Payment Processor Disputes · Memo

Stripe Rolling Reserves: Enforceability and What Actually Moves Them

Counsel for merchants caught in a Stripe hold ask the same questions in the same order, so I am going to lay out the analysis I run on these matters and the points where the Stripe ToS, the federal payment-network rules, and California's UCL actually have leverage.

A rolling reserve on a payment processor account is a contractual mechanism that withholds a percentage of a merchant's settled volume for a defined period to cover anticipated chargebacks, refunds, and fraud losses. In Stripe's ecosystem, reserves are imposed under the Stripe Services Agreement and the relevant Payment Terms. The merchant's first reaction to a reserve announcement is almost always shock, because the cash-flow effect on a young business can be terminal. The legal analysis is more nuanced than the shock would suggest. The contract gives Stripe broad discretion. The discretion is not unbounded. And the practical work is at the operations layer, not the litigation layer.

Where the contractual authority comes from

The Stripe Services Agreement reserves to Stripe the right to suspend, restrict, or terminate access, and to require a reserve, where Stripe determines in its discretion that the account presents elevated risk. The agreement is governed by California law (Stripe is headquartered in San Francisco, and the choice-of-law clause points to California for most merchants). The merchant typically clicks through the agreement at onboarding without negotiation. That makes the agreement a contract of adhesion under California law, which does not by itself make it unenforceable but does invite scrutiny under Civ. Code section 1670.5 unconscionability principles and under the implied covenant of good faith and fair dealing.

The implied covenant is doing most of the work in the cases I have seen. California recognizes that the covenant operates to prevent a party from using the contract to deprive the other party of the benefit of its bargain. Carma Developers (Cal.) Inc. v. Marathon Development California Inc., 2 Cal. 4th 342 (1992), is the standard citation. Where Stripe has discretion under the agreement, the discretion must be exercised reasonably and in good faith. That is the legal lever that lets a merchant argue a reserve is excessive even though the contract gives Stripe the right to set one.

What 'reasonable' looks like

Network rules and processor industry practice converge on a few reference points. For a low-risk merchant category code (MCC), a reserve above ten percent of rolling settled volume held for more than ninety days is on the high end. For higher-risk MCCs (digital goods, subscriptions in certain verticals, online sales of regulated goods), reserves of twenty to thirty percent for one hundred eighty days are within industry tolerance. Reserves at fifty percent of volume held for two hundred ten days, which I have seen on enforcement matters, are at the edge of what any underwriter would defend in a deposition.

The Visa Core Rules and the Mastercard Rules both speak to reserve sizing in the merchant acquirer relationship. Stripe's relationship with the underlying card networks is through bank partners, not directly, so the network rules do not control Stripe's contractual posture. But the rules do define what 'commercially reasonable' looks like, and they form the implicit benchmark a court or arbitrator would use.

The McGill rule and the arbitration clause

The Stripe Services Agreement contains an arbitration clause that requires individual arbitration and waives class actions. Under McGill v. Citibank N.A., 2 Cal. 5th 945 (2017), an arbitration clause cannot bar a California plaintiff from seeking public-injunctive relief. The clause has been the subject of litigation, and the case law has tightened on what does and does not survive McGill. The practical implication for a merchant: a UCL claim seeking public-injunctive relief (an injunction against the reserve practice as it applies to similarly situated merchants) has been argued to survive the arbitration clause. The narrower the public-injunctive-relief request, the harder it is to characterize as a private restitutionary claim that the clause does push to arbitration. Counsel pursuing this approach should expect a fight on the threshold motion.

What tactics actually move reserves

In the matters I have worked, three categories of leverage have moved a reserve. The first is documentation of why the underwriting analysis is wrong. Stripe's risk model is automated and conservative. When the merchant can show the model misclassified the business (wrong MCC, wrong customer concentration assumption, wrong chargeback baseline), the reserve gets revisited. The documentation has to be operational: customer cohorts, refund rates, chargeback ratios, settlement velocity, and any external risk indicators (D&B reports, processor history with other providers). This is unglamorous evidence work, and it is where the volume of the wins happens.

The second is escalation past customer-support layer to legal. The frontline support team at any major processor is not authorized to release a reserve. The risk-operations team is, on a case-by-case basis. The legal team is, when the matter has been positioned as a litigation candidate. A letter from an attorney that articulates a specific legal theory (UCL public-injunctive, breach of implied covenant, conversion as to amounts held beyond the contractually defined period) and that demonstrates familiarity with the relevant section of the Services Agreement tends to move correspondence into a different inbox. I have written enough of these to know what does not work, which is a generic complaint letter that recites general principles without the contract analysis.

The third is the regulatory complaint. The California Department of Financial Protection and Innovation (DFPI) has authority over money-transmission and payment-processor consumer-protection issues in California. A complaint to the DFPI is not a fast remedy. It is a useful escalation lever in the right matter. The CFPB, post-2025 reorganization, has more limited scope on commercial payment-processor disputes but is still a relevant audience for systemic-practice complaints. Filing a complaint should be a considered move, not a reflex.

Where I draw the line on litigation

I have argued in cases that an excessive Stripe reserve, held beyond a commercially reasonable period and without a documented underwriting rationale, violates the implied covenant and is recoverable as breach-of-contract damages, with UCL injunctive relief available. Outcomes depend on facts, on the size of the reserve, on the merchant's documented losses, and on the arbitration-clause posture. I do not promise outcomes. I do say that the analysis is real and the leverage is meaningful when the documentation supports it.

The litigation calculus turns on the merchant's loss model. If the reserve is $200,000 and the merchant can document that the held funds caused the business to lose a key customer, miss payroll, or take on emergency capital at unfavorable terms, the consequential damages are real. If the merchant cannot document that, the recoverable amount may be limited to interest on the held funds, which is rarely worth the cost of litigation. The economics often push toward a negotiated release rather than a contested proceeding. That is fine. The negotiated release is usually the goal.

The drafting note for SaaS counsel

For SaaS counsel advising on the merchant agreement before signing, two clauses to read carefully. First, the reserve and settlement-hold sections. Look for the specific basis on which a reserve may be imposed, the maximum percentage and duration, and the notice obligations. The good language is bilateral: Stripe has the right to impose a reserve, and the merchant has the right to documentation of the underwriting rationale on request. Second, the dispute-resolution clause. The arbitration provision constrains the forum and remedy in any future dispute. If the merchant operates a high-volume business and a freeze would be existential, considering an alternative processor whose terms are more negotiated may be a better use of effort than fighting a Stripe arbitration clause after the fact.

Sergei Tokmakov, Esq., CA Bar #279869. This memo is attorney commentary on legal questions and is not legal advice. Reading it does not create an attorney-client relationship. Past matter outcomes depend on facts and the responding party; nothing here is a prediction of result.