Private-Label Credit Cards: The Retailer Economics, Borrower Costs and Credit Risks Behind the Store Card

The central question: what does the store card actually finance?

A private-label credit card connects three businesses: selling merchandise, managing a customer relationship and lending money. The checkout discount is the visible part. Behind it sit a revolving receivable, a funding obligation, a distribution contract and a set of incentives that can pull the retailer, lender and customer in different directions. Understanding the product requires following those cash flows rather than treating the retailer’s logo as a complete description of the credit.

This report focuses on U.S. consumer private-label cards and the adjacent retail co-brand market. It separates observed market data from issuer examples and explicitly hypothetical calculations. Sources were checked through October 4, 2026; the broad market observations are older than the latest company filings. Neither a successful loyalty program nor a profitable lender, by itself, demonstrates that the financing improves household welfare.

The product can be useful when a customer obtains a genuine discount on a purchase already planned, understands the repayment terms and avoids costs that exceed that benefit. The same design can become expensive when the customer revolves at the ordinary APR, mistakes deferred interest for unconditional 0% financing, or borrows more because approval and a discount arrive together. The distinction is behavioral and contractual, not simply whether the card has a familiar brand.

Closed-loop, co-brand and dual-card are different architectures

Regulation Z defines a private-label credit-card account as one usable only at a single merchant or an affiliated merchant group. That is a restriction on acceptance, not a statement that the retailer owns the loan. “Open-end” describes reusable credit; “closed-loop” describes where it can be spent. A card can therefore be open-end credit and closed-loop payment at the same time. [1]

Retail co-brand cards can run on a general-purpose network and be used away from the sponsoring merchant. Synchrony also distinguishes Dual Cards, with partner-channel and outside-network uses, and reports interchange income on outside-partner Dual Card transactions and general-purpose co-brands. Its product set includes private label, co-brand and installment credit. The company is consequently not a pure proxy for closed-loop cards. [4]

For analysis, keep four separate labels: the customer-facing brand, the legal creditor, the acceptance network and the repayment contract. A retailer-branded debit product has no revolving card loan merely because its rewards resemble those of a credit card. A fixed installment loan at the same checkout also remains a different contract. This classification should come before comparisons of interest rates, loss rates or market share.

Scroll horizontally to see all columns.

Adoption: a large installed base with a declining account count

The CFPB’s December 2025 report puts year-end 2024 open private-label accounts at 185 million, versus 608 million general-purpose accounts. Its comparable private-label series peaks at 290 million in 2018. These are accounts, not unique consumers or annual purchases. The account series uses the Consumer Credit Information Panel; other report metrics draw on large-issuer account data plus specialized issuers, or mass-market issuer surveys. Those universes are not interchangeable. [2]

Original arithmetic using those rounded counts puts private-label accounts at about 23.3% of the combined 2024 account total and the 2018–2024 decline at about 36.2%. Neither calculation establishes the share of households using these products. One person may have several cards, an open account may be inactive, and a shift from private label to co-brand can change the category without eliminating a retailer relationship.

A useful adoption funnel runs from eligible shoppers to applications, approvals, first purchases, repeat purchases and sustained active accounts. Growth at one stage can mask deterioration at another. A holiday application drive that adds dormant accounts is different from growth in repeat use. Likewise, stronger billed balances can reflect slower repayment rather than more customers choosing the product.

Price and access: compare the same population and the same date

For 2024, the CFPB reports average APRs of 31.3% on private-label cards and roughly 25% on general-purpose cards. Its mass-market issuer data show approval rates of 54% and 41%, respectively; the share of accounts making only minimum payments was about 20% versus 15% in its account-level analysis. These are separate measures from different datasets, not the characteristics of a single representative borrower. [2]

Higher aggregate approval does not prove that every applicant is more likely to qualify for a store card. Credit-score mix, applications solicited, channel, requested line and lender policy all influence the observed rate. The right access test compares otherwise similar applicants and then follows their payment outcomes. Approval without sustainable repayment can be a weak form of inclusion.

An advertised purchase APR, interest actually collected and the customer’s all-in cost also answer different questions. Promotions and grace periods can lower collected yield below a headline APR; late fees or other charges can raise total customer cost. Average APR is therefore neither the bank’s realized return nor a personalized quote. Current applicants need the actual agreement and offer, not a dated market average.

The partnership contract is the economic engine

The CFPB’s 2024 retail-card study describes issuers typically handling underwriting and servicing while retailers market the card and run loyalty programs. Agreements can allocate data rights, exclusivity, operating responsibilities, financial compensation and termination rights. Some arrangements also involve merchant support for promotional financing. These are observations from reviewed contracts, not uniform terms imposed on every program. [3]

Synchrony says many large-partner programs share economics above contractual thresholds. Crucially, the formula can use agreed expense measures rather than actual expenses, so a rise in the bank’s real cost does not necessarily reduce partner payments. Its funding includes deposits, securitizations and unsecured notes. Those disclosures describe how one large issuer operates, not a universal merchant price list. [4]

The commercial negotiation therefore has more dimensions than a revenue-share percentage. A larger merchant payment may come with a longer exclusive term, marketing commitments or a different risk allocation. A seemingly cheaper promotion can require lower acceptance margins elsewhere. Comparing two bids requires the same customer mix, approval policy, rewards promise, service standard, expected life and exit assumptions.

Risk ownership and economic sensitivity must also be separated. A merchant can be affected by lower profit-sharing income even when the bank owns the receivables. Conversely, a lender may retain credit losses but possess contractual levers that redistribute some earnings impact. The operational question is who is legally obliged to do what; the economic question is whose cash flow changes when the portfolio deteriorates.

Issuer economics: a transparent illustration, not an industry estimate

Consider an entirely hypothetical annual portfolio with average receivables of $100. Assume $25 of net customer finance income, after reversals; $4 of funding expense; $8 of net credit losses; $6 of operating and fraud costs; and $3 of partner compensation. The remaining $4 is a simplified pretax contribution before any separate capital charge. Every input is an assumption. It is not an estimate of a named issuer’s margin.

The denominator matters: this illustration uses average balances, not purchase volume. Merchant fees may be quoted on new purchases, funding expense depends on funded balances and time, and account-servicing costs may be driven by active accounts. Combining percentages without converting them to dollars on a consistent basis produces a meaningless margin. Promotional balances also need their own yield and duration assumptions.

Now raise assumed losses from $8 to $12, holding everything else constant. If partner compensation stays fixed, contribution falls to zero. If a hypothetical contract reduces the $3 payment to $1, contribution falls to $2 instead. That is partial risk sharing, not loss elimination. Actual thresholds, floors, caps and expense definitions determine the result; the example deliberately does not infer confidential contract terms.

For reported earnings, use the company’s allowance and provision framework instead of treating realized losses as a second expense on top of the provision. Reserve building can make current expense exceed current charge-offs; reserve release can do the opposite. A cash-performance view and a GAAP earnings view can both be useful, but they must be reconciled rather than mixed.

Scroll horizontally to see all columns.

Retailer economics: incrementality is harder to prove than card spending

Macy’s reported $328 million of net credit-card revenue for the 26 weeks ended August 1, 2026, compared with $306 million a year earlier, and $9.548 billion of net sales. The card figure is net program revenue, not purchase volume or the bank’s gross interest income. The disclosure covers Macy’s card relationships; it does not isolate private-label-only economics. [7]

Original calculation: $328 million divided by $9.548 billion is approximately 3.44%. That illustrates scale relative to net sales, not a retail operating margin or a profit share. Card revenue can matter disproportionately when merchandise operating margins are thin, but attributing all of it to incremental value requires considering the costs and customer behavior the program changes.

The causal merchant test is counterfactual: how much profitable merchandise would have been sold without the card? Existing loyal shoppers may be the most likely to apply, making their larger baskets evidence of selection rather than evidence that credit created those purchases. A credible analysis uses controlled offers or carefully matched cohorts, includes returns and discounts, and measures repeat contribution after rewards and funding support.

Suppose a program moves a purchase from another accepted payment method onto the store card without changing what the customer buys. It may alter acceptance costs, program income and customer data while adding no merchandise sales. If it brings a future purchase forward, this quarter’s lift can be next quarter’s shortfall. Card penetration alone cannot distinguish these mechanisms.

Underwriting: the initial limit is only the first decision

Regulation Z requires consideration of the consumer’s ability to make required minimum periodic payments when an account is opened or a credit limit is increased, using income or assets and current obligations. The rule permits several information sources and has additional provisions for younger applicants. Meeting this minimum-payment test is not a promise that a customer will clear a promotional balance before its deadline. [8]

A disciplined analytical design separates identity verification, credit eligibility, affordability and line assignment. A legitimate identity can still be unaffordable; a high score can coexist with a short-term cash squeeze; a manageable first purchase can be followed by additional draws. The reusable line creates future exposure that a one-purchase approval does not fully describe.

Monitor cohorts by application channel, starting line, score band, promotional term and merchant category. Ask whether apparent improvement survives when the mix is held constant. Tightening can reduce losses while lowering approvals and merchant sales. Conversely, growth can temporarily dilute delinquency ratios with young accounts that have not had time to miss enough payments. Cohort seasoning is essential to interpreting the trade-off.

Credit access should also be evaluated alongside explanations and fair treatment. Regulation B sets adverse-action notification requirements, including specific reasons or the applicable process for obtaining them. A complex model or retailer-branded application does not make those obligations disappear. For governance, decision explanations need to reflect the actual factors driving the action. [16]

Fraud is a separate problem from willingness or ability to repay

The OCC’s credit-card handbook discusses fraud detection, identity-theft red flags, security controls, investigation and reporting. It emphasizes understanding fraud by type rather than relying only on an aggregate loss total. Supervisory material is a control framework; it is not evidence that any particular retailer or issuer has suffered a specific incident. [9]

For a private-label program, an analytical threat map should distinguish stolen identity at application, account takeover, fraudulent purchases, refund abuse and misuse by an authorized participant. Closed acceptance narrows where a card can be spent; it does not prevent goods from being resold or an account from being compromised. Controls at checkout and in account recovery must work together.

Friction has two costs: letting a bad transaction through and rejecting a legitimate one. The relevant evaluation measures confirmed losses, review expense, false declines, recovery time and customer harm. A lower fraud rate achieved by rejecting an entire channel can be commercially expensive. A higher approval rate bought by relaxing identity controls may create losses that are invisible in a conventional credit-score comparison.

Synchrony’s Q2 2026 filing excludes third-party fraud losses and unpaid finance charges and fees from its net charge-off measure; those third-party fraud losses are recorded elsewhere in expense. Thus credit charge-offs alone do not measure the complete cost of failed transactions and accounts. An analyst needs a reconciliation before comparing issuers or products. [5]

Reading 2026 issuer performance without calling it the market

For Q2 2026, Synchrony reported $1.364 billion of net charge-offs and a 5.43% annualized rate on average loan receivables including held-for-sale balances, versus 5.70% a year earlier. Its credit-card product category was 5.34%; that category is broader than private label. These are issuer-specific observations, not the national loss rate for store cards. [5]

Bread Financial reported a Q2 2026 net loss rate of 6.98%, compared with 7.88% a year earlier. Its definition annualizes net principal losses divided by average credit-card and other loans, using daily average balances. It also reported $316 million of net principal losses and $313 million of provision expense; the difference reflects a $3 million reserve release. These measures cover its reported portfolio rather than a private-label-only cohort. [6]

The two reports support a narrow conclusion: their published loss measures improved year over year. They do not establish that one lender underwrites better, or that a given store card is safer. Product mix, borrower distribution, recovery policies, portfolio purchases and sales, and accounting classifications can all affect the comparison.

Do not compare an annualized loss rate on average receivables with BNPL losses divided by annual originations and call the lower percentage safer. One measures a stock financed through time; the other measures a flow of newly issued loans. A meaningful comparison uses consistent principal, exposure duration, vintage maturity, fraud treatment and recoveries. Delinquency also needs an explicit overdue threshold and either an account or balance denominator.

Deferred interest: the deadline changes the payoff

A true 0% promotional APR and deferred interest are different. Under the former, interest attributable to the zero-rate period is not retroactively imposed. Under a deferred-interest arrangement, interest accrues conditionally and can become payable if the promotional conditions are not satisfied. Regulation Z’s advertising interpretation explicitly makes this distinction. [10]

The CFPB warns that required minimum payments usually do not pay off a deferred-interest purchase before the promotional period ends. The offer and agreement determine the deadline and applicable consequences; the ordinary statement due date should not be assumed to be the promotional expiration date. The practical task is to identify the promotional balance and the amount needed to extinguish it on time. [11]

An original simplified illustration shows the discontinuity. Start with a $1,200 purchase, 12 equal monthly periods and a 30% annual rate modeled as 2.5% monthly simple accrual on each month’s opening principal. Pay $99 at each month-end. After 12 payments, $12 of purchase principal remains. The sum of the opening balances is $7,866, so modeled deferred interest is $196.65. If all that interest becomes payable, the remaining amount is $208.65 rather than $12.

This is a teaching model, not a quotation from a card agreement. It assumes no other purchases, fees, compounding or payment-allocation complications; real cards generally use daily calculations and contract-specific terms. Its lesson is the payoff threshold: being nearly finished need not mean nearly all finance charges are avoided. For the deeper mechanics, see this publication’s existing deferred-interest and payment-allocation research.

Scroll horizontally to see all columns.

Payment allocation and the grace period are separate moving parts

Regulation Z generally directs payments above the required minimum to the highest-APR balance first. For allocation purposes, a deferred-interest balance is treated as zero-rate during its promotion, with special treatment in the last two billing cycles. The issuer may also follow a consumer’s qualifying allocation request. The rule does not require the minimum-payment portion to be allocated in the same way. [12]

This is why an account-level payment can be misleading. A customer may pay more than the minimum yet reduce a different balance from the one approaching its deadline. The right reconciliation follows each promotional balance, its expiration and the payment allocation actually posted. A lender’s statement or portal should make that account arithmetic inspectable.

A grace period is another concept: when available and its conditions are met, it can allow purchase interest to be avoided by paying the full relevant balance on time. The CFPB notes that grace periods are not universally required and can be lost when balances are carried, affecting new purchases as well. A promotion does not automatically resolve how the rest of the account is treated. [13]

For a basic cost screen, hypothetical rewards of $50 on a $1,000 purchase are offset by $50 of simple interest on a $1,000 average interest-bearing balance for two months at 30% annually. Actual results depend on average balances, grace periods and repayment. The point is to compare dollar benefits with dollar costs over the expected holding period, rather than comparing a one-time reward percentage directly with an annualized interest rate.

Funding, liquidity and the merchant relationship can deteriorate together

A revolving book creates a timing problem: the lender pays for purchases before collecting the balances, and customers decide how quickly to repay within the contract. Funding must be available for that interval. Portfolio yield alone cannot answer whether the lender can refinance itself or fund additional draws during stress.

Original stress analysis should combine slower repayments, higher losses, funding repricing and weaker merchant traffic rather than shocking each in isolation. A merchant’s deterioration can hurt new account acquisition and repeat use at the same time that existing borrowers encounter financial pressure. A large partnership can consequently be a distribution concentration and a credit concentration, even when individual customer balances are small.

A securitization analysis adds its own layer: inspect eligible receivables, cash-allocation rules, seller interests, performance triggers and servicing obligations in the actual documents. It is not enough to observe that an issuer uses asset-backed funding. Trigger definitions and available liquidity determine whether cash that once supported new lending must instead amortize investors’ balances.

The useful decision metric is risk-adjusted contribution through the cycle after operating costs, partner payments and capital needs, paired with a funding survival test. A program that appears attractive in a benign year can become fragile if its contract commits cash to partners faster than its own funding and credit costs can adjust.

Competition with BNPL: compare the job, the term and the customer

The OCC’s short-term BNPL bulletin concerns loans with four or fewer installments and no finance charge; it explicitly excludes longer or interest-bearing loans from that particular scope. It describes merchant-discount economics and cautions about repayment visibility, returns, disputes and operational risk. That narrow product is a useful comparator, not a definition of everything marketed as BNPL. [18]

A pay-in-four product can compete with a store card for a single discretionary purchase. A long deferred-interest promotion can compete with monthly installment financing for a larger durable good. A co-brand can compete for everyday spending away from the merchant. Combining these contests into one market-share statistic hides the distinct customer jobs and funding durations.

The merchant’s choice is an optimization problem across conversion, subsidy, approval, returns, repeat visits and program income. A checkout lender can broaden the financing menu while cannibalizing some store-card usage. Coexistence is therefore plausible: the right offer depends on purchase size, customer eligibility and repayment preference. A new financing button does not prove the incumbent program is obsolete.

To test substitution, measure the change in total profitable merchant sales and customer debt across products, not simply the growth of the new product. Track whether existing cardholders switch, whether previously declined customers gain sustainable access and whether customers accumulate obligations across providers. Product-specific approval and loss data, measured on comparable cohorts, would be stronger evidence than a provider’s conversion claim.

Consumer protections: established card rules still matter

For covered billing errors, Regulation Z generally requires the creditor to receive written notice at its designated address within 60 days after transmitting the first statement showing the error. The creditor normally must acknowledge it within 30 days unless resolved, and resolve it within two complete billing cycles, no later than 90 days. Protections concerning disputed amounts apply while the process runs; undisputed obligations remain. The exact notice requirements and exceptions matter. [14]

The card rules also restrict unsolicited issuance and address unauthorized-use liability. A closed-loop acceptance model does not by itself remove those protections. A merchant return policy, an issuer billing-error procedure and a network dispute process are different mechanisms; the absence of a general-purpose network does not mean the customer has no statutory rights. [15]

The CFPB’s compliance page states that its 2024 Credit Card Penalty Fees Final Rule was vacated by court order on April 15, 2025. This report therefore does not describe an $8 late-fee safe harbor from that rule as operative. Older articles or issuer planning discussions about the rule need to be read as historical material. [17]

Regulation is a boundary for product design, not a substitute for customer comprehension. A disclosure can be technically present while the surrounding sales conversation emphasizes only a discount. Useful oversight examines consent, promotional explanations, accessible servicing, accurate statements and complaint resolution throughout the customer journey.

A decision-ready scorecard for a private-label program

The following is an original diligence framework rather than a claim that public sources disclose every field. Start with product and customer outcomes, then reconcile the financial results. A scorecard limited to approvals, spending and current earnings can miss the very risks that later erase the program’s value.

Scroll horizontally to see all columns.

What remains unknown, and what would change the conclusion

Public evidence does not establish a current, universal merchant fee, a standard revenue-share contract or a single private-label return on equity. Important contract terms are confidential or redacted. Issuer portfolio disclosures commonly mix product types. The market report’s historical data do not supply a live October 2026 national dashboard. Those are limits to inference, not blanks to fill with a convenient industry average.

The most useful additional evidence would be comparable merchant-level cohorts showing incremental sales after all discounts, promotion payoff outcomes by customer risk, separately identified fraud losses and private-label-only credit performance. A matched analysis of shoppers offered a card versus an installment alternative would illuminate competition better than comparing headline company growth rates.

The durable conclusion is that private-label credit is a negotiated financing and distribution system. Its value depends on whether merchant economics, lender risk-adjusted returns and customer repayment outcomes hold together over time. An attractive checkout offer can be entirely real; evaluating it still requires reading the contract, following the balance and measuring who ultimately pays.

Sources

1. CFPB Regulation Z §1026.58: private-label account definition; current text checked October 4, 2026Official textBack to text: ↑
2. CFPB, December 2025: Consumer Credit Card Market, especially pages 10–12, 22, 41–44, 71 and 114; historical 2024 observationsOfficial source · PDFBack to text: ↑1↑2
3. CFPB, December 18, 2024: The High Cost of Retail Credit Cards; historical partnership and market analysisOfficial sourceBack to text: ↑
4. Synchrony 2025 Form 10-K: program arrangements, card types and funding; filed February 6, 2026Filing / reportBack to text: ↑1↑2
5. Synchrony Q2 2026 Form 10-Q: credit quality and charge-off definitions, quarter ended June 30, 2026Filing / reportBack to text: ↑1↑2
6. Bread Financial, July 23, 2026: Q2 results and metric definitionsSourceBack to text: ↑
7. Macy’s Q2 2026 Form 10-Q: credit-card net revenue, 26 weeks ended August 1, 2026Filing / reportBack to text: ↑
8. CFPB Regulation Z §1026.51: ability-to-pay requirements; current text checked October 4, 2026Official textBack to text: ↑
9. OCC Comptroller’s Handbook: Credit Card Lending, April 2021, including fraud and portfolio risk managementOfficial source · PDFBack to text: ↑
10. CFPB Regulation Z official interpretation §1026.16: deferred interest versus actual 0% APROfficial textBack to text: ↑
11. CFPB: How a purchase with no interest if paid in full within 12 months works; reviewed January 22, 2024Official sourceBack to text: ↑
12. CFPB Regulation Z §1026.53: allocation of payments, including deferred-interest balancesOfficial textBack to text: ↑
13. CFPB: What is a grace period for a credit card?; reviewed September 23, 2024Official sourceBack to text: ↑
14. CFPB Regulation Z §1026.13: billing error resolutionOfficial textBack to text: ↑
15. CFPB Regulation Z §1026.12: issuance and unauthorized-use protectionsOfficial textBack to text: ↑
16. CFPB Regulation B §1002.9: adverse-action notificationsOfficial textBack to text: ↑
17. CFPB credit-card penalty-fee compliance page: April 15, 2025 vacatur; checked October 4, 2026Official sourceBack to text: ↑
18. OCC Bulletin 2023-37: risk management of short-term BNPL lending; updated March 20, 2025Official sourceBack to text: ↑