.
Subscribe & Share
🔗 Subscribe: Was this newsletter forwarded to you? Signup to receive The Middleman’s Cut every Monday.
📤 Share: Know someone reviewing a PBM contract? Forward this issue.
Tools and Resources
────────────────────────────────────────
Employers pay far more than they should to middlemen most executives have never heard of. Change the status quo and reap the benefits. Employers who act are recovering 20% to 30% of pharmacy spend, and returning it to wages and benefits.
Schedule an introductory call with us.
