Nematrian
/

Stress Testing and Back-testing for UCITS [10]

Go to: Summary | Previous | Next   
Bullet points include: Will also want exceptionally adverse outcomes not to clump through time. Might imperil solvency/credibility more than if occurring apparently randomly through time. i.e. we also want independence between any two elements of the hit sequence. Can test for this, e.g. by creating a contingency table of It versus It-h (e.g. It-1). Or by analysing distribution of times between consecutive hits, which should exhibit certain statistical features if hits are independent. Contingency Table: number of occurrences. It-h = 0. It-h = 1. It = 0. It = 1.

NAVIGATION LINKS
Contents | Prev | Next | Library

Desktop view | Switch to Mobile