Predicting Individual Corporate Bond Returns

Abstract

This paper finds positive evidence of return predictability and investment gains for individual corporate bonds from 1976 to 2020. First, we provide a comprehensive study for multiple predictive models and find random forest is the best. Considering bonds issued by private companies, we find their useful return predictors differ from bonds issued by public companies. The predictability of public bond returns is more sensitive to the T-Bill rate, equity market return and volatility, but private bonds deliver higher investment gains. Second, for 20 macro predictors and 20 bond characteristics considered for return prediction, we find the macro predictors contain additional information. Finally, based on return forecast, market-timing and cross-sectional long-short strategies are profitable and economically significant net of the transaction cost.

Keywords: Bond Characteristics, Individual Corporate Bonds, Macro Predictors, Return Predictability, Private Company Bonds.

JEL Classification: C55, C58, G0, G1, G17.