Five Ways Your Risk Reports are Probably Failing You

It’s not a question of IF your risk reports contain one of these critical errors. It’s a question of WHEN they will manifest as a reporting result you can’t explain.

In a meeting with an investor or client? When the market gets shocked by an unforeseen event? During a risk committee, investor, or regulatory review? Your annual job review?

To compound the issue, these type of errors usually DO NOT appear as errors in risk system report logs. They’re ghosts, viruses. And they attack when you least expect them to.

How will you deal with the following critical concerns when they arise?

1. Investors, Risk Committee or Regulators request up-to-date model inventory and mapping rule documentation to verify how your firm is modeling risk in your portfolios.
   Model mapping rule documentation should indicate how models are configured and applied to your portfolio holdings. Without it, can you quickly and easily explain how securities are properly represented by risk models? If you can’t easily review position mapping to risk models, you have no assurance that positions missing required terms weren’t just mapped to Cash by default, or if Fixed Income and Commodity ETFs were incorrectly mapped to Equity models. These and more hidden situations could wildly affect the trustworthiness of risk reporting from your systems.
2. Your developer asks what proxy rules to use for a new fixed income position where data is missing.
   If your answer is, “I don’t know, what’s everything else mapped to?” then you have a problem. Ambiguous or overgeneralized proxy rules mask and distort risk. In this example, Fixed Income instruments mapped to the same Fixed Income index as a proxy will cause an additional problem by inadvertently substituting fixed income risk with equity risk. Another of the many examples occurs when all equities are mapped to the same index if market data is missing. A clear set of proxy rules with well-defined logic will mitigate risk reporting mistakes in these types of situations.
3. Your CIO notices commodity futures spreads are suddenly out-of-whack, though you’ve not changed your position, and the market has not been shocked. He needs an answer now for an investor call.
   You are likely using the wrong contract in a commodity calendar futures spread, and one leg expired early. Risk systems will allow you to use almost any market data in a model. For example, you can apply a US Government curve to a Turkish or Greek sovereign CDS, although this is obviously a big mistake. Using US Government market data for corporate debt may appear fine when looking at DV01 but it distorts VaR and stressed scenarios. What tests are you performing to validate models other than market values and Greek statistics?
4. You’ve just added new call options and puts to your portfolio. Easy, right?
   The job doesn’t end with adding the positions. Unfortunately, it’s not always as easy as mapping them to defaults in the risk system. Often, these defaults will erroneously enter call options as puts and vice versa. Which, obviously, is problematic for close-to-the-money options, but also will confusingly manifest in stressed scenarios for deep out-of-the-money options. And we can all agree we don’t want any more surprises in a stressed environment.
5. Your year-end report shows your inverse ETFs hedges are behaving much differently than you had planned.
   In fact, they’re performing exactly opposite in a stressed environment, throwing off the actionable information your firm has used to manage real portfolios as markets entered a period of stress. It is very easy to setup inverse ETFs so you get exactly the wrong behavior in a stressed scenario, since they are traded as a long position. For example, a short S&P 500 exposure via SH will appear as long equity exposure and therefore show a gain when you move equities up in a stress test, exactly the opposite of what you expect.

Effective risk reporting starts with diligent model risk management.

Until you tackle this beast, your risk reports are likely riddled with mistakes that undermine the confidence in your risk process and cause skepticism from portfolio managers, investors and regulators.

Relying on error logs, market values and simple greek statistics for validation of risk modeling can allow these types of problems to persist in risk reports. A more reliable method is needed to achieve a robust risk reporting process.

Red Swan Risk specializes in diagnostic testing and monitoring that can detect these types of common mistakes and provide greater transparency into the risk modeling process.

If you are concerned about your risk modeling process and want to build more confidence in your risk reports, please contact us for a free analysis.