Fraud rate in basis points expresses fraud losses or incidents as a share of volume, where one basis point is 0.01 percent. Stating it in bps normalizes the number so teams can compare it across products, over time, and against peers.
What is fraud rate in bps, in plain English?
Fraud rate is how much fraud you are taking, sized against how much you are processing. Because raw fraud dollars mean nothing without context, teams divide fraud by volume and express the result in basis points. One basis point is one hundredth of a percent, so a fraud rate of 10 bps means fraud equal to 0.10 percent of volume, or one dollar of fraud for every thousand dollars processed.
Using bps instead of a percentage keeps small, important differences readable. The gap between 8 bps and 15 bps is easy to talk about, where the equivalent decimals, 0.08 percent versus 0.15 percent, are easy to gloss over. That is why fraud, payments, and risk teams almost always quote fraud in bps.
It is the headline number teams target, report to leadership, and benchmark against competitors. But a bps figure is only as meaningful as the definitions behind it, which is where most confusion lives.
The definitions that change the number
Two teams can both say "our fraud rate is 12 bps" and mean completely different things. The choices behind the number matter as much as the number itself.
Choice | One way to count | A different way to count |
What counts as fraud | Attempted fraud, including blocked | Approved fraud that got through |
Unit | Count of fraud incidents | Dollar value of fraud |
Denominator | All volume processed | Only approved volume |
Timing | Recent period, disputes still pending | Matured cohort, fraud surfaced |
Because disputes lag, recent periods understate the true rate. Read fraud bps on matured cohorts, where late fraud has come in, and always state the denominator plainly so the figure can be trusted and compared.
What it looks like in practice
In practice
A risk lead reports fraud at 6 bps to the board and gets congratulated. A month later a peer benchmarking group publishes a median of 14 bps for the same product, and leadership asks why the company looks so much safer than everyone else.
The answer is definitional. The 6 bps used total processed volume in the denominator and only counted fraud that had already been confirmed. The peer figure used approved volume and matured cohorts. Once the team recomputed on the same basis, its rate landed near 13 bps. Nothing had changed except the definition, and the original comparison had been meaningless.
Why the denominator decides everything
Fraud bps looks like one clean number, but it is a fraction, and both the top and the bottom are negotiable. Counting attempted versus approved fraud, incidents versus dollars, and total versus approved volume can each move the figure by a factor of two or more. A team can look best-in-class or alarming on the same underlying losses, purely by how it defines the ratio.
Timing compounds this. Because chargebacks and reports lag, a recent period always looks cleaner than it will once the cohort matures. That is why a trustworthy fraud rate comes with its definitions attached: what counts as fraud, what sits in the denominator, and how mature the cohort is. Without those, a bps figure is a talking point, not a measurement.
What to watch in the data
- State the denominator. Total volume, approved volume, or count-based rates are all different; a bps number without its denominator cannot be compared to anything.
- Attempted versus approved. Counting blocked attempts inflates the rate; counting only what got through deflates it. Know which one is being quoted.
- Immature cohorts. A low recent rate usually reflects disputes that have not landed yet, not genuine safety.
- Mixed products. A blended rate across products can hide a single risky line; segment before you trust the headline.
- Apples-to-oranges benchmarks. A peer figure computed on a different basis tells you nothing; align definitions first.
Quick questions
What does one basis point equal?
One basis point is 0.01 percent, or one part in ten thousand. A fraud rate of 25 bps means fraud worth 0.25 percent of volume, roughly one dollar in every four hundred.
Why use bps instead of a percentage?
Fraud rates are small numbers where a tiny percentage difference is a big deal. Basis points keep those differences readable, so 9 bps versus 16 bps is clearer than 0.09 percent versus 0.16 percent.
Why do two teams quoting the same bps disagree?
Because the definitions differ: attempted versus approved fraud, count versus dollars, total versus approved volume, and how mature the cohort is. The number is only comparable when those choices match.
Why does a recent fraud rate look too good?
Disputes and reports lag the fraud by weeks, so recent periods have not surfaced their full losses yet. Reading matured cohorts corrects for this understatement.
Is fraud rate the same as loss rate?
They are close but not identical. Loss rate typically reflects realized net losses after recoveries, while fraud rate may count gross fraud or incidents. Always check which is being measured.
Go deeper
- FFIEC BSA/AML Examination Manual ↗ — The manual US examiners use to assess BSA and AML programs.
- FTC Consumer Advice: Scams ↗ — US consumer guidance on current scams and fraud, and how to report them.