Derivatives Valuation Model Governance Failure

This is the most dangerous risk in derivatives - not market risk, but Model Risk: the loss you suffer because the model you used to price, hedge and control risk was wrong, manipulated, or not understood by governance.

RBI defines it in line with Fed SR 11-7: Model Risk = risk of loss from decisions based on incorrect or misused model outputs.

1. What is Valuation Model Governance?

A derivative has no observable market price for many OTC contracts. You calculate the price.

Valuation Governance = 4 lines of defense:

A. Model Development

• Clear purpose, theoretical basis, assumptions
• Input data quality - yield curves, volatility surfaces, correlation

B. Independent Model Validation

• Validator cannot be developer or trader
• Tests: conceptual soundness, back-testing, benchmarking to alternative models, sensitivity

C. Implementation & Use

• No Excel manual copying. Version control. Access controls.
• Trader cannot override model without Risk sign-off

D. Ongoing Monitoring & Audit

• Daily P&L attribution, IPV - Independent Price Verification
• Breach of risk limits must be escalated, not hidden

Failure in any one is governance failure.

2. How Governance Fails - 6 Typical Patterns

1. Input Manipulation / Mark Manipulation: Trader inputs own volatility to show profit
2. Flawed Model Design: Value-at-Risk model underestimates tail risk
3. Lack of Understanding: Board does not understand Greek exposures
4. Override of Controls: Risk limits breached multiple times, yet no corrective action
5. No Independent Validation: Desk that makes money also validates model
6. Lack of Transparency: There was a lack of transparency and understanding of the derivatives products. A 'no risk of failure' attitude predominated

Research shows well-governed firms are more likely to use derivatives to hedge rather than to speculate, weak internal governance structures are more likely to take a view with derivatives.

3. Major Global Case Laws / Enforcement Cases

Case 1: JPMorgan London Whale - The Textbook Governance Failure

Facts: Bruno Iksil, trader in London CIO, took enormous positions in credit derivatives, earning nickname London Whale. By early 2012, Iksil's trading book had breached JPMorgan's risk limits multiple times, yet the bank failed to enforce corrective measures. A flawed risk model downplayed the exposure, creating a false sense of security.

Loss: $6.2 Billion.

Governance Failures Identified by Senate & FCA:

• JPMorgan was solvent and liquid - its $2.3 trillion balance sheet and fortress capital position held. What failed was governance: the bank's ability to know and understand the risks it was actually running. This maps directly onto Basel's Three Pillars where London Whale is a Pillar III governance failure, not Pillar I or II
• VaR model changed from 95% to new model that halved VaR overnight, without validation
• P&L manipulation via using trader's marks vs mid-market
• Lack of oversight and proper governance allowed for these trades to go through without being properly reviewed or flagged
• FCA fined JPMorgan Chase Bank NA £137.6 million for serious failings connected to its CIO. The breaches related to $6.2 billion trading loss

Legal Outcome: SEC, OCC, FCA actions. Underlines that model risk management failure is personal liability for Board.

Case 2: Deutsche Bank - LIBOR & Self-Regulation Failure

Deutsche Bank has paid $18B+ in fines across: LIBOR rate manipulation ($2.5B), mirror trading scheme ($10B laundered, $630M fine), mortgage-backed securities fraud. Deutsche Bank represents complete failure of self-regulation model: every compliance system failed, every internal control was overridden.

Valuation issue: LIBOR was the key input for valuing trillions of interest rate swaps. Manipulating LIBOR = manipulating valuation model.

Case 3: Vestia vs Deutsche Bank, Barclays etc. - Ultra Vires Derivatives [English High Court 2014]

Dutch social housing corp Vestia entered into €23bn derivatives. Court determined whether:

1. various derivatives entered into by Vestia were outside scope of its constitutional documents under Dutch law i.e. ultra vires
2. derivatives were void under English law due to lack of capacity and authority

Principle: If company lacks capacity to enter derivatives for speculation, contract void. This is governance failure at client suitability level.

Case 4: SEC vs Yorkville & Alderman - Valuation Process Failure

SEC enforcement focusing on valuation issues. Both cases examine closely purported conduct of valuation committees and how those committees performed their assigned roles.

SEC criticized allegedly deficient internal processes and alleged failures to implement meaningful guidance - i.e., model governance on paper but not in practice.

Case 5: AIG Financial Products (2008) - Model Failure

AIG sold CDS protection on subprime using Gaussian Copula model that assumed house prices would not fall nationally. No independent validation, no collateral stress test. Result: $99bn loss, US Government bailout $182bn. Auditor case - trustees required expert opinions on management decisions on continued investment in derivatives which ultimately caused firm to fail, and on accounting treatment recommended for derivatives by auditors .

4. Indian Case Laws on MTM Valuation & Governance

Indian tribunals are now creating jurisprudence on MTM.

1. The Principle of Prudent Accounting - MTM Loss Allowable

Bombay HC held mark to market loss in respect of forward contracts claimed as loss from Business income cannot be disallowed as the forward contracts were secured to cover variation in foreign exchange rate which would impact its business of import and export of diamonds .

Tribunal: MTM loss is an actual and ascertained liability and only payment falls on future date, loss is real and accurately determined. View that provision for MTM losses on foreign currency swaps is contingent liability is totally wrong .

2. CBDT Instruction 3/2010 vs Judicial View

CBDT Instruction No.3/2010 laid down that whenever there is no sale or settlement, losses on financial instruments on MTM basis are notional losses and contingent .

But courts rejected blanket disallowance. In Prime Woven Ltd, it is held that mark to market losses on account of trading in derivatives are allowable . Tribunal quashed revision orders under section 263, finding AO and PCIT had incorrectly treated settled losses as notional and speculative without proper appreciation of facts .

Given Enercon's proven import hedges and consistent method accepted in other years, MTM loss was allowable. Unrealised losses recognised under prudent accounting allowable even if not finally crystallised .

3. Rajshree Sugars & Chemicals Ltd vs Axis Bank (Madras HC 2008-2011)

First major mis-selling case. Axis Bank sold exotic derivatives - Range Accrual, KIKO - to sugar company without explaining MTM risk. Loss Rs 35 cr. Court held bank owed duty to explain valuation model. Settlement reached after RBI intervened. Led to RBI Circular on Customer Suitability and Appropriateness.

This is India's London Whale at SME level - 2007-09 saw 100+ companies lose $2bn due to model opacity.

5. What Good Governance Looks Like - RBI & SEBI Checklist

After these failures, RBI Master Direction 2021 mandates:

1. Board-approved Model Risk Policy
2. Three lines: Front Office = owner, Market Risk = validator, Internal Audit = assurance
3. Independent Price Verification (IPV) daily
4. Model Inventory: Every pricing model registered, with Tier 1/2/3 criticality
5. No Excel model for >$10M exposure without IT controls
6. Stress on Model Assumptions: What if correlation goes to 1? What if liquidity dries?
7. Documentation: Why Black-Scholes not Local Vol? Source of vol surface?

Failure to do this is now considered lack of technical expertise and adequate procedures to develop and document common stock warrant analysis on applicability of ASC 815 Derivatives and Hedging - a material weakness in ICFR under SOX.