Aggregate Risk Data. But What About Risk Insight?


The fist time when this came to my mind was when I thought about Big Data. What About Big Insight?. With the conclusion: computation is a strong concept especially, when it is combined with data management intelligently.


Regulators are assessing that many financial market participants are still inactive and risk data aggregation and risk reporting represents a fiasco (just waiting to happen?). Basel Committee on Banking Supervision released a consultative document 2012: Principles for Effective Risk Data Aggregation and Risk Reporting.

Predict what I get from what/how I do ... and control towards objectives. 

Provocatively speaking, I recall the great Total Quality Management movement (my factory automation time). But most of the  certificates were granted based on the formal aspects of procedures and reporting, not so much on methodologies, models, procedural instances, ... .

But we went already into inline quality development. With the first step of predicting quality from the process/manufacturing features and then control features towards quality requirements. Aerospace industry forced suppliers already then to prove exactly this.

Regulators want data and reports that allow them to make cross-firm comparisons. Market data positions, trades and collateral, exposures, ... aggregated by clients, industry, territories ...?

Yes, I agree it is not enough to have data in Excel books.

This is why we have integrated valuation and data management before 2008. Our UnRisk FACTORY keeps "who did what and how with which data" evident. Complete over the complete history. And it continuously produces "cubes" of risk results across instruments, models, portfolios, risk factors, ... You can view them in the FACTORY web front end, but also process them.

In our understanding risk data aggregation needs a vast variety of mathematical/statistical functions, data analysis capabilities and dynamic visualization to provide insight into the dynamics of the return at risk behavior. With the objective to control deal structures towards the preferred profile. This is what we provide with UnRisk-Q - the programming power behind UnRisk.

Risk technology should provide more than data sets that allow for cross-firm comparisons?

Picture from sehfelder