Creating and validating risk models [22]

Go to: Summary | Previous | Next   
Bullet points include: Estimate correlation of factors based either on equity return, spread change or default rate E.g. set country / industry weights to correspond to relative shares of name’s turnover in that country / industry and set other parameters to enforce ‘reasonable’ average correlations etc. Or if name has observable equity value or spreads then regress observed returns or spread changes against relevant country / industry indices Results then quite sensitive to holding period and, more subtly, to choice of universe and / or weights ascribed in indices to individual countries / sectors How should we choose between the various approaches?

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile