Social Sharing
Extended Viewer
In the Market - April 2010
Apr 15, 2010
Early Warning Case Study: European Divergence
Discuss Sovereign Default more in the Risk Client Community
The European sovereign debt crisis can be used as a topical case study to illustrate an integrated top-down and bottom-up approach to early warning.
- Top-down: Future oriented, macro-view, strategic. Identification of longer-term structural risks based on fundamental analysis, and quantification of potential losses via stress scenarios.
- Bottom-up: Present oriented, micro-view, tactical. Short term volatility analysis based on traded market factors, VaR and related measures.
RiskMetrics presented European Divergence as one of two hypothetical stress themes in the September 2008 European Client Event (see A Decade of Risk Management Lessons), highlighting the vulnerability of Portugal, Italy, Greece, and Spain. Chart 1 shows a plot of widening sovereign credit spreads for the riskiest European countries based on 5 year CDS prices
Chart 1: Three Waves Of Increasing European Sovereign Risk
Chart 2 shows 95% confidence VaR backtesting bands for Greece CDS, and highlights three dates when VaR outliers marked significant spread widening, reversing a period of relative calm.
Chart 2: VaR Outliers Provide Initial Early Warning
Chart 3 shows three red flags that indicate unusual VaR clustering (as measured by an excession of the the expected 5% level over a 100 day period). Rising VAR outliers show a trend of escalating of volatility, and reconfirm the initial VaR outlier warning signal. Markets have now tipped into fat-tailed non-normal conditions, which often cascades into more extended drawdowns as seen below. Note especially the extreme outlier clustering associated with most recent selloff, peaking at 25 outliers over 100 trading days on March 19 2010.
Chart 3: Greece CDS Spreads compared to VaR outlier clustering
The EU/IMF Greek rescue plans package announced over the weekend has averted a market breakdown, and Greece CDS spreads tightened by a record 62 bps to 364 on April 12 2010 and volatility and outliers are now likely to mean revert to more moderate levels. But even assuming a complete bailout of Greece (which represents a mere 2.5% of the Eurozone), the cycle of sovereign risk is far from over. This period is reminiscent of the March ’08 Bear rescue, which only provided temporary relief until Lehman’s Sep ’08 bankruptcy precipitated a global meltdown. While regulators are now determined not to let systemically important institutions fail, sovereigns are more difficult to rescue and will require extensive global cooperation. As Harvard economist Ken Rogoff put it in a 15 Sep 2009 interview with the U.K. Telegraph: "By guaranteeing everything governments have transformed a financial crisis into a debt crisis. Who guarantees the governments?" Take advantage of improving market sentiment to build robust portfolios that can weather the next wave of systemic risk.
Summary: How Do We Know It’s Systemic Risk?
Given that we expect to see a 95% confidence daily VaR outlier once a month, the main challenge is to distinguish signal from noise. In other words, when are outliers early warning as opposed to expected random shocks?
- From a top-down perspective, look for structural vulnerabilities (e.g., Greece’s mounting fiscal deficits, Portugal and Spain’s debt fuelled real estate binge, Italy’s anemic economy and persistent deficits).
- From a bottom-up perspective, closely monitor key market factors (e.g., credit spreads) and look for telltale endogenous risk patterns: Initial VaR outliers which often reverse a period of relative calm, reconfirmation of this signal through clustering of outliers, escalating volatility, increased momentum, and contagion (as illustrated in the above charts).
This process is analogous to seismologists first identifying fault lines where pressure builds (structural risk), and then paying close attention to early warning signals (precipitating foreshocks) which can amplify to major phase transitions (major ruptures/earthquakes).
For additional information on this case study, please visit the RiskMetrics Stress Testing Community discussion. For a more comprehensive presentation of the RiskMetrics Early Warning approach, please see "Integrated Risk Management - Early Overview." If you have any questions, please email alan.laubsch@riskmetrics.com.