Cross-Currency Heath-Jarrow-Morton Framework in the Multiple-Curve Setting

This paper establishes a general cross-currency Heath-Jarrow-Morton framework within a multiple-curve setting that incorporates collateralization, arbitrary market indices, and instantaneous basis spreads to simultaneously model both forward-looking IBOR and backward-looking overnight rates for multi-currency interest rate portfolios.

The Gist

Imagine the global financial system as a massive, bustling marketplace where countries trade money, loans, and promises of future payments. For decades, this marketplace ran on a few simple rules. But recently, the rules have changed, the currency exchange booths have been upgraded, and the "trust" between banks has become a bit more complicated.

This paper by Alessandro Gnoatto and Silvia Lavagnini is essentially a new, super-flexible rulebook for how to price these financial deals in this new, messy world.

Here is the breakdown using everyday analogies:

1. The Problem: The "Broken" Marketplace

Imagine you want to borrow money.

• The Old Way: You could borrow from a bank (unsecured) or borrow from a friend who trusts you so much you don't need collateral (secured). In the past, the interest rates for these two were almost the same.
• The New Reality: After the 2008 financial crisis, banks stopped trusting each other as much. Now, if you borrow from a bank, you have to put up collateral (like a house or cash) to be safe. This created a "wedge" or a gap between the rate you pay for a risky loan and the rate you pay for a safe, collateralized loan.
• The Currency Mess: Now, imagine you are a US company borrowing in Euros. You have to deal with:
  • The US interest rate (which is changing from LIBOR to SOFR).
  • The Euro interest rate (which is still using EURIBOR).
  • The exchange rate between Dollars and Euros.
  • The fact that your collateral might be held in a third currency (like Swiss Francs).

The old math models were like a single-lane road. They couldn't handle all these different currencies, different types of interest rates (some looking forward, some looking backward), and different collateral rules all at once. They kept breaking down.

2. The Solution: The "Universal Adapter"

The authors built a Heath-Jarrow-Morton (HJM) framework. Think of this as a universal power adapter for the financial world.

Instead of building a new model for every specific currency or every specific type of loan, they created one giant, flexible mathematical engine that can plug into any situation.

• The "Abstract Index": Imagine a smart thermostat that can measure temperature, humidity, and pressure all at once. The authors created an "abstract index" that can measure any financial rate, whether it's a classic bank rate (like EURIBOR), a new overnight rate (like SOFR), or even something totally different like the price of electricity or temperature.
• The "Collateral Currency": In the old days, if you traded in Euros, your collateral was usually in Euros. Now, you might trade in Euros but hold your collateral in Dollars. The authors' model treats the collateral currency as just another variable in the equation, like a spice in a recipe. You can swap the spice, and the math still works.

3. The "Cross-Currency Basis": The Hidden Fee

One of the biggest discoveries in recent years is the Cross-Currency Basis.

• The Analogy: Imagine you are buying a car in Germany but paying in Dollars. The dealer says, "The price is €50,000." But because the exchange rate is tricky and banks are nervous, they secretly add a "nervousness fee" to the price.
• In the paper, this "nervousness fee" is called the basis spread. The authors developed a way to model this fee dynamically. They realized this fee isn't static; it changes every second based on how risky the market feels. Their model tracks this fee like a weather vane, allowing traders to predict how much extra it will cost to swap currencies.

4. The "Benchmark Transition": Changing the Engine While Driving

The financial world is currently undergoing a massive transition. The old engine (LIBOR) is being removed, and new engines (SOFR, ESTR) are being installed.

• The Challenge: Some contracts were signed 10 years ago using the old engine. Some new contracts use the new engine. Some contracts are hybrids.
• The Paper's Magic: The authors' framework is like a hybrid car engine that can run on both old fuel and new fuel simultaneously. It allows a bank to value a portfolio that contains:
  • Old contracts paying LIBOR (which might now be replaced by SOFR).
  • New contracts paying SOFR.
  • Contracts paying EURIBOR.
  • All while the collateral is in a different currency.

5. Why Does This Matter? (The "xVA" Problem)

Banks need to calculate something called CVA (Credit Valuation Adjustment). This is basically a "safety tax" they have to pay to cover the risk that their trading partner might go bankrupt.

• The Analogy: Imagine you are managing a portfolio of 1,000 different insurance policies. Some are for fire, some for flood, some for theft. Some are in English, some in French. To calculate your total risk, you can't just add them up; you have to see how they interact.
• The authors' framework allows banks to run massive computer simulations (Monte Carlo) to calculate this safety tax accurately, even when the portfolio is a chaotic mix of different currencies, different interest rate types, and different collateral rules.

Summary

Think of the financial market as a giant, multi-lingual orchestra.

• Before: The conductor (the model) could only handle one instrument (one currency) playing one song (one interest rate).
• Now: The orchestra has instruments playing different songs, in different languages, with some musicians wearing safety harnesses (collateral) in different colors.
• This Paper: It provides the new sheet music and the conductor's baton that allows the entire orchestra to play together in harmony, ensuring that the price of every note (every financial contract) is fair, accurate, and safe, no matter how complex the mix becomes.

It is a foundational tool that ensures the global financial system doesn't crash when the rules change, by giving mathematicians and traders a single, robust language to describe a very complicated reality.

Technical Summary

Here is a detailed technical summary of the paper "Cross-Currency Heath-Jarrow-Morton Framework in the Multiple-Curve Setting" by Alessandro Gnoatto and Silvia Lavagnini.

1. Problem Statement

The paper addresses the complexities arising in modern interest rate markets due to two major structural shifts:

1. The Benchmark Transition: The shift from unsecured interbank offered rates (IBORs like LIBOR, EURIBOR) to risk-free rates (RFRs like SOFR, ESTR) and the coexistence of forward-looking and backward-looking rates.
2. Cross-Currency Dislocations: The persistent violation of Covered Interest Rate Parity (CIRP) observed since the 2008 financial crisis. This necessitates the use of cross-currency basis spreads to price instruments, as collateralized derivatives (priced via OIS curves) no longer align with unsecured lending rates (IBOR curves).

The Core Challenge: Existing models often struggle to simultaneously handle:

• Multiple currencies with distinct collateral conventions.
• The coexistence of forward-looking rates (e.g., EURIBOR) and backward-looking compounded rates (e.g., SOFR).
• Arbitrary fixing and payment adjustments (e.g., in-arrears, delayed payments).
• The need for a unified framework to value large portfolios of multi-currency derivatives (e.g., for xVA calculations) where different legs rely on different market conventions.

2. Methodology

The authors construct a general Heath-Jarrow-Morton (HJM) framework within a multiple-curve, multi-currency setting driven by a general Itô semimartingale.

A. Market Setup and Pricing Measure

• Multiple Currencies: The model considers $L$ currency areas.
• Collateralization: It incorporates a "cash-collateral" convention where collateral is posted in a specific currency ($k_3$), distinct from the currency of the cash flows ($k_2$) and the domestic currency ($k_0$).
• Pricing Measure: The framework operates under a domestic risk-neutral measure $Q^{k_0}$. Unlike traditional models where a single bank account is the numeraire, here, different risky assets are discounted by their own asset-specific cash accounts (repo accounts).
• No Arbitrage: The absence of arbitrage is ensured via the "No Asymptotic Free Lunch with Vanishing Risk" (NAFLVR) condition, enforced through a repo constraint that links risky asset holdings to their specific funding accounts.

B. Modeling Components

The framework models three primary families of term structures:

1. Domestic Collateral Discount Curves: Zero-coupon bonds (ZCBs) collateralized in the same currency ($B^{k_0, k_0}$).
2. Cross-Currency Basis Curves: Modeled as multiplicative spreads between domestic bonds with domestic collateral and domestic bonds with foreign collateral ($Q^{k_0, k_3} = B^{k_0, k_3} / B^{k_0, k_0}$).
3. Foreign Collateral Discount Curves: Derived from the product of the domestic curve and the basis spread.

Key Innovation: Abstract Indices
Instead of modeling specific rates (like LIBOR or SOFR) directly, the authors introduce Abstract Indices ($I^{k_2}$). These indices are defined by:

• Fixing Date ($T_f$): When the rate is observed.
• Payment Date ($T_p$): When the cash flow occurs.
• Period: The accrual period $[T_s, T_e]$.
• Adjustments: Arbitrary fixing ($\delta_f$) and payment ($\delta_p$) lags.
  This abstraction allows the model to encompass:
• Forward-looking IBORs (fixing at start).
• Backward-looking RFRs (compounded over the period).
• In-arrears rates.
• Commodity indices.

C. Dynamics and Drift Conditions

The dynamics of forward rates and spreads are driven by an Itô semimartingale $X$. The authors derive HJM drift conditions to ensure the martingale property of discounted bond prices under the appropriate forward measures.

• Drift Condition for Basis Spreads: A novel drift condition is derived for the cross-currency basis spread $q^{k_0, k_3}$. It depends on the local exponent $\Psi$ of the driving semimartingale and the integrated volatilities of both the domestic collateral curve and the basis spread itself.
• Drift Condition for Index Spreads: The model introduces a multiplicative spread $S$ between the forward index rate and a "simple forward collateral rate" (analogous to an OIS FRA). The drift condition for this spread accounts for the convexity adjustments arising from the interaction between the index volatility, the FX rate, and the collateral curves.

D. Measure Changes

The paper rigorously defines measure changes between:

• Spot measures ($Q^{k_0}$).
• Forward measures ($Q^{T, k_0, k_3}$) associated with specific maturities and collateral currencies.
• Foreign spot measures ($Q^{k}$).
  Crucially, the framework handles the Quanto adjustment required when modeling foreign rates under a domestic measure, explicitly deriving the drift adjustments caused by the correlation between FX rates and interest rates.

3. Key Contributions

1. Unified Framework for Benchmark Transition: The model is the first to simultaneously accommodate forward-looking rates (like EURIBOR) and backward-looking compounded rates (like SOFR) within a single cross-currency HJM framework.
2. General Abstract Indices: By defining forward contracts on "abstract indices" with arbitrary fixing and payment adjustments, the framework covers all current market conventions (standard, in-arrears, delayed) without needing separate models for each.
3. Rigorous Cross-Currency Basis Modeling: The paper provides a general HJM derivation for cross-currency basis spreads in a semimartingale setting, generalizing previous work (e.g., Fujii et al.) and explicitly linking the basis dynamics to the collateral rates of both currencies.
4. Collateral Dislocation Handling: It formally integrates the "multiple-curve" phenomenon (IBOR vs. OIS) with the "cross-currency" phenomenon (different collateral currencies), showing how the basis spread acts as the bridge between them.
5. Application to Complex Swaps: The framework is applied to price Cross-Currency Swaps (CCS), including legacy contracts transitioning to new benchmarks (e.g., USD leg moving from LIBOR to SOFR/AMERIBOR) and resetting (Mark-to-Market) swaps.

4. Results

• Pricing Formulas: The authors derive explicit pricing formulas for:
  • Zero-coupon bonds with arbitrary collateral currencies.
  • Forward contracts on abstract indices.
  • Cross-currency swaps (both constant notional and resetting).
• Drift Conditions: They establish the necessary and sufficient conditions (drift conditions) for the forward rates and spreads to be martingales under the relevant measures. These conditions ensure the model is arbitrage-free.
• Martingale Properties: It is proven that the multiplicative spread between an abstract index forward and the collateral curve forward is a martingale under a specific forward measure, extending the results of Cuchiero et al. (2016) to the multi-currency and backward-looking setting.
• Numerical/Practical Implication: The framework demonstrates that for a portfolio of multi-currency derivatives, the valuation of Credit Valuation Adjustment (CVA) and other xVAs requires a joint simulation of all risk factors (interest rates, FX, basis spreads) to capture non-linearities arising from netting agreements.

5. Significance

This paper provides the theoretical foundation for the next generation of interest rate modeling required by the post-LIBOR era.

• Practical Relevance: It solves the immediate industry problem of valuing portfolios that contain a mix of legacy LIBOR contracts and new SOFR/ESTR contracts across different currencies.
• Flexibility: The "abstract index" approach allows the model to adapt to future market innovations without structural changes.
• Completeness: By treating the market as incomplete (due to the lack of traded interest rate assets) and using a martingale modeling approach with specific funding constraints, it offers a more realistic representation of market mechanics than complete market models.
• Calibration Path: The authors outline a cascade calibration procedure: calibrate single-currency curves first, then use them to calibrate FX and cross-currency basis models, providing a roadmap for implementation in large-scale Monte Carlo simulations used for risk management (xVA).

In summary, Gnoatto and Lavagnini have constructed a robust, mathematically rigorous, and highly flexible framework that unifies the disparate elements of modern cross-currency markets, enabling accurate pricing and risk management of complex derivatives in a post-benchmark reform environment.