Making Sense of Risk Reports

We’re doing more and more business with firms that are less and less familiar with absolute risk measures.  Many institutional investors are familiar with Relative risk measures, like Tracking Error and Beta, but they are much less familiar with things like VaR, Stress Tests, Correlations, and all the other analytics that are standard fare for risk managers.  As Investor Analytics works with more of these institutional investors, its become clear to me that they could use a hand in interpreting their risk reports.  Most of our competitors (which my marketing department has told me it’s never a good idea to mention by name — see, I can learn something) don’t offer interpretation or consulting services.  They just produce the reports and send them out.  But I think it’s much more valuable to have a guide on how to make sense of them.

For me, interpreting a risk report is much easier than teaching someone to do the same.  I recently looked at one client’s reports and concluded that they were reasonably diversified but that they were putting quite a bit of weight into their domestic equity investments, which made up the bulk of their risk even though it didn’t make up the bulk of their investments.  Another client’s report told me that they were well positioned in typical markets but that as soon as volatility rose, they were in for some problems.  When each of these companies asked me to tell them how I came to that conclusion, it took me quite a while to explain that many of the different measures we provide pointed in that direction but that there isn’t just one result that lets me conclude anything.  It’s like triangulation or one of those ubiquitous police investigation shows: you need several analyses to point together to get a reliable picture.  No one piece is the clincher.  But the clients don’t want to hear that: they want to know some rule of thumb that goes like “if this number exceed 80%, then blah blah blah.”  Unfortunately, there is no magic bullet.  Instead, there’s subtle interpretation and lots of different clues as to what’s going on.

Weathermen are almost never asked to explain their forecasts on the air for a very good reason: it’s impossible to really explain their stuff in a 10-second soundbite all the while keeping a smile.  If you ask a physician for an explanation into a complex diagnosis, she and you had better have plenty of time to get into it or you won’t get a satisfying answer.  It’s the same with risk information.  For example, one analytic I promote a lot is stressing VaR when correlations go to one.  Let’s say your portfolio’s VaR is $10M, but if all the correlations are set to one that the VaR spikes to $30M.  A common question is “is this good or bad?” My answer is that it’s a two-edged sword.  On the one hand, your current portfolio shows signs of taking advantage of diversification because the VaR is much lower than if you made similarly sized investments into just as volatile securities that were highly correlated.  So that’s good.  But if markets start behaving like they did during the crisis in 2007-2009 when just about everything lost money together, then your portfolio has no natural hedges and its risk is much higher.  In other words, just when you need your portfolio’s hedges to work is precisely when they don’t.  I guess it’s like having a parachute that opens at sea-level but not at altitude.  Not exactly useful, and hence the two-edges sword.

On the positive side of things, I’ve been pleasantly surprised by how many of our clients have recently begun to accept this message – that they need to spend more time on risk and they need to dedicate a person to learn how to interpret and explain the risk analyses.  If they’re not going to hire a dedicated “Chief Risk Officer” then someone already on staff needs to spend some time learning how to do this.  In my opinion, that’s a much better plan than the alternative.  As a service provider to these funds, I think it’s time my firm started providing more help in interpreting the results.  The way I see it, it’s a win-win: we get more satisfied clients and they get more benefit from our risk services.

One Response to Making Sense of Risk Reports

  1. Glenn says:

    Great post. Of course, in many respects, it’s the cross-asset class diversification/portfolio effects that we are trying to address with more holistic measures of risk in the first place, to get away from the limited dimensionality of asset class specific measures, and yet, something does seem to get lost in the translation. I think that part of the problem is that the implication’s of VaR are much more fundamental to how we do modeling, and in their own way, are more limiting than many people realize. Even with “fatter” tails, it’s all based on correlation, and as such it assumes a great many things about the behavior of markets that aren’t necessarily true. I mean, if all actors in a model aren’t simply utility maximizing machines, hopping from equilibrium point to equilibrium point, varying only in their possession of information, then the use of correlation to understand the behavior of a given actor in that system is much less useful. I think these measures are better used to start a conversation about risk, not end one. Risk comes from the behavior of individual market participants pursuing their subjective goals under varying and subjectively evaluated constraints and conditions, not from monte carlo simulations.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s