QuarcHome
Model catalogueBack to the model index

Jarrow–Lando–Turnbull rating migration

Uses a rating-state Markov chain to link migrations and default to credit spreads.

How it works

Ratings are states of a Markov chain, with default typically absorbing. A term structure of credit spreads is linked to risk-neutral rating transitions, recovery and the default-free curve. Empirical historical transitions can inform the structure of the model. Pricing probabilities need calibration to observable bond or credit derivative prices.

Worked example

Use a two-state annual chain with 98% chance of staying performing and 2% chance of entering default, which is absorbing. With time-homogeneous transitions, two-year survival is 0.98² = 96.04%, so cumulative default probability is 3.96%. A multi-rating chain sums over all surviving paths through the intermediate rating state.

Validation and limitations

Check row sums, default absorption and the term structure implied by the calibrated transitions. Test whether one transition law prices more than the instruments used in calibration. A Markov model assumes the current state contains enough information. Rating history and sector can improve migration forecasts.

Research sources