Test Methodology

In line with EMIR KELER CCP Risk Management completes stress tests, back tests daily and analyses their results and prepares sensitivity tests at least monthly.

The stress tests simulate with scenarios the effect the already occurred (historic) or possibly occurring (hypothetic) scenario events would have on the value of the current open market interest and the level of risk that KELER CCP runs in case of default by clearing members. In line with the requirements of EMIR the default fund has to cover the higher of the highest risk or the sum of the second and third higher risks (max (1;2+3)) that result from the daily stress tests of KELER CCP. The stress test methodology is detailed in the methodology document published.

Methodology
Stress test results are published each month, please click here to view these results.


Back tests check at portfolio level the extent to which the margining model of KELER CCP can cover the portfolio level variation margin due to the change of prices only. Back tests examine the changes in the value of the position only from one clearing day to the other, the change in the amount of the positions is not covered by the calculation. Regulations require 99% confidence level annually, this is backtested daily. Details of the back test methodology can be found in the methodology document published.

Methodology
Back test results are published each month, please click here to view these results.


Sensitivity tests are run at product and portfolio level. At product level we examine what effect the changes in certain parameters (level of confidence and holding period) of the calculated risk measure would have on the product initial margin parameter. At portfolio level we examine with the historic and hypothetic stress test scenarios what effect the % increase/decrease of margin levels and the termination of spread discounts have on the level of risks to be covered by the default fund. Details on the sensitivity test methodology can be found in the methodology document published.

Methodology