The Verification Venue · a money myth two in three Americans hold
The Balance That Buys Nothing
You've been told to carry a small balance on your credit card to build your score. In a 2022 LendingTree survey, 65% of Americans believed it, and 79% of Gen Z. It is false. The interest you pay for it buys exactly zero points.
Here is the mechanism the myth misses. Your credit score doesn't watch your bank account. It reads the one balance your card reports to the bureaus, usually the balance on your statement date. Whether you then pay that statement off in full (owing the bank nothing, paying $0 interest) or carry it (paying interest for months), the number the bureau saw is the same. Same reported balance → same utilization → same score. The carrying only adds interest.
Set your own numbers and watch the two paths run side by side. Everything below recomputes live.
Reported utilization = statement balance ÷ credit limit = 10.0%
▸ Pay in full
Reported utilization
10.0%
Approx. score band
…
Interest paid, 12 months
$0
▸ Carry the balance
Reported utilization
10.0%
Approx. score band
…
Interest paid, 12 months
$240
Score band: identical · the only difference is $240 of interest that bought 0 points.
Slide the APR to 0% and the two columns become the same page. Slide it back up and only the right column moves, because interest is not an input to a credit score. There is no version of these controls where paying interest raises the band. That is the whole answer.
The one place the fear has a grain of truth
Being right for the wrong reason is still wrong, but the myth grew from a real seed, and the complete answer has to name it. It isn't "carry debt." It's a marginal, optional wrinkle about reporting a non-zero balance, and it never costs a cent of interest.
Credit scoring rewards activity. If every card you own reports $0 (no revolving activity at all), some FICO and VantageScore models read a hair less information about you than if one card reported a tiny balance. The documented gap is small and model-dependent: myFICO says a zero balance is simply not penalized, and Experian notes that keeping utilization at 0% "won't have much additional benefit" over a small single-digit balance. Neither bureau publishes a point figure; independent scoring experiments that toggled one file between 0% and 1% have reported a single-digit dip on the order of ~9 points. The fix, called AZEO (All Zero Except One), is to let a single card report a small balance, and then pay that statement in full.
Compare the two edge cases
Reported utilization
0.0%
Approx. score effect
−9 pts
Interest cost
$0
Both cost $0 interest. The AZEO edge is bought with a reported balance you pay off in full, never with a carried one. This is a single-digit-to-low-double-digit tweak on an 850-point scale, optional and model-dependent, not the "carry a balance" rule the myth teaches. Chasing it by revolving debt would just recreate the myth in miniature.
Statement balance vs. carried balance: the thing nobody separates
The confusion lives in one word doing two jobs. The reported (statement) balance is a snapshot the bureau sees once a cycle; it drives your score. The carried (interest-accruing) balance is what you still owe after the due date; it drives your interest. You can have a healthy reported balance and pay zero interest; that's what "pay in full" is. The myth quietly swaps the second for the first.
| Reported balance | Utilization | Score band | Interest / yr |
|---|
The check: every number recomputed in front of you
Two paths, one utilization. The utilization and interest are exact arithmetic from your inputs; the score band is an explicitly approximate map (see the caveats). The proof doesn't rest on the band's exact value; it rests on the band being a function of utilization only, so both paths share it while the interest column diverges.
Reproduce it offline: node research/carrying-a-balance-credit-score/verify-carrying-a-balance-credit-score.mjs. It recomputes utilization, the 12-month interest (two independent ways), the band's path-independence, and its monotonicity, and exits 0 only if all pass.
What's exact here, what's approximate, and every free choice
Exact. Reported utilization = statement balance ÷ credit limit. The 12-month interest on a carried balance held flat is S · APR, which is identical to summing twelve monthly charges of S · APR/12; the page shows that month strip, and the verifier checks both ways agree. Pay-in-full interest is exactly $0. None of this depends on any scoring model.
Approximate, and labelled so. The score band is a coarse, illustrative map from utilization to a tier and a rough point-delta. Real FICO and VantageScore models are proprietary and non-linear; the exact points depend on your whole file (history, mix, inquiries, the other 70% of the score) and on the model version. Never read the band as an exact score. Its only load-bearing job is to be the same on both paths.
Model-dependent generalizations, not laws. "Amounts owed" (which includes utilization) is ~30% of a FICO score, and the elite target is under ~10% utilization; both are standard FICO guidance, not physical constants. The "under 30%" rule of thumb and the "under 10% for 800+" target are conventions that shift by model. FICO and VantageScore weight and bucket utilization differently.
The AZEO / all-zero nuance is marginal and optional. The all-zero dip is documented as small and model-dependent (myFICO: zero is "not penalized"; Experian: 0% "won't have much additional benefit" over a small single-digit balance; neither states a point figure). We show −9 pts as a representative figure (the single-digit gap independent 0%-vs-1% scoring experiments have reported), not a guarantee. It is bought with a reported balance paid in full, never a carried one. Overstating it would recreate a sibling myth ("you must micromanage a small balance"), so: it's a tweak, not a rule.
Free choices you can change. Interest uses monthly compounding on a balance held flat; real cards use the average-daily-balance method with daily compounding, so an unpaid balance costs slightly more than S · APR (we show that steady figure as an honest floor). We treat utilization as a single aggregate figure; per-card utilization also matters to real models. APR is a fixed purchase APR with no promotional 0% period.
The survey number, dated. 65% (and 79% for Gen Z) is the 2022 LendingTree study of 1,323 U.S. consumers. An earlier 2021 LendingTree survey reported 45%; a 2023 U.S. News survey reported 48%. We cite 2022 by year to keep the figure pinned to its source.