Anait A. Karamian Lasso, Ridge and ElasticNet in Models of the Formation of the Yield Spread of Russian Corporate Bonds

DOI: 10.38050/2078-3809-2025-17-4-99-122

Abstract

The article is devoted to the analysis of models for the formation of the yield spread of Russian corporate bonds in the non-financial sector, which is important for understanding the factors determining the difference in the profitability of financial instruments. The purpose of the work is to build models of spread formation using Lasso, Ridge and ElasticNet methods and to draw a conclusion based on the results obtained about the applicability of these methods, which have not previously been used in the Russian literature to solve such problems. The research methodology is based on a comparative analysis of two samples: data for the period from 01.01.2014 to 23.02.2022 and for 24.02.2022 to 31.12.2023, which makes it possible to assess the stability of models in different market conditions. The results showed that, despite the partial ability of the methods to select economically significant factors, their application for the Russian corporate bond market in the non-financial sector is limited. This is due to the presence of illogical dependencies and difficulties in interpreting the coefficients. In further research, it is advisable to consider alternative approaches, since the development of an accurate model will allow market participants to manage risks more effectively and optimize returns in accordance with individual risk appetite.

Keywords: Lasso, Ridge, ElasticNet, spread yield, corporate bonds.

JEL: G12.

For citation: Karamian, A.A. (2025) Lasso, Ridge and ElasticNet in Models of the Formation of the Yield Spread of Russian Corporate Bonds. Scientific Research of Faculty of Economics. Electronic Journal, vol. 17, no. 4, pp. 99-122. DOI: 10.38050/2078-3809-2025-17-4-99-122.

Download