Factor credit models
The third and last type of quantitative credit models that we want to mention briefly are factor models. In contrast to both the structural and reduced-form models, the factor model does not attempt to model default.
Rather, linear regression is used to assess the relative richness or cheapness of individual credits. The factor model attributes current spreads of a broad universe of issues to certain common factors that should be appropriate to explain differences in credit spreads. Credit quality, for example, may be represented by a rating and a leverage factor. Debt to EBITDA ratios have proven a particularly reliable indicator for future relative spread movements.
Duration and implied equity volatility of the issuer as well as a dummy variable for the sector might be further explanatory variables. The residual from the regression is taken as an indicator for the richness or cheapness of individual securities. Empirical studies show that factor models may be a valuable tool to generate relative value ideas, especially for short time horizons.
Comments are currently closed.