Scenario analysis and reverse stress-testing key building a risk model

Scenario analysis and reverse stress testing are two techniques that risk managers should focus on when developing a risk model for an investment portfolio, according to an industry expert.

Speaking at the SunGard APT 'Meeting The 2012 Risk Challenges' briefing, SunGard head of research Dr Laurence Wormald believes one of the biggest challenges for risk managers is being able to demonstrate the value of risk management. 

Adding value to the investment process is having the ability to tell a good story about risk, and the three pillars for determining risk are risk measurement, risk attribution and scenario analysis, he said.

"Thinking about what might happen if the market is shocked is the best way to think about risk management," Wormald said.

"You need a way of looking at the non-linear pay-off associated with the derivatives that make up an investment portfolio. That's a real challenge for the linear factor model providers."

Wormald believes an important technique in risk modelling is scenario analysis, which involves using both historical scenarios that have been observed over the past 30 years and forward looking scenarios - both of which focus on shocks to the market.

It's a framework that focuses on shocks to levels, volatility and correlation, and works across any asset class, he said. 

"Forward looking models should be based on observable factors.

"Some things don't make sense economically, they don't make sense from a macro point of view - we're wasting our time trying to build scenarios that don't obey the basic laws of the global economy," he added.

Another vital tool in the scenario analysis approach is what Wormald describes as "reverse stress-testing" which helps risk managers understand what part of a portfolio is experiencing loss.

"It (reverse stress-testing) turns the conventional question around - instead of saying, 'how much will I be hurt under a particular scenario?', it asks 'what could hurt me the most?'," he said.