Blog

Why every 'average credit card APR' you read is a different number

Why Fed figures for accounts charged interest, all accounts and advertised-rate trackers differ, with the 2026 numbers and a checklist for reading any average.

Open three articles about credit card rates in the same week and you can find three averages: one near 22 percent, one near 21 percent, one near 20 percent. None of them is wrong. Each answers a different question. One counts only accounts that were actually charged interest. One counts every card account, whether or not the holder paid interest. One averages the rates that card issuers advertise on their websites. Put those side by side without the labels and the "average credit card APR" looks like a number that cannot make up its mind. This article explains what each figure measures, gives the latest published values with their dates and sources, shows with arithmetic why the numbers diverge, and ends with a checklist for reading any average before you let it shape a decision. It also shows how little a national average tells you about the rate on your own card.

The figures as of November 2026

Three published measures matter most. Here is what they said, with sources and dates.

The Federal Reserve's G.19 consumer credit release (released November 6, 2026) reports two APR series for commercial banks:

  • Accounts assessed interest: 22.36 percent in the third quarter of 2026, 22.15 percent in the second quarter of 2026 and 22.83 percent in the third quarter of 2025. Annual figures were 22.89 percent for 2024 and 22.32 percent for 2025.
  • All accounts: 21.19 percent in the third quarter of 2026, 20.94 percent in the second quarter of 2026 and 21.39 percent in the third quarter of 2025. Annual figures were 21.58 percent for 2024 and 21.22 percent for 2025.

Bankrate's weekly tracker, which is a private publication and not an official statistic, showed an average advertised APR of 19.71 percent in October 2026, below its own record of 20.79 percent on August 14, 2024, according to a news report that carried its figures. It averages the midpoints of advertised APR ranges across 111 cards from the 50 largest issuers.

One more number often appears in articles: the highest the Fed's accounts-assessed-interest series reached, reported by one news outlet as 23.37 percent in August 2024. That record comes from a secondary source, so treat it as reported, not confirmed.

Put the three current figures in a row and the spread is plain:

Measure Latest value Period
Fed: accounts assessed interest 22.36 percent Q3 2026
Fed: all accounts 21.19 percent Q3 2026
Bankrate: average advertised APR midpoint 19.71 percent October 2026

From top to bottom the range is 22.36 - 19.71 = 2.65 percentage points. Before explaining why, note that the quarters and months do not line up either, which is one more reason to compare carefully. The timeline of how these figures moved from 2024 to 2026 is in How average card APRs moved from 2024 to 2026 and what drove them.

Difference one: who is in the average

The Fed's two series differ by which accounts they include. "Accounts assessed interest" counts only accounts on which the issuer charged interest during the period, which means accounts that carried a balance past the grace period. "All accounts" also counts accounts that paid in full and owed no interest. Those accounts still have an APR, and it is typically lower than the APR on accounts that carry balances, partly because borrowers who pay in full are often those with the strongest credit, who are offered the lowest rates.

The result: the "all accounts" average is lower, by 22.36 - 21.19 = 1.17 percentage points in the third quarter of 2026. The accounts assessed interest are the ones paying the higher rates, so the average of that subset is higher. If you carry a balance, 22.36 percent is the more relevant benchmark. If you want the typical rate on the card in a typical wallet, 21.19 percent is closer.

A small illustration

To see the mechanism, take a made-up group of six card accounts and a simple average. The Fed's actual series are computed from reported data, and its release notes explain the method and any weighting, so the example is only to show direction.

Account APR Paid in full this period?
A 18 percent Yes
B 19 percent Yes
C 21 percent Yes
D 24 percent No
E 26 percent No
F 28 percent No
  • All six accounts: (18 + 19 + 21 + 24 + 26 + 28) / 6 = 136 / 6 = 22.67 percent
  • Only the three assessed interest (D, E, F): (24 + 26 + 28) / 3 = 78 / 3 = 26.00 percent

The all-accounts average is lower because accounts that paid in full have lower rates in this example. The gap in the example, 3.33 points, is bigger than the real one, since real data have more variation, but the direction is the same as in the published series.

Difference two: actual versus advertised

The Fed's numbers describe what issuers charged on real accounts. Bankrate's describes what issuers advertise for new accounts. These are not the same population.

An advertised APR is a range. A card might show a range of 17.99 percent to 27.99 percent, variable, in its schumer box. Bankrate takes the midpoint of each range, which for that card would be (17.99 + 27.99) / 2 = 22.99 percent, and averages the midpoints across cards. Whether a given applicant is offered the low end, the high end or the middle depends on their credit file and the issuer's models. Many applicants with weaker credit land toward the top, and the average of midpoints does not know how many applicants land where.

Two further differences widen the gap:

  • Advertised rates cover cards open to new applicants. Existing accounts opened years ago may carry different rates, and their rates have moved with the index since.
  • Not every card is in the sample. A tracker built on 111 cards from the 50 largest issuers leaves out smaller issuers and probably many store and credit union cards, whose rates can differ. A Fed series that covers commercial banks excludes credit unions and non-bank lenders.

So 19.71 percent and 22.36 percent are not in conflict. One says that, in October 2026, the midpoints of the advertised ranges for a particular set of cards averaged 19.71 percent. The other says that, in the third quarter of 2026, commercial bank accounts that were charged interest averaged 22.36 percent. The second is higher in part because accounts that carry balances tend to be on higher rates than the midpoint of a range that includes strong-credit offers.

Two applicants, one advertised range

Make the midpoint problem concrete. Say a card advertises a variable APR of 17.99 to 27.99 percent, and two people are approved. The strong-credit applicant, Xena, is offered 17.99 percent. The applicant with a thinner file, Yusuf, is offered 27.99 percent. The midpoint that a tracker records is 22.99 percent, and neither of them pays it.

On a $3,000 balance carried for a year (a simplification):

  • Xena: $3,000 x 0.1799 = $539.70
  • Yusuf: $3,000 x 0.2799 = $839.70
  • Difference: $300.00 a year, from the same card

A tracker that records 22.99 percent for this card has told you the middle of the offer, not what either borrower faces. This is why a published average of advertised ranges is a statement about the market's menu, not about anyone's bill. It is also why the schumer box on your own offer, which states the range and the rate type, matters more than any average.

Difference three: the clock

The Fed reports quarterly and annual values; Bankrate reports weekly. A quarter's value blends three months, so a rate change in the middle of the quarter shows up only partly. The Federal Reserve raised the federal funds target range to 3.75 to 4.00 percent on September 16, 2026, and the bank prime rate rose from 6.75 to 7.00 percent on September 17. Most card APRs are variable and tied to prime, so those cards' rates move within one or two billing cycles. A tracker measured in October may already show the change, while the quarter ending in September shows only about two weeks of it. That is another reason you should never compare a weekly figure with a quarterly one without noting the dates. What happens to your card's APR in the weeks after a Fed rate change traces how a policy move reaches your statement.

Within the Fed's own series, the quarter-to-quarter moves are small: the accounts-assessed-interest series went from 22.15 percent in the second quarter to 22.36 percent in the third, a rise of 0.21 points, and from 22.83 percent a year earlier to 22.36 percent, a fall of 0.47 points. Year over year it is lower, quarter over quarter it is higher. Which one a headline picks changes its story.

Difference four: mean, median, weighted

An average can be computed in several ways, and it can change the number without changing a single account. A simple mean treats each account equally. A balance-weighted mean gives more weight to accounts with larger balances, whose rates can differ from those on small balances. A median is the middle account. Each is legitimate; each yields a different value. When an article quotes "the average APR" without saying how it was computed, you cannot tell which it is. The source's own notes usually say, which is why the Fed's release notes and any tracker's methodology page are worth a minute.

What the national average says about your card

Very little. It describes a distribution, and you are one point in it. Issuers set rates for individual accounts based on prime rate plus a margin that depends on the applicant's risk, the card type and the issuer's strategy, covered in My APR is 'prime plus' a margin. How do I work out what I will pay? and Why is my card's APR higher than the national average I keep reading?. A card that looks high against a headline average might be ordinary for its type, and a card below the average might still be costly for the borrower who carries a balance on it.

What the number is useful for is a rough sanity check. If your card's purchase APR, shown in your cardholder agreement and your statement, is several points above the Fed's accounts-assessed-interest figure, that is information: it may reflect your credit profile, the type of card or a penalty rate, and it may be worth a call or a look at alternatives.

What a few points cost

The arithmetic of a difference is more useful than the difference itself. Say you carry a $4,000 balance for a year with no change, a simplification, since real balances move.

  • The gap between the two Fed measures in the third quarter of 2026 is 1.17 points. On $4,000, 1.17 percent is $46.80 a year.
  • Say your own APR is 27.99 percent, which is 27.99 - 22.36 = 5.63 points above the accounts-assessed-interest average. On $4,000, 5.63 percent is $225.20 a year, or about $18.77 a month.

A rate gap of a few points matters at a four-figure balance, and an average that moves by a fraction of a point between quarters does not change what you owe by much. The credit card interest calculator shows what any rate costs on any balance, and the APR, daily rate and APY converter turns an APR into a daily or monthly rate.

A short worksheet: where does your card sit?

You can do this in ten minutes with a recent statement.

  1. Find your purchase APR. It is on your statement, usually near the interest charge calculation, and in the cardholder agreement. Write it down with the statement date. Remember that it is probably variable and may have changed since you opened the account.
  2. Pick the matching benchmark. If you carry a balance, use the Fed's accounts-assessed-interest figure (22.36 percent for the third quarter of 2026). If you pay in full, the rate matters little and the all-accounts figure (21.19 percent) is the closer comparison.
  3. Take the difference. Say your APR is 25.24 percent. 25.24 - 22.36 = 2.88 points above the benchmark.
  4. Convert it to dollars. On a balance of $2,500 carried for a year, 2.88 percent is $72.00. That is the yearly cost of being 2.88 points above the benchmark at that balance, a number you can weigh against the effort of asking for a lower rate or moving the balance.
  5. Decide what the gap means. A gap can reflect your credit profile, a card type that carries higher rates or an old penalty rate. If the gap is large and your credit has improved, a call to the issuer or a look at other offers can be worth it. Do not apply for several cards at once on the strength of a headline average; each application can involve a hard inquiry.

The point of the exercise is that the benchmark is a ruler, not a verdict.

How to read any "average credit card APR"

Before you let a number into a decision, ask:

  1. Who published it? The Federal Reserve, a trade group, a rate-comparison website, a news outlet repeating one of them?
  2. What does it count? Accounts charged interest, all accounts, advertised ranges, new offers?
  3. Which date or period? A quarter, a year, a single week? Is it the latest?
  4. Whose cards? Commercial banks only, the largest issuers, a sample of cards?
  5. How is it computed? A simple mean, a weighted mean, a median, a midpoint of ranges?
  6. Is it a rate you could get? An average of what people pay is not an offer.
  7. Does it say "APR" or something else? A card's APR is a variable rate; a fee or a promotional rate is a different thing.

If a source does not answer at least the first four, treat its figure as an anecdote.

A note on rate caps and averages

Proposals to cap card rates at 10 percent have appeared in 2025 and 2026, and as of the coverage found for July 2026 none had passed; a cap is proposed, not law, as of that date. When an article argues about what a cap would do, it will quote one of the averages above, and which one it picks shapes the story. A cap of 10 percent against 22.36 percent is a gap of 12.36 points; against 19.71 percent, it is 9.71. Read the label before you read the claim.

Key takeaways

  • The Fed's G.19 reported 22.36 percent on accounts assessed interest and 21.19 percent on all accounts for the third quarter of 2026, and Bankrate's advertised midpoint average was 19.71 percent in October 2026: three different questions, three different numbers.
  • Accounts assessed interest average higher because the accounts that carry balances tend to be on higher rates than accounts that pay in full, which is why the all-accounts figure is lower.
  • Advertised APRs are ranges, and what an applicant is offered depends on their credit, so an average of midpoints is not what people actually pay.
  • Compare only like with like: same source, same measure, same period, and note whether a figure is quarterly or weekly.
  • A national average is a sanity check, not a forecast of your rate; the number that matters is the one in your cardholder agreement, and a few points on a $4,000 balance is real money, as in the $225 example.

Terms used in this article

Related tools

Latest on this topic