Podcast: Richard Martin on improving credit migration models

My guest for this edition of Quantcast is Richard Martin, independent consultant and visiting professor at Imperial College London – and currently the most published author in Risk.net’s Cutting Edge pages.

Martin discusses his most recent paper, Credit migration: generating generators, in which he proposes improvements to classic Markovian models.

Credit migration describes the evolution of upgrades and downgrades of bond ratings. It is commonly modelled using a matrix of parameters, known as generator matrix. But this approach has a known problem: a large number of parameters can lead to an unstable calibration, potentially making it difficult to identify the optimal solution.

Martin’s proposal is to use a tridiagonal matrix that has significantly fewer parameters to specify. He also plugs in a stochastic time change, which allows multiple upgrades and downgrades to happen over the same time period. The risk-neutral calibration is also adapted to include not only information on the prices, but spread volatility.

The result is a more sensible and realistic setting to capture the probabilities of upgrades and downgrades.

Index

00:00 Intro

01:00 What is credit migration?

04:45 The standard Markovian approach to credit migration

06:20 The alternative matrix generator

09:24 A time-varying matrix

13:05 Risk-neutral calibration

15:55 How has Covid-19 affected credit migration?

To hear the full interview, listen in the player above, or download. Future podcasts in our Quantcast series will be uploaded to Risk.net. You can also visit the main page here to access all tracks, or go to the iTunes store or Google Podcasts to listen and subscribe.