To attract assets and satisfy both clients and regulators, every buy-side investment strategy needs to tell a good story. But when it comes to risk management, is the story always as complete as it could be?
Traditionally, buy-side institutions have used factor models to pinpoint, through attribution, the sources of risk in a portfolio or fund. Sell-side investment banks, on the other hand, have more typically used simulation tools that are better-suited to
manage the large numbers of derivatives on their books.
Nowadays, of course, many buy-side firms invest in the same derivatives and structured products as investment banks. Like the sell side, they are finding that, in the case of fat tails, simulation models provide downside risk information in a way that factor
models simply cannot.
For a complete view of market risk, however, firms also need to know what can happen when markets are shocked. Here, the buy side can again learn from the sell side – by using stress tests as well as simulation tools. The key is for factor models to drive
simulations, which themselves are based on a coherent macro perspective, rather than random processes. It is not enough, however, to consider a couple of obvious stress scenarios: many plausible scenarios will be required to capture all the risk for even a
moderately complex set of funds.
Buy-side firms, like banks, also need to think about reverse stress testing, and focus on the scenarios that could most hurt their specific investment strategy. Factor models are a good starting point for identifying appropriate stress tests, as they provide
insight into the factors driving investment strategy and risk. But model-defined stress tests risk not being specific enough to the positions held within a portfolio. So, organizations should be prepared to stress correlations rather than just pricing factors
Meanwhile, it is critical to remember how dominant macro risk can be, as the financial crisis so clearly demonstrated. Risk models must not only treat macro factors like interest and foreign exchange rates, commodity prices and inflation coherently, but
also capture correlations between and across multiple asset classes. As we have witnessed, shocks to credit conditions affect all asset classes. So, it is advisable to implement a coherent multi-class model that can be extended to new classes as a firm grows.
First, though, the buy side needs to emulate the sell side and establish more flexible, open risk management systems. These in turn must use a full range of indicators, from VaR to volatility – and not as absolute measures, but as part of a holistic risk
management approach based on scenario analysis, tail risk and stress testing.
The buy side is already learning to combine simulation and factor model approaches, consolidate risk systems and introduce an integrated approach to risk management. But when firms can realize the advantages of enterprise-wide market risk systems, and tell
the whole story of risk on a portfolio, investors are even more likely to buy into their strategies.