As hedge funds increasingly employ new approaches to asset allocation there has been a rise in the use of Risk Parity funds. The vulnerability of Risk Parity funds is that they allocate by risk rather than by value. That is actually a bet on whether equities will outperform the return of additional risk adjusted asset class allocations utilizing leverage. As you may come to believe after reading this short article, “Risk Parity” is an objective to satisfying a Liability Driven Investment Policy acknowledged or not. RISK PARITY should actually be termed, “RETURN PARITY.”
The idea behind the growing interest in Risk Parity funds touted mostly by hedge funds is to have an “all weather” approach equalizing risk from all asset classes. Whenever leverage is employed implicitly through derivatives or explicitly through borrowing, certain financial measuring metrics are manipulated out of an apples and oranges comparison, for example, the Sharpe ratio. One thing in common to both Risk Parity funds and traditional asset allocation models is correlation.
The biggest selling criteria for hedge funds is touting correlation. Many investors – even large institutions – see correlation as a risk metric though it is not. An asset may have high or low correlation discrete from its probability of providing a targeted return. Correlation itself says nothing to the risk of not achieving a targeted return. Correlation is not scaled, correlation assumes linearity and correlation is a moving target. Most of your adviser’s charts predicate recommendations of asset allocation on correlation that assumes consistency to reflect a “mean-variance” optimized portfolio. For hedge funds, the use of leverage, derivatives and hedging requires higher moment analytics at the least.
The real question is, “What is the best objective to achieve the highest probability of achieving the return target?” For pensions, an underfunded status could be back-breaking for private organizations to have to remedy. If you had $100 and needed to win $20 would you bet the $100 all on one race on a single 1-5 favorite horse? Would you bet $10 per race split into $5 each on two horses to win? Or would you bet $1 on a 20-1 long shot and pocket the other $99? As you can see, there are many combinations to achieve the return. That is the real risk – the risk of not achieving the return. That is the paradigm behind Liability Driven Investment Policy. Comparatively, whether an institution adopts so-called “risk parity” solutions is really an attempt to optimize the probability of achieving a targeted return consistently in periods of differing capital market scenarios. There are as many approaches to risk parity, the first being to define risk – not as easy as the efficient mean-variance charts illustrate.
Lower rates have caused heightened sensitivities not present previously among asset classes. This is seen in lower spreads and premiums in credit and interest rate sensitive securities, as well as exaggerated ROE and Operating Income from lower cost of capital and labor costs. None of this increased risk is captured in chaining return observations. In fact, the longer the chain of observations, the more likely that correlation becomes statistically significant as it is a function of degrees of freedom. So for certain risk parity models, it is important to assume low or no correlation of added asset classes for it to work. Other risk parity models rely on risk factors rather than return observations.
The answer to the question, “What is the best objective to achieve the highest probability of achieving the return target?” is that there is no one objective. Risk parity is great for certain asset classes and bad for others. Hedge funds are still the best place for the applications of risk parity partly due to the fact that the use of leveraged investment vehicles is conducive to risk parity.
Hedge funds take advantage of anomalies of pricing in one way or another. For instance, in my DITMo Strategy, the investment philosophy is that Deep-In-The-Money options are overpriced. Convertible arbitrage will take advantage of gamma trading, or credit mispricing. Hedge fund managers may see that correlation is not apparent except in lagging data or that there is a geopolitical or economic impact. Hedge fund managers make credit, basis, theme, trend, special situation and volatility investments that, in themselves, are discrete asset classes conducive to risk parity. Other asset classes that base value on a longer view of discounted cash flow would be less conducive to risk parity models. The few exceptions would be where there are embedded options in bonds or structured products. Risk parity models should be customized to differing asset classes rather than applying risk parity as a homogeneous asset class allocation tool. Rather than correlation, mean-variance or chaining return observations, scenarios of factor risk are the superior risk parity paradigms. Any model of asset allocation should be used as an objective within the parameters of Liability Driven Investment Policy to a stated targeted return.
Pj de Marigny