Вы находитесь на странице: 1из 4

RISK AND OTHER DARK MATTERS

[Presentation to Hedge Fund Industry Executives at an Infovest21 Roundtable on Risk


Transparency vs. Risk Translucency, October 12, 2004, at the Princeton Club in New
York City.]

Thank you. I’m here to talk about “Managing Risk in Complex Environments” in relation
to the overall theme of our panel, “Risk Transparency vs. Risk Translucency”.

My title for this talk is “Risk and Other Dark Matters. To me, risk is too dark a concept to
be illuminated by light metaphors alone. I think of it as something that hides in the dark,
like the so-called dark matter of the universe, which along with an even more mysterious
dark energy, is said to account for more than 90% of the mass of the universe. What we
see is a sliver of what is there. So with many forms of risk. Yet, in each case, we can infer
existence of hidden complexity from the behavior of what is visible.

When I think about the many guises assumed by risk in our business, I have to admit that
neither transparency nor translucency spring to mind. Murk is more like it. Risk is a
complex, murky thing, hard to see, harder still to grasp. Risk management is a phrase
hovering uncomfortably close to hubris. Sort of like the phrase “portfolio optimization”
which seems far too optimistic a goal for a branch of the dismal science.

As Chief Risk Officer for a registered investment advisor that manages multiple hedge
funds, I am faced daily with this murky complexity and with the task of making it
sufficiently understandable to be reasonably manageable and to let everyone, myself
included, get at least an occasional good night’s sleep!

There’s an immediate paradox here. Complexity cannot be understood simply (without


destroying its nature, which is the opposite of simple). Yet no one—not even risk
managers—can visualize, let alone manage, risk in its ineffable totality. We must model
it, and by modeling it, risk hiding the very things that make risk so risky! So we hide the
complexity by reducing it to something simple, preferably a number, whether that
number be VaR, or a Sharpe Ratio, or the ever-elusive “alpha”.

There is nothing wrong with doing this, it is the only way we can approximate the risks
we are trying to tame. Yet the danger remains that we forget this, and confuse the map
with the territory. It is good to remember this as we forge ahead with our best efforts in
quantitative modeling and systems infrastructure work.

When it comes to risk reporting, there is (and should be) a lively debate about the optimal
degree of information sharing between portfolio managers and investors. We’ve grown
accustomed to framing this discussion in optical terms, calling full sharing “risk
transparency” and zero sharing “risk opacity”. Within this metaphor, everyone agrees that
risk opacity is a bad thing. What is debated, generally, is whether and under what
circumstances complete risk transparency is preferable to a middle ground, sometimes
called risk translucency. A basic rationale for stopping short of full transparency is what

1
Marc M. Groz
© 2001-2004
All rights reserved.
RISK AND OTHER DARK MATTERS

we might call “the unbearable lightness of complexity”, more popularly referred to as


“information overload”.

I have participated in this debate for nearly ten years, going back to an analysis of the
original Risk Standards for Money Managers and Institutional Investors in 1996 and
more recently, as a member of the Investor Risk Committee of the International
Association of Financial Engineers. In case you’re not familiar with it, the Committee’s
white paper on Hedge Fund Valuation is available at www.iafe.org. This work was cited
by the SEC as a valuable source for hedge fund managers and institutional investors.

An important assumption should be brought to light here. Hedge funds cannot share risk
information that they do not themselves possess. So we must ask, how good is the state of
the art in modeling investment risk?

Many would argue that we know exactly what we mean by investment risk. Investment
risk is the annualized standard deviation of returns. Don’t like that definition? How about
“Investment risk is the 1 day 1% value at risk. No good? Well, let’s try “downside
deviation” or “conditional VAR” or maximum drawdown or…

OK, you may say. So there are a lot of different definitions of investment risk, but as long
as the investor and manager agree, why should this matter?

Actually, this wasn’t quite my point. I wasn’t talking about how we measure investment
risk, but how we identify it as investment risk as opposed to say operational risk. An
example illustrates: if a portfolio declines in value overnight, we would usually
characterize the loss as an instance of investment risk. If, however, the portfolio should
have been hedged but wasn’t because of system malfunction, data entry error, or because
of a mismatch between model and reality, then we might be more inclined to view the
outcome as an instance of operational risk.

This leads us to a more fundamental question: how well-defined is the category of


investment risk, in relation to things like counterparty, operational, and other business
risks? In other words, do we even know what we are trying to model?

We are now brought face to face with a basic conceptual problem in modern quantitative
finance, the lack of a reliable framework or “metamodel” with which to organize our
overall approach to risk management. For some (primarily in academia) the efficient
market hypothesis still serves this purpose.

Most practitioners reject the efficient market hypothesis as, at best, a good approximation
that describes the behavior of markets that they don’t want to play in. But EMH is silent
when it comes to characterizing markets where hedge funds can deliver the magical letter
“alpha”.

2
Marc M. Groz
© 2001-2004
All rights reserved.
RISK AND OTHER DARK MATTERS

Where, precisely, does this alpha originate? Whether we are investors, managers, or
intermediaries, we need to understand this better than we do today. In some cases, what is
called “alpha” is probably just a run of good luck. In other cases it is an artifact of a mis-
identified benchmark. (Sometimes what looks like “alpha” is just a lack of correlation or
a miscalculation of beta.) In yet other cases, what is sold as “alpha” is the nefarious
product of inside information.

Yet beyond luck, beyond bad benchmarking, beyond the exploitation of inside
information, there are, I believe, a good number of hedge funds generating genuine alpha.
But we have no accepted theory of the origin and structure of deviations from market
efficiency, hence no general theory of the origin of genuine alpha.

Some of us get our alpha by being quicker than the competition to identify and capture
the inefficiencies that EMH assumes cannot exist for long. (Some of these inefficiencies
seem to be quite persistent, relating to structural market inefficiencies or regulatory
arbitrages.) In other cases, we may be trading identifiable investment risk for other,
harder to quantify risks. These may include other investment risks concealed in the “fat
tails” or other arcana associated with the true distribution of returns. Perhaps investment
risk is being exchanged for counterparty risk (for example, with illiquid derivatives
contracts). Perhaps investment risk is re-emerging in the form of operational risk, as
when apparent out-performance is a chimera achieved through neglect of necessary
infrastructure, including weak systems architecture and engineering, bad/non-existent
policies and procedures, and poor personnel training and retention.

Examples of this sort led me, several years ago, to formulate a principle of conservation
of risk. We begin by postulating that total risk is a conserved quantity, like momentum or
mass/energy in physics. Adoption of this postulate means that risk equations for closed
systems will always balance by definition. Conservation laws of this type are not so much
self-evident statements as they are injunctions telling scientists what to look for and how
to interpret what they observe. Their value is heuristic.

The Conservation of Risk is an injunction to find the risks that will balance a set of risk
equations in what we assume is a closed system. In some situations this simplifying
assumption should work well enough. In other circumstances, it will be more natural to
model total risk as changing within an open system. The description of systems as open
or closed is meant to suggest that Conservation of Risk belongs in a context together with
information theory and thermodynamics.

Conservation of Risk can be seen as a generalization of EMH, in which market


inefficiencies may be related to hidden risks. EMH implicitly regards the markets as a
closed system. Conservation of Risk goes a step beyond the EMH framework, expanding
the system to encompass exogenous factors whose existence provides a richer framework
for modeling the “dark matter” aspect of risk.

3
Marc M. Groz
© 2001-2004
All rights reserved.
RISK AND OTHER DARK MATTERS

An important consequence of this framework is that if we believe that the risk is


transformed, rather than eliminated, then we must continue to track and manage it. And
we must think again, and carefully, about what forms of risk we wish to retain and
manage, versus the forms of risk we transform, sell, or insure against.

Absent such a framework, it is far too easy for hedge fund managers to pretend to
themselves and others that they have gotten rid of investment risk when all they’ve really
done is buried it out of sight. For instance, as we know, VaR does not offer guidance on
tail events. And it is known that the value of historical return data even when it is
abundant is precisely what EMH calls into doubt.

As an example of how to apply this framework, let us consider the interplay among the
categories of investment, counterparty, and operational risk. While these types of risk can
be defined somewhat artificially so that there is no overlap between them, they have a
disturbing habit of ignoring such niceties and blurring one into the other. Consider a
computer systems error that gives rise, in turn, to a cascade of operational problems,
counterparty failures, and market events. Analogous cascades can (and have) begun with
counterparty or market mishaps. Like the Great Chicago Fire, all it takes is a kick from a
well-positioned cow.

Of course, the Conservation of Risk is powerless to avert such disasters; rather, it is a


systematic reminder that we are constantly trading off one risk against another. It alerts
us to look for these tradeoffs, and to seek out those risks that best match our management
abilities, rather than resting with the illusion that we have found a riskless strategy.

The fantasy of the riskless strategy dies hard. We find a way to beat the market, but
forget about the counterparty, operational, and model risk we have created to do so. It
would be better if we learn to search systematically and to make translucent if not
transparent the hidden tradeoffs in the strategies we rely upon, whether as managers,
intermediaries, or investors. (We can even try to apply this concept to everyday life.)

By becoming aware of such tradeoffs early in the game, we are more likely to benefit
from our decisions, and less likely to be haunted by their unintended consequences.

4
Marc M. Groz
© 2001-2004
All rights reserved.

Вам также может понравиться