Time-Varying Correlations for Credit Risk: Modelling, Estimating and Stress Testing

Oleg Burd

The burst of the US housing bubble in summer 2007 triggered distress of financial systems around the world. Confronted with the comovements of prices across different asset classes, which were far beyond the scope predicted by risk models, the vast number of financial institutions were unable to weather the financial storm and were faced with unprecedented losses and write-downs. One of the most striking analyses of these events is by Alan Greenspan (2008)

In line with the time-honoured observation that diversification lowers risk, computers crunched reams of historical data in quest of negative correlations between prices of tradeable assets; correlations that could help insulate investment portfolios from the broad swings in an economy. When such asset prices, rather than offsetting each other’s movements, fell in unison on and following August 9 last year [2007], huge losses across virtually all risk asset classes ensued.

And further referring to the source of the problem

The most credible explanation of why risk management based on state-of-the-art statistical models can perform so poorly is that the underlying data used to estimate a model’s structure are drawn generally