Many defined benefit (DB) pension schemes are seeking ways to regain a fully funded status by increasing contributions and maximising investment returns. However, they are also concerned about controlling risk and ensuring that contribution levels are kept under control. Most schemes cannot afford to move to a ‘low risk’ 100% bond strategy that closely matches assets to liabilities as they need to generate additional returns. So asset allocations in the UK, for example, typically remain biased towards equities. But pension funds should be able to benefit from a more diversified approach by introducing a greater balance between equity risk and alternatives, including hedge funds. Given that the majority of pension funds now make strategic asset allocation decisions using asset-liability modelling (ALM), hedge funds should be included in this framework.
At present, hedge funds are not typically a major part of pension fund portfolios. There are no comprehensive statistics available, but we estimate that less than 1% of UK pension assets are allocated to hedge funds in total. Of the pension schemes that have invested in these funds, most have only allocated between one and 5% of their assets, although a minority have invested significantly more.
One of the main barriers to hedge fund investment is lack of knowledge and experience, although this is starting to change with the help of investment advisers and hedge fund providers. Pension funds also have genuine concerns about transparency and investor protection, which institutional providers are trying to address through product design.
Some pensions funds are apprehensive too about liquidity and the high fees charged by hedge funds compared to traditional investment products. However, exponents of hedge funds will point out the liquidity is usually better than private equity, and that fees fairly represent the quality of investment management that hedge funds provide.
Characteristics of a hedge fund
When pension funds first look at hedge funds they often find an asset which is difficult to define. They also see a large, diverse universe which includes some funds that are by no means ‘hedged’. This can be confusing, so rather than seeking a definition of hedge funds it is more useful to think about their key characteristics:
o They are heavily skills-based
o They have considerable flexibility when managing assets, and may employ short-selling and use leverage
o They focus on absolute returns or a benchmark relative to cash, for example, they may have a target of 10% a year or cash +6% a year rather than a clearly defined benchmark such as MSCI. Market neutral funds sometimes have a target relative to cash.
Because the hedge fund universe is so diverse, it is often said that there are as many hedge fund investment strategies as there are hedge funds. This means that they make money in many different ways, although it is useful to divide the returns broadly into investment skill (alpha) and market (beta). Unfortunately, it is not always obvious or possible to separately identify beta from the alpha return. The most obvious beta returns are those related to the equity and credit markets, but there are others such as liquidity risk and event risk that hedge funds have exposure to.
Because returns come from both alpha and beta, hedge funds are not a pure asset class in the sense that bonds or equities are. Depending on which hedge fund you choose, you will get some exposure to one or more pure asset classes plus exposure to active management. It is somewhat irrelevant whether hedge funds are an asset class or not, because the key questions for pension funds are whether hedge funds can match liabilities, offer higher returns, and/or offer any diversification benefit.
For any defined pension liability, it is possible to identify an asset portfolio that most closely replicates the liability. Such a portfolio is commonly referred to as the ‘least risk’ portfolio, and for DB pension schemes normally contains a mixture of fixed interest and inflation-linked bonds. For hedge funds to have any matching characteristics, they would need to exhibit consistent high correlation to the least risk portfolio. The chart below demonstrates that the correlation of hedge funds to inflation-linked bonds is far too low for this asset class to have any matching properties to pension fund liabilities.
The returns achieved by hedge funds are often highly attractive to investors. But hedge fund data is notoriously problematic, and not necessarily a good guide to future performance. Historical returns are also wide ranging, reflecting the diverse universe of hedge funds. Pension funds therefore need to assess whether the best active managers can only be accessed via hedge funds, and whether these managers have the ability to add value, after their considerable fees and adjusting for risk, over the liability matching asset class (bonds).
Hedge funds are often marketed on their diversification benefits. But for diversification to be a justification for investment, their ability to generate return must be evident. The diversification benefit is also often overstated, as illustrated below. This chart shows that the correlation of many hedge fund strategies to equities tends to increase when equities perform badly.
If hedge funds are to be included in asset-liability modelling, the modelling clearly needs to allow for investment strategy risk (for example, equity risk relative to liabilities) as well as active manager risk, as both are present in hedge funds. The total risk of the assets relative to the liabilities is comprised both of strategic and active risk.
A pragmatic approach is required when building models and setting assumptions. Because hedge funds are so wide ranging, it is often useful to group funds together rather than trying to separately model the different strategies, which is likely to be complex and, ultimately, spurious. This ‘grouping approach’ also reflects the practice of pension funds to invest in a diversified fund of hedge funds composed of a number of underline hedge fund strategies.
Consistency of assumptions is key to avoid ‘garbage in, garbage out’. For example, there is no point in modelling if the assumption for equities is a return of 8% with 15% volatility when for hedge funds it is a return of 9% with 10% volatility. The results of the modelling are known in advance (allocate all non-matched assets to hedge funds and none to equity). Mean-variance optimisation is often used in asset-liability modelling. Because of the skewness and ‘fat-tails’ of hedge fund returns (they don’t follow the normal ‘bell-shaped’ distribution that optimisation assumes), mean-variance optimisation is not wholly reliable. One way to overcome this might be to use sensitivity tests to check robustness.
Although quantitative tools such as asset-liability modelling can help pension funds decide how much to invest in hedge funds, qualitative judgement is required. As pension funds increasingly look for alternative sources of return, hedge fund investment is likely to become more widespread. The potential role of hedge funds in alpha transport and liability benchmark products is also likely to grow.
Robert Howie is a senior investment consultant at Mercer Investment Consulting, and is head of hedge fund research in Europe