Field note

Banks count fraud dollars and miss the exception bill

Every failed payment kicks off a chain of manual work that rarely shows up in a fraud dashboard. Across the top 15 US banks, dispute processing alone runs about $3 billion a year - and the wire flagged at $50K costs the same to investigate whether it is real or not.

Aug 13, 2025 · Navin Agrawal · Payments · 3 min read

Banks count fraud dollars and miss the exception bill

Visual brief

Visual brief

Banks count fraud dollars and miss the exception bill

As of August 2025

When a payment fails, the reversal is the cheap part. The expense is everything after: the research, the compliance review, the correspondent-bank back-and-forth, the escalation to save the relationship.

Banks meter fraud to the dollar and leave exception handling as an unmeasured cost center. It is one of the largest operating costs in payments that almost nobody puts a number on.

Start with the one figure that is well sourced. McKinsey put dispute processing across the top 15 US banks at roughly $3 billion a year, at $10 to $50 per dispute. Disputes are only a slice of exceptions - ACH returns, wire recalls, and RTP or FedNow errors that often have no clean resolution path all land on the same operations teams - so the true exception bill sits above that line, not below it.

Payment exceptions are the cost center banks ignore (as of August 2025): dispute processing across the top 15 US banks runs about $3 billion a year at $10 to $50 per dispute, and disputes are a subset of all exceptions (McKinsey, 2018); a next-generation operating model can remove 25 to 40 percent of dispute cost (McKinsey, 2018); and a complex wire exception can cost an estimated $200 to $500 in internal labor regardless of the transaction amount (field estimate).
Fraud gets a dashboard. The bigger, quieter number is what it costs to work every exception by hand.

Dispute spend

$3B

annual dispute-processing cost across the top 15 US banks (McKinsey, 2018).

Per dispute

$10-50

cost per individual dispute, a subset of all payment exceptions (McKinsey, 2018).

Reducible

25-40%

of dispute cost a next-generation operating model can remove (McKinsey, 2018).

The cascade

A flagged $50K wire does not resolve itself. Customer service fields the call, compliance runs a manual review, the correspondent bank gets looped in, legal weighs a recall, and management steps in to protect the relationship.

The fix is operational, not heroic

The banks that win here are not preventing every exception. They resolve them faster with less manual work: automated categorization and routing, straight-through processing for common return codes, real-time customer notifications, and investigation templates with the compliance steps built in. McKinsey’s own figure is that a next-generation operating model can take 25 to 40 percent out of dispute cost. A single complex wire exception can still run an estimated $200 to $500 in internal labor regardless of the amount, which is exactly why automating the routine ones pays.

Was this useful?

Choose once.

Related Posts

View All Posts »
A chartered bank just put a stablecoin on card settlement

A chartered bank just put a stablecoin on card settlement

SoFi Bank put SoFiUSD on Mastercard's settlement layer. Authorization still rides the card rails and the merchant sees nothing change, but the interbank fund movement happens on-chain, around the clock, with no batch window and no correspondent float. For the first time, issuers and acquirers get to choose their settlement medium.

Spend limits are not delegated payment authority

Spend limits are not delegated payment authority

An invoice agent hits a transient timeout, retries, and two payments clear to the same vendor. The model calls it success. Ops calls it a week of cleanup. Spend limits and roles are a speed bump. Real delegated authority is a delegation object, policy on every attempt, ledger-bound idempotency, and an instant kill switch.

The hard part of payments is access, not the API

The hard part of payments is access, not the API

A payroll rollout taught me the friction is never the integration. It is funding caps, routing permissions, and what can clear directly versus what has to detour through a sponsor bank. That is why the Fed's new payment-account RFI matters - access is becoming a design decision, not just a policy one.

ACH's most boring 10 characters just became policy

ACH's most boring 10 characters just became policy

From March 2026, Nacha standardizes the ACH Company Entry Description as PAYROLL or PURCHASE for those payment types. It sounds trivial, but a free-text field becoming a dependable category is what finally lets reconciliation automation stick on a rail that still moves tens of billions of payments a year.