White Paper 01 · Updated August 2026
The New Economics of Card-Not-Present Risk
Why fraud, disputes, false declines and customer friction must be managed as one economic system.
What the evidence says
The Federal Trade Commission reported that consumers filed 2.6 million fraud reports in 2024 and reported losing more than $12.5 billion — 25% more than in 2023. Just as telling: the share of people who reported losing money rose from 27% in 2023 to 38% in 2024. Report volume held broadly steady while conversion into loss climbed. These are consumer-reported losses, not merchant losses, but they describe the pressure landing on consumer-facing commerce.
On the dispute side, Visa has stated that approximately $11 billion of charges were disputed with U.S. issuers in the cited year, against $7.2 billion in 2019. Mastercard, separately, cites a forecast that global merchant chargeback costs may reach $42 billion by 2028, with nearly half reported as fraudulent. That $42 billion is a forecast, not an observed result, and should be treated as directional planning input only.
FTC-reported consumer fraud losses, 2024
Consumers reported more than $12.5B lost to fraud in 2024 — 25% more than 2023. FTC did not publish the comparison year as a charted value here, so only the 2024 total and the stated year-over-year change are shown.
- 2024 reported consumer fraud losses$12.5B+
2.6 million fraud reports filed in 2024.
- Year-over-year change vs 2023+25%
Stated by the FTC as a percentage increase, not a charted dollar comparison.
View data as a table
| Measure | Reported figure |
|---|---|
| 2024 reported consumer fraud losses | $12.5B+ |
| Year-over-year change vs 2023 | +25% |
Share of fraud reports involving a monetary loss
The proportion of people reporting fraud who said they lost money rose year over year.
- 202327%
- 202438%
View data as a table
| Year | Share reporting a loss |
|---|---|
| 2023 | 27% |
| 2024 | 38% |
Executive implications: the loss waterfall
Most fraud programmes are still reported as a single ratio. That ratio is one line in a waterfall of eight costs, and every control moves more than one line at a time. A tightened rule that removes approved fraud almost always adds false declines, review labour and support contacts. The programme looks better; the business is worse.
| Line | What it measures | Typical direction when fraud rules tighten |
|---|---|---|
| Approved fraud | Confirmed fraud losses on approved transactions | Down |
| False declines | Good demand rejected by policy, model or issuer response | Up — usually the largest hidden line |
| Authentication abandonment | Customers lost during 3DS challenge or step-up | Up |
| Manual-review expense | Analyst hours, tooling, queue latency | Up |
| Dispute handling | Representment labour, evidence assembly, win-rate outcomes | Mixed |
| Network exposure | Monitoring-programme ratio risk and remediation effort | Mixed — depends on dispute mix, not fraud alone |
| Support cost | Contacts generated by declines, holds and disputes | Up |
| Lifetime-value damage | Cohort retention loss from wrongly blocked customers | Up, and rarely measured |
The discipline is not complexity for its own sake. It is that no single owner currently holds all eight lines, so improvements are declared in one function while the cost lands in another. Optimising fraud rate alone can and does worsen total economics.