Riskdata, the leading provider of risk management solutions to the financial marketplace, has used a "Shock VaR" calculation which aims to overcome the temporary under–estimation of risk that typically happens to VaR calculations during strong market crises. Value–at–Risk (VaR) is the measure of risk for assessing capital adequacy requirements to market operators.
Given the crucial importance of Shock VaR at this time, from today, 9th October, Riskdata will start daily publishing of VAR indicators for the leading European, American and emerging equity indices on its website – www.riskdata.com. The purpose of this disclosure is to provide objective quantitative information on the volatility and potential fluctuations of global markets. In addition the analysis will cover: major Government bond yields, Government bond yield spreads, major commodities, major currencies, Var of a Risk Budgeted "Global Portfolio" of Equity indices, and Implied Volatilities Across asset classes.
The information comprises One Day Shock Var, One Day Long Term Var and their combined relationship. Riskdata will rely on its unique technology known as RiskTicker to calculate shock and Long Term VaR for each leading market index including Dow Jones, DAX, FTSE 100, NASDAQ, S&P 500, DJ Euro Stoxx 50, CAC 40, RTS, SMI, Nikkei 225, S&P ASX, Bovespa, Sensex, and Shanghai Composite.
Ingmar Adlerberg, CEO of Riskdata commented: "Riskdata will be providing this information as long as the market turbulence persists, to give market players an insight into changing risks of these markets. We believe that transparency and analytical rigour is the best way to counter the ongoing market crisis."
Riskdata´s analysis shows that world´s leading exchanges have quite varied levels of ShockVar at present. These range, at the 99.9% confidence level, from –20.3% for the Russian market, to –11.9% for the Nasdaq, –10.2% for the FTSE 100, –9.2% for the Dow, –8.6% for the DAX and just 5.1% for the Australian ASX index.
Traditional "long–term" VaR calculations are meant to be relatively stable through time. Over the long run (several years), back–tests of Riskdata´s long–term VaR, for example, show that the frequency of "exceptions", i.e. when the actual index returns drop down more than the long–term VaR, is approximately equal to the specified frequency (e.g. 1% of the time for the 99% VaR). The stability of this type of VaR calculation makes it very useful for capital adequacy requirements. However, a crucial weakness is that this type of VaR approach can overestimate risk in calm periods and underestimate it in highly volatile environments.
Riskdata´s "Shock VaR" is a much more reactive estimate that accounts for market changes of regime. It aims to anticipate increases/decreases in VaR by using the micro–signals that sometimes are present during market shocks. Back–tests are performed on much shorter time periods, and the frequency of exceptions during the turmoil is thus more in–line with the specified boundaries. It is less stable than the long–term VaR, as it may increase by a factor of 2 within days of a shock, or anticipated shock, as well as drop down to its initial value if the market volatility is back to long term levels. The Shock VaR is a Riskdata proprietary algorithm. It is a form of heteroskedastic model with a short term volatility estimate (1 month), specially adapted for better fat tail estimation and anticipation.
For example, when looking at Riskdata´s one month ahead ShockVaR estimates for the FTSE 100 index, an increased level of the ShockVaR was visible over the past year (roughly doubling), and a further substantial increase in ShockVaR occurred around September 9th, just a few days before major market turmoil. However, although the ShockVar attempts to exploit all the pre–signals it finds, it cannot always properly anticipate shocks - particularly as there are situations that are real surprises. It is thus a necessary, but not sufficient, risk estimate.
