Many providers of financial services believe reality is much too complicated to handle per se.
Instead they spend most of their resources promoting models designed to simplify reality while capturing "most" of the risks.

 

At RiskServers we have always considered models as the biggest flaw in the measure of Financial-Risk.
This is why RiskServers puts huge amount of effort into designing highly accurate valuation techniques that  offer the best and most flexible financial engineering and numerical procedures; are completely adaptable to a wide variety of assumptions and methodologies and yet can be tweaked to benchmark values down to the last cent !

As model risk is part of the risk engine kernel there are numerous frameworks to control what and how valuation takes place.

As many frameworks and technologies are proprietary, we can only offer a general overview of possibilities?
If you require more specific information regarding aspects of model risk, please contact product support here.

Levels of Model Risk Control.

RiskServers offers multiple layers of control in the measure of model risk.

 

Asset Valuation : For each asset in your portfolio, you can switch between any number of pricing algorithms you desire.
The standard approach used by many customers is to run a differential analysis with the industry standard pricing algorithms that are delivered with the engine and one or multiple in-house or proprietary valuation algorithms used by your company.

Methodology Control. The Risk Management Framework provides multiple market and credit building blocks to control the type and kind of stochastic movements or price sensitivities applied to your analysis.  

Risk Factor Controls: Every variable that affects the forward projection of a risk factor can be modified in the analysis

Assumption Controls: Assumptions can be modified in every analysis, which means you can actually measure the impact that each set of assumptions will have on your results

 

 

this as the biggest flaw of financial-risk-management.

 ve always followtaken the opposite strongly believe the opposite: we never take shortcuts or impose simplifications.

 of reality is not an acceptable shortcut. This is why we have spent a lot of time and effort providing the frameworks that  empower users to impose the exact same patterns that take place in the marketplace.

 

 followed the much more complicated taken the opposite view. We believe it is more important to  precisely what takes strongly felt this is the taken the opposite
The idea of the model is that the simplification is a good representation of the reality that takes place in the marketplace.

 

At RiskServers our philosophy

 Models are a simplification RiskServers does not promote models.

 

Measure the Risk of your Models by design:

You can manage the risk of different models by giving each asset a different pricing routine.

This is done by including the model field with the name of your pricing routine.

When you assign a "model" field to your trade definition, the engine will override the industry standard pricing algorithm with the one you have supplied.

This feature is controlled at the Trade level, which  means you can assign different pricing routines to the same instrument through each trade definition.
You can also switch between proprietary / industry standard  models through the setup definition.

Details:

Every Trade includes a model entity or tag:

<model>MyPropPricing</model>

or

Name Value Tag or HTTP.

MODEL MyPropPricing

The MyPropPricing method or function is then picked from one of the following C and C++ APIs:

1. A shared Memory Hookup that catches every risk factor used to value and project asset prices.
    Pro: Easy to program. See the OpenAPI specs.

2.  A shared Memory Hookup that catches every risk factor used to value and project asset prices.
    Pro: Easy to program.

 A plug-in module
Con:Requires packaging code into dll.





©2009 RiskServers SA. All rights reserved.