Increasingly, market practitioners are postulating that mandates in future should have a direct linkage to the fund’s liabilities and a much lower tolerance to downside risk. Such mandates will have little similarity to the ‘benchmark plus’ mandates of the past, and will make quite different demands on managers.
Leaving aside the difficulty of actually finding matching assets, State Street Global Advisors CEO Alan Brown argues that the added value products of the future will incorporate explicit deviations from the benchmark, essentially going short benchmark assets and going long some other form of risk. “Pension funds these days are long equities and short bonds, versus the liability-matched benchmark, which to most eyes is quite an aggressive and risky stance,” comments Brown. A lower-risk proposition might be to include ‘equity’ risk by assuming the credit spreads of corporate debt, either in the physical markets, or via credit derivatives. Alpha could be generated from equities, not by taking beta risk, as currently, but by adding market-neutral equity exposure, with the possibility of alpha generation derived from manager skill.
Should a fund wish to take beta risk, it should first question whether equity beta offers the highest return-to-risk ratios. Equity might offer the highest returns, but at an unacceptable level of risk relative to the returns that could be obtained by leveraging lower-yielding assets. Products focusing on absolute return must explicitly consider this trade-off and figure out a way of directing money to those assets with the highest expected return per unit of risk, so as to meet return targets.
A scheme’s tolerance to risk, being defined as deviation from the liability-matched benchmark, should vary according to the strength of the sponsor. Theoretically, a sponsor should be prepared to underwrite any loss that results from benchmark deviation, and its readiness to do so will scale the extent of the deviation. But for sponsors who are prepared, in the final analysis, to close a scheme, deviation from the liability-matched benchmark is in some senses a one-way bet. The sponsor benefits from contribution holidays if the bet does well, and has the option to close the scheme if it doesn’t. The sponsor’s readiness to exercise this option must be considered in deciding whether there is real value in benchmark deviation, as opposed to making a gamble that is only to the sponsor’s gain.
To have a realistic product offering, fund managers must make a strong case that certain beta bets are worth making, that these bets can be managed dynamically so as to reduce the risk of loss and that they can also generate alpha consistently through exploiting market inefficiencies. Chris Woods, SSGA’s head of hedge fund strategies, contends, “Structural inefficiencies, such as in the currency markets, will persist because the majority of players operate with no reference to investment objectives. In other markets the inefficiency derives from an information deficit, such as in small cap equities.” Using the CSFB Tremont indices as a proxy for the returns from various hedge fund strategies, Woods discovered that credit, equity market-neutral and ultimately global macro hedge funds were the biggest sources of returns as one increased one’s tolerance of shortfall risk.
SSGA has products within the alpha space, via its range of hedge and overlay funds. It has three market-neutral strategies, in European, US and Japanese equities, three fixed income programmes, exploiting differentials in country risk, yield curve and duration, and a currency overlay fund. SSGA has client money invested in each of these strategies and reasonably long track records.
Woods is combining these according to their information ratios as a first pass at an optimised absolute return product. The package of strategies could sit alongside a liability-matched portfolio as a source of additional returns.
SSGA is also developing a process to allocate dynamically to risky assets according to the investment horizon and the extent of the surplus. Comments Woods, “a static allocation may be acceptable in the very long term, but if the fund cannot take short-term losses, then some ongoing re-allocation is necessary, perhaps using portfolio insurance techniques”.
A number of other houses have developed and sold the absolute return concept, and their products already factor in a variable allocation to risky assets according to the risk tolerance of the client. One also incorporates an allocation to its own hedge fund product.
GMO’s Global Allocation Absolute Return (GAAR) product grew out of its concern over the incongruity between how it ran institutional mandates, versus the way its managers ran their own family money. As Boston-based Ben Inker, CIO of quantitative developed equities, relates, “The benchmark really didn’t come into it. What mattered was ensuring that there was enough money in the bank to provide for a comfortable retirement. Coincidentally, at this time GMO was arguing for money to be taken out of equities, but our institutional mandates forced us to make equity allocations against our better judgement.”
The crunch came in 1999 when GMO proposed ‘where to hide’ portfolios to its clients. Only one client made the switch at the end of 2000 after the equity bubble had already burst. Inker argues, “as soon as one employs a professional manager the motivation shifts from ‘how best to manage this money’ to ‘how best to keep my job’. This prevented a lot of clients from coming on board in the initial stages.” Now that trustees are expressing more concern about maintaining the value of funds’ assets, GMO finds that more funds are adopting the GAAR approach, at least for a portion of the plan. GMO now has funds from 10 clients, amounting to $750m (e665m) invested.
GMO allocates GAAR money to any of the 12 asset class categories on which it makes return forecasts, maintaining a fixed 20% allocation towards its own multi-strategy hedge fund. Allocations to each asset class will depend on its current risk and return forecasts, and so the following proportions are subject to change. The low-risk GAAR fund targets a sub 5% volatility and now contains 50% in inflation-protected government bonds, 22% in international equities (mainly small caps) and 8% in real estate investment trusts (REITs). As the targeted volatility increases, a higher weighting to risky assets is absorbed. The medium-risk GAAR offering, the most popular among clients, targets a sub 9% volatility. Here the inflation-protected element is just 18%, the equity component has risen to 42%, and there is a 5% allocation to emerging market debt. The higher-risk GAAR product, with a sub 12% volatility, contains no inflation-protected bonds and is 61% in equities, both small and large cap, international and emerging. GMO has monitored the performance of dummy portfolios since September 1999, and real money since December 2000, and the performance of each strand from inception has been +50% for the higher-risk, +44% the medium-risk and +37% for the low-risk option, compared with the S&P500 down 26% and a global balanced benchmark down 11%, over the same period.
Inker contends that the portfolios in many ways are easier to manage than the traditional balanced portfolio, whose allocations are set with reference to a benchmark. “Making allocations up to maximum limits based on one’s real convictions is a good deal easier than having to weight relative to a benchmark with a maximum tracking error constraint,” comments Inker. “A fairly simple mean variance technique, with upper limits, gives rise to the portfolio with the targeted risk/return combination.”
The Immuno product, offered by Union Investment, takes as its starting point the risk tolerance of the client, from which a maximum initial allocation away from benchmark assets is determined. The lowest acceptable net asset value (NAV) of the portfolio at the end of the investment period is determined, and the difference between this and the NAV that would be obtained from an investment in benchmark bonds, is the risk budget. For a fund with a target of no capital loss over the period, the risk budget is essentially the risk-free rate. Risk may be taken in the form of equity risk, interest rate risk or foreign exchange risk, depending on where the managers see the most potential for gain. If the risk budget is high enough, Union devotes some portion towards active management, as opposed to solely index allocations.
Union Investment fund manager Jens Gottsman in Frankfurt reports that over the eight years that Union has operated the strategy, the probability of shortfall has always been below 0.1% with a high level of confidence. It achieved this by allocating only towards liquid assets within the spectrum of European stocks and bonds and dynamically re-allocating according to the available risk budget as the fund’s value shifts. Funds run using this strategy amount to E8bn across 195 separate client portfolios, mainly banks, insurers and pension funds.
As Gottsman relates, “the asset allocation is dynamically re-adjusted by Immuno across multiple asset classes. Hedging initially with derivatives and then unwinding into cash reduces the risk of a market gap, which might force us to put all the assets back into bonds. This happens to between 1–2% of portfolios, whereas in standard constant protection portfolio insurance it occurs maybe half of the time. There is a clear relationship between the risk budget and the outperformance of the fund over the risk-free rate. The typical capital guaranteed mandate effectively allocates 200 basis points of risk for a likely outperformance of 100bps. This compares with the required 20% risk budget to be completely in equities, for a potential 400bps of additional return.”