Predicting the yield spreads on the Russian debt market

Abstract

The article provides an empirical study of factors which make a significant impact on the yield spread of corporate bonds on Russia’s market. Unlike most studies based only on IPO data, the current study covers primary and secondary markets data of 2010–2019. The aim of the study is to test the method allowing to build a model capable of predicting the yield spread with maximum approximation to actual value. The author develops a regression model applying a cross-validation method, a procedure for empirical assessment of the model generalizing ability. An important prerequisite is the inclusion of a limited number of regressors with high explanatory power in the model, stability over time, and explicit economic interpretation. The paper confirms the key hypothesis on high degree of issuer's rating on the yield spread which is a new step in the study of the Russian market. The findings prove that the yield spread is determined mainly by the level of risk corresponding to the degree of the issuer's reliability. The issuer's industry affiliation, the size of the company, the MSCI stock index has also a significant impact on the spread. Among the advantages of the proposed model is its relative simplicity, explicit economic interpretation and stable response to various data which determine the practical significance of the work. The approaches to building and testing the model applied in this work can be useful in further studies aimed at developing predictive models of this class.