Risk managers are making a more conscious effort to not only understand the virtues, but also the limitations, to risk models as they adjust to life under greater regulatory scrutiny, in particular the AIFM Directive in Europe.
Indeed, such is the impact of regulation that it has in effect transformed the risk management function from something that was previously confined to portfolio management into something altogether more holistic.
Alan Picone (pictured) is a managing director at Duff & Phelps' Kinetic Partners division. He has a broad range of expertise on risk management and given that Kinetic Partners was one of the first firms to obtain a third party AIFM license in Luxembourg, Picone is well versed in understanding the changing dynamics of risk management that hedge funds now face under the Directive.
"We are acting as an AIFM on behalf of external hedge fund managers to help them meet the risk management and compliance challenges of AIFMD and as such are well placed to see what the impact of AIFMD is having on managers. One of the things that we are seeing is a fairly profound mutation of the risk management function," says Picone.
The upshot to AIFMD is that managers are increasingly relying on more systems, processes and procedures, and indeed staff, to remain compliant. The role of risk management must now embrace all dimensions of the fund's value chain: portfolio risk management, operational risk management, liquidity risk as well as the management of risks that have been delegated to third party providers; this is the risk oversight function, which has become an integral part of a manager's aggregate risk framework.
As Picone points out, this requires understanding the processes and risks of the brokers, custodians, administrators and any other companies that support a hedge fund's operations.
"When you look at risk management from this perspective, and you add to that such things as liquidity management rules, concentration risk rules under the Directive, you can see that the risk management function has become very far reaching as opposed to what it was in the past," says Picone, who adds: "Another key observation is that risk management in the past was viewed as an ex post control, meaning that decisions were taken and then risk management was applied afterwards to measure the potential impact of those trading decisions.
"The AIFMD changes this fundamentally, such that ex ante risk management becomes critical. We see more interaction of the risk management function with the portfolio management function; these have always been embedded to some extent in hedge funds but not to the degree that risk management becomes, so to say, intrusive. It requires a number of critical tasks such as risk attribution, margin-at-risk impact and so on."
In other words, ex ante risk has become more systematic. Some of the larger hedge funds have long had in place robust processes to embed risk management as much as possible into the portfolio decision making process at the pre-trade level. Under AIFMD, all managers are required to adopt this mindset; it is a cultural shift.
A good example of where ex ante risk management has become a major focus would be systematic CTAs. Given that these strategies employ sophisticated trading algorithms to detect market signals and put on positions in the portfolio, there is less involvement by risk teams in the decision making process. However, where risk teams get more deeply involved is in validating the risk models used by the algorithms.
"I'm generalising here but what we've seen is that the role of risk managers in terms of validating trading models has become much more important. There is more caution being applied to test the resilience of trading models," says Picone.
Caution is no bad thing when it comes to risk management. Nobody is naïve enough to expect risk models to provide a crystal ball into future volatility and whilst AIFMD has amplified the risk function, expectations need to be kept in check. One element of this caution is by being more attuned to the limits of risk models; that is to say, understanding the qualitative risks and trying to avoid being reliant on models that are too simplistic to measure risk.
"The Black-Scholes formula for pricing options was a major breakthrough but it was qualitative and mathematically driven and as such it was viewed as a Golden Rule. It became almost Holy Grail-like," remarks Picone. "All risk modelers now understand that it is far from reality. Models come with limitations as they try to describe the complexities of reality.
"Whereas in the past, such models would have been seen as the ultimate weapon to describing reality, today they are viewed more as a starting point. Models are more prone to be reviewed and critiqued. I think risk managers are more inclined to understand the limitations of their models, and the risks attached, as opposed to believing that their model is a fair description of the world. This is a change in philosophy. Risk management has to work within limits and I think everyone now recognises those limits; risk managers are applying more rigorous governance to their risk models, which you wouldn't have seen 10 or 20 years ago."
Regardless of whether the risk model is ex ante or ex post, the implied assumptions mean that it can only ever produce a risk estimate. In that sense, risk management is like trying to predict the weather; some may say it is even harder, given the true complexity of financial markets. Risk managers are required to somehow quantify the level of risk the fund portfolio is exposed to, without having a complete understanding of the nature of the market.
It is a fantastic challenge.
"Risk managers know that whenever they develop a risk model it is only ever likely to describe some element of reality, but not the full reality. In the past, this acknowledgement of the limitations to risk models was not communicated; the classic example being Black-Scholes. When you look at the underlying assumptions they are completely outside the realm of empirical reality and you have to wonder why this model has been used for so long. The assumptions have nothing to do with reality. The reason it has been used so extensively is precisely because the model is simple.
"My point is that there is now a more conscious attitude towards acknowledging not only the virtues but also the limits of risk models. Risk managers are increasingly mindful when using models to understand, first and foremost, what they don't do rather than what they actually do. And that's an important paradigm shift," posits Picone.
In some respects, the AIFMD has helped push forward the cultural adoption of risk management and broaden its application. Model risk is an important evolution precisely because risk managers do not want to get blinded by the VaR number; nobody wants another Long Term Capital Management event to occur.
"There's now a need to more accurately capture the specific features of financial markets. If you want to benefit from market inefficiencies you need to look at numerous factors and lots of data. For quantitative and systematic strategies, the detection of inefficiencies and the decision making programs are becoming increasingly influenced by risk managers.
"For other strategies, such as equity long/short, risk managers will have a growing influence in guiding the portfolio manager(s) during the investment process. They will provide additional insights and tools to help the front office with respect to risk allocation, risk attribution etc. Portfolio managers want to better understand the risk budget," concludes Picone.
With greater scrutiny of the risk numbers and a willingness to question the limitations of risk models, hedge funds are working harder than ever to operate safely in today's financial markets.