Has the Proportion of Subprime Credit Card Accounts that Revolve their Balances Changed from 2015 to 2024?

Introduction

The credit card industry is one of the largest and most utilized financial industries in the United States, serving more than 200 million customers. According to CFPB’s 2025 Consumer Credit Card Market Report, purchase volume on consumer credit cards reached $3.6 trillion in 2024, while total balances exceeded $1.2 trillion (CFPB, 2025). Consumers who typically pay their balances in full each month are commonly referred to as “transactors”, while those who carry their balances month-to-month and incur interest charges are referred to as "revolvers." Credit cards serve two primary functions for consumers: convenient transitions and short-term borrowing. These two functions are bundled into the same product, yet have drastically different financial consequences for consumers in their respective categories. For transactors, the annual percentage rate (APR) on their credit card is effectively irrelevant; for consumers who revolve balances, that same APR determines the compounding cost of every dollar they borrow. 

The distinction between these two groups is not spread evenly across the American population; it is correlated with credit score, income, and financial stability. Cardholders with subprime credit scores, generally considered to be below a FICO score of 660, revolve their accounts at much higher rates than their prime and superprime counterparts. They also tend to encounter broader spreads, lower credit limits, and higher late fees. All of these factors mean that the experience of revolvers is categorically and quantitatively different from the average credit market statistics that are usually presented. 

Differences between the two credit groups are very significant, and the question less understood is whether the subprime segment has followed the same trajectory as the broader market. Specifically, has the proportion of subprime accounts that revolve changed in the past decade? And if so, in what direction and for what reasons? Using CFPB data, this paper finds that the recent improvements in the revolve rates are limited to prime or higher credit scores, potentially due to an increase in the number of prime or higher card accounts, not necessarily consumer-level payment behavior. The current publicly-available data is at the account level, making it difficult to get a clear view of changes in consumer payment behavior.

Definitions and methods

Key Terms

Transactor: A categorization for a credit card account that pays off its entire balance in full every time its statement is due. This excludes inactive accounts (accounts that have had no balance for three or more consecutive months). 

Revolver: An account that carries over some portion of its balance each month, resulting in interest accrual. 

Revolve Rate: The share of active accounts that are classified as revolvers within any period. 

Credit tier classifications: Deep subprime (<580) and subprime (580–619) tiers per CFPB definitions. Below-prime refers to both, as the CFPB sometimes reports them jointly. 

Persistent Debt: When an account's interest and fee charges are over half of the actual total payment amount during a calendar year. 

Data Sources

CFPB Consumer Credit Card Market Report (2015): Figure 18 provides a CFPB’s Credit Card Database (CCDB)-based time series of revolving rates by credit score tier from 2008 to 2015. 

CFPB Consumer Credit Card Market Report (2021): Figure 8 provides a Y-14+-based grouped bar chart of revolve rates by credit score tier (superprime, prime, near-prime, subprime, deep subprime, and overall) for each year from 2015 through 2020. The 2021 report also provides direct textual confirmation that there was no significant change in revolving rates for subprime and deep subprime accounts from 2018 levels in the 2019–2020 period.⁸

CFPB Consumer Credit Card Market Report (2025): Figures 49–50 provide a Y-14+-based quarterly time series and annual snapshot of revolve rates by credit tier from 2019 through 2024. Figure 52 provides aggregate delinquency data from the CFPB’s Consumer Credit Information Panel (CCIP). Figures 26, 45, and 51 provide the effective interest rate, minimum payment, and “persistent debt data,” respectively.

Federal Reserve Survey of Household Economics and Decisionmaking (SHED) (2015 and 2024/2025): Annual surveys capturing consumer-level self-reported credit card payment behavior. The 2015 survey's Table 15 provides income-stratified payment behavior; the 2024 survey provides the household-level aggregate revolving rate cited by Alexandrov (2025).

Results and Analysis

Figure 1 is the most complete publicly reconstructed time series of general-purpose credit card revolving rates by credit score tier for 2008–2024, made by combining other time series graphs showing revolving rates. 

Two findings stand out immediately. Firstly, below prime revolving rates remain elevated throughout the 2015 to 2024 period. Deep subprime rates between 83-90%, and subprime account rates between 79% and 86%. Excluding the pandemic-era lows in 2020-2021, rates remain relatively consistent among all subprime categories.

Second, and equally important, is the variance in the improvement, which increased dramatically among the different credit tiers. Prime cardholders revolve rates dropped by nearly 29 percentage points, whereas subprime cardholders rates dropped 3%. Not all tiers improved equally.

Credit Score Tier	2015 Revolve Rate	2024 Revolve Rate	Net Change (pp)
Deep subprime (<580)	89%	88%	−1
Subprime (580–619)	86%	83%	−3
Near-prime (620–659)	83%	72%	−11
Prime (660–719)	78%	49%	−29
Superprime & prime plus (720+)	45%	31%	−14
Overall (GP accounts)	65%	49%	−16

The data in Table 1 support the conclusion that subprime revolving rates have declined modestly over the decade, especially among prime and above-prime tiers. Deep subprime accounts declined by approximately 1 percentage point, and subprime accounts by approximately 3 percentage points within the Y-14+ dataset. This improvement is, however, dwarfed by the massive improvements among prime cardholders (−29 percentage points) and super prime cardholders (−14 percentage points), who drove the majority of the aggregate market decline of −16 percentage points.

The CFPB's 2021 report comments that “there was no significant change in revolving rates for subprime and deep subprime accounts from 2018 levels”, through the 2019–2020 period. This asserts the idea that the change apparent could be a result of the post-pandemic period rather than a steady long-run trend. The pandemic increased repayment rates across all credit tiers as a result of subsidies, but it is notable that below prime borrowers returned to pre-pandemic levels, while prime and super prime cardholders shifted much more durably towards lower revolving. 

The 2024 cross-sectional snapshot from Figure 50 of the CFPB 2025 report provides the most precisely reported current-period data. Figure 6 below shows revolve rates by credit tier for both general-purpose and private-label cards.

The revolve rate difference between deep subprime and superprime cardholders is 68 percentage points on general-purpose cards. Even after the decline in subprime revolving rates documented in Figure 1, below-prime borrowers remain in an almost entirely different financial situation from superprime cardholders. The market-wide average revolve rate of 49% hides the massive difference in rates. The metric is heavily weighted down by the large number of superprime accounts (which revolve at only 20%) and upward by the much smaller number of below-prime accounts (which revolve at 83–88%). (CFPB, 2025 CARD Act Report, Figure 3; general purpose accounts totaled roughly 50 million for superprime and 69 million for prime plus cardholders, versus about 11 million for subprime and 21 million for deep subprime, at year-end 2023). The average cardholder who revolves about half the time describes very few actual consumers.

SHED

The Federal Reserve's SHED surveys show a more consumer-level view. The 2015 SHED found that among adults with at least one credit, around 42% always paid in full, meaning 58% carried a balance at least once during the year. By the 2024 SHED survey, this household-level revolving rate declined to approximately 46%. 

The SHED income-tier data shows lower-income households tend to have the highest revolving rates in both periods, which is consistent with the common correlation often found between income, credit score, and revolving behavior. The aggregate improvement from 58% to 46% remains relatively consistent with the CFPB data, but it is important to recognize that this improvement likely came from higher-score borrowers in each category. 

Delinquency

While revolving rates and delinquency are distinct statistics, delinquency rates can provide some insight into the true state of the consumer credit market. Revolvers carry balances but continue to at least pay minimum payments, while delinquent accounts do not make any payments. Figure 2 shows aggregate 60+ day delinquency rates for general-purpose and private-label cards from 2014 to 2024.

Delinquency rates fell to a multi-decade low of 1.4% in 2021, mostly because of pandemic stimulus support, before rebounding to 4.0% in 2023, the highest level since the fallout of 2009. Although delinquency has fallen since then, it remains elevated compared to pre-pandemic levels. The CFPB attributes most of this increase to a laxer practice of underwriting new vintages, starting around 2021-2022, when issuers began to expand credit to riskier consumers during the post-pandemic period (Fulford & Gibbs, 2024). This underwriting-driven explanation is not without dispute, however. Using Federal Reserve Y-14M data through December 2024, Stavins (2025) finds that the erosion in underwriting standards was temporary: credit scores for new accounts opened in 2021 and 2022 initially fell below pre-pandemic levels, but had already begun reversing later in 2022, and by the end of 2024 the average credit score of new account holders exceeded its pre-COVID level. This suggests that looser underwriting is unlikely to fully explain delinquency that remains elevated through 2024, and that some of the apparent “risk shift” may have been a temporary pandemic-era artifact rather than a durable change in underwriting practice.

Additionally, higher delinquency rates mean that more accounts are failing to make minimum payments. Since those accounts were likely revolving before they became delinquent, some of the decrease in subprime revolving rates could be attributed to a reclassification of accounts from revolving into delinquent accounts.

Account vs. Consumer Level 

One major issue with everything discussed thus far is the fundamental difference in how statistics are measured. All CFPB revolve rate data is at the account level. The CFPB 2025 report demonstrates that subprime consumers typically hold fewer credit cards than superprime consumers. who hold significantly more accounts, on average (CFPB, 2025 CARD Act Report, Figure 4). This means that for a superprime consumer with six cards who revolves on one account, the account level data would show around 17% revolving, whereas at the consumer level, the consumer would be 100% revolving. The account-level data dramatically understates the true revolving behavior in the market. 

For a subprime consumer, who typically holds only one or two cards and revolves on one of them, the account-level and consumer-level revolve rates are much closer together. This has an important, and somewhat counterintuitive, implication: because superprime consumers hold far more cards per person than subprime consumers (CFPB, 2025 CARD Act Report, Figure 4), a meaningful share of the low 20% superprime account-level revolve rate is mechanical – it reflects a revolving consumer’s balance being averaged in with several non-revolving transactor cards, not necessarily less revolving behavior. Subprime consumers, holding fewer cards, do not benefit from this same dilution, so their account-level rate sits closer to their true consumer-level rate. Correcting for this suggests the reverse of what a naive reading of Figure 6 implies: the 68 percentage-point account-level gap between subprime and superprime likely overstates, rather than understates, the true difference in how often subprime and superprime consumers carry a balance at the consumer level. Put differently, some of superprime cardholders’ apparently low revolving rate is an artifact of how many cards the average superprime consumer holds, rather than evidence that they revolve balances less often than the account-level statistic suggests.

This problem can be even more significant if you factor in how much the credit market has changed over the past decade. As the prevalence of reward cards and signup bonuses exploded in the recent era, the average number of cards per consumer grew, with growth concentrated among prime and superprime consumers: CFPB data show that nearly two-thirds of superprime cardholders hold three or more credit cards, compared to just 42% of deep subprime cardholders (CFPB, 2025 CARD Act Report, Figure 4; see also CFPB Consumer Credit Trends, Origination Activity.) Along with this, the typical superprime cardholder steadily increased the number of cards that they used as a transactor card. Someone may open and hold a few different cards for the sole purpose of gaining rewards points and benefits, and only makes small and easy-to-pay-off purchases to benefit from the rewards. If they carried a balance on one card in 2015, and still carry a balance on that card in 2024 but now hold three additional rewards cards they pay off in full, statistically, it would show that they revolve less at the account level, despite continuing to revolve on their original card. Every single one of these new transactor accounts drags down the total account-level revolving rate, even when borrowing behavior never changed at all.

This changes the entire central finding of this paper. The headline result from Figure 1 and Table 1 was that prime revolve rates fell by 29 percent and superprime by 14 percent, while subprime barely moved. But a very meaningful portion of that prime and superprime decline is likely not people revolving their accounts less, but people holding more cards. Simply put, the denominator grew. Subprime consumers, who never got this dilution effect, saw little change, and their account level number would likely be much closer to the consumer level reality. The comparison, that nearly every credit tier improved except subprime, is therefore partially comparing a larger social and financial behavior change rather than an assessment of financial strength. 

Conclusion

The goal of this paper was to answer one question: has the proportion of revolving subprime credit card accounts changed from 2015 to 2024? The short answer is yes, but only slightly, and not in a way that means much for cardholders actually carrying the debt. 

According to the Y-14+ data, subprime account revolving rates fell by around 3 percentage points over the past decade, with deep subprime falling about 1%. Both of these tiers remained highly elevated, sitting at around 83 to 90 percent early every year. What did change, however, was the rest of the credit card market. Prime revolve rates dropped 29%, and superprime dropped 14%. The supplemental data muddies the water, however, with delinquency rates and the proportion of borrowers only making minimum payments climbing in the subprime category climbing higher. 

The single most important takeaway of this research is that every trend and tier comparison presented in this paper is built on account-level data, which overexaggerates the higher credit tiers over the lower ones. The small 3 percentage point decrease in the lower tiers is likely representative of the real trend, but the dramatic declines everywhere else are probably not. Subprime revolving behavior has stayed essentially the same over the past decade, and the appearance of a broadly improving market is likely a product of how the data is measured instead of a change in consumer behavior.

References

Alexandrov, Alexei. 'What Drives Changes in Credit Card Borrowing Costs? Monetary Policy, Metrics, and Fees.' SSRN Working Paper No. 6637518 (2025). Sponsored by the Consumer Bankers Association.

Ausubel, Lawrence M. 'The Failure of Competition in the Credit Card Market.' The American Economic Review 81, no. 1 (1991): 50–81.

Agarwal, Sumit, Souphala Chomsisengphet, Neale Mahoney, and Johannes Stroebel. 'Regulating Consumer Financial Products: Evidence from Credit Cards.' The Quarterly Journal of Economics 130, no. 1 (2015): 111–164.

Consumer Financial Protection Bureau. The Consumer Credit Card Market (2013). http://files.consumerfinance.gov/f/201309_cfpb_card-act-report.pdf.

Consumer Financial Protection Bureau. The Consumer Credit Card Market (December 2015). http://files.consumerfinance.gov/f/201512_cfpb_report-the-consumer-credit-card-market.pdf.

Consumer Financial Protection Bureau. The Consumer Credit Card Market (December 2017). https://files.consumerfinance.gov/f/documents/cfpb_consumer-credit-card-market-report_2017.pdf.

Consumer Financial Protection Bureau. The Consumer Credit Card Market (September 2021). https://files.consumerfinance.gov/f/documents/cfpb_consumer-credit-card-market-report_2021.pdf.

Consumer Financial Protection Bureau. The Consumer Credit Card Market Report (2025). https://files.consumerfinance.gov/f/documents/cfpb_consumer-credit-card-market-report_2025.pdf.

Consumer Financial Protection Bureau. Making Ends Meet in 2024. November 2024.

Federal Reserve Board. Report on the Economic Well-Being of U.S. Households in 2015. Board of Governors of the Federal Reserve System, May 2016.

Federal Reserve Board. Report on the Economic Well-Being of U.S. Households in 2024: Fact Sheet. Board of Governors, 2025.

Federal Reserve Board. Report on the Economic Well-Being of U.S. Households in 2025: Fact Sheet. Board of Governors, 2026.

Nevins, Joanna. 'Sticker Price vs. Reality: Why APR Doesn't Tell the Whole Story of Credit Card Costs.' Policy Brief, May 14, 2026.

Stavins, Joanna. 'Are Recently Issued Credit Card Accounts More Risky?' Federal Reserve Bank of Boston, Current Policy Perspectives (April 2, 2025).

UK Financial Conduct Authority. Credit Card Market Study: Persistent Debt and Earlier Intervention. Policy Statement PS18/04, February 2018.