Risk Engine

DFA 2020 – A Complex Scenario Generation Tool

Neural Network Based Risk Measure

Beyond VaR Tool

DFA 2020 – A Complex Scenario Generation Tool

ERM-Economic Risk Capital System

ERM-Economic Risk Capital System

For more details please write to info@rsquarerisklab.com & copy to mail@energyday2030.com

The INPUT & OUTPUT of the Pilot-DFA (either for Non-life or Reinsurance), upon completion of the project with the prospective collaborator or client we should be able to feed customer data and give them the customized Excel (or MATLAB or GAUSS or JAVA) based output and report based results. R-square RiskLab doesn’t share the source code with the collaborators and clients. The extensions are NOT included for the Pilot-DFA however we explore the possibility to contemplate additional questions after first delivery.

Neural Network Based Risk Measure:

R-square RiskLab is developing Neural Network based Risk Neutral Measure Tool for ENERGY & FINANCE. One of the implications is to measure Short Term Future Price Strategy of Assets (Oil to Structured Products).

Beyond VaR Tool:

Are present risk measures and capital requirements adequate for Finance sector? Short of 100% reserves this is the best we can do.  The scary thing is the concept of “adequate” capital requirements.  If people think such a thing exists then they are simply deluding themselves.  “Adequacy” here is a relative concept. Even 100% reserves. depends from leverage. Leverage 15/20 times and even 100% does not make you safe. The risk measure that is in use in the financial industry is Value-at-Risk or VaR.  Relevance of VaR, the expected value  is an indication. When negative this is not good. The quantile  or  is defined as the number  q_α=inf{x ├|P[ξ≤x]┤>α}. Sometimes we get  P[ξ≤q_α ]= α

Mathematically speaking this is a quantile at accepted failure level (say 0.1% or 0.5%).  The required capital is then calculated in order to have a failure probability lower than the VaR-level.  No concern of what happens below that level.  This is the same as given a free option to the company. To put it in an extreme way:  in case you go bankrupt, make sure it is a very big bankruptcy.  The lower you might get, the higher you can come with the expected benefit. Positive parts and negative parts are in equilibrium. Some have created VaR-like measures which integrate the tail (or the square of the tail, etc.) to distinguish the cases as you note.  This does give more information, and even can buy one out of the failure at subadditivity. There is a limit, however, to the information content of a single point (or two or three points) on the real line.  A complete function (like a distribution function) provides infinitely (literally) more information.

This is a unique website which will require a more modern browser to work!

Please upgrade today!