Why reserves matter
Reserves are the bank's or processor's way of saying: "We trust you, but not for free." A program with material chargeback risk, ACH-return exposure, or fraud history will be asked to park cash \u2014 sometimes weeks or months of transaction volume \u2014 in a non-operating account where it earns nothing and absorbs losses first if the program breaks.
For a small fintech, reserves are a working-capital tax. For a large one, they can be the difference between a viable program and an uneconomic one.
When reserves get re-priced
Reserves are typically set at signing, then re-priced when (a) actual loss experience comes in higher than projected, (b) regulators tell the bank to tighten risk posture across its programs, or (c) the sponsor's risk committee rotates and the new committee re-reads the file. The second case has been particularly common in the post-2023 regulatory environment.
Common pitfalls
- Treating reserves as one-time. They are renegotiated annually at most banks, and the operator who walks in with bad chargeback numbers walks out with bigger reserve requirements.
- Modeling on day-one reserve, not steady-state. Build your unit economics on the realistic year-three reserve, not the one your sponsor offered at launch.
- Confusing reserves with holdbacks. Reserves are a standing collateral pool; holdbacks are time-based withholds on specific transaction flows. Both can apply.
Reserves are negotiable, but only with operational data \u2014 model them honestly before you sign.