Revenue yield is not the return that matters
A consumer-credit program can report a high and still destroy economic value. Funding, acquisition expense, servicing, fraud, credit losses and capital all consume the apparent spread. , or RAROC, is a framework for relating an explicitly defined profit measure to the capital allocated to support the risk. It is a management construct; institutions use different definitions.
The Federal Reserve’s May 1998 study of internal credit-risk models describes using allocated capital and hurdle rates in lending decisions. [1] That historical research is useful for the concept, not as a statement of today’s capital rules. Current regulatory capital requirements must be evaluated under the applicable framework, such as the OCC’s Part 3 for institutions in its scope. [2]
Keep expected loss and unexpected loss distinct
Expected loss is the average loss anticipated for a defined exposure and horizon. A simple illustration uses probability of default multiplied by loss given default and exposure at default, but revolving utilization, recoveries, timing and segmentation complicate real portfolios. The economic profit calculation should charge for expected loss once, on a consistent basis.
Capital supports adverse outcomes beyond the base expectation and other risks included in the institution’s allocation methodology. Economic capital, regulatory minimums, leverage constraints and management buffers answer different questions. A program can appear attractive on a narrowly modeled economic-capital basis while consuming scarce regulatory or balance-sheet capacity elsewhere in the bank.
Avoid subtracting both a full capital charge and then comparing the remaining profit divided by capital with the same hurdle unless the metric is deliberately defined that way. and economic profit are related views: one expresses a ratio, the other subtracts a required return from profit. Mixing them can double count the cost of capital.
A hypothetical one-year calculation
Assume $10 million of average funded exposure, $1.5 million of interest and fee revenue, $450,000 of funding cost, $300,000 of operating and acquisition expense, $400,000 of expected credit and fraud loss, and $50,000 of other program expense. The resulting pretax risk-adjusted profit is $300,000. Assume $1.5 million of allocated capital. Pretax is 20%.
With an assumed 18% pretax hurdle, the required profit is $270,000 and economic profit is $30,000. Every figure is hypothetical. The calculation excludes taxes and assumes all revenues, expenses and average exposures cover the same one-year period. Comparing that 20% with an after-tax hurdle would be invalid without conversion.
Now increase expected loss by $200,000 while holding the other assumptions constant. Profit falls to $100,000 and to 6.7%. If the risk also requires $1.8 million of capital, RAROC falls further to 5.6%. A small margin above the original hurdle offers little protection against errors in loss estimates or an adverse funding environment.
Scroll horizontally to see all columns.
| Annual item | Assumed amount |
|---|---|
| Revenue | $1,500,000 |
| Funding cost | −$450,000 |
| Operating and acquisition expense | −$300,000 |
| Expected credit and fraud loss | −$400,000 |
| Other expense | −$50,000 |
| Pretax risk-adjusted profit | $300,000 |
| Allocated capital | $1,500,000 |
| Pretax | 20% |
Price the cash flows, not just the headline APR
is a customer disclosure measure; realized portfolio revenue depends on balances, payment behavior, promotional periods, fees, waivers and losses. Merchant-subsidized financing may generate substantial upfront revenue while customer interest remains low. That revenue should be matched to acquisition costs, future servicing obligations and refund exposure rather than treated as costless margin.
For amortizing loans, compare discounted lifetime cash flows with a consistent capital path. A one-year accounting ratio can favor products with early fees and back-loaded losses. Revolving accounts require assumptions about future draws and payment rates. The useful pricing model makes those assumptions visible and shows which ones matter most.
The OCC’s retail-lending handbook connects lending strategy with risk management, portfolio analysis and controls. [3] A return model should therefore inform a decision within approved risk appetite, not override legal restrictions or affordability concerns. A high modeled return does not make an otherwise impermissible price or practice acceptable.
Why the denominator deserves challenge
Allocated capital can make a weak program look strong if diversification credits are too generous or stress correlations are understated. A merchant vertical concentrated in one employment market may diversify less than its many individual accounts suggest. Similarly, multiple funding sources can depend on the same capital-market conditions when stress arrives.
Use several views: standalone risk, contribution to the portfolio, regulatory constraints and a stressed capital requirement. Explain which view governs approval and why. A portfolio-level benefit should not become a permanent subsidy that hides deteriorating unit economics. Revisit allocations when product mix, underwriting or funding arrangements change materially.
Governance and commercial tradeoffs
Recommended controls include independent reconciliation of model inputs to finance and servicing records, documented loss assumptions, versioned hurdle rates and a record of exceptions. The current April 2026 interagency model-risk guidance is the appropriate supervisory reference for material models in scope. [4] Model governance should be proportionate to use and risk, including simpler tools that materially influence pricing.
Raising price can improve modeled margin but reduce acceptance, change borrower mix or increase merchant subsidy demands. Tightening credit can reduce losses while increasing acquisition cost per booked account. Shorter terms can reduce duration but raise required payments. Evaluate these second-order effects instead of changing one spreadsheet cell and assuming everything else stays fixed.
Evidence that earns confidence
The most persuasive evidence is a bridge from forecast to realized results by vintage: revenue, balances, losses, recoveries, expenses and capital use. Track errors separately so favorable funding conditions do not conceal weak underwriting. Run reverse stresses to identify the loss rate or funding cost at which the program falls below its hurdle.
A changed capital rule, new loss experience or a revised funding structure could change the conclusion. The objective is not to produce a precise-looking percentage. It is to establish whether the program pays for its full economics and remains worthwhile when reasonable assumptions move against it.
Sources
- Federal Reserve System, Credit Risk Models at Major U.S. Banking Institutions, May 1998; historical conceptual researchOfficial source · PDFBack to text: ↑
- eCFR, OCC 12 CFR Part 3, current capital framework; reviewed September 27, 2026Official textBack to text: ↑
- OCC, Retail Lending handbook, October 2021; reviewed September 27, 2026Official source · PDFBack to text: ↑
- Federal Reserve, SR 26-2, April 17, 2026Official sourceBack to text: ↑