Quant BuffetRelax, Not Over Thinking

Market Timing Corporate Bonds with Machine learning – Random Forests

Log in to collect

Academic paper

Predicting Individual Corporate Bond Returns

AuthorsXin He; Guanhao Feng; Junbo L. Wang; Chunchi Wu

Institute
  • HKCity University of Hong Kong
  • University of Science and Technology of China
  • Hunan University
  • ?City University of Hong Kong (CityU)
  • ?University of Science and Technology of China (USTC)
  • MOCity University of Macau
  • City University of Seattle
  • ?City University of Hong Kong (CityU) - Department of Economics and Finance
  • ?Dept. of Economics and Finance, City Univ. of HK
  • University at Buffalo, State University of New York
  • Buffalo State University
  • ?SUNY at Buffalo - School of Management
  • ?The State University of New York (SUNY) at Buffalo - School of Management

Strategy in a nutshell

This strategy predicts corporate bond returns using machine learning models, including random forest, lasso, and partial least squares. A panel approach combines bond-specific characteristics and aggregate economic predictors. The models are evaluated with out-of-sample R² metrics and Fama-MacBeth regressions, identifying the most influential predictors for public and private bonds across different rating and maturity groups. Market-timing is implemented by longing bonds with positive return predictions and shorting those with negative predictions, creating long-short portfolios optimized for risk-adjusted performance.

Economic rationale

The strategy’s effectiveness stems from improved data availability, long-span historical datasets, and the adoption of advanced machine learning techniques. By capturing nonlinear relationships and differentiating public versus private bond behaviors, it delivers robust return forecasts, enabling better-informed investment and risk management decisions across the corporate bond market.

Backtest performance

Annualised return19.7%
Volatility11.45%
Sharpe ratio1.72