Paper: arXiv 2311.12575

Authors: Gijs Mast, Fang Fang, Xiaoyu Shen, Marnix Brands

Abstract

This paper introduces a novel approach for computing netting–set level and counterparty level exposures, such as Potential Future Exposure (PFE) and Expected Exposure (EE), along with associated sensitivities. The method is essentially an extension of the Fourier-cosine series expansion (COS) method, originally proposed for option pricing. This method can accommodate a broad range of models where the joint distribution of involved risk factors is analytically or semi-analytically tractable. This inclusivity encompasses nearly all CCR models commonly employed in practice. A notable advantage of the COS method is its sustained efficiency, particularly when handling large portfolios. A theoretical error analysis is also provided to justify the method’s theoretical stability and accuracy. Various numerical tests are conducted using real-sized portfolios, and the results underscore its potential as a significantly more efficient alternative to the Monte Carlo method for practical usage, particularly applicable to portfolios involving a relatively modest number of risk factors. Furthermore, the observed error convergence rates align closely with the theoretical error analysis.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 7.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper uses advanced Fourier analysis and spectral methods with complex error analysis (high math), and validates the approach with large-scale numerical tests on real-sized portfolios, showing direct performance comparisons and convergence rates (high empirical rigor).

Research Flowchart

  flowchart TD
  A["Research Goal<br>Develop faster exposure calculation<br>method for CCR"] --> B["Methodology<br>Extend COS method to exposures"]
  B --> C["Data/Inputs<br>Real-sized portfolios<br>Risk factor distributions"]
  C --> D["Computational Process<br>Fourier series expansion<br>vs Monte Carlo simulation"]
  D --> E["Key Findings<br>Significant efficiency gain<br>vs Monte Carlo<br>Consistent error convergence"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a COS‑tensor framework with dimension‑reduced cosine expansions as a deterministic alternative to Monte‑Carlo for credit exposure quantification.

Source sentence
This paper initiates a series of studies on a COS-tensor framework, as an efficient alternative to MC for large and liquid portfolios characterized by a modest number of dominant risk factors but a large number of trades.

Assumes 4

  • Model parameters are calibrated under the measure relevant for the application. model
    Source sentence
    we do not prescribe which measure should be used, but instead assume that model parameters are calibrated under the measure relevant for the application.
  • Each risk factor Zi,t is a deterministic function of a corresponding state variable Yi,t. model
    Source sentence
    Each risk factor Zi,t is assumed to be a deterministic function of a corresponding state variable Yi,t, namely Zi,t = ϕi(Yi,t, t),
  • The joint density or joint characteristic function of the state vector Yt is available in closed or semi‑closed form. model
    Source sentence
    Within the COS–tensor framework, we assume that either the joint density or the joint characteristic function of Yt is available in closed or semi-closed form.
  • Risk factors follow one‑dimensional continuous‑time Markov processes with Brownian motions and instantaneous correlations. market
    Source sentence
    we assume that each component Zi,t, 1 ≤i ≤2N +1, evolves as a one-dimensional continuous-time Markov process of the form dZi,t = µi(t, Zi,t) dt + σi(t, Zi,t) dWi,t, where {Wi,t}

Data

  • Netting sets comprising tens of thousands of trades driven by seven risk factors synthetic unknown
    10,000‑derivative portfolio with 7 risk factors
    Source sentence
    Numerical experiments on netting sets comprising tens of thousands of trades driven by seven risk factors demonstrate that the proposed method achieves relative errors below 0.1% while requiring only a small fraction of the runtime of MC simulation.

Costs

Not applicable The paper reports no trading or execution result.

Finds 2

  • The COS‑CPD method achieves relative errors below 0.1% while requiring only a small fraction of the runtime of Monte‑Carlo simulation. performance
    Source sentence
    the proposed method achieves relative errors below 0.1% while requiring only a small fraction of the runtime of MC simulation.
  • Fourier‑domain training of the CPD expansion yields speed and accuracy improvements of more than two orders of magnitude compared with gradient‑based training in the physical domain. performance
    Source sentence
    Relative to gradient-based regression in the physical domain, this Fourier-domain training yields speed and accuracy improvements of more than two orders of magnitude in the training stage.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 20 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (132)

  1. The XVA Challenge: Counterparty Risk, Funding, Collateral, Capital and Initial Margin (2020)
  2. Counterparty risk under correlation between default and interest rates (2007)
  3. A guide to modeling counterparty credit risk (2007)
  4. XVA: Credit, Funding and Capital Valuation Adjustments (2015)
  5. Portfolio compression (2022) — link
  6. Counterparty risk in credit markets (2008)
  7. Interest rate models-theory and practice: with smile, inflation and credit (2006)
  8. Spectral Tensor‑Train Decomposition (2016) — link
and 124 more
  1. Brownian motion, martingales, and stochastic calculus (2016)
  2. Mathematical Modeling and Computation in Finance: With Exercises and Python and Matlab Computer Codes (2019)
  3. Pricing interest-rate-derivative securities (1990)
  4. Martingales and stochastic integrals in the theory of continuous trading (1981)
  5. A short remark on feller’s square root condition (2011)
  6. A novel pricing method for European options based on Fourier-cosine series expansions (2009) — link
  7. Chebyshev and Fourier spectral methods (2001)
  8. Monte Carlo methods in financial engineering (2004)
  9. Coherent measures of risk (1999)
  10. Efficient computation of exposure profiles for counterparty credit risk (2014)
  11. A general gaussian interest rate model consistent with the current term structure (2012) — link
  12. Which measure for PFE? The risk appetite measure A (2015)
  13. Speed-up credit exposure calculations for pricing and risk management (2021) — link
  14. A new approach for American option pricing: The Dynamic Chebyshev method (2019)
  15. Efficient Least Squares Monte-Carlo Technique for PFE/EE Calculations (2021) — link
  16. Efficient Monte Carlo counterparty credit risk pricing and measurement (2014)
  17. Chebyshev Methods for Ultra-efficient Risk Calculations (2018) — link
  18. Function approximation for option pricing and risk management Methods, theory and applications. (2020)
  19. Efficient estimation of sensitivities for counterparty credit risk with the finite difference Monte Carlo method (2016)
  20. Pricing financial instruments: The finite difference method (2000)
  21. Advanced estimation of credit valuation adjustment (2017)
  22. A stochastic mesh method for pricing high-dimensional American options (2004)
  23. Key Trends in the Size and Composition of OTC Derivatives Markets in the Second Half of 2020 (2020)
  24. The standardised approach for measuring counterparty credit risk exposures (2014)
  25. An equilibrium characterization of the term structure (1977)
  26. On the term structure of interest rates (1978)
  27. An intertemporal general equilibrium model of asset prices (1985)
  28. On the estimation of credit exposures using regression-based Monte Carlo simulation (2008)
  29. The stochastic grid bundling method: Efficient pricing of Bermudan options and their Greeks (2015)
  30. Counterparty credit exposures for interest rate derivatives using the stochastic grid bundling method (2016) — link
  31. Efficient computation of exposure profiles on real-world and risk-neutral scenarios for Bermudan swaptions (2016)
  32. A benchmark approach of counterparty credit exposure of Bermudan option under L 'e vy process: The Monte Carlo-COS method (2013)
  33. Brownian motion (2010)
  34. Changes of numeraire, changes of probability measure and option pricing (1995)
  35. Risk quantification and allocation via inversion from the Fourier domain (2022)
  36. The Fourier Cosine Method for Discrete Probability Distributions — → 2410.04487
  37. Pricing early-exercise and discrete barrier options by Fourier-cosine series expansions (2009)
  38. Is Gauss quadrature better than Clenshaw–Curtis? (2008)
  39. An optimal lognormal approximation to lognormal sum distributions (2004) — link
  40. A Systematic Approach to Pricing and Hedging of International Derivatives with Interest Rate Risk (1995)
  41. A multi-currency model with FX volatility skew (2005) — link
  42. On cross-currency models with stochastic volatility and correlated interest rates (2012)
  43. FX options and structured products (2017)
  44. Foreign exchange option pricing: A Practitioner’s guide (2011)
  45. On the Gibbs phenomenon and its resolution (1997) — link
  46. Applied analysis prentice hall (1956)
  47. On the application of spectral filters in a Fourier option pricing technique (2015) — link
  48. A New and Efficient Fourier Method for Risk Quantification and Allocation of Credit Portfolios (2022)
  49. Advanced mathematical methods for scientists and engineers I: Asymptotic methods and perturbation theory (1999)
  50. Monte Carlo methods for security pricing (1997)
  51. On the approximation of the sum of lognormals by a log skew normal distribution (2015) — link
  52. Credit Valuation Adjustment risk: targeted final revisions (2019)
  53. Option pricing and CVA calculations using the Monte Carlo-Tree (MC-Tree) method (2022) — link
  54. Derivatives and systemic risk: Netting, collateral, and closeout (2006)
  55. Deep xVA Solver: A Neural Network–Based Counterparty Credit Risk Management Framework (2023) — link
  56. The stochastic collocation Monte Carlo sampler: highly efficient sampling from ‘expensive’distributions (2019)
  57. Low-rank tensor approximation for Chebyshev interpolation in parametric option pricing (2020)
  58. Deep learning for CVA computations of large portfolios of financial derivatives (2021)
  59. The deep parametric PDE method and applications to option pricing (2022)
  60. Sample recycling method–a new approach to efficient nested Monte Carlo simulations (2022) — link
  61. Fast and efficient nested simulation for large variable annuity portfolios: A surrogate modeling approach (2020) — link
  62. Efficient exposure computation by risk factor decomposition (2018)
  63. Total value adjustment for a stochastic volatility model. A comparison with the Black–Scholes model (2021)
  64. Explicit expressions to counterparty credit exposures for Forward and European Option (2020) — link
  65. Analytical Expressions to Counterparty Credit Risk Exposures for Interest Rate Derivatives (2022) — link
  66. Option Volatility Trading Strategies (2012)
  67. Consultative Document, Credit Valuation Adjustment risk: targeted final revisions (2019) — link
  68. Accelerating CVA and CVA Sensitivities Using Quasi-Monte Carlo Methods (2020)
  69. Importance sampling and stratification for value-at-risk (1999) — link
  70. Family of Spectral Filters for Discontinuous Problems (1991) — link
  71. Precise option pricng by the COS method - How to choose the truncation range (2022) — link
  72. Error estimates for Clenshaw-Curtis quadrature (1971) — link
  73. Neural network expression rates and applications of the deep parametric PDE method in counterparty credit risk (2023) — link
  74. The multidimensional cos method for option pricing — → 2307.12843
  75. Tensor rank is NP-complete (1990) — link
  76. Neural Networks with Exact Correspondence to Dimension-Reduced Cosine Expansions and for Expectation Calculations (2025)
  77. Afternotes on numerical analysis (1996)
  78. Tensor rank and the ill-posedness of the best low-rank approximation problem (2008)
  79. Importance sampling for portfolio credit risk (2005)
  80. Parallel cross interpolation for high-precision calculation of high-dimensional integrals (2019) — link
  81. Quasioptimality of maximum-volume cross interpolation of tensors (2014) — link
  82. Chebyshev interpolation for parametric option pricing (2018)
  83. Valuing American options by simulation: a simple least-squares approach (2001)
  84. Quasi-Monte Carlo methods in finance (1998)
  85. Multilevel Monte Carlo path simulation (2008)
  86. Multi-index Monte Carlo: when sparsity meets sampling (2016)
  87. Functional quantization for numerics with an application to option pricing (2005)
  88. The Wiener–Askey polynomial chaos for stochastic differential equations (2002)
  89. Sparse grids (2004)
  90. Analysis of Individual Differences in Multidimensional Scaling via an N -Way Generalization of the E ckart– Y oung Decomposition (1970)
  91. Option valuation using the fast Fourier transform (1999)
  92. Fractional Fourier transform methods for credit portfolio loss distributions (2023)
  93. Machine learning for exposure computation in counterparty credit risk (2021)
  94. Deep learning for solving high-dimensional partial differential equations (2019) — link
  95. Sparse grid method for highly efficient computation of exposures for xVA (2022)
  96. Efficient parallel Monte-Carlo techniques for pricing American options including counterparty credit risk (2024)
  97. Numerical smoothing with hierarchical adaptive sparse grids and quasi-Monte Carlo methods for efficient option pricing (2023)
  98. Sequential importance sampling and resampling for dynamic portfolio credit risk (2012)
  99. Neural Networks for Portfolio-Level Risk Management: Portfolio Compression, Static Hedging, Counterparty Credit Risk Exposures and Impact on Capital Requirement — → 2402.17941
  100. Chebyshev interpolation for parametric option pricing (2018)
  101. Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations (2017)
  102. Supervised Learning and Canonical Decomposition of Multivariate Functions (2021) — link
  103. A practical randomized CP tensor decomposition (2018)
  104. Breaking the curse of dimensionality, or how to use SVD in many dimensions (2009)
  105. Tensor-train decomposition (2011)
  106. TT-cross approximation for multidimensional arrays (2010) — link
  107. Tensor decompositions for feature extraction and classification of high dimensional datasets (2010)
  108. Adam: A method for stochastic optimization (2014) — link
  109. Improvement of the cross-entropy method in high dimension for failure probability estimation through a one-dimensional projection without gradient estimation (2021)
  110. Tensor Decompositions and Applications (2009) — link
  111. CP Decomposition for Tensors via Alternating Least Squares with QR Decomposition (2021)
  112. A literature survey of low-rank tensor approximation techniques (2013)
  113. Numerical operator calculus in higher dimensions (2002)
  114. Hierarchical Tucker decomposition of tensors (2010)
  115. A new data format for matrices and tensors (2009)
  116. The sample size required in importance sampling (2018)
  117. Concrete mathematics: a foundation for computer science (1994)
  118. Low rank tensor recovery via iterative hard thresholding (2017)
  119. A Probabilistic Theory of Pattern Recognition (1996)
  120. Representing and Recovering Structured High-Dimensional Data: Fast Dimension Reduction, Recovery Guarantees, and Neural Network Representation (2023)
  121. Function spaces with dominating mixed smoothness characterization by differences (2006)
  122. Approximation theory and approximation practice, extended edition (2019)
  123. Notes on rademacher complexity (2020)
  124. Some applications of concentration inequalities to statistics (2000)

3 of 132 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.