AN ANALYSIS OF SOVEREIGN CREDIT RATINGS USING RANDOM FOREST

Authors

  • Oliver Takawira
  • John W. Muteba Mwamba

Keywords:

Sovereign Credit rating, Random Forest, Decision Tree, Machine Learning and Macroeconomic variables

Abstract

The study’s objective was to build a novel approach for analysing sovereign credit ratings. Quarterly data from 1999 to 2020 were analysed using a random forest model generated from decision trees. Macroeconomic indicators and sovereign credit ratings (SCR) ’’were used from the three major credit rating agencies, Fitch, Moody’s, and Standard & Poor’s. The random forest model is a machine learning methodology that analyses and forecasts data using categorisation algorithms. The random forest classifier and analyser fared admirably well when classifying and analysing sovereign credit ratings. The data imply that the most relevant variables for estimating and ranking credit ratings are household debt to disposable income, exchange rates, and inflation. The data indicate that increases in economic metrics such as Real Effective Exchange Rates, Gross Domestic Product Growth, Household Debt to Disposable

Downloads

Published

2022-04-16