Risk Accounting vs. Traditional ERM and FAIR: What Does Each Measure?
- 1. Why comparing risk methodologies requires care
- 2. What is distinctive about Risk Accounting?
- 3. Side-by-side comparison: what can each approach tell a bank?
- 4. Worked example: one bank, different questions
- 5. Should Risk Accounting replace RCSA, loss databases or scenario models?
- 6. Governance, auditability and regulatory positioning
- 7. Frequently asked questions
- 8. Explore the method and evidence
Risk Accounting, enterprise risk management (ERM) and FAIR address related but different questions. ERM provides governance and prioritisation across enterprise risks; FAIR models risk scenarios in financial terms; Risk Accounting proposes a consistent unit for current non-financial-risk exposure. Their outputs are not directly interchangeable, and a well-governed institution may use these approaches together.
1. Why comparing risk methodologies requires care
Operational risk management has several legitimate measurement questions. An internal auditor may ask whether a payment process complies with required controls. A chief risk officer may ask whether current exposure is increasing as the institution expands. A cybersecurity leader may ask what a ransomware scenario could cost. A regulator may ask whether a bank has adequate governance, capital and operational resilience arrangements. One risk rating cannot answer all four questions.
It is therefore misleading to compare Risk Accounting, enterprise risk management (ERM), risk and control self-assessment (RCSA), FAIR, ORX services and regulatory capital requirements as if each were a rival formula. They occupy different categories:
- ERM is an enterprise-wide governance and decision-making discipline. It establishes objectives, risk appetite, roles, policies, escalation and reporting. An ERM programme can incorporate qualitative and quantitative techniques.
- Risk matrices and conventional RCSAs are assessment and prioritisation tools commonly used within an ERM framework. Their ordinal rankings are not necessarily ratio-scale measurements.
- FAIR (Factor Analysis of Information Risk) is a quantitative analytical model used to estimate financial consequences of cybersecurity and operational risk scenarios. It is not confined exclusively to cyber risk. [3]
- ORX is an operational-risk services and research organisation providing, among other things, loss data exchange, scenario resources, standards, practitioner collaboration and benchmarking. It is not a single risk-scoring formula. [4]
- Risk Accounting is a proposed method for quantifying inherent and residual current non-financial-risk exposure using a common method-defined unit, the Risk Unit (RU). [1][2]
The relevant question is not simply “Which is best?” but “What decision is required, what is being measured, and how defensible is the measurement?”
2. What is distinctive about Risk Accounting?
Risk Accounting, developed by Peter J. Hughes, begins with business activity rather than with a hypothetical loss event. In the published banking method, exposure is classified by product and principal risk family. The model separates (i) product-specific structural exposure, (ii) an approved weighting of daily new business, and (iii) assessment of mitigation effectiveness. [1]
For a given principal risk family, the summarised inherent-risk relationship is:
where EUF is the approved Exposure Uncertainty Factor for family r, and VBW is the discrete Value-Based Weighting selected under the applicable daily-new-business value table. RUs are not dollars or euros. The value of daily new business is used to select a band, not multiplied directly by an arbitrary factor.
Assessed mitigation derives from Enhanced Risk and Control Self-Assessment (E-RCSA) inputs. At the correct reporting scope:
Here ASA denotes the Actual Score Aggregate and MSA the Maximum Score Aggregate. Both must correspond to the same governed assessment perimeter. Larger-scope RMIs are recomputed from aggregated ASA and MSA, rather than obtained by unweighted averaging of component percentages. [1, pp. 140–141, 163–168]
This makes Risk Accounting primarily an exposure-measurement proposition. The separate research challenge of estimating monetary expected losses requires credible empirical calibration and validation. Neither an RU figure nor the equations above establishes a reserve, a capital requirement, or a forecast of loss events.
3. Side-by-side comparison: what can each approach tell a bank?
| Criterion | ERM with qualitative RCSA / risk matrix | FAIR model | Risk Accounting | ORX tools and services |
|---|---|---|---|---|
| Primary role | Governance, risk identification, assessment and prioritisation | Financial quantification of defined risk scenarios | Quantification of current non-financial-risk exposure | Shared data, standards, research, benchmarking and scenario support |
| Core question | Which risks matter and who manages them? | What financial loss could arise, and with what modeled frequency or magnitude? | What method-defined inherent and residual exposure is carried by current activity? | What do peer events, data and industry practice tell us? |
| Typical outputs | Risk registers, control findings, ratings, actions | Financial distributions or ranges and scenario analyses | Inherent RUs, RMI and residual RUs | Loss-data benchmarks, scenario information, taxonomies and analytical resources |
| Primary inputs | Workshops, assessments, control evidence, indicators | Defined scenarios and relevant frequency/magnitude assumptions and evidence | Product classifications, approved EUFs, daily new business / VBW, effective-dated E-RCSA inputs | Contributed data, industry taxonomies, reports and practitioner research |
| Treatment of controls | Assessment and oversight; can be qualitative or quantitative | May affect parameters and scenario assumptions | Explicit control-assessment contribution through RMI | Depends on the particular resource or programme |
| Aggregation | Depends on the measures used | Requires a defensible model and treatment of dependence | Subject to common RU definitions, mapping and weighted aggregate rules | Depends on dataset and use case |
| Central limitation | Ordinal ratings do not become cardinal merely because numbers are assigned | Output quality depends on scenario scope, assumptions and calibration | Economic interpretation, independent validation, consistent taxonomy and monetary calibration require scrutiny | Peer data and tools do not automatically represent an institution's current exposure |
This table compares different functions, not regulatory approval levels or independent performance test results. Many organisations can benefit from all four.
A note about ordinal scales
A red/amber/green rating may be useful for escalation and management communication. But converting red, amber and green to numbers 3, 2 and 1 does not demonstrate that “red” contains three times the exposure of “green”. Arithmetic operations on such codes can create false precision unless the underlying scale is quantitatively defined and validated. That criticism applies to misused ordinal arithmetic, not to every purpose of a risk matrix.
A note about FAIR
FAIR's stated focus on financially derived outputs is a strength when the institution needs to compare scenario-based loss exposure in monetary terms. Risk Accounting does not yet supply the same output merely by reporting an RU value. Calling FAIR “cyber-only” or implying that a Risk Unit replaces a monetary distribution would distort both methods. [3]
4. Worked example: one bank, different questions
Consider the Commercial Loans (Trade Receivables) case published in Hughes's Risk Accounting (2023). The book explicitly identifies its underlying data as fictitious and simplified. The example permits a calculation of Risk Accounting's exposure quantities; it does not contain the additional scenario frequency, severity and dependence inputs needed to generate a valid FAIR financial-loss result. [1, pp. 144, 148–149, 163–168]
Step 1 — Establish the published product activity. Daily new business is US$147 million, which selects VBW = 33.1 from the published Value Table.
Step 2 — Establish the approved exposure categories. Published EUFs are Processing 10, Lending 8, Selling 2, Funding 6 and Interest Rate 6. Trading is not applicable in this case. The applicable EUFs sum to 32, producing:
Step 3 — Use the product-level control-assessment totals. For the whole Commercial Loans product, the published Actual Score Aggregate is 28,956,542 and the Maximum Score Aggregate 42,593,080:
Step 4 — Determine the residual exposure at the same scope.
Interpretation by discipline: An ERM programme could use this result alongside risk appetite and control observations; a FAIR analysis would require a separately specified loss scenario and empirical or expert-supported monetary parameters; ORX peer information could help challenge relevant assumptions or scenarios. The Risk Accounting figure itself states current model-defined exposure under its inputs. None of the latter exercises has been performed by this numerical illustration.
5. Should Risk Accounting replace RCSA, loss databases or scenario models?
No. In Risk Accounting, E-RCSA is a central input. It replaces arbitrary traffic-light coding within the method's assessment mechanism with governed benchmark scores, category weights and mappings, but it still depends on credible evidence and professional challenge. Existing RCSA processes may provide control inventories, control owners and tests relevant to an E-RCSA transition.
Historical loss data serves a different role: it describes realized outcomes and can support the statistical tests required to investigate the monetary significance of residual RUs. Scenario analysis can explore extreme but plausible conditions; it is particularly relevant when observed losses are sparse or future systems change. Audit findings, key risk indicators, incidents and threat intelligence can likewise complement a current-exposure view.
A sound implementation might combine these disciplines into a layered decision process:
- Identify and govern the risk with ERM, accountabilities and controls.
- Measure current activity-linked exposure, if an RU method version and valid inputs are available.
- Challenge the state of controls through evidenced assessments and independent testing.
- Model particular outcomes, where warranted, with an appropriate financial risk model.
- Compare realized incidents and peer data against earlier assessments and hypotheses.
- Reassess decisions and any claim of cross-method agreement using observed results rather than labels alone.
No automatic conversion between an ERM heat-map position, a FAIR dollar estimate, an ORX peer-loss amount and an RU balance is warranted.
6. Governance, auditability and regulatory positioning
Published Risk Accounting specifies an arithmetic and reporting chain whose reproducibility can be checked. Institutional adoption would require an approved model version, documented product and risk-family mappings, reliable activity data, effective-dated controls evidence, reviewable category weights, aggregation rules, exception management and independent challenge. Auditability of an input trail is not identical to proof that the measure predicts losses or improves decisions. [1]
For banks, the Basel Committee's Principles for the Sound Management of Operational Risk describe the governance and operational-risk-management expectations relevant to this discussion. BCBS 239 addresses data aggregation and risk reporting capabilities; it does not endorse or require RUs. The European Union's CRR3 introduced a Business Indicator Component-based Pillar 1 operational-risk capital calculation from January 2025. An RU result does not replace that legal capital formula. [5][6][7]
A further EU development is the EBA consultation launched on 26 August 2026 on draft RTS for institutions' operational-risk management frameworks under Article 323 CRR, with responses due 31 December 2026. That is a consultation, not an enacted requirement to use Risk Accounting. [8]
The most defensible claim is therefore narrower and testable: Risk Accounting offers a way to represent exposure and assessed mitigation in a consistent, traceable unit under a specified method, and invites empirical testing of its decision usefulness and economic interpretation. Comparative superiority would require matched data, evaluation criteria and independent evidence.
7. Frequently asked questions
Is Risk Accounting the same as enterprise risk management?
No. ERM is the overarching discipline for governing and managing enterprise risks. Risk Accounting is a particular method proposed for measuring present non-financial-risk exposure within such a programme.
Does FAIR quantify only cybersecurity risk?
No. FAIR's own description covers cybersecurity and operational risk and emphasizes financially derived risk outcomes. Its measurement objective differs from the RU-based current-exposure metric. [3]
Can a Risk Unit be directly converted into a FAIR expected-loss estimate?
Not without a separate, validated empirical relationship and compatible measurement scope. An RU is not a currency amount, a loss-frequency parameter or an expected-loss estimate by definition.
Are Risk Units mandated by Basel or the EBA?
No. Basel and EU regulations establish regulatory obligations, including operational-risk management and capital rules; they do not currently prescribe Risk Accounting's RU as their mandatory calculation unit. [5][7]
8. Explore the method and evidence
Start with What Is a Risk Unit?, or read How to Quantify Operational Risk for a broader overview of available measurement approaches. See how Inherent Risk differs from Residual Risk, and how Enhanced RCSA informs the Risk Mitigation Index.
For a concrete application, explore Transaction-Level Risk Accounting and the public Payment Orders controlled synthetic demonstrator. The high-level sequence appears on The Method. Institutions and researchers interested in controlled evaluation can contact RASB.
Return to the Risk Accounting Knowledge Base for all nine topics.
Sources
- Hughes, Peter J. Risk Accounting: The Complete Guide to Quantifying and Accounting for Non-Financial Risks (Grosvenor House Publishing, 2023), especially pp. 136–141, 144, 148–149, 163–168. A published fictitious example is used; no live institution-level measurement is claimed. Bibliographic record.
- RASB. Risk Accounting and The Method — public conceptual distinction between current exposure, monetary valuation and accounting recognition.
- FAIR Institute. FAIR Frequently Asked Questions — stated scope, scenario-oriented financial quantification and relationship to other assessment methods.
- ORX. ORX Services and Risk Management Resources Explorer — loss-data exchange, scenario practices, taxonomies and risk-management resources.
- BCBS. Principles for the Sound Management of Operational Risk — current Basel consolidated guidelines, ORR10 (2026 presentation of the underlying principles).
- BCBS. Principles for effective risk data aggregation and risk reporting (BCBS 239).
- EBA. Risk Assessment Report — December 2025 — explanation of the CRR3 business indicator component context, applicable from January 2025.
- EBA. Consultation on draft operational-risk-management RTS — 26 August 2026 — proposal status and closing date.
Explore Methodology & Empirical Research
Compare these definitions with the mathematical specifications on The Method or examine transaction-level operational risk quantification in the Payment Orders Study.
