Scope

  • Client Risk Reporting
  • Black-box service vendor replacement
  • 25,000 Client Portfolios
  • Cash, Equity, Bonds & Options

Requirements

  • Historical Value at Risk (VaR)
  • Fixed Stress Tests
  • No disruption to existing systems

Benefits

  • Showcased savings of over €400k per year
  • More transparency for Risk Managers when answering client questions
  • Improved data quality through automation and transparency
  • Improved operational efficiency

Delivery

  • PoC in 3 weeks
  • Remote, fully managed service
  • Production ready in 4 months

Use Case Details

The private bank was looking to replace a vendor service which had proved to be very expensive and unreliable, especially when risk results required explaining. The results were used in client reports and they needed the ability to understand the drivers of their risk exposures.

The bank had a large number of portfolios but relatively simple asset classes. The computation requirements were therefore trivial and as it turned out didn’t require distributed computation, as a single computation node was able to compute the required VaR and Stress Tests in a few seconds.

Implementation:

The existing vendor had an XML format to transmit daily position feed. We mapped this feed to RiskMine’s XML position feed very easily.

RiskMine Chronos was used to automatically download and manage the necessary historical time series required for VaR scenarios, directly from Bloomberg via the build-in Per-Security integration.

RiskMine Scenarios was used to create historical Value at Risk scenarios (1-day holding period, 250 day scenario window, exponentially weighted returns) and multiple pre-defined stress tests.

The compute requirements were minimal and only required a single Wildfire node to complete the total computation in under a minute.

Post-processing of the results produced other risk metrics such as marginal VaR and diversification benefit in an output format identical to what was already being consumed by the Bank.

The end result, including data and software costs showed a saving of over EUR 400k per year for the Bank.