Type a keyword to filter the 26 learning sections.
LEARNING SECTION 01
What is operational risk, and why does a bank need capital for it?
Operational risk arises when people, processes or systems fail, or external events cause losses. Basel’s definition includes legal risk and excludes strategic and reputational risk. Fraud, processing errors, cyber incidents, mis-selling claims and disruption can damage a bank even when borrowers repay and market prices remain stable.
Failure to repay
A borrower does not meet its debt obligations.
Adverse price movement
A trading position loses value as market factors change.
Failure in execution or control
An unauthorised payment, failed system or legal breach creates a loss.
Capital is the loss-absorbing cushion. Controls reduce incident likelihood; resilience limits disruption; insurance can transfer some financial consequences. Each contributes differently to protection. Holding capital does not make an insecure payment process acceptable.
LEARNING SECTION 02
Which seven loss-event categories should students recognise?
| Basel event category | Illustrative banking event |
|---|---|
| Internal fraud | An employee circumvents controls and diverts funds. |
| External fraud | A third party uses stolen credentials for fraudulent payments. |
| Employment practices and workplace safety | A workplace discrimination claim or employee injury creates liability. |
| Clients, products and business practices | Mis-selling, unsuitable advice or a compliance breach leads to redress. |
| Damage to physical assets | Flood or fire damages premises and equipment. |
| Business disruption and system failures | A payment platform outage interrupts critical transactions. |
| Execution, delivery and process management | Incorrect settlement instructions or a reconciliation failure causes a loss. |
A cyber incident can generate fraud, legal, processing and disruption consequences. Classify consistently using the event’s substance and the applicable taxonomy; avoid treating “cyber” as an automatic eighth Basel event type. Group losses with a common root cause and preserve their individual accounting impacts.
Source: Basel II comprehensive framework · operational-risk section, paragraphs 644–683 · BCBS Principles for the Sound Management of Operational Risk · March 2021
LEARNING SECTION 03
How did operational-risk capital evolve under Basel?
| Stage | Approach / reform | Why it matters |
|---|---|---|
| 1988 · Basel I | No separate explicit Pillar 1 operational-risk charge | The Accord focused primarily on credit risk. |
| 1996 · market-risk amendment | Explicit trading-book market-risk capital | This was not an operational-risk model framework. |
| 2004 / 2006 · Basel II | BIA, TSA / permitted ASA, and AMA | Operational risk became an explicit Pillar 1 capital category. |
| 2009 · Basel 2.5 | Primarily trading-book and securitisation reforms | No separate operational-risk “Basel 2.5 formula” replaced the Basel II approaches. |
| 2010–11 · initial Basel III | Stronger capital quality and buffers | The operational-risk measurement approaches were not immediately replaced. |
| 2014 / 2016 · consultations | Review of standardised approaches; proposed standardised measurement approach | Addressed weak risk sensitivity and model comparability. |
| December 2017 · final Basel III reforms | One new standardised operational-risk approach replaces BIA, TSA and AMA | Combines a business proxy with internal loss experience where applicable. |
| March 2020 / January 2023 | BCBS target deferred by a year to 1 January 2023 | Domestic implementation dates differ. |
| 2021 · risk management and resilience | Revised sound-management principles and resilience principles | Capital sits alongside governance and continuity planning. |
| India · 2023 / 2025 / 2026 / 2027 | 2023 directions consolidated into 2025 instructions; October 2026 amendment sets April 2027 implementation | Use the final domestic rules and commencement date. |
Source: BCBS history · Basel III: Finalising post-crisis reforms · December 2017 · BCBS Basel III implementation deferral · March 2020 · RBI Forthcoming Instructions Directions, 2025 · Chapter IV
LEARNING SECTION 04
Which approaches belong to Basel II and Basel III?
BIA
Apply 15% to the average of positive annual gross income.
TSA / ASA
Use business-line indicators with prescribed coefficients.
AMA
Approved internal measurement of annual operational losses.
Stronger capital, continuing methods
Applicable Basel II measurement approaches could remain until the local replacement took effect.
One standardised framework
BI → BIC; loss history → LC → ILM; capital = BIC × applicable ILM.
The two uses of “standardised” refer to different methods. Basel II TSA allocates income across eight business lines. Basel III SA uses financial-statement BI components and marginal BIC buckets. A bank does not freely choose BIA or AMA after the replacement SA becomes applicable.
Source: Basel II comprehensive framework · operational-risk section, paragraphs 644–683 · Basel III: Finalising post-crisis reforms · December 2017
LEARNING SECTION 05
How do frequency, severity, expected loss and tail risk fit together?
Many small errors create high-frequency, low-severity losses. A major fraud or litigation event can create a low-frequency, high-severity tail. Average losses and catastrophic losses therefore require different attention.
E[L] = E[N] × E[X]
Var(L) = E[N] × Var(X) + Var(N) × E[X]²
A compound-Poisson teaching model
Assume N is Poisson with mean 20 events per year. Each event has mean severity ₹0.10 crore and standard deviation ₹0.20 crore, independent of other events and N. Expected annual loss = 20 × 0.10 = ₹2.00 crore. Variance = 20 × 0.04 + 20 × 0.01 = 1.00 crore²; standard deviation = ₹1.00 crore.
Annual loss VaR at 99.9% is a loss quantile with a modelled 0.1% upper-tail probability; ES is the average loss beyond the 99.9% VaR. If several scenarios sit exactly at VaR, include only the share needed to make up the worst 0.1% tail. These moments do not determine a reliable 99.9% quantile for a heavy-tailed loss distribution. A normal approximation can materially understate tail risk. Choosing a severity distribution, modelling dependence and simulating aggregate losses require separate assumptions.
Figure 1 · Average loss and the extreme tail
LEARNING SECTION 06
Basel II · Approach 1: the Basic Indicator Approach (BIA)
BIA uses a broad income proxy rather than a loss-distribution model. Let GI be annual gross income under the regulatory definition. Exclude zero and negative GI years from both numerator and denominator; use only positive years in the three-year window.
Operational RWA = 12.5 × KBIAn is the number of positive-GI years in the last three years.
Positive years only
| Year | GI | Included in BIA? |
|---|---|---|
| Year 1 | 1,000 | Yes |
| Year 2 | −200 | No |
| Year 3 | 1,200 | Yes |
K = 15% × 1,100 = ₹165.00 crore
RWA = 12.5 × 165 = ₹2,062.50 crore
Dividing by three would incorrectly dilute the proxy. If no year has positive GI, the simple formula has no positive observations; the laboratory displays a zero proxy charge and flags the case for supervisory review rather than dividing by zero.
Gross income is not net profit. The Basel II measure starts from net interest income plus net non-interest income, before provisions and operating expenses, excluding realised gains or losses from sales of banking-book securities, extraordinary or irregular items, and insurance income, subject to the regulatory accounting adjustments. Use the regulatory definition before applying 15%.
Source: Basel II comprehensive framework · operational-risk section, paragraphs 644–683
LEARNING SECTION 07
Basel II · Approach 2: the Standardised Approach (TSA)
TSA divides activities into eight business lines. Multiply each line’s annual GI by its beta, add the results within each year, floor the annual total at zero, then average the three annual totals. Negative contributions can offset positive contributions within a year; a negative annual aggregate does not offset another year.
| Business line | Beta |
|---|---|
| Corporate finance | 18% |
| Trading and sales | 18% |
| Retail banking | 12% |
| Commercial banking | 15% |
| Payment and settlement | 18% |
| Agency services | 15% |
| Asset management | 12% |
| Retail brokerage | 12% |
Source: Basel II comprehensive framework · operational-risk section, paragraphs 644–683
LEARNING SECTION 08
Case 1 · Calculate TSA across all eight business lines
| Business line | Year 1 GI | Beta | Year 1 contribution |
|---|---|---|---|
| Corporate finance | 100 | 18% | 18.00 |
| Trading and sales | 200 | 18% | 36.00 |
| Retail banking | 400 | 12% | 48.00 |
| Commercial banking | 300 | 15% | 45.00 |
| Payment and settlement | 80 | 18% | 14.40 |
| Agency services | 60 | 15% | 9.00 |
| Asset management | 50 | 12% | 6.00 |
| Retail brokerage | 40 | 12% | 4.80 |
The Year 1 weighted aggregate is 18 + 36 + 48 + 45 + 14.40 + 9 + 6 + 4.80 = ₹181.20 crore. Assume the completed weighted aggregates for Year 2 and Year 3 are −₹20 crore and ₹210 crore.
= ₹130.40 crore
RWA = 12.5 × 130.40 = ₹1,630.00 crore
The negative Year 2 amount is floored after aggregation across lines. Do not floor every business-line contribution separately, and do not divide by two. These data differ from the preceding BIA example; the totals do not measure a capital saving from changing approach.
LEARNING SECTION 09
Where did the Alternative Standardised Approach (ASA) fit?
Under Basel II, national supervisors could permit ASA. It retained the standardised structure but replaced GI in retail and commercial banking with a prescribed proxy based on loans and advances. It was not a bank’s unapproved election and is not an additional current Basel III approach.
Commercial proxy charge = 15% × 0.035 × LAcommercialLA is the prescribed three-year average outstanding loans and advances measure; other business lines follow the applicable standardised treatment.
The two substituted business lines
Assume eligible three-year average LA of ₹10,000 crore in retail and ₹8,000 crore in commercial banking. Retail contribution = 0.12 × 0.035 × 10,000 = ₹42 crore. Commercial contribution = 0.15 × 0.035 × 8,000 = ₹42 crore. Together they contribute ₹84 crore before including the other lines and completing the applicable ASA calculation.
Source: Basel II comprehensive framework · operational-risk section, paragraphs 644–683
LEARNING SECTION 10
Basel II · Approach 3: the Advanced Measurement Approaches (AMA)
AMA required supervisory approval and quantitative and qualitative standards. The bank combined internal loss data, relevant external data, scenario analysis, and business-environment / internal-control factors. It had to capture a one-year loss horizon at a 99.9% confidence standard. There was no single universal Basel AMA formula.
Internal + external data
Internal history alone can miss rare but severe events.
Scenario analysis
Experts assess plausible tail events and changing exposures.
Governance and validation
Measurement must connect to actual risk management.
Expected versus unexpected loss in a model
Assume an approved model produces annual 99.9% loss VaR of ₹500 crore and annual expected loss of ₹40 crore. VaR − EL is ₹460 crore as an economic-capital illustration. Under legacy AMA, the regulatory charge generally covered EL and UL unless the bank demonstrated adequate capture of EL in its internal business practices; subtracting ₹40 crore was not automatic.
Historical AMA insurance recognition was subject to eligibility conditions and capped at 20% of the AMA capital requirement. This historical allowance is not a blanket insurance deduction under the final Basel III SA.
Source: Basel II comprehensive framework · operational-risk section, paragraphs 644–683
LEARNING SECTION 11
Why did Basel replace AMA and the old standardised methods?
A simple income percentage can be insensitive to controls and actual losses. Internal models can be more tailored, yet differences in data, scenarios, assumptions and modelling choices make outputs difficult to compare. The final reform replaced this menu with one standardised method, improving consistency while retaining a role for loss history where the rules require it.
| Earlier weakness | Final framework response |
|---|---|
| Gross income could poorly represent business complexity | A broader Business Indicator with three components |
| Internal-model outputs varied across banks | Prescribed marginal BIC coefficients and ILM formula |
| Rare events were difficult to estimate consistently | Historical loss component rather than a bank-specific tail model for Pillar 1 |
| Poor data could obscure the risk | Loss-data standards, supervisory review and disclosure |
The reform standardises regulatory capital. It does not make internal scenario analysis, economic-capital modelling or control assessment unnecessary. Those remain useful for risk appetite, stress testing and Pillar 2.
Source: Basel III: Finalising post-crisis reforms · December 2017
LEARNING SECTION 12
Basel III SA · What is the complete calculation sequence?
ILDC + SC + FC
Business Indicator · BI12% / 15% / 18%
Business Indicator Component · BICAverage annual net losses × 15
Loss Component · LCLC relative to BIC, subject to regime
Internal Loss Multiplier · ILMORC = BIC × ILM
Operational RWA = 12.5 × ORCLEARNING SECTION 13
Step 1 · How are ILDC, SC and FC combined into BI?
| Component | Calculation using three-year averages | Reason |
|---|---|---|
| ILDC · interest, leases and dividend | min[average |interest income − interest expense|, 2.25% × average interest-earning assets] + average dividend income | Broad intermediation proxy with a cap on the interest term. |
| SC · services | max[average fee income, average fee expense] + max[average other operating income, average other operating expense] | Income and expense indicate the scale of service activity. |
| FC · financial | average |net trading-book P&L| + average |net banking-book financial P&L| | Profits and losses both reflect financial activity. |
Take absolute values year by year before averaging the relevant net items. Apply the SC maxima to the three-year averages, not separately to each year and then average. Net trading and banking-book P&L are separate absolute-value series. Regulatory constituent definitions determine which accounting entries belong in each component.
Why absolute values come first
Trading-book P&L of +₹100, −₹80 and +₹40 crore gives average absolute P&L = (100 + 80 + 40) / 3 = ₹73.33 crore. Taking the absolute value of the average instead gives ₹20 crore and understates this component.
LEARNING SECTION 14
Case 2 · Build BI directly from three years of financial data
| Item | Year 1 | Year 2 | Year 3 | Required average |
|---|---|---|---|---|
| Interest income | 1,000 | 1,100 | 1,200 | 1,100 |
| Interest expense | 600 | 700 | 800 | 700 |
| Absolute net interest | 400 | 400 | 400 | 400 |
| Interest-earning assets | 15,000 | 16,000 | 17,000 | 16,000 |
| Dividend income | 15 | 20 | 25 | 20 |
| Fee income | 180 | 200 | 220 | 200 |
| Fee expense | 40 | 50 | 60 | 50 |
| Other operating income | 20 | 30 | 40 | 30 |
| Other operating expense | 30 | 40 | 50 | 40 |
| Net trading-book P&L | 100 | −80 | 40 | Absolute average: 73.3333 |
| Net banking-book financial P&L | 30 | −20 | 10 | Absolute average: 20 |
SC = max(200,50) + max(30,40) = 240
FC = 73.3333 + 20 = 93.3333
BI = 380 + 240 + 93.3333 = ₹713.3333 crore
Under RBI’s first bucket, BIC = 12% × 713.3333 = ₹85.60 crore. With the applicable first-bucket ILM of 1, ORC is ₹85.60 crore and operational RWA is ₹1,070.00 crore. Retain full precision until the final display.
LEARNING SECTION 15
Step 2 · How does the marginal BIC structure work?
| Bucket | Basel BI range | RBI BI range · ₹ crore | Marginal coefficient |
|---|---|---|---|
| 1 | Up to €1 billion | Up to 8,000 | 12% |
| 2 | Above €1bn to €30bn | Above 8,000 to 2,40,000 | 15% |
| 3 | Above €30 billion | Above 2,40,000 | 18% |
+ 0.15 × min[max(BI−8,000,0),2,32,000]
+ 0.18 × max(BI−2,40,000,0)BI and BIC are in ₹ crore. Basel thresholds are expressed in euros; the RBI thresholds are prescribed domestic amounts, not a daily FX conversion.
₹12,000 crore BI
BIC = 12% × 8,000 + 15% × 4,000 = ₹960 + ₹600 = ₹1,560 crore. Applying 15% to the entire ₹12,000 would give ₹1,800 crore and is wrong.
₹3,50,000 crore BI
BIC = 0.12 × 8,000 + 0.15 × 2,32,000 + 0.18 × 1,10,000 = ₹55,560 crore. The effective rate is 15.87%; the marginal rate on the top slice is 18%.
Source: OPE25 · standardised approach, updated 5 July 2024 · RBI Forthcoming Instructions Directions, 2025 · Chapter IV
Figure 2 · Marginal coefficients rise in steps
LEARNING SECTION 16
Step 3 · How do operational losses become LC?
LC = 15 × average annual loss
A ten-year history
| Annual net losses · ₹ crore | Total | Annual average | LC |
|---|---|---|---|
| 80, 100, 120, 140, 160, 180, 200, 220, 140, 160 | 1,500 | 150 | 2,250 |
The ₹1,500 crore is a ten-year sum. LC = 15 × 150 = ₹2,250 crore, not 15 × 1,500. Each annual amount must already reflect the eligible loss population and recovery treatment.
Loss data are not a discretionary list of unusual expenses. Occurrence, discovery and accounting dates, gross losses, recoveries, root causes and event grouping need consistent definitions. A short or poor-quality history cannot simply be treated as a zero-loss record.
Source: OPE25 · standardised approach, updated 5 July 2024 · RBI Forthcoming Instructions Directions, 2025 · Chapter IV
LEARNING SECTION 17
Step 4 · How does the Internal Loss Multiplier work?
| LC / BIC | Formula ILM | Interpretation |
|---|---|---|
| 0 | 0.5413 | Mathematical value; eligibility and domestic rules govern its use. |
| 0.5 | 0.8297 | Lower losses relative to the business proxy. |
| 1 | 1.0000 | Loss component equals BIC. |
| 2 | 1.2411 | Higher losses relative to BIC. |
Under the Basel default treatment, bucket 1 uses ILM = 1. National discretion can alter the loss-data application, including setting ILM = 1 for all banks. RBI’s prescribed regime uses BIC alone in bucket 1 and loss-sensitive capital for qualifying bucket 2 and 3 banks; it does not set ILM to 1 for every bank.
Complete the ₹12,000 crore BI example
BIC = ₹1,560 crore; average annual net loss = ₹150 crore; LC = ₹2,250 crore. Then LC/BIC = 1.4423, ILM = 1.117996 and ORC = 1,560 × 1.117996 = ₹1,744.0739 crore. Operational RWA = ₹21,800.9234 crore. The laboratory reproduces these results.
Source: OPE25 · standardised approach, updated 5 July 2024 · RBI Forthcoming Instructions Directions, 2025 · Chapter IV
Figure 3 · Loss experience scales capital nonlinearly
LEARNING SECTION 18
Which losses and recoveries enter the calculation?
| Item | Treatment principle |
|---|---|
| Gross loss | Capture before recoveries; include eligible direct charges, write-downs and settlements. |
| Recovery | Reduce eligible losses when payment is actually received; an expected insurance receivable is not cash recovered. |
| Provision followed by settlement | Include the initial operational-loss provision and only the incremental settlement impact later. |
| Routine maintenance, improvements and insurance premiums | Do not automatically classify these as operational losses. |
| Credit-related operational event | Apply the credit / operational boundary rules to avoid double counting credit losses already captured in credit RWA. |
| Market-related operational event | Operational failures connected with trading remain relevant operational losses. |
| Common root cause | Group the event consistently while retaining accounting-year impacts. |
A recovery and a later settlement
An eligible gross fraud loss is ₹100 crore. Received recoveries are ₹20 crore; a further ₹10 crore insurance claim is still unpaid. Net recorded loss for this illustration is ₹80 crore, not ₹70 crore.
Separately, a legal event generates a ₹50 crore provision in Year 1 and settles for ₹60 crore in Year 2. Record ₹50 crore in Year 1 and the ₹10 crore increment in Year 2; recording 50 + 60 would double count.
Source: Basel Framework · OPE10 and OPE25 · RBI Forthcoming Instructions Directions, 2025 · Chapter IV
LEARNING SECTION 19
Which rules apply in India, and from when?
The RBI issued its operational-risk master direction on 26 June 2023 and consolidated the forthcoming framework in Chapter IV of the 28 November 2025 Directions. The Second Amendment dated 23 October 2026 sets 1 April 2027 as the implementation date and adds financial-statement mapping and clarifications. Read the base Chapter IV together with that amendment.
| Topic | RBI treatment |
|---|---|
| Before commencement | Applicable existing capital rules, including BIA, continue for covered banks until the replacement comes into force. |
| Covered institutions | The Forthcoming Instructions define their scope; exclusions include SFBs, LABs, Payments Banks and RRBs. |
| BI thresholds · ₹ crore | 8,000 and 2,40,000; marginal rates 12%, 15%, 18%. |
| Bucket 1 | ORC = BIC. |
| Buckets 2 and 3 with qualifying loss history | ORC = BIC × ILM. |
| Loss window | Normally ten years; a first-time ILM bank with five or more available high-quality years uses that available history, subject to the rules. |
| Insufficient / non-qualifying data | BIC-based minimum and possible supervisory uplift; a short history is not proof of low risk. |
| October 2026 clarifications | LC uses losses net of recoveries; indicative BI mapping and treatment of damaged / destroyed assets are clarified. |
The revised approach replaces the old calculation for banks to which it applies. The prescribed financial-statement mapping is indicative: classify entries according to the regulatory descriptions where accounting presentation differs. Regulatory status on this page is dated 2 November 2026.
Source: RBI Forthcoming Instructions Directions, 2025 · Chapter IV · RBI Second Amendment Directions · 23 October 2026
LEARNING SECTION 20
Step 5 · Convert the charge into RWA and test all capital ratios
Total RWA = credit RWA + market RWA + operational RWA
CET1 ratio = CET1 / total RWA
Tier 1 ratio = (CET1 + AT1) / total RWA
Total capital ratio = (CET1 + AT1 + Tier 2) / total RWA
The 12.5 factor is the reciprocal of Basel’s 8% conversion standard. RBI’s higher minimum total-capital percentage does not make the conversion become 1 ÷ 9% ≈ 11.11. The prescribed conversion remains 1 ÷ 8% = 12.5. Buffers and higher requirements are applied to the resulting RWA denominator.
An operational-capital increase affects every ratio
Start with credit RWA ₹800 crore, market RWA ₹100 crore and ORC ₹8 crore: operational RWA = ₹100 crore; total RWA = ₹1,000 crore. CET1 ₹80 crore, AT1 ₹15 crore and Tier 2 ₹20 crore yield 8.00%, 9.50% and 11.50%.
If ORC rises to ₹12 crore while capital stays unchanged, total RWA becomes ₹1,050 crore. The ratios fall to 7.62%, 9.05% and 10.95%. To restore RBI baseline ratios including the full CCB, one qualifying mix is +₹4.00 crore CET1, +₹0.75 crore AT1 and +₹1.00 crore Tier 2, totalling ₹5.75 crore. Adding ₹5.75 crore CET1 alone also satisfies all three.
This isolates the denominator effect. A newly recognised operational loss may also reduce earnings and CET1, subject to accounting, tax and regulatory adjustments. Do not assume the RWA increase is the only impact.
| Baseline including full 2.5% CCB | CET1 | Tier 1 | Total |
|---|---|---|---|
| Basel | 7.0% | 8.5% | 10.5% |
| RBI | 8.0% | 9.5% | 11.5% |
Source: RBI Prudential Norms on Capital Adequacy Directions · Basel III: Finalising post-crisis reforms · December 2017
LEARNING SECTION 21
How does a large loss affect future capital?
One event enters a ten-year window
For the ₹12,000 crore BI bank, hold BIC at ₹1,560 crore. Base ten-year losses total ₹1,500 crore, giving average ₹150 crore. If an additional eligible ₹500 crore event enters that same ten-year dataset, the average becomes ₹200 crore; LC becomes ₹3,000 crore.
New ORC = ₹1,911.6576 crore
Increase in ORC ≈ ₹167.5837 crore
Increase in operational RWA ≈ ₹2,094.7963 crore
The ₹500 crore event is a loss, not a ₹500 crore regulatory capital charge. Its future charge effect is nonlinear and depends on BI, the rest of the loss history, received recoveries and the applicable regime. The example holds the window and other inputs fixed; rolling out an older year changes the result.
LEARNING SECTION 22
How do controls, KRIs and operational resilience complement capital?
| Tool | What it adds | Illustration |
|---|---|---|
| RCSA | Assess inherent risk, control effectiveness and residual risk | Review payment approvals and reconciliation controls. |
| Key risk indicators (KRIs) | Monitor exposure or control deterioration | Failed-payment rate, unresolved breaks and critical access exceptions. |
| Scenario analysis | Examine severe but plausible events | Multi-day vendor outage with customer remediation and legal costs. |
| Loss and near-miss reporting | Learn from events and control weaknesses | A recovered payment error still reveals a control failure. |
| Third-party risk management | Assess dependencies and concentration | Identify common cloud / payment-service dependencies. |
| Resilience testing | Maintain critical operations within disruption tolerances | Test recovery, backup integrity and alternative processing. |
| Independent assurance | Challenge controls and data quality | Internal audit checks event completeness and access governance. |
A lower loss-sensitive capital requirement may follow a better loss record, but the relationship is delayed and influenced by business growth. With ILM fixed at one, stronger controls still matter even if Pillar 1 capital does not immediately fall. Financial capital cannot restore a failed system during an outage.
Source: BCBS Principles for the Sound Management of Operational Risk · March 2021 · BCBS Principles for Operational Resilience · March 2021
What must be disclosed under Pillar 3?
Basel DIS60 provides annual disclosures: ORA covers the qualitative framework; OR1 covers historical operational losses; OR2 covers BI and its subcomponents; OR3 shows BIC, ILM, minimum capital and operational RWA. OR1 applies to buckets 2–3 even if national discretion fixes ILM at one, and to approved loss-sensitive bucket 1 banks. Domestic templates and implementation rules must also be followed.
LEARNING SECTION 23
Interactive laboratories · reproduce the worked cases
Lab 1 · BIA and negative years
Lab 2 · Basel III BIC, LC, ILM and capital
Choose the prescribed thresholds and a loss-data treatment. The loss-sensitive mode assumes eligible data; this laboratory does not validate data quality.
Lab 3 · Operational RWA and the three capital ratios
Targets: RBI baseline plus full CCB, 8% / 9.5% / 11.5%. Excludes other buffers, Pillar 2, LR and further output-floor effects.
LEARNING SECTION 24
How can students reproduce the calculations in Excel?
Operational Risk: Excel Companion Workbook
Download the companion workbook to study operational-risk calculations alongside this article. Keep Excel percentages as rates and check the currency unit shown on each worksheet.
Worksheets: README · BIA · SA · Capital · BI_Case2 · TSA_Case1 · ILM_Curve · Large_Loss
Basel II Operational Risk Workbook
Practise the historical BIA, TSA, ASA and AMA approaches alongside the worked examples. These Basel II methods provide the foundation for understanding the Basel III standardised approach taught later in this article.
Tab guide: 2_BIA — Basic Indicator Approach; 3_TSA — business-line betas and annual floors; 4_ASA — Alternative Standardised Approach; 5_AMA_LDA — loss distribution, expected loss, VaR and insurance credit; 8_Q&A — revision questions and answers.
The following formulas are read directly from this downloadable workbook. The existing table below is a separate Excel reconstruction of the article’s BIA, Basel III and capital-ratio examples.
| Sheet / cell | Workbook formula | Purpose |
|---|---|---|
| 3_TSA!G10 | =$C10*D10 | β × GI for one business line and year |
| 3_TSA!G20 | =MAX(G18,0) | Floor each year’s weighted total at zero |
| 3_TSA!C23 | =SUM(G20:I20)/3 | TSA capital charge; always divided by 3 |
| 4_ASA!C16 | =$C$11*$C$9*$G$11 | Retail charge: β × 3.5% × average loans and advances |
| 5_AMA_LDA!E21 | =AVERAGE('6_AMA_Sim_Data'!$G$7:$G$10006) | Expected loss from simulated annual losses |
| 5_AMA_LDA!E22 | =_xlfn.PERCENTILE.INC('6_AMA_Sim_Data'!$G$7:$G$10006,$C$15) | 99.9% VaR of simulated annual losses |
| 5_AMA_LDA!C43 | =MIN(C41,C42) | Insurance credit, capped at 20% of the eligible charge |
AMA simulation is a teaching illustration; a 99.9% sample percentile depends on the simulated years and does not by itself establish regulatory model approval. Insurance credit remains subject to eligibility conditions as well as the cap.
Keep the three panels on separate worksheets. Currency inputs are ₹ crore unless a different unit is explicitly selected. Store rates as Excel percentages; keep full precision in formulas and format output cells to two decimals.
| Sheet / cell | Input or formula | Purpose |
|---|---|---|
| BIA!B2:B4 | 1000; −200; 1200 | Annual GI inputs |
| BIA!B6 | =IF(COUNTIF(B2:B4,">0")=0,0,AVERAGEIF(B2:B4,">0")) | Average positive years |
| BIA!B7 | =15%*B6 | BIA charge |
| BIA!B8 | =12.5*B7 | Operational RWA |
| SA!B2 / B3 | 12000 / 150 | BI / annual average eligible loss |
| SA!B4 / B5 | 8000 / 240000 | RBI thresholds |
| SA!B7 | =12%*MIN(B2,B4)+15%*MIN(MAX(B2-B4,0),B5-B4)+18%*MAX(B2-B5,0) | Marginal BIC |
| SA!B8 | =15*B3 | Loss Component |
| SA!B9 | =IF(B2<=B4,1,LN(EXP(1)-1+(B8/B7)^0.8)) | Eligible loss-sensitive regime; bucket 1 fixed at one |
| SA!B10 | =B7*B9 | ORC |
| SA!B11 | =12.5*B10 | Operational RWA |
| Capital!B2:B7 | 8; 800; 100; 80; 15; 20 | ORC / credit RWA / market RWA / CET1 / AT1 / Tier2 |
| Capital!B9 | =12.5*B2+B3+B4 | Total RWA |
| Capital!B10 | =B5/B9 | CET1 ratio |
| Capital!B11 | =(B5+B6)/B9 | Tier 1 ratio |
| Capital!B12 | =SUM(B5:B7)/B9 | Total ratio |
| Capital!B14:B16 | 8%; 9.5%; 11.5% | Selected combined thresholds |
| Capital!B18 | =MAX(0,B14*B9-B5,B15*B9-SUM(B5:B6),B16*B9-SUM(B5:B7)) | Additional CET1 alone |
LEARNING SECTION 25
Revision questions · reveal the answers
Does operational risk include legal and reputational risk?
The Basel definition includes legal risk and excludes strategic and reputational risk. An operational event can still have reputational consequences.
Why is BIA divided by two in the example?
Only two of the three annual GI observations are positive. The negative year is excluded from both the sum and the count.
Does TSA also drop a negative year?
No. Floor the annual sum of business-line weighted contributions at zero, then divide the sum of three annual results by three.
Are Basel II TSA and Basel III SA the same?
No. TSA uses eight business-line income proxies and betas. Basel III SA uses BI, marginal BIC coefficients and the applicable ILM treatment.
Can AMA still determine Pillar 1 capital after the replacement SA applies?
The final Basel III reform removes AMA as a Pillar 1 capital approach. Internal models can remain useful for risk management and Pillar 2.
Why is BI not net profit?
It is a prescribed business-scale proxy combining intermediation, services and financial activity. Expenses and negative P&L can contribute to its size.
Practice: BI is ₹12,000 crore. What is BIC?
12% × 8,000 + 15% × 4,000 = ₹1,560 crore. The 15% coefficient applies only to the upper slice.
Practice: ten-year losses sum to ₹1,500 crore. What is LC?
Average annual loss is ₹150 crore. LC = 15 × 150 = ₹2,250 crore.
What happens when LC equals BIC?
The formula ILM is ln(e−1+1) = 1. For eligible loss-sensitive treatment, ORC equals BIC.
Does zero recorded loss imply zero capital?
No. Formula ILM at LC = 0 is about 0.5413 where applicable, and bucket 1 uses ILM = 1 under the default treatment. Missing or ineligible data do not establish a zero-loss record.
Can an unpaid insurance claim reduce the loss dataset?
A recovery is recognised for this purpose when payment is received, subject to the applicable loss-data rules. A claim receivable is not the same as a received recovery.
When does the revised RBI operational-risk SA commence?
The 23 October 2026 amendment sets 1 April 2027 for Chapter IV and its amendments. Read the scope, base directions and amendment together.
Why is RWA multiplied by 12.5 despite RBI’s 9% minimum total ratio?
The Basel charge-to-RWA conversion is 1 ÷ 8% = 12.5; it does not become 1 ÷ 9% ≈ 11.11. Apply RBI capital percentages and buffers to the resulting RWA.
Is a higher operational-risk charge the only effect of a major fraud?
No. A recognised loss can reduce profit and CET1, while future capital requirements may also rise through the loss history. Operational and reputational disruption can continue separately.
LEARNING SECTION 26
Sources and connected learning
Reference date: 2 November 2026. Historical approaches are labelled by regime. The examples isolate calculation mechanics and do not replace the complete regulatory instructions. India’s base directions must be read with subsequent amendments.
- Basel II comprehensive framework · operational-risk section, paragraphs 644–683
- Basel III: Finalising post-crisis reforms · December 2017
- Basel Framework · OPE10 and OPE25
- OPE25 · standardised approach, updated 5 July 2024
- RBI Forthcoming Instructions Directions, 2025 · Chapter IV
- RBI Second Amendment Directions · 23 October 2026
- Basel DIS60 · operational-risk Pillar 3 disclosures
- RBI Prudential Norms on Capital Adequacy Directions
- BCBS Principles for the Sound Management of Operational Risk · March 2021
- BCBS Principles for Operational Resilience · March 2021
- BCBS history
- BCBS Basel III implementation deferral · March 2020