KomFi. Academy

hard · FRM Part 2 Credit Risk

A CDO tranche desk uses the Gaussian copula with a single correlation parameter to map a CDS index to tranche prices. Empirically the desk observes a 'correlation smile': calibrating each tranche separately to its market price yields a higher implied correlation for the equity (0–3%) and senior tranches than for the mezzanine tranches. A risk manager must explain why a single base correlation surface (rather than compound/implied correlation) is preferred for interpolating bespoke tranches.

What is the strongest technically correct justification?

  1. Base correlation is preferred because compound correlation can be non-monotonic and non-unique for mezzanine tranches, whereas base correlation parametrizes a sequence of equity tranches that is monotone in detachment and admits unique calibration and arbitrage-free interpolation
  2. Base correlation is preferred because it directly fits the physical-measure default correlation observed in long-run historical rating-transition and default-timing data, entirely removing the practical need to calibrate any risk-neutral copula model to market tranche quotes.
  3. Compound correlation is rejected because it always systematically overstates senior-tranche implied correlation relative to observed market quotes, which historically mispriced the full index tranche stack relative to the arithmetic sum of its constituent tranche prices.
  4. Base correlation is preferred because it forces the implied copula correlation parameter to remain exactly constant across the entire capital structure, thereby eliminating the correlation smile altogether and restoring a clean, fully tractable single-factor Gaussian copula pricing model.

Sign up free to see the explanation and track your rank →

More FRM Part 2 Credit Risk practice

All Credit Risk questions · FRM Part 2 study guide · FRM Part 2 glossary · FRM Part 2 flashcards