Paper: arXiv 2408.02273

Authors: Preetha Saha, Jingrao Lyu, Dhruv Desai, Rishab Chauhan, Jerinsh Jeyapaulraj, Philip Sommer, Dhagash Mehta

Abstract

The trading ecosystem of the Municipal (muni) bond is complex and unique. With nearly 2% of securities from over a million securities outstanding trading daily, determining the value or relative value of a bond among its peers is challenging. Traditionally, relative value calculation has been done using rule-based or heuristics-driven approaches, which may introduce human biases and often fail to account for complex relationships between the bond characteristics. We propose a data-driven model to develop a supervised similarity framework for the muni bond market based on CatBoost algorithm. This algorithm learns from a large-scale dataset to identify bonds that are similar to each other based on their risk profiles. This allows us to evaluate the price of a muni bond relative to a cohort of bonds with a similar risk profile. We propose and deploy a back-testing methodology to compare various benchmarks and the proposed methods and show that the similarity-based method outperforms both rule-based and heuristic-based methods.

Complexity vs Empirical Score

  • Math Complexity: 7.0/10
  • Empirical Rigor: 8.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced machine learning mathematics, including supervised similarity learning with CatBoost, multi-output regression, and custom proximity metrics, while the methodology includes a deployed back-testing framework and specific performance comparisons against benchmarks, demonstrating high empirical rigor.

Research Flowchart

  flowchart TD
  A["Research Goal: Relative Valuation<br>of Municipal Bonds"] --> B["Data & Inputs<br>Large-scale Muni Bond Dataset"]
  
  B --> C{"Methodology: CatBoost Algorithm"}
  
  C --> D["Computational Process:<br>Learn Bond Similarities & Risk Profiles"]
  
  D --> E["Framework:<br>Supervised Similarity-based Relative Value"]
  
  E --> F["Evaluation:<br>Back-testing vs. Rule/Heuristic Methods"]
  
  F --> G{"Key Findings/Outcomes"}
  
  G --> H["Superior Performance:<br>CatBoost outperforms traditional benchmarks"]
  G --> I["Deployment:<br>Validated model for relative value assessment"]

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 data‑driven supervised similarity framework for municipal bonds using a CatBoost model to evaluate relative value.

Source sentence
We propose a data-driven model to develop a supervised similarity framework for the muni bond market based on CatBoost algorithm.

Assumes 2

  • Bonds ranked higher by the similarity‑based ranking will self‑correct and deliver higher returns in an efficient market. market
    Source sentence
    we expect these bonds to self-correct in an efficient market.
  • Cohorts of k‑nearest neighbors based on the learned CatBoost similarity are more representative of a bond’s risk profile than simple rule‑based cohorts. model
    Source sentence
    we argue that such cohorts are more representative of the target bonds in terms of similar risk profiles and

Data

No dataset named in the text.

Costs

Not modeled A trading or execution result is reported without costs.

Finds 3

  • CatBoost regression significantly outperforms elastic net, decision tree, and random forest baselines on both OAS and yield targets. performance
    Source sentence
    The results show that overall CatBoost significantly outperforms all other models.
  • The similarity‑aware relative‑value method outperforms heuristic‑based approaches across five years of market scenarios and sometimes beats simple rule‑based methods. performance
    Source sentence
    our proposed solution of using similarity-aware relative value outperforms the heuristic-based approach in different market scenarios over five years.
  • CatBoost performance remains consistent and stable across multiple snapshot periods. performance
    Source sentence
    The results show that the model is overall consistent and stable over different scenarios.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 10 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 1 paper in the archive

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 (45)

  1. Can an unsupervised clustering algorithm reproduce a categorization system? — → 2408.10340
  2. Case-based Explainability for Random Forest: Prototypes, Critics, Counter-factuals and Semi-factuals — → 2408.06679
  3. Open Set Recognition for Random Forest (2024) — link
  4. Company Similarity Using Large Language Models — → 2308.08031
  5. Quantifying Outlierness of Funds from their Categories using Supervised Similarity — → 2308.06882
  6. Do Municipal Bond Dealers Give their Customers “Fair and Reasonable” Pricing? (2023) — link
  7. How Trading Analytics and Data Science Can Improve Investment Outcomes (2022) — link
  8. Supervised similarity learning for corporate bonds using Random Forest proximities (2022) — link
and 37 more
  1. Geometry- and Accuracy-Preserving Random Forest Proximities (2022) — link
  2. Systematic Pricing and Trading of Municipal Bonds (2021) — link
  3. On Robustness of Mutual Funds Categorization and Distance Metric Learning (2021) — link
  4. Fast TreeSHAP: Accelerating SHAP Value Computation for Trees (2021) — link
  5. Fund2Vec: mutual funds similarity using graph learning (2021) — link
  6. Interpretable Machine Learning (2019) — link
  7. Municipal Bond Markets (2019) — link
  8. Optuna: A Next-generation Hyperparameter Optimization Framework (2019) — link
  9. A Comprehensive Analysis of Traditional Clustering Algorithms On Corporate Bond Data (2018) — link
  10. CatBoost: unbiased boosting with categorical features (2017)
  11. A Unified Approach to Interpreting Model Predictions (2017) — link
  12. Fundamental Analysis Works (2015) — link
  13. The Municipal Securities Rulemaking Board (2015) — link
  14. Risk and Information in the Municipal Bond Market (2014)
  15. The Muni Bond Spread: Credit, Liquidity, and Tax (2014) — link
  16. Gradient boosting machines, a tutorial (2013) — link
  17. Trading the Fixed Income, Inflation and Credit Markets: A Relative Value Guide (2011) — link
  18. Scikit-learn: Machine Learning in Python (2011) — link
  19. The Fundamentals of Municipal Bonds (2008)
  20. Valuation waves and merger activity: The empirical evidence ☆ (2005) — link
  21. The Elements of Statistical Learning: Data Mining, Inference, and Prediction (2004) — link
  22. Who Is My Peer? A Valuation-Based Approach to the Selection of Comparable Firms (2002) — link
  23. Equity Valuation Using Multiples (2002) — link
  24. Investment Valuation: Tools and Techniques for Determining the Value of Any Asset (1995)
  25. The Handbook of Municipal Bonds (1994)
  26. Pushing bonds over the edge: Investor demand and municipal bond liquidity (2023)
  27. Pushing Bonds Over the Edge: Monetary Policy and Municipal Bond Liquidity (2023) — link
  28. markets. Annual Review of Financial Economics , (2019)
  29. Municipal Bond Liquidity and Default Risk (2017)
  30. Social norms and corporate peer effects (2012) — link
  31. pandas: a Foundational Python Library for Data Analysis and Statistics (2011)
  32. SylvanGFeldsteinandFrankJFabozzi (2008)
  33. Random Forests (2001) — link
  34. ICAIF ’24, November 14–17, 2024, Brooklyn, NY, USA
  35. Machine Learning-based Relative Valuation of Municipal Bonds
    1. Towards Enhanced Local Explainability of Random Forests: A Proximity-Based Approach
  36. Relativevaluationwithmachinelearning

4 of 45 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.