Generating the Term Structure of Interest Rates with Diffusion Modelss

Fukunishi Yosuke, Haorong Qiu, Akihiko Takahashi and Fan Ye
Additional contact information
Fukunishi Yosuke: Graduate School of Economics, The University of Tokyo
Haorong Qiu: Graduate School of Economics, The University of Tokyo
Akihiko Takahashi: School of Interdisciplinary Mathematical Sciences and Graduate School of Advanced Mathematical Sciences, Meiji University
Fan Ye: Graduate School of Economics, The University of Tokyo

No CIRJE-F-1262, CIRJE F-Series from CIRJE, Faculty of Economics, University of Tokyo

Abstract: This study introduces a novel generative modeling framework for simulating the term structure of interest rates. In recent years, generative models have achieved significant progress in image generation and are increasingly being applied to finance. To the best of our knowledge, this is the first study to apply a generative model—specifically, a diffusion model—to the term structure of interest rates. We used a diffusion model that incorporates conditional generation mechanisms and v-parameterization. The training dataset consists of spot yield curves constructed from daily overnight index swap (OIS) rates using cubic Hermite splines. As base conditioning variables, we use short-term interest rate and macroeconomic indicators such as changes in consumer price indexes (CPIs). Our empirical analysis covering the period from 2015 to 2025 demonstrates that our model successfully reproduces the level and shape of yield curves corresponding to historical macroeconomic conditions and short-term interest rate environments. Additionally, when incorporating further conditioning variables related to quantitative easing policies, monetary base, current account balances, and nominal gross domestic product (GDP), we find that the inclusion of quantitative easing indicators notably enhances the model’s performance relative to the base conditioning case. Our model with quantitative easing indicators shows that it can generate synthetic yield curves that effectively represent actual ones even in out-of-sample generation over a horizon of approximately six months. In consideration of practical applications, we also examine generation outcomes from difference-based learning and an indirect generation method based on the Nelson-Siegel-Svensson (NSS) model. Both methods show comparable performance in reproducibility in terms of the level and shape of the yield curves. This framework offers several practical applications, including scenario analysis to support bond investment strategy formulation and risk management under stress scenarios

Pages: 18 pages
Date: 2025-12
New Economics Papers: this item is included in nep-mon
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.cirje.e.u-tokyo.ac.jp/research/dp/2025/2025cf1262.pdf (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:tky:fseres:2025cf1262

Access Statistics for this paper

More papers in CIRJE F-Series from CIRJE, Faculty of Economics, University of Tokyo Contact information at EDIRC.
Bibliographic data for series maintained by CIRJE administrative office ().