Modelling market implied ratings using LASSO variable selection techniques

Making accurate predictions of corporate credit ratings is a crucial issue to both investors and rating agencies. In this paper, we investigate the determinants of market implied credit ratings in relation to financial factors, market-driven indicators and macroeconomic predictors. Applying a variable selection technique, the least absolute shrinkage and selection operator (LASSO), we document substantial predictive ability. In addition, when we compare our LASSO-selected models with the benchmark ordered probit model, we find that the former models have superior predictive power and outperform the latter model in all out-of-sample predictions.