Rating Migrations: The Effect of History and Time

Abstract

We examine the effect of rating history and the passage of time on the rating migration hazard for corporate debt issuers. Controlling for industry effects and the evolution of business and political cycles, the results consistently show that the next change of rating depends more strongly on rating history than it does on the current rating. However, there are significant interactions between the main effects of rating history and the duration of the current rating. The result is substantial decay in the effects of rating history the longer a rating remains unchanged. This decay effect is stronger for downgrades and for ratings in the speculative category.

National Culture and Corporate Rating Migrations

“…First, previous studies suggest that financial institutions exhibit different rating migration dynamics compared with industrial firms (Nickell et al 2000;Lando and Skodeberg 2002). Second, financial institutions and 4 For the evidence of downward momentum in rating migration dynamics, see Altman and Kao (1992); Carty and Fons (1994); ; Bangia et al (2002); Lando and Skodeberg (2002); Güttler and Wahrenburg (2007); Figlewski et al (2012); Dang and Partington (2014) utilities operate in highly regulated environments and are closely monitored by their respective regulatory agencies. This suggests less latitude for corporate managers in these sectors to exhibit flexibility in decision making.…”

“…This study applies a survival analysis framework and develops Cox's dynamic hazard model (Cox 1972) to examine the effect of culture on the probability of corporate rating migrations. Previous studies suggest that upgrade and downgrade follow different dynamics (Figlewski et al 2012;Dang and Partington 2014). Thus, upgrade and downgrade are treated as competing risks, and separate models are estimated for these two migration outcomes.…”

“…The advantage of this method is that it does not require any assumptions about the nature or the structure of the dependence (Allison 1995, p. 242). Second, the estimation of the stratified hazard model, as in as Dang and Partington (2014), takes into account the sequence of rating changes. Each stratum includes ratings that have the same number of prior changes.…”

“…Current rating grade The current rating (start rating) for the rating transition being observed. Carty and Fons (1994); Figlewski et al (2012); Dang and Partington (2014) Investment rating boundary The dummy takes the value of one if the current rating is in the investment grade boundary (BBB−, BBB, BBB+) or zero otherwise Carty and Fons (1994); Johnson (2004); Dang and Partington (2014) Junk rating boundary The dummy takes the value of one if the current rating is in the speculative (junk) grade boundary (BB−, BB, BB+) or zero otherwise Carty and Fons (1994); Johnson (2004); Dang and Partington (2014) Logarithm of age since first rated This variable takes the value of one if a firm had experienced a big downgrade of at least three rating notches as of the start of the current rating, and zero otherwise Carty and Fons (1994); Dang and Partington (2014)…”

“…This variable takes the value of one if a firm had experienced a big upgrade of at least three rating notches as of the beginning of the current rating, and zero otherwise Dang and Partington (2014) Rating volatility This is the average number of migrations per year over a firm's rating history. It is calculated as the number of migrations a firm had experienced as of the beginning of the current rating divided by age since first rated.…”