Anthony Malakian: More Than Just Window Dressing

“Since the financial crisis, there’s been a whole industry created to come up with new names for Value at Risk (VaR). A lot of this is window dressing for investors.”

That is the sentiment of one hedge fund manager who left a top-tier fund to start his own shop shortly after the financial crisis took hold in 2008. This month Waters took a look at risk management practices on the buy side, focusing on the models and analytics used by portfolio managers and traders.

For the piece, I spoke with numerous hedge fund managers and founders, CTOs, traders, and risk managers. Trying to find a universal way to manage risk is useless, but one thing is clear: Investors are being more forceful in terms of which risk management capabilities they expect.

Knee-Jerk
But while investors are now asking more questions about risk, they might not be the most intelligent questions, according to one fund manager, which has led to a knee-jerk reaction by the hedge fund industry at large, where funds are now wrapping pretty little bows around their risk models.

“A lot of investors, honestly, don’t know what they’re talking about,” he says. “One of the problems I’ve had as a commodities trader is that I’ve had to deal with very equity-oriented risk measures that people want to expose—I’ve seen some very inappropriate risk measures applied to our business and in managed futures, in general.”

Investors might not always ask the right questions, but at least they are asking questions, which, as a minimum, warrant empirical, quantifiable responses.

But isn’t there something to be said for the axiom that there’s no such thing as a stupid question?

Data is king—that much is understood. Yet, how firms go about cleansing and distributing data tends to vary. A risk platform or model that is being fed incorrect data is a dangerous proposition. As the first fund manager noted: “I think that the problems that are attributed to VaR are actually problems of either incorrect, or frankly, dishonest input into the model.”

Disregarding the possibility of malfeasance, which cannot easily be caught in a real-time, fast-paced environment, the manager’s point regarding VaR as a useful risk measure when used correctly, is a good one. Risk systems and models are tools. As is true of any tool, these systems are only as good as their own robustness—or, in this case, data correctness—and that of the user wielding it.

One interesting observation I made while researching this piece—and this is true for more than risk-related subjects—is that there is a different mindset between the technology providers and the technology users—the traders and portfolio managers. Of the technology users I spoke with, most seemed preoccupied with their own brain power and insight, and their ability to hire top-level quants to develop models. The data feeding the models was just expected to be correct.

While vendors and CTOs might talk a good game about there being continuity and integration between IT and the front office, that feeling is not always reciprocated by front-office workers. The technology is just supposed to work and the data is supposed to be clean. End of story.

Separation Anxiety
But this separation of powers might be changing, and that’s thanks to the investors. The questions being raised by clients and prospective clients are forcing even traders and portfolio managers to develop a better grasp of their risk technologies and processes. And that means more than simply handing over a PDF with a rundown of risk factors—they now have to show that they can couple VaR calculations with stress tests. They also need to demonstrate to investors that they have the ability to calculate and understand tail risk and anticipate black swan events. Investors might not always ask the right questions, but at least they are asking questions, which, at a minimum, warrant empirical, quantifiable responses—not mere window dressing.