Sadefo

VaR for Quadratic Portfolio’s with Generalized Laplace Distributed Returns: 

VaR for Quadratic Portfolio’s with Generalized Laplace Distributed Returns SADEFO KAMDEM Jules Laboratoire de Mathématique (UMR 6056 Reims), France.

Slide2: 

Slide3: 

Slide4: 

Slide5: 

Proposal Derive rigorous analytic upper/lower bounds for Quadratic Value-at-Risk of multi-asset portfolios. Sharp bounded for distribution function of the quadratic form of GLD random vectors. Critique Everything in finance computable by (straightforward or clever) MC, so why bother ? Proposal Critique

Defense:: 

Defense: Computational speed Theoretical insight: parameter dependence (portfolio loads, stress testing) effects of parameter uncertainty on statistical parameters Checks on MC computations Mathematically interesting problem

Usual P&L approximation: 

Usual Pandamp;L approximation

Slide8: 

Slide9: 

Sharp Rate of Convergence

Slide10: 

Slide11: 

Slide12: 

Slide13: 

Slide14: 

Slide15: 

Slide16: 

Slide17: 

Slide18: 

Slide19: 

Slide20: 

Slide21: 

Slide22: 

Slide23: 

Slide24: 

Slide25: 

Slide26: 

Slide27: 

Slide28: 

Slide29: 

Slide30: 

Slide31: 

Slide32: 

Slide33: 

Slide34: 

Slide35: