Risk-adjusted return on capital (RAROC)

Executive Summary

In most banks, Risk-adjusted return on capital (RAROC) still functions as a reporting metric, a number reviewed periodically but rarely used to steer real decisions. That is increasingly inadequate. In a world of tighter capital, margin pressure, and rising balance-sheet constraints, RAROC is evolving from a performance metric into the operating language through which banks connect pricing, underwriting, funding, risk, and capital allocation.

The central argument of this paper is straightforward. The next generation of winning banks will not be the fastest to harvest balance-sheet growth or the most sophisticated models. They will be the institutions that allocate scarce capital with greater economic discipline, and that requires RAROC to serve not as a reporting metric but as the bank’s operating language.

But RAROC is only as credible as the foundations beneath it. Weak transfer pricing, inconsistent capital assumptions, and arbitrary cost allocations can create an illusion of precision while quietly distorting decisions. The issue is no longer whether banks measure profitability, but whether they measure it consistently enough to allocate scarce balance-sheet capacity effectively.

Why RAROC Matters Again

Why a metric most banks already report has become central to running the bank – and where the cracks in current frameworks show.

Risk-Adjusted Return on Capital was originally designed to answer a simple but powerful question: is the bank generating sufficient return for the level of risk and capital it is consuming? For years, many institutions treated it primarily as a reporting metric – useful for performance reviews but peripheral to real decision-making. That is changing. As capital becomes scarcer and balance-sheet constraints intensify, banks are rediscovering RAROC as a strategic tool for pricing, underwriting, portfolio steering, and capital allocation.

The cleanest banking interpretation is the standard one: RAROC equals risk-adjusted return divided by economic capital. The numerator is allocated revenues – net of matched funding costs, expected losses, operating costs, and taxes – and the denominator is the capital required to absorb unexpected losses at the bank’s chosen confidence level, calibrated to the credit rating the institution wishes to support.

Used well, RAROC measures capital efficiency. But efficiency is only one side of the profitability equation. A smaller business may generate very high RAROC while contributing little absolute profit, whereas a mature franchise with moderate RAROC may create substantial shareholder value on a scale. Leading banks, therefore, evaluate both RAROC and Economic Profit together – Economic Profit being the risk-adjusted return less than the product of economic capital and the hurdle rate. Capital efficiency and absolute value creation answer different questions, and either alone misleads.

RAROC and RORWA are not the same measure

Many banks treat Return on Risk-Weighted Assets (RORWA) and RAROC as interchangeable indicators of capital efficiency. They are not. RORWA measures profitability against regulatory capital and is useful for external benchmarking and regulatory capital management. RAROC measures profitability relative to internally assessed economic capital and is better suited for pricing, portfolio steering, and strategic capital allocation.

The distinction matters more under the Basel III endgame and the output floor, which are narrowing the gap between regulatory and economic capital for many institutions. As regulatory capital constraints tighten, banks increasingly need both lenses – RORWA to understand their regulatory standing, and RAROC to understand their underlying economics of risk and capital deployment. The divergence between the two often reveals deeper structural realities inside the balance sheet. Highly diversified retail businesses may appear economically attractive under RAROC, yet remain constrained by regulatory capital requirements. Concentration-heavy wholesale portfolios may appear efficient from a regulatory perspective, but they generate weaker economic returns once concentration and relationship risk are recognized.

The hidden fragility of profitability frameworks

The challenge is rarely the ratio itself – most banks can calculate RAROC. The deeper problem is consistency. Frameworks fail not because the methodology is unsophisticated, but because they fail to hold foundational assumptions steady across pricing, performance management, stress testing, and capital planning. None of these is hard to diagnose. What makes them persistent is that each is governance-resistant – they live inside business-line negotiation rather than central methodology, and they tend to coexist.

The Foundations That Must Hold

Four foundations – numerator, FTP, economic capital, attribution – against the 2026 risk and regulatory backdrop.

A credible RAROC framework rests on four foundations: a disciplined numerator that distinguishes accounting profit from economic profit; an FTP layer that prices funding, liquidity, optionality, and behavioral maturity transparency; an economic-capital base that is built independently of regulatory weights; and a defensible cost attribution that does not load every transaction with every overhead. How each of these shapes decisions – Basel III endgame and the output floor, International Financial Reporting Standard 9 (IFRS 9) expected loss, climate risk, and the standardized-approach reality that dominates Asian and Middle Eastern banking.

The numerator: accounting profit is not economic profit

Provisions under IFRS 9 are designed to reflect point-in-time credit deterioration and forward-looking macroeconomic expectations. RAROC is fundamentally an economic profitability measure intended to support pricing, underwriting, and capital allocation over longer horizons. Problems emerge when banks mix accounting and economic loss concepts inconsistently across pricing, performance management, stress testing, and capital planning. In benign conditions, point-in-time provisioning artificially inflates profitability; during downturns, it mixes otherwise viable businesses around structurally unattractive – producing procyclical decision-making at precisely the wrong point in the cycle.

The issue is less about methodology than consistency. Whichever expected-loss definition a bank chooses, that definition must flow through pricing models, RAROC scorecards, capital plans, and stress tests. A bank that uses through-the-cycle for pricing, point-in-time for accounting, and an undocumented hybrid for stress testing does not have an expected-loss methodology – it has an expected-loss accident.

The FTP layer: where RAROC lives or dies

Revenue attribution and transfer pricing are inseparable. A bank without a credible FTP framework lacks a credible profitability measurement. In practice, many of the largest distortions in RAROC originate not from credit modeling but from inconsistent funding assumptions, poorly allocated liquidity costs, and an opaque treatment of deposit franchise value.

A RAROC-grade FTP framework must allocate funding, liquidity, optionality, and deposit-franchise value transparently and consistently across businesses. Where FTP becomes opaque or negotiated, profitability measurement quickly loses credibility as a capital-allocation tool.

The challenge is most acute in businesses with significant optionality and behavioral uncertainty. Long-term fixed-rate assets, revolving portfolios, and deposit-rich franchises can all appear materially more or less profitable depending on how funding, liquidity, and behavioral maturity are treated. A simple diagnostic reveals whether an FTP framework is genuinely decision-useful: if business-unit profitability moves materially quarter-to-quarter because “FTP changed”, management should be able to explain precisely why. Where FTP outcomes are opaque, negotiated, or poorly understood by the deal team, it should be taken for granted that profitability measurement quickly loses credibility as a capital-allocation tool.

The most contested FTP question inside many banks is deceptively simple: who owns the economic value of deposits? Deposit-gathering businesses argue that stable customer balances are a strategic funding advantage that should enhance relationship profitability. Treasury functions often view those balances as centrally managed liquidity and stressed reserve resources. Profitability frameworks fail when the same economic benefit is recognized multiple times – or not recognized at all. As balance-sheet constraints intensify, clear ownership of deposit franchise value becomes increasingly important to customer strategy, business-line performance, and capital allocation.

Economic capital, with or without IRB

Most RAROC literature implicitly assumes IRB. That is correct for European, UK, US, and Australian banks – and largely wrong for the bulk of the Asian and Middle Eastern market. The majority of Saudi Central Bank (SAMA) regulated banks, all Indian public-sector banks, most GCC banks, and most Association of Southeast Asian Nations (ASEAN) banks operate on the Standardized approach for credit risk. They do not need Internal Ratings-Based (IRB) approval to run RAROC; they need a defensible loss distribution. The two are not the same thing.

The single most common conceptual error in Standardized Approach (SA) banks is to use SA risk weights as a proxy for economic capital. This produces a “RAROC” that is mathematically just the inverse of Return on Risk-Weighted Assets (RORWA), adding no new information beyond what the regulatory framework already provides. Economic Capital (EC) for internal models must be built independently of regulatory weights, using internal data, external benchmarks, and pragmatic modeling. For an IRB bank running RAROC against pure economic capital, the floor introduces a regulatory-economic capital boundary-constraint level.

Climate as a capital-allocation issue

Climate risk is shifting from a disclosure and compliance question into a capital-allocation question. Three transmission channels are now visible in RAROC frameworks: Probability of Default (PD) and Loss Given Default (LGD) overlays for high-emitting sectors flow through expected loss and economic capital; all others are often hidden in mortgage and Commercial Real Estate (CRE) books, where coastal real estate, flood-zone residential, and drought-exposed agriculture see increased LGD and concentration overlays recalibrate sector-and geography-concentration portfolios. Transition-sensitive sectors may become structurally more capital-intensive over time, while physical climate risk is beginning to reshape assumptions around collateral haircuts in mortgage and Commercial Real Estate markets, where coastal real estate, flood-zone residential, and drought-exposed agriculture see increased LGD. And concentration overlays recalibrate sector-and geographic concentration portfolios. Transition-sensitive sectors may become structurally more capital-intensive over time, while physical climate risk is beginning to reshape assumptions around collateral quality and geographic concentration. The more important shift, however, is conceptual: climate overlays are gradually moving profitability frameworks from static historical-loss assumptions toward more scenario-based, forward-looking views of risk.

Cost attribution: three layers, not one

Cost attribution is the second most common reason RAROC frameworks lose credibility, behind FTP. When shared cost is ignored, RAROC is overstated. If shared cost is forced into transaction-level RAROC, good businesses look bad. If shared cost is allocated and then reallocated against a clean activity-based driver, it belongs to the transaction, customer, and BU RAROC. Layer one is directly attributable cost – origination, servicing, collection. Relationship Manager (RM) compensation, deal-specific structuring – charged directly to the deal. It belongs to the transaction, customer, and BU RAROC. Layer two is controllable shared cost – IT systems, operations, compliance, risk management – driven by business activity and allocated against a clean activity-based driver. It belongs to the Business Unit (BU) level RAROC, segment economics, and performance scorecards. Layer three is non-controllable overhead – CEO office, enterprise projects – which should be excluded from decision-useful RAROC and reflected only in a fully loaded P&L for strategic planning and the ROE walk.

The consequences are concrete. The same loan can produce a RAROC of 17.3% on direct costs alone, 11.8% when controllable shared costs are added, and 9.4% on a fully loaded basis. The right answer is dual reporting: a decision-useful RAROC of 12% for pricing and performance management, and a fully loaded view at all three layers for strategic planning and group-level capital allocation. Mixing the two collapses the framework into a set of negotiated numbers.

The Forward-Looking Lens

Why one-year RAROC systematically misprices long-tenor business – and what to do about it.

The argument that RAROC should be forward-looking rather than ex-post is widely accepted in principle but seldom implemented in practice. Lifetime RAROC requires a real operational build: lifetime PD curves, behavioral Exposure at Default (EAD), forward FTP curves, and an EC profile that evolves with the portfolio. Each is a genuine investment, and most banks have stopped short. The cost of stopping short, as the next pages show, is systematic mispricing of long-tenor business.

A 25-year mortgage tells the clearest story.

The contrast between one-year and lifetime RAROC is most visible on long-tenor amortizing assets. Take a SAR 1 million, 25-year fixed-rate residential mortgage at 5.80%.

The mortgage creates value over its life. The early-year drag from acquisition cost reverses as the loan seasons, EC declines as the balance amortizes, and lifetime expected loss, though larger in absolute terms, sits against a much larger Net Present Value revenue stream. A bank that prices and originates based on one-year RAROC will systematically reject value-creating business and accept short-tenor business that looks good in year one and poor over the life.

The pattern is not uniform across products.

The mortgage is a particularly clean illustration because the gap between one-year and lifetime view is so wide. But the pattern is not always favorable – for some products, it is harsher than the one-year view; for others, the two converge.

The credit card line is the one most pricing committees need to see. A one-year RAROC of 28% routinely justifies aggressive growth budgets in retail; lifetime view, which captures attrition, and reward-cost loading, is closer to 19.5% – still good business, but a fundamentally different scaling story. Project finance is the mirror image: a one-year view looks unfavorable; the lifetime view at 14.2% clears the hurdle. Banks that originate and decide on one-year RAROC are systematically miscalibrated relative to their own product mix.

Three rules for the long-tenor playbook

From Deal to Relationship to Group

How the unit of analysis changes the answer – across retail and wholesale, customer and conglomerate.

RAROC plays out very differently across retail and wholesale, and at the deal, customer, and group levels. Treating retail and wholesale as variants of the same calculation – which many banks do – is the source of much of the cross-business friction RAROC frameworks produce. Retail is a price-economics problem, wholesale is a relationship and concentration-economics problem. The cross-business comparison is engineered on the apples-to-apples on the dimensions that drive the ratio, even when the underlying machinery differs. Eight design choices must be standardized before any cross-business comparison is meaningful.

Retail as pool economics, wholesale as relationship economics

Segment behavior, vintage effects, utilization, attrition, acquisition cost, prepayment, collections, and lifetime customer behavior drive retail RAROC. No individual small loan warrants its own standalone decision. Retail RAROC is measured at the product-by-segment-by-vintage level, with the key design choices being how acquisition cost is amortized, how attrition is modeled, and how revolver-balance decay is treated in credit cards.

The credit card line in any retail RAROC scorecard deserves disproportionate scrutiny. A pre-tax RAROC north of 80% is mathematically defensible but headline-grabbing, and is heavily influenced by behavioral assumptions on revolver attrition, charges-off recoveries, and reward-cost loading. Banks should present credit card RAROC alongside its lifetime, attrition-focused economics. The analysis must extend further still: a small subsidiary loan is best evaluated against the parent group’s wealth, not the subsidiary’s standalone economics. A walk-up from transaction to full relationship makes the point most starkly.

The term loan looks unattractive standalone. But once the full relationship wallet is included, RAROC is 30.5% – a completely different conclusion. This single example carries the central insight of customer-level RAROC, developed below.

Wholesale RAROC is fundamentally different. Individual exposures are larger, idiosyncratic risk matters more, and sector or name concentration can materially move economics capital. The walk-up from transaction to full relationship wallet is the point most starkly: the single largest source of cross-business friction RAROC frameworks produce. Concentrations-heavy wholesale portfolios may appear economically attractive under RAROC, yet remain constrained by regulatory capital requirements. The challenge is not building separate frameworks; it is creating a unified profitability language that allows capital to be allocated consistently across both.

Engineering comparability: the eight-point bridge

Cross-business RAROC comparisons are the single most contentious moment in any bank’s capital-allocation cycle. Retail says wholesale is over-rewarded for diversification. Wholesale says retail uses pooled FTP that flatters returns. Treasuries says everyone is using wrong optionality assumptions. The honest framing is that retail and wholesale are built differently because the underlying businesses are different, and that the cross-business comparison is engineered on the apples-to-apples on the dimensions that drive the ratio, even when the underlying machinery differs. Eight design choices must be standardized before any cross-business comparison is meaningful.

The effect is non-trivial. Applied to the same two exposures – a 25-year retail mortgage and a 5-year wholesale CRE term loan – the raw 1-year numbers tell one story: wholesale 14.5%, retail 11.0%. The fully-bridged comparable numbers tell another: retail 17.0%, wholesale 14.3%. Neither view is correct in isolation; both are internally consistent given one’s design choices. What matters is that capital-allocation decisions are made on the bridged view, not the raw one.

The institutional answer is a single Risk-Adjusted Performance Measurement (RAPM) Methodology Council that owns the bridge – composition and mandate discussed in Part VI. Without a single accountable owner, the bridge erodes into negotiated numbers within two budget cycles.

Customer-level RAROC: the missing liability dimension

Most RAROC implementations focus exclusively on the asset side – the loan, the exposure, the credit risk. This is a fundamental oversight, particularly when RAROC is measured at the customer or relationship level. A corporate customer does not simply borrow from the bank. They also deposit money, maintain operating accounts, use trade-finance facilities, hold guarantees, and generate fee income. The customer’s liabilities create significant economic value that the bank should recognize: an FTP credit on deposits reflecting the funding the bank would otherwise have to raise in wholesale markets; a liquidity offset where stable operational deposits qualify as core deposits under Liquidity Coverage Ratio (LCR) and Net Stable Funding Ratio (NSFR) and reduce the bank’s liquidity buffer; and a capital benefit where stable deposit franchises reduce the EC attributable to liquidity and funding risk.

If customer-level RAROC ignores these contributions, it systematically undervalues the bank’s most valuable relationships. Take “IndustryCo” – a corporate client with a SAR 50M term loan at 7.00%, a SAR 20M revolving facility (60% utilized), SAR 30M of current-account balances paying 1.50%, a SAR 15M time deposit paying 3.50%, and SAR 0.8M of annual fee income.

If the bank evaluates IndustryCo only on asset-side RAROC, it undervalues the client by nearly 17 percentage points. In the worst case, the bank declines to renew a loan and pushes a valuable full-service client to a competitor.

Three implementation cautions are worth flagging. Not all deposits are equal – operational balances are stickier than excess liquidity parked temporarily, and FTP credits should reflect this with higher credit for stable, core deposits. The behavioral-versus-contractual distinction must be modeled and validated: a demand deposit with zero contractual maturity may behave like a 2- to 5-year liability, and the FTP framework needs an actual behavioral model rather than a flat assumption. And double-counting must be policed actively – if treasury already captures the full deposit margin centrally, the business unit should not also receive credit; the framework needs a single, clear policy on who owns the deposit franchise value.

A clarification on netting versus separate attribution. Some banks’ net assets and liabilities at the customer level before computing RAROC. This is conceptually wrong – it obscures the risk on the asset side. The correct approach is to compute asset-side RAROC and add the liability-side contribution as a transparent, separately reported line item.

The conglomerate dimension

A nuance that asset-side RAROC misses, and that even relationship-level RAROC under-treats, is particularly acute in markets dominated by family-owned conglomerates and state-owned groups – Indian, GCC, and most ASEAN banking. The question is: when the borrower is a subsidiary of a larger conglomerate, what is the right unit of analysis?

In practice, the “customer” in customer-level RAROC has at least three valid definitions, each of which yields a different answer. The legal entity – the borrower of record – is the right unit for credit decisions, regulatory single-name limits, and contractual enforcement. The risk group, defined under Basel large-exposure rules, the Reserve Bank of India (RBI) group-exposure norms, or SAMA’s connected-counterparty regime, is the unit for concentration limits, regulatory aggregation, and single-name capital. The relationship wallet – all entities under common beneficial ownership that the bank does  business with, including treasury, FX, trade finance, mandates, and employee banking – is the unit for relationship strategy, pricing leverage, capital allocation, and RAROC.

The numbers move materially across the three definitions. “AlphaGroup” is a regional industrial conglomerate. The bank is being asked to renew an SAR 80M loan to a real-estate subsidiary. The deal team’s standalone analysis returns a  RAROC of 9.2%, below the 12% hurdle. Walking up  the group structure tells a different story.

The standalone subsidiary loan looks unattractive. The relationship wallet is highly attractive. Decline the subsidiary loan, and the bank is likely to lose the group.

Two operational principles follow. First, aggregate up for risk; evaluate up for economics. Risk concentration – single-counterparty limits, sector aggregation, large-exposure rules – must always be calculated at the group level. This is regulatorily required and non-negotiable. Economic value should also be evaluated at the group level for relationship decisions, with explicit transparency about which entity generates each line of revenue. The numbers must be auditable. Second, do not quietly migrate the subsidiary’s PD or LGD up to parent quality. The temptation in conglomerate lending is to take comfort from the parent’s creditworthiness without formal documentation. This is dangerous. If the bank intends to rely on parent support, it should be documented as a parental guarantee, internal credit substitution, keep-well letter, or comfort letter – not as a vibes-based PD adjustment. A large share of GCC and Indian non-performing-loan stories trace back to exactly this mistake: implicit “group halo” PDs that turned out not to reflect actual support when the subsidiary defaulted (IL&FS, NMC Health, and several GCC construction groups in 2015–2020).

Field note: the conglomerate haircut

Several leading regional banks now apply an explicit “conglomerate haircut” to RAROC numbers presented at the credit committee – typically a 50 to 150 bps reduction in claimed relationship RAROC, depending on the strength of cross-default protection and the degree of formal versus implicit group support. The haircut is not punitive; it is a governance device that forces the credit committee to debate the actual structure of group support, rather than waving it through under the heading of “strong relationship”.

Architecture: How RAROC Is Built

The deal-level mechanics, beyond credit risk, the hurdle rate, and the diversification choice.

The mechanics of RAROC are well understood. What separates frameworks that drive decisions from those that sit decoratively in a finance pack is the discipline of four design choices: how the deal-level calculation handles funding costs, expected losses, and capital; how non-credit risks are layered in; how the hurdle rate is set; and how diversification is treated. None is conceptually new. All are routinely mishandled.

The deal-level calculation

At its core, RAROC asks whether the return generated by a transaction, customer, or portfolio justifies the amount of scarce capital being consumed. The answer depends heavily on how non-credit risks are layered in, how the hurdle rate is set, and how diversification is treated. A point-in-time worked example is useful for explaining the mechanics, but it masks four design choices that change the answer materially in practice.

A point-in-time worked example like this is useful for explaining the mechanics, but it masks four design choices that change the answer materially in practice. The single-period view does not capture the lifetime economics of a 5-year loan (Part III). The static 12% assumption does not credit portfolio diversification – appropriately, since the deal team should not be rewarded for portfolio effects they do not control. And the through-cycle expected loss in the numerator sits alongside a separate IFRS 9 P&L bridge for Stage 1, 2, and 3 movements. The standard EC denominator does not credit portfolio diversification – appropriately, since the deal team should not be rewarded for portfolio effects they do not control. And the through-cycle expected loss in the numerator sits alongside a separate IFRS 9 P&L bridge for Stage 1, 2, and 3 movements.

Beyond credit risk

For pure loan pricing, a credit-only RAROC can serve as a first-pass screen. For business-line performance measurement, capital allocation, or incentive design, limiting the denominator to credit-risk capital is too narrow. A serious framework runs a tiered approach: at transaction level, credit-risk EC plus FTP/liquidity plus direct cost is enough for origination approval; at business-line level, market and Interest Rate Risk in the Banking Book (IRRBB) capital overlays are added in for performance management and incentive design; at enterprise level, diversification benefits, management buffers, and Pillar 2 add-ons are layered in for group capital adequacy and strategic planning.

Two risk types deserve specific attention. Operational risk is genuinely difficult to allocate below the business-line level – revenue-based allocation creates perverse incentives by penalizing growth. At the same time, activity-based proxies require data that most banks do not have. Concentration and name concentration and can materially move economic capital. Where does the diversification benefit on the appropriate level. They should not be ignored simply because they are hard to allocate granularly.

Trading-book RAROC requires explicit treatment of XVAs. A trading-book RAROC that ignores CVA, KVA, and MVA systematically overstates the economics of long-tenor uncollateralized derivatives. Banks running material derivative books should integrate XVA charges directly into the pricing-grade RAROC; the standalone treatment of XVAs as a separate “desk” is a transitional arrangement, not a target state.

The hurdle rate

A RAROC framework lives and dies by its hurdle rate. Set it too low, and the bank approves value-destroying business. Set it too high, and it rejects profitable business in favor of competitors. Most banks anchor to a CAPM-derived cost of equity, cross-check against market-implied estimates, and benchmark to peers – settling on a single group hurdle reviewed annually. The methodology debate is well-trodden. The harder question is whether the hurdle should vary by business unit.

A single group hurdle is simpler and reflects the fact that equity is fungible at the group level. BU-specific hurdles are more economically correct because business risk profiles genuinely differ – investment banking is riskier than retail deposit-gathering – but they are politically harder and require sophisticated capital-allocation governance to defend. The pragmatic middle path adopted by several leading banks is a single group hurdle plus a documented “strategic premium” on businesses where execution risk or scale-up requirements genuinely warrant a higher bar, governed by the RAPM Council and reviewed annually.

Diversification: standalone or diversified capital

Diversification is one of the most consequential and least transparent design choices in RAROC. The bank’s actual economic capital is well below the sum of the standalone capitals across business units because risks are imperfectly correlated. Where does the diversification benefit go?

The practical answer is to use both standalone and diversified views, but for clearly different purposes. Standalone capital is the right denominator for pricing and frontline performance – a deal team negotiating a single transaction does not control the diversification benefit, and rewarding them for it produces moral hazard. Diversified capital is the right denominator for enterprise capital allocation and BU-level Economic Profit – group capital is genuinely fungible, and allocation decisions should be made on the marginal cost of capital, which reflects diversification. The “diversification reserve” – the gap between the sum of standalone EC and group EC – should be visible, governed, and reviewed annually. Burying it produces the worst of both worlds: standalone numbers that double-count, or diversified numbers that obscure who gets credit.

The methodological choice – proportional, Euler, or component – is technical. The governance choice – who owns the methodology, who signs off, when it changes – is what determines whether the framework survives contact with the budget cycle.

Steering the Bank

Capital allocation, self-sufficiency, compensation, governance, and the implementation journey.

This is where most RAROC programs either translate into management impact or die at the first contact with the budget cycle. The mechanics are now in place. The question is whether the framework actually drives decisions – capital allocation, growth assessment, compensation, governance – or sits decoratively in a finance pack.

RAROC as a capital-allocation engine

RAROC is most valuable when capital is scarce. But the bank should focus on marginal, forward-looking RAROC rather than average historical RAROC. The real question for capital allocation is which incremental growth proposals generate the best risk-adjusted return, given all constraints. Bank Alpha illustrates the point. It has SAR 2 billion of deployable capital. Three business units submit growth proposals.

A business with moderate average RAROC but high execution risk, rising concentration, or marginal economics should not be prioritized over a business with higher marginal profitability and lower hidden constraints. RAROC sets the ranking, but it never decides allocation alone – concentration, liquidity, regulatory, and strategic overlays do the binding work, as the table illustrates. RAROC should sit within that broader portfolio-steering framework rather than operate as a standalone ranking.

Self-sustaining business units

A strong RAROC indicates that a business generates an acceptable risk-adjusted return. It does not, on its own, prove that the business can fund its own growth. A second metric – Capital Self-Sufficiency, defined as retained earnings from a BU divided by the incremental capital required for planned growth – answers whether the business can expand without drawing on group capital. They are different questions, and either alone misleads.

Corporate banking has a solid RAROC (16.0%) but is capital-hungry. Retail is the star: strong RAROC and self-sustaining. A mature bank should assess both dimensions; a high-RAROC business that consistently consumes group capital is a different proposition from one that funds its own growth.

Linking RAROC to compensation

A RAROC framework that does not influence variable compensation does not run the bank. It is documentation. Mature banks embed RAROC and Economic Profit in BU scorecards alongside revenue and risk metrics, tie a defined slice of variable compensation (typically 20 to 35%) to RAROC and EP outcomes, and hold a portion back against multi-year realization and clawback where ex-post returns underperform what was claimed at origination. The structural design is straightforward: the failure modes are not.

The same design that ties pay to RAROC creates incentives to game it. Mature programs counter this by weighting realized over-ex-ante RAROC with clawback capping RAROC at 35% of variable comp so it cannot dominate scorecards, and placing methodology authority firmly outside business-line negotiation. The last is what the RAPM Council independence is fundamentally designed to protect.

Governance: the RAPM Council

The difference between a bank running a credible RAROC framework and one running a decorative one is almost always governance, not methodology. The institutional center of gravity is a Risk-Adjusted Performance Measurement Council, chaired independently of P&L responsibility – typically by Strategy, the CRO, or a dedicated RAPM head, and not by the CFO if the CFO owns financial reporting. The Council comprises heads of major BUs, the CRO or deputy, the CFO or deputy, the head of model risk, and the head of finance or planning. It meets monthly on methodology issues, quarterly on BU RAROC trends, and annually for recalibration of methodology. It reports to Group ALCO and the Board Risk Committee – not buried inside Finance.

The Council owns the four foundations document, the eight-point comparability bridge between retail and wholesale, the hurdle-rate methodology and any business-specific premia, the diversification methodology and the diversification reserve, the cost-attribution layering and dual-reporting policy, the annual external benchmarking review, and the methodology change protocol – including impact testing before any change goes live. It approves methodology changes that move BU RAROC by more than 50 bps.

Data and technology

Methodology and governance are necessary but not sufficient. Most RAROC programs that fail in implementation fail because the data layer cannot support them. Decision-level RAROC requires unified data across transactions, risk parameters, costs, FTP, and capital – and the binding constraints are usually consistent customer hierarchy across systems and activity-based cost granularity, not the calculation engine itself.

The single most common mistake in RAROC technology programs is to build the calculation engine before the source-of-truth layer. This produces a fast, sophisticated engine running on inconsistent inputs – a more elegant version of the same problem.

The journey: from where banks are, to where they need to be

A RAROC program is multi-year. Four maturity stages describe where a bank sits today.

In our experience, most banks self-assess one stage higher than they actually operate. A practical diagnostic: a bank is at Stage 3 if and only if a deal team would price differently because of RAROC, and a sub-hurdle business has been actively repriced or wound down in the past 12 months. If neither is true, the framework is at Stage 1 or 2, regardless of how it appears on slides.

The journey from Stage 1 to Stage 4 is multi-year, and the sequence matters more than the speed. Banks that try to install RAROC-linked compensation before FTP and EC have settled find that the framework collapses under political pressure. Banks that try to extend RAROC into climate, lifetime, and full-relationship views before the core methodology is consistent produce numbers that do not stand up to first contact with a credit committee. The right starting point is a credible RAPM Council with clear authority, a documented methodology, and three to five visible quick wins – repricing or repositioning decisions that demonstrate the framework actually changes behavior. The build of FTP, EC, cost attribution, and the comparability bridge follows. Comp linkage, climate overlays, and full-lifetime RAROC come last, once the foundations are established.

Three principles for the journey

The Five Questions RAROC Should Answer

A credible RAROC framework gives the bank crisp, defensible answers to five questions every senior committee asks – but few banks answer with the same number twice. If the framework cannot answer all five consistently, it is not yet running the bank.

A bank running RAROC at Stage 4 can answer all five questions on a single page, sourced from a single canonical methodology owned by the RAPM Council. A bank at Stage 1 or 2 can typically answer one or two, and is silent, evasive, or politically negotiated on the rest. The honest test of a RAROC framework is not how sophisticated the methodology is. It is whether the same number falls out of every committee.

Conclusion: RAROC 2.0

The argument is not that RAROC is a perfect measure. No single metric can capture the full complexity of bank economics. The argument is that RAROC – when built on the four credible foundations and embedded in real management processes – is the most powerful internal decision tool available to a bank.

It tells management which activities truly earn more than their cost of risk, which businesses consume capital without earning enough, which units can fund their own growth, and where scarce balance-sheet capacity should be deployed next. It does this with a single, consistent language that connects the front line to the C-suite – but only when the four foundations hold: a disciplined numerator, a credible FTP layer, a robust economic-capital base (achievable even on Standardized Approach), and defensible attribution across revenues, costs, and risks.

The banks that will win the next cycle are not the ones that measure RAROC the most precisely. They are the ones that run the bank through it – using it to price deals, steer portfolios, allocate capital, align incentives, and translate genuine value creation into management action.

That is RAROC 2.0.

Glossary of Terms

Terms used throughout this paper. Where multiple definitions exist in regulatory or industry practice, we use the convention applied in this document.