The continuing bear market and declining interest rate environment have challenged the ability of many pension funds to meet both current and future obligations. Concurrently, the successes of many hedge fund products have caught pension fund trustee’s attention; and may offer a method to diversify this equity risk premium bet. While considering an allocation to hedge funds is clearly prudent, making a decision is by no means straightforward.
To ALM or not to ALM? Traditionally, asset allocation decisions result from an assessment of the forecasted performance of competing investment opportunities. Since these forecasts often borrow heavily from historical data and reliable hedge fund data is elusive, the ability to understand and forecast behaviour of hedge fund opportunities in a manner useful to trustees lags our ability to describe the behaviour of established, traditional asset classes and the interactions among them – and we expect this to remain true for the foreseeable future. Some shortcomings of hedge fund data are:
o survivor bias,
o impact of performance fees,
o stale prices,
o data gaps,
o diversity within styles,
o complexity of strategies.
Recognising that hedge fund forecasting is difficult and likely to remain so, we do not recommend trustees expend the bulk of their resources on attempting to create more precise hedge fund forecasts. Rather, we believe the focus should be on accommodating this difficulty into the development of an appropriate hedge fund exposure.
The differing levels of confidence in forecasts of different investment opportunities naturally lead to a multi-stage decision framework in which the quantifiable aspects of investment returns can be balanced with the more qualitative, or at least less quantifiable, aspects.
The answer? Two-stage asset allocation. The traditional approach to asset allocation involves the estimation of expected return and volatility levels for each asset class as well as correlation coefficients to describe their interaction. Typically this information filters through a quantitative process which provides the “optimal” allocation to the various asset classes.
While such an approach is usually adequate for simple investment problems using traditional asset classes, it easily becomes an overly-precise extrapolation of uncertain data for complex investment questions such as the appropriate allocation to hedge funds.
In areas where investors have fairly high confidence in their forward-looking assumptions (such a traditional asset classes) we recommend using the traditional tools. In areas of lower confidence such as hedge funds, we believe the traditional approach should be augmented with additional analysis. To that end, we recommend a two-stage decision framework.
Stage 1: Develop an allocation to broadly-defined core asset classes (eg, domestic and non-domestics equities and fixed income categories). In this stage a quantitative methodology evaluates the benefits of various allocations to these asset classes by balancing the investor’s risk tolerance and preferences with fund goals (funded ratios, contribution levels, etc). Optimisation tools are usually appropriate to evaluate this decision. This high level decision defines the benchmark that any later-stage allocation must exceed.
Unfortunately, this stage does not provide trustees much insight into hedge fund allocation – our stated purpose. To move toward a hedge fund allocation, trustees must develop a hedge fund forecast consistent with those of other investment opportunities. A straightforward way to do this is by linking hedge fund return behaviour to other market factors. However, setting a forecast is as much art as skill and should incorporate known fund history along with prospectus information and assessments of manager skill. As such, our focus shifts to simulation and qualitative reasoning in Stage 2.
Stage 2: Address allocation issues such as the active/passive decision, exposures to asset class segments and the allocation to alternative asset classes. Simulation models are often more appealing than optimisation as we delve into the lower level decisions of this stage. A hallmark of the Stage 2 analysis is a reliance on lower confidence data. We confirm and roughly measure the advantages of shifts from the Stage 1 default/passive allocation.
The qualitative features of a hedge fund investment must be given at least as much weight as imprecisely quantified characteristics. Rather than relying on an optimiser to provide the correct allocation within a fraction of a percent, qualitatively investigating how a hedge fund allocation will improve the investor’s ability to achieve stated goals is a more insightful approach.
Hedge funds are not for everyone. Some trustees may not be comfortable with hedge funds, or are not willing or able to invest resources into evaluating managers or fund of funds providers. For these funds a zero allocation is best. It is better not to invest than to invest poorly.
For investors who believe that hedge funds are an attractive opportunity, our first rule of thumb is to have a 5-10%. Below this level of exposure, an allocation to hedge funds will not have a meaningful impact on portfolio performance. Above this range, the organisational and investment risks associated with hedge fund investing may become onerous. Exactly where in this 5-10% range a pension fund should fall is often a matter of judgment and finesse. Allocations in excess of 10% are rare and generally limited to high net worth individuals and well-staffed endowments.
Considerations on the level of hedge fund exposure are:
o Funding level of the scheme
o Current level of ‘risky’ assets
o Sophistication level of Trustees with respect to complex strategies,
o Resource constraints for finding, combining and monitoring managers
o New fangled tools and old fashioned common sense.
We encourage most investors to consider an allocation to hedge funds. This is an exciting investment opportunity with considerable promise. However, we also advise investors to go into the strategy with eyes wide open and not to leap at an allocation level based on the very precise machinations of a mean-variance analysis. The hedge fund investor should use quantitative information in a judicious manner, combine it with appropriate levels of qualitative research, and recognise that this investment category, while promising, is still maturing.
Steve Murray and Leola Ross are with Frank Russell Company in London