Making Risk Additive: The Alpha and Beta of Risk Attribution and Risk Delta
May 16, 2012 6:02 PM
Institutional investors face the challenge of understanding changes in risk. Did a recent increase in risk come from a shift into more aggressive positions, a spike in market volatility, or a loss of diversification? Which of the institutional investor's positions or managers drove the change? Which parts of the market became more risky?
Correlation Risk Attribution (CRA) and Risk Delta are two innovations that provide insight into sources of risk and into sources of changes in risk forecast.
Correlation Risk Attribution identifies three key drivers of portfolio risk: exposures to risk factors, the volatility of risk factors, and the correlation between risk factors. The additive nature of CRA means that contributions to total and active risk can be summed up across these three components, as well as across dimensions such as asset classes, managers, or markets within a portfolio.
The Risk Delta methodology uses CRA as a foundation, and decomposes changes in risk forecasts across periods into changes due to:
- exposures to risk factors
- changes in risk factor volatilities, and
- changes in correlations between risk factors
As with CRA, Risk Delta decompositions are additive, and contribution can be aggregated to zoom in and out, anywhere from the holding-level to the asset class level, depending on required granularity.
One application of Risk Delta for plan sponsors is to aggregate changes at the manager level. This means that changes in monthly risk forecasts are explained as coming from the change of each manager’s weight in the plan, the change in each manager’s standalone volatility, and the change in each manager’s correlation to all other managers in the plan.
- Clients Only - Risk Contribution is Exposure times Volatility times Correlation
- Clients Only - Delta-Sigma Attribution: Understanding Differences in Risk