Creating and validating risk models [17]

Go to: Summary | Previous | Next   
Bullet points include: Models usually have a large number of statistical parameters that need to be estimated, so subject to maxim: “Garbage in garbage out” Basel II/III and Solvency II both emphasise importance of parameterisation Both prefer data-driven parameterisations rather than judgementally derived parameterisations Even if in practice regulators have themselves introduced various judgemental overrides E.g. VaR multiplier, stressed VaR, illiquidity premium adjustment, contra-cyclical adjustments Here we illustrate with reference to parameterisation of correlations

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile