- Black dots: Lambda time series
- Red dots: VarCoVaR time series
Lambda time series is shifted for better comparison, VarCoVaR time series without shifting
- a better pronounced behaviour (than the old FRM version)
- the Pearson correlation between both time series: 0.968
- a clear visual correlation with the Lambda structure
- that market heights and lows are better captured
- a clear, concise statistical definition and derivation: conditional variance of CoVaR
- a clear economic interpretation: volatility of the (Co)VaR according to the covariates
- in contrast to the abstract Lambda index the VarCoVaR scale, which is the value domain of the (Co)VaR
- that the Lambda value domain does not start with zero: clear disadvantage in comparison with VarCoVaR. "No risk" should start with zero.
- a better spread between extreme values
- more stability over different companies but still different and individual characterizations
- the incorporation of the whole market volatility and not only Lambdas (based on Betas) from L1-Q-regression
- the financial crisis from 2008 much better pronunciated (peak value)
- also the volatility and uncertainty of the CoVaR of any given company/asset
- many interesting other features waiting for being analysed, investigated and evaluated...
- Black dots: Lambda time series
- Red dots: VarCoVaR time series
- Green dots: Squares of euclidean norm of the Beta vectors (from L1-Q-regression)
