What Is a Risk Unit (RU)?
A Risk Unit (RU) is the common unit of measurement proposed by the Risk Accounting method for expressing inherent and residual exposure to non-financial risk. It enables calculations, comparisons and aggregation under consistent methodological rules. It measures exposure, not money, incident probability, expected loss or regulatory capital.
On this page
8 sections
- 1. Why does non-financial risk need a common unit?
- 2. How is a Risk Unit calculated?
- 3. Worked example: a commercial-loan portfolio
- 4. How is a Risk Unit different from a risk score or loss estimate?
- 5. When is an RU calculation auditable?
- 6. What has been demonstrated, and what remains under investigation?
- 7. Frequently asked questions
- 8. Explore the method and related concepts
- 1. Why does non-financial risk need a common unit?
- 2. How is a Risk Unit calculated?
- 3. Worked example: a commercial-loan portfolio
- 4. How is a Risk Unit different from a risk score or loss estimate?
- 5. When is an RU calculation auditable?
- 6. What has been demonstrated, and what remains under investigation?
- 7. Frequently asked questions
- 8. Explore the method and related concepts
1. Why does non-financial risk need a common unit?
A financial institution can reconcile accounts in currency, count payments, identify its customers and measure the financial value of its assets. Its non-financial risk information is often less readily comparable. A bank might have an operational loss database, a five-point risk register, control-testing outcomes, scenario estimates and an inventory of audit findings. Each contributes useful information, but they do not automatically express the same underlying quantity.
For example, a risk rated “high” for one department may not represent twice the exposure of a risk rated “medium” in another. The numbers 4 and 2, when applied to ordinal categories, are labels—not necessarily quantities that can be meaningfully multiplied or added. Conversely, the fact that no fraud loss has been recorded does not demonstrate that the institution currently carries no fraud exposure.
The Risk Unit was developed by Peter J. Hughes to address a particular measurement question: how can the non-financial-risk exposure associated with current business activity be represented in a common, auditable unit? In his published method, the answer combines the structural characteristics of an activity, its scale and a structured assessment of mitigation effectiveness. [1][2]
The resulting quantities are:
- Inherent risk (IR): exposure before the effects of mitigation are considered, expressed in RUs.
- Residual risk (RR): exposure remaining after mitigation has been assessed, also expressed in RUs.
An RU is therefore not a claim about the cost of a future event. It is a method-defined exposure measure. Converting that measure to an estimated monetary loss would require an independently justified empirical calibration. Deciding whether any monetary estimate qualifies for recognition in financial statements is a further, separate accounting question. [1][3]
2. How is a Risk Unit calculated?
The published banking method begins by specifying a product or business activity and the relevant principal risk family. Its fundamental relationship for inherent exposure in a particular family is:
Here:
- EUF — Exposure Uncertainty Factor: the approved factor representing structural exposure characteristics of the product in a specified principal risk family. It is not a probability of loss.
- VBW — Value-Based Weighting: a method-defined weighting selected from a discrete value-band table using the applicable daily new business measure. It is not a direct multiplication of the dollar amount or transaction count.
- IR — Inherent Risk: the calculated exposure before assessed mitigation. Where multiple applicable risk families are measured, the compatible family results are combined according to the method's aggregation rules.
Mitigation is assessed through Enhanced Risk and Control Self-Assessment (E-RCSA), using the applicable benchmarks, scores, category weights and business-component mappings. From these governed assessments, the method calculates:
ASA is the Actual Score Aggregate; MSA is the Maximum Score Aggregate. The denominator must be valid and nonzero. When RMI is rolled up from components to larger reporting scopes, ASA and MSA are aggregated and the RMI is recomputed; component percentages are not indiscriminately averaged. [1, pp. 140, 164–168]
Residual exposure is then calculated at the matching scope:
This complement is important. An RMI of 100% implies zero residual RUs within this calculation, not a demonstrated absence of all future losses. An RMI of 0% leaves inherent exposure unreduced by assessed mitigation. In both cases, the meaning depends on the quality, relevance and version of the underlying inputs and controls assessment.
3. Worked example: a commercial-loan portfolio
The example below follows Hughes's published Commercial Loans (Trade Receivables) case. It is fictitious and deliberately simplified, as the book explicitly states; it demonstrates the arithmetic, not actual risk measurement at a named institution. All input figures and control aggregates are taken from the published example, rather than invented for this article. [1, pp. 144, 148–149, 163–168]
Step 1 — Determine business scale and VBW
The example records US147 million in daily new business for this product. Under the published value-band schedule, that amount maps to VBW = 33.1. The underlying dollar figure selects a weighting; the calculation does not multiply US147 million by an arbitrary score.
Step 2 — Calculate inherent RUs by principal risk family
| Principal risk family | Published EUF | VBW | Inherent RUs |
|---|---|---|---|
| Processing | 10 | 33.1 | 331.0 |
| Lending | 8 | 33.1 | 264.8 |
| Selling | 2 | 33.1 | 66.2 |
| Funding | 6 | 33.1 | 198.6 |
| Interest rate | 6 | 33.1 | 198.6 |
| Total, applicable families | 32 | — | 1,059.2 |
Trading does not apply in this published product example. The EUFs are case-specific published assignments; they must not be copied automatically to other products or classifications. [1, pp. 148, 163]
Step 3 — Calculate the matching product-level RMI
The whole-product report gives an ASA of 28,956,542 and an MSA of 42,593,080:
This is not the 63.2% RMI reported elsewhere in the book for the narrower Transaction Processing part of the Commercial Loans workflow. A different reporting perimeter has a different numerator, denominator and result. [1, pp. 164–168]
Step 4 — Determine residual RUs
Apply the full-precision RMI ratio before rounding:
The published summary reports approximately 1,059 inherent RUs, 68.0% RMI and 339 residual RUs. The extra decimal precision above is provided to make the calculation independently reproducible. [1, p. 166]
Management interpretation: the method presents the volume/structure-related exposure separately from the assessed mitigating effect of controls. It can therefore help identify whether changes in measured residual exposure arise from a changing activity profile or a changing mitigation assessment. The figure of 339 RUs does not predict 339 losses, US$339 of financial harm or an accounting provision.
4. How is a Risk Unit different from a risk score or loss estimate?
| Criterion | Other risk approaches | Risk Accounting's RU |
|---|---|---|
| Principal question | Risk priority, incidents observed, potential loss severity or a loss distribution, depending on the approach | Current method-defined non-financial-risk exposure |
| Typical output | Risk category, event count, currency amount, distribution or scenario estimate | Inherent and residual Risk Units |
| Treatment of business activity | Varies by tool and model | Explicit input mapped into the approved VBW |
| Treatment of controls | Narratives, control tests, maturity scales, model parameters or scenario assumptions | Governed E-RCSA assessment summarized through RMI |
| Aggregation | Depends on unit and methodological assumptions | Possible only under compatible RU definitions, mappings and aggregation rules |
| What the number does not alone prove | Depends on measure | Future loss frequency, financial severity, monetary equivalence or regulatory capital |
These methods need not compete. FAIR describes a model for financially quantified cybersecurity and operational risk, while ORX supports multiple operational-risk tools, including loss-data and scenario practices. The distinctive contribution proposed by Risk Accounting is the common current-exposure metric and its explicit activity–mitigation calculation chain. [4][5]
5. When is an RU calculation auditable?
The arithmetic is deterministic if the same valid inputs, classifications and method version are used. An audit or independent review must go further and establish whether those inputs are substantiated. For a traceable RU record, a reviewer should be able to recover:
- Scope: product, risk family, business component, reporting date, period and method version.
- Activity: source of daily new business, currency normalization, completeness checks and approved VBW band.
- Exposure classification: relevant EUF table, supporting rationale and approval history.
- Mitigation assessment: E-RCSA criteria, observed evidence, category weights, effective date, ASA and MSA.
- Output and roll-up: calculation version, aggregation path, rounding convention, reconciliations and later adjustments.
Auditability is a property of the documented measurement process, not a guarantee that every judgment is correct or that the measure has proved its ability to predict economic outcomes. Independent evaluation must test control evidence, assessor bias, data reliability, changes to mappings and the stability of the RU's interpretation across reporting scopes. [1][2]
Regulatory boundary. The Basel Committee's Principles for the Sound Management of Operational Risk and BCBS 239 underscore governance, risk assessment and data quality. They do not require RUs. Similarly, an RU amount cannot be treated automatically as an IAS 37 provision or an operational-risk regulatory capital requirement. [6][7][8]
6. What has been demonstrated, and what remains under investigation?
The published formulas and examples provide evidence of specified calculation mechanics. They do not, by themselves, establish independent empirical accuracy, predictive usefulness or a stable monetary conversion across institutions and risk categories. Those questions require appropriately designed studies using observed data, transparent comparisons and independent replication.
RASB's Payment Orders Risk-Exposure Study is a controlled synthetic demonstrator. Its 522 daily observations illustrate the behavior of the measurement chain under recorded baseline conditions and exploratory adjustments. They are not actual commercial-bank observations and do not prove loss prediction, accounting recognition or superiority to other operational-risk methods. [9]
This distinction matters because RASB's research proposition is better served by an explicit boundary between measuring exposure, empirically relating exposure to monetary losses, and arguing for any eventual accounting application.
7. Frequently asked questions
Is a Risk Unit equal to one dollar or euro?
No. A Risk Unit is a method-defined measure of non-financial-risk exposure, not currency. Any translation into a monetary-loss estimate would require a separately validated calibration to suitable loss data.
Can Risk Units be added across business units?
The method is designed to support aggregation where common definitions, mappings and reporting rules apply. Aggregate RMI values must be recomputed from their underlying score aggregates at the correct scope; simple averaging of percentages is not a valid substitute.
Does a low number of residual RUs mean a bank is safe?
No. It indicates a lower calculated exposure within the selected method, scope and assessed inputs. It does not establish that future incidents cannot occur, that all controls are effective or that the measured value is predictive of monetary loss.
Are Risk Units required by banking or accounting standards?
No. RUs are part of the Risk Accounting methodology proposed and researched by RASB. Existing Basel and financial-reporting frameworks have their own requirements; an RU result does not, by itself, satisfy or replace them.
8. Explore the method and related concepts
Start with How to Quantify Operational Risk to see where Risk Units fit in the overall measurement problem. Then examine Inherent Risk vs. Residual Risk to understand how activity exposure and mitigation are kept separate.
For the full methodological sequence, see The Method; for a transparent controlled illustration, explore the Payment Orders research demonstrator. Institutions and researchers interested in independent testing can collaborate with RASB.
Return to the Risk Accounting Knowledge Base for the full glossary.
Additional perspectives: See Risk Accounting vs ERM and FAIR for the relationship to other risk methodologies, and Transaction-Level Risk Accounting for how exposure measurement relates to identifiable banking activity.
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. The fictitious banking case and calculations are reproduced in paraphrase with arithmetic shown independently. Bibliographic record.
- Risk Accounting Standards Board. The Method — public framework description.
- RASB. Risk Accounting — separation of risk measurement, monetary valuation and accounting recognition.
- FAIR Institute. Frequently Asked Questions — stated model scope and monetary emphasis.
- ORX. Risk Management Resources Explorer — breadth of operational-risk practices.
- BCBS. Principles for the Sound Management of Operational Risk — FSI executive summary.
- BCBS. BCBS 239: Principles for effective risk data aggregation and risk reporting.
- IFRS Foundation. IAS 37: Provisions, Contingent Liabilities and Contingent Assets.
- RASB. Payment Orders Risk-Exposure Study (2025–2026) — controlled synthetic study and stated limitations.
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.
