Paper: arXiv 2511.17892
Authors: Xiang Gao, Cody Hyndman
Abstract
We develop an arbitrage-free deep learning framework for yield curve and bond price forecasting based on the Heath-Jarrow-Morton (HJM) term-structure model and a dynamic Nelson-Siegel parameterization of forward rates. Our approach embeds a no-arbitrage drift restriction into a neural state-space architecture by combining Kalman, extended Kalman, and particle filters with recurrent neural networks (LSTM/CLSTM), and introduces an explicit arbitrage error regularization (AER) term during training. The model is applied to U.S. Treasury and corporate bond data, and its performance is evaluated for both yield-space and price-space predictions at 1-day and 5-day horizons. Empirically, arbitrage regularization leads to its strongest improvements at short maturities, particularly in 5-day-ahead forecasts, increasing market-consistency as measured by bid-ask hit rates and reducing dollar-denominated prediction errors.
Complexity vs Empirical Score
- Math Complexity: 9.2/10
- Empirical Rigor: 7.5/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced mathematics, including the Heath-Jarrow-Morton framework, stochastic differential equations, Kalman/particle filters, and neural network integration with explicit arbitrage constraints, warranting a high complexity score. It also demonstrates strong empirical rigor with backtesting on U.S. Treasury and corporate bond data, measuring performance across horizons and maturities, and reporting metrics like bid-ask hit rates and dollar-denominated errors.
Research Flowchart
flowchart TD A["Research Goal:<br>Arbitrage-Free Bond &<br>Yield Curve Forecasting"] --> B["Data & Inputs:<br>US Treasury & Corporate Bonds"] B --> C["Methodology:<br>Neural State-Space Model<br>with HJM Constraints"] C --> D["Computational Process:<br>LSTM/CLSTM + Kalman/Particle Filters"] D --> E["Training:<br>Arbitrage Error<br>Regularization AER"] E --> F["Key Findings:<br>Increased Accuracy<br>Bid-Ask Alignment<br>Short Maturity Improvement"]
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 an arbitrage‑free deep learning framework for yield‑curve and bond‑price forecasting that embeds HJM no‑arbitrage drift restrictions into a neural state‑space architecture with Kalman, extended Kalman and particle filters and an explicit arbitrage error regularization term.
Source sentence
We develop an arbitrage-free deep learning framework for yield curve and bond price forecasting based on the Heath-Jarrow-Morton (HJM) term-structure model and a dynamic Nelson-Siegel parameterization of forward rates.
Assumes 5
- The forward‑rate model obeys a no‑arbitrage drift restriction derived from the HJM framework. model
Source sentence
Our approach embeds a no-arbitrage drift restriction into a neural state-space architecture by combining Kalman, extended Kalman, and particle filters with recurrent neural networks (LSTM/CLSTM).
- The state variable Xt follows a mean‑reverting stochastic differential equation. model
Source sentence
We define Xt as extended Vašíˇcek process dXt = κt (θt −Xt) dt + σtdWt
- Only fixed‑coupon, non‑callable, non‑convertible bonds with remaining maturity between 3 months and 30 years are used. data
Source sentence
We retain fixed-coupon, non-callable, non-convertible bonds with remaining time-to-maturity between 3 months and 30 years, yield-to-maturity under 700 bps, and an absolute YTM–coupon difference under 500 bps.
- The decay parameter λ is selected by grid‑search minimizing out‑of‑sample RMSE and then kept fixed across the sample. statistical
Source sentence
We select the decay parameter λ by a grid search minimizing out-of-sample RMSE over a rolling validation window and keep it fixed across the sample for stability.
Show 1 more
- Observation noise for price forecasting follows a multivariate generalized Gaussian distribution (MGGD). model
Source sentence
Because price errors exhibit heavy tails and occasional outliers, we model observation noise with a multivariate generalized Gaussian distribution (MGGD), which nests Gaussian…
Data
- FINRA‑TRACE and proprietary Treasury data feed market-prices commercial
Source sentence
Raw data were obtained from FINRA-TRACE supplemented by proprietary Treasury data feed from a commercial vendor. Our data comprise daily clean prices, yields to maturity, coupon rates/frequencies, instrument type, convertibility/callability flags, and issue/maturity dates for U.S. Treasuries and corporate bonds from 20
Costs
Not applicable
Finds 5
- Arbitrage regularization yields the largest predictive gains at short maturities and the 5‑day horizon, improving bid‑ask hit rates while keeping MAE/RMSE competitive. performance
Source sentence
Empirically, AER delivers the largest gains at short maturities and at the 5-day horizon, improving bid–ask hit rates while keeping MAE/RMSE competitive (Section 5).
- The arbitrage‑regularized model shows strongest improvements for short‑maturity forecasts, especially 5‑day‑ahead, reducing dollar‑denominated prediction errors. predictive-accuracy
Source sentence
Arbitrage regularization leads to its strongest improvements at short maturities, particularly in 5-day-ahead forecasts, increasing market‑consistency as measured by bid‑ask hit rates and reducing dollar‑denominated prediction errors.
- Particle filtering with MGGD observation noise achieves high effective sample size and stable forecasting performance despite heavy‑tailed price errors. performance
Source sentence
We conclude that the particle filter with multivariate generalized Gaussian distribution is very efficient and stable for bond prices forecasting in both training and testing data.
- For corporate bond data, the Kalman filter outperforms EKF and PF models, but overall errors are larger than for Treasury data due to limited sample and omitted credit risk factors. predictive-accuracy
Source sentence
In the case of corporates the yield data contains seems to contain more information than the corporate bond prices and we find that the KF model significantly outperforms the EKF and PF models.
- Different filters trade off accuracy and computation: KF in yield‑space is fastest, EKF offers a good accuracy‑runtime balance, PF is most demanding but handles non‑Gaussian residuals. other
Source sentence
For practical use, different filters trade accuracy for computation. A KF in yield-space is fastest but requires extracting yields from prices. An EKF in price-space offers a good accuracy–runtime balance. A PF in price-space with MGGD observation noise and systematic resampling is the most computationally demanding but
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 33 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 (46)
- Deep Learning for Bond Yield Forecasting: The LSTM‐LagLasso (2025) — link
- Real-world models for multiple term structures: a unifying HJM semimartingale framework — → 2411.01983
- Yield curve extrapolation with machine learning (2024) — link
- Forecasting the yield curve: the role of additional and time‐varying decay parameters, conditional heteroscedasticity, and macro‐economic factors (2024) — link
- Machine Learning and the Yield Curve: Tree-Based Macroeconomic Regime Switching (2024) — link
- VolGAN: A Generative Model for Arbitrage-Free Implied Volatility Surfaces (2024) — link
- A Smooth Shadow-Rate Dynamic Nelson-Siegel Model for Yields at the Zero Lower Bound (2024) — link
- Multiple yield curve modeling and forecasting using deep learning (2024) — link
and 38 more
- The geometry of multi-curve interest rate models — → 2401.11619
- Transformers Can Solve Non-Linear and Non-Markovian Filtering Problems in Continuous Time For Conditionally Gaussian Signals (2023) — link
- Bayesian KalmanNet: Quantifying Uncertainty in Deep Learning Augmented Kalman Filter (2023) — link
- Term structure modeling with overnight rates beyond stochastic continuity (2023) — link
- A stochastic control perspective on term structure models with roll-over risk (2023) — link
- Particle Filter with LMMSE Importance Sampling (2022) — link
- Learning Forecast-Efficient Yield Curve Factor Decompositions with Neural Networks (2022) — link
- KalmanNet: Neural Network Aided Kalman Filtering for Partially Known Dynamics (2021) — link
- Arbitrage-Free Neural-SDE Market Models (2021) — link
- Model-Based Deep Learning (2020) — link
- Deep Arbitrage-Free Learning in a Generalized HJM Framework via Arbitrage-Regularization (2017) — link
- Machine Learning for Better Models for Predicting Bond Prices (2017) — link
- Parameter Estimation For Multivariate Generalized Gaussian Distributions (2013) — link
- Price Dispersion in OTC Markets: A New Measure of Liquidity (2011) — link
- A stochastic processes toolkit for risk management: Geometric Brownian motion, jumps, GARCH and variance gamma models (2007) — link
- Global Yield Curve Dynamics and Interactions : A Dynamic (2007) — link
- Credit risk modeling with affine processes (2007) — link
- On the Existence of Finite‐Dimensional Realizations for Nonlinear Forward Rate Models (2001) — link
- Long Short-Term Memory (1997) — link
- Testing Term Structure Estimation Methods (1996)
- Estimating and Interpreting Forward Interest Rates: Sweden 1992-1994 (1994) — link
- Common Factors Affecting Bond Returns (1991) — link
- Autoencoder-Based Risk-Neutral Model for Interest Rates (2024) — link
- Yield Curve Forecasting using Deep Learning Nelson-Siegel Model (2023) — link
- Nonlinear Kalman filtering in affine term structure models (2014)
- LIQUIDITY AND CREDIT RISK PREMIA IN GOVERNMENT BOND YIELDS (2012)
- The affine arbitrage-free class of Nelson–Siegel term structure models (2011)
- NBER WORKING PAPER SERIES THE MACROECONOMY AND THE YIELD CURVE: A DYNAMIC LATENT FACTOR APPROACH (2004)
- The Rodney L . White Center for Financial Research Forecasting the Term Structure of Government Bond Yields (2003)
- Filtering in finance (2003)
- Monte Carlo strategies in scientific computing (2001)
- A No-Arbitrage Vector Autoregression of Term Structure Dynamics with Macroeconomic and Latent Variables (1999)
- Monte Carlo filter and smoother for non-Gaussian nonlinear state space models (1996)
- Bond pricing and the term structure of interest rates: A new methodology for contingent claims valuation (1992)
- Parsimonious modeling of yield curves (1987) — link
- Deep learning of data-driven Heath–Jarrow–Morton models
- Stochastic control, numerical methods, and machine learning in finance and insurance
- Machine learning for interest rates: Using auto-encoders for the risk-neutral modeling of yield curves
2 of 46 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.