The investment industry is going through a period where products increasingly are being split up into their building blocks and then reassembled into tailor-made composite investments to suit client needs. The trend seems to point towards 'sandwich investments', designed to reflect investors' true benchmarks and maximise the chances of adding performance relative to these benchmarks. Overlays have a long history of being an additional building block in investments. More recently, the more traditional currency overlays have been joined by overlays covering a wide range of markets, commonly referred to as global tactical asset allocation overlays or GTAA. These overlays are now moving into the limelight as ideal building blocks for a wide range of portfolios where they can be used to add alpha as well as to manage betas.

The development of derivatives and the use of leverage have provided a wide range of building blocks for investments. Traditionally, these tools were associated with risk and complexity and conservative investors were initially reluctant to use them. However, as derivatives have become more mainstream and it also has become increasingly clear that they can be used to increase the efficiency of portfolio management, views have started to change. With the insight that increased efficiency means having to use less risk to reach a return target, different types of derivatives are now seen more as a way to decrease rather than to increase risk.

Some of the most important derivatives used as building blocks for composite or sandwich investments are futures and total return swaps which replicate market index returns and can be used to construct synthetic benchmarks. Forwards, futures and swaps have also allowed investors to take active risk in a number of markets without having to use almost any cash. This is the toolkit needed for efficient asset allocation and the construction of optimal investment portfolios.

True benchmarks and active risk

A good starting point when constructing an optimal sandwich investment is a well defined 'true benchmark' reflecting future liabilities. For insurance companies or pension plans this can be quite straight forward. For many companies and high net worth individuals it might be more complicated to assess what will be needed for future investments and consumption. Historically, 'investment benchmarks' sometimes have been rather different from true benchmarks. Examples can be found among insurance companies and pension funds using an investment benchmark with a large equity component, while their true benchmark based on their liability profile might be closer to an inflation-adjusted domestic bond portfolio. This was not viewed as a problem as long as the investment benchmark performed better than the true benchmark, the portfolios and schemes were in surplus, most of the peer group had similar investment benchmarks and the accounting rules did not insist on mark-to-market. However, with the equity bear market a few years ago, the surpluses gone and regulatory changes towards mark-to-market rules, things have changed and it has become increasingly clear that any deviation from the true benchmark should be considered as an active investment risk and as such should be compared with other forms of active risk taking.

These developments are likely to fast-forward the change from investment benchmarks to true benchmarks and increase the demand for replication of the true benchmarks in the most cost efficient way, which in many cases will boost the interest in synthetic benchmarks created with derivatives. It will also make it more important for investors to use their active risk budget in the most efficient way, which, in turn, is likely to speed up the interest in diversified alpha investments, such as broad-based overlays, fund of hedge funds and other investments with high risk adjusted returns and low correlation to market indices.

Separate alpha generation from beta

To construct the best possible building blocks it is important to look at the creation of the benchmark or beta separately from the generation of the outperformance or alpha. Based on the view that risk-adjusted returns are a function of skill and opportunity sets, there is no reason to limit the search for the best alpha generation to the asset class of the benchmark. On the contrary, an unconstrained search for the best alpha generation often leads to a diversified investment in a number of areas unrelated to the benchmark.

The way the betas are generated will have implications for the flexibility on the alpha side. If investment constraints prevent the use of synthetic benchmarks or if derivatives are not available for the true benchmark, the cash will be needed for the benchmark portfolio. In these cases the active risks are often taken in the same asset class as the benchmark. A typical example is a total portfolio with benchmark allocations to domestic and international equity and fixed income and the active risks expressed as active mandates in some of these areas. Here overlays can become the entry point for separation of alpha and beta and allow the investor to introduce an additional alpha source with low correlation to the benchmark and other active returns. This is how currency overlays have been used for a long period of time and this is where most of the current interest for GTAA overlays is coming from. With synthetic benchmarks created with futures and swaps, the flexibility on the alpha side increases and the alpha investment can now be funded or unfunded, as the cash is not locked in the beta generation.

A combination of synthetic benchmarks and the cash invested in an alpha source, such as fund of hedge funds, is commonly referred to as 'portable alpha', even though it probably should be called 'portable beta', given that the beta is added on to the funded alpha source. Technically, it is also possible to construct derivatives such as total return swaps on fund of hedge funds investments, even though this very often becomes less cost efficient than a direct cash investment into the fund of hedge funds and a synthetic benchmark. Overlays have a large role to play in these scenarios as well. First, they provide an opportunity to manage the residual betas in the fund of hedge fund investment. Secondly, the overlays continue to be a valuable source of additional uncorrelated cost-efficient alpha, which can reduce the need for leveraging the fund of hedge funds investment.

Overlays to manage residual betas

Currency overlay managers have for a long time made a distinction between passive and active management. The passive management is associated with the selection of a currency benchmark or optimal hedge ratio and the job is to keep the exposures at this neutral point. Active management is the deviation from this neutral point with the purpose of adding returns. An example could be an international equity portfolio hedged 50% into the domestic base currency and the manager then taking active currency positions in a number of different currencies in order to add alpha.

This currency overlay example involves two different things, reducing unwanted residual exposures, the currency risk embedded in the equity portfolio, and adding an uncorrelated alpha, the active management of a number of currencies. This framework is now being widened to new types of portfolios, including alternative portfolios, and to new areas of overlay alpha creation, such as GTAA. The reason there is a need to manage the currency exposures in many portfolios is that the currency exposures often come as 'part of the package'. Buying a stock on an unhedged basis means expressing a view on the company as well as the domestic currency of the company. The same often happens with many hedge fund strategies, where the pure alpha is mixed with different types of residual betas.

The clearest example is long/short equity managers with a long bias. Many of these managers aim to capture around 23 of the upside of the equity market and only 13 of the downside. This can be viewed as a combination of pure alpha and some equity beta exposure. This happens in many other hedge fund strategies as well, even though it is less explicit.
 

Removing the beta component of a single active manager can be very difficult, unless the manager is a more traditional long only manager. For most single hedge funds it requires a level of information which is unrealistic for an overlay manager to receive. However, the beta exposures tend to be more stable in most diversified fund of hedge funds investment, which makes it easier to estimate the betas from the underlying hedge fund managers and pass on these estimates to the overlay manager to act upon. One example can be that the overlay manager is instructed to passively reduce global equity betas by 50%.

GTAA overlays to add alpha

One lesson from currency overlay is that it is better to have a long-term strategic view on the benchmark and to base this primarily on total portfolio risk considerations and not to confuse changes in the benchmark with active management. It is easy for an investor with a fully hedged benchmark to become very tactical and short-term when the domestic currency is falling relative to the currency of a foreign investment and start changing the benchmark to partly or totally unhedged. However, this often means trading a single exposure infrequently, in large size and with limited information, which is not the best environment for alpha generation. It is better to take the implicit risk budget of these benchmark changes and use it for active professional currency overlay management in a broad range of currencies. The same lesson has also been learnt in other types of asset allocation. Initially, global tactical asset allocation was more strategic than tactical and mostly focused on changes between asset classes, the big calls on the shifts between equities, bonds and cash. However, with more focus on an appropriate long-term strategic benchmark, most modern GTAA overlays primarily create alpha at the country level within a number of asset classes, as this broadens the opportunity sets and adds diversification benefits.

Looking at most modern GTAA programmes, currency and fixed income overlays still very often comprise the core. Country allocations with equities are often also included and sometimes also other asset classes such as commodities. In some cases equity market neutral equity allocation at the stock level within countries is also part of the overlay, even though this more commonly is referred to as the hedge fund strategies equity market neutral or statistical arbitrage. Some of the most successful GTAA overlays use a cross-sectional multi-strategy approach, which means that they constantly hold positions in different relative size across a number of asset classes and countries. This approach leads to increased diversification benefits between asset classes, countries and over time. The bottom line is that GTAA now should be seen as a diversified multi-asset class overlay rather than a few calls on the global equity and fixed income markets.

Putting it all together

Looking at the different parts of a composite investment, we would like to introduce the 'sandwich approach' to investment (see Figure 1).

At the bottom we place the generation of benchmarks or betas, which can be done synthetically with futures and swaps or by using direct cash investments. The middle part is the primary alpha source. If the bottom part is unfunded, ie. not tying up cash, there is full flexibility to choose between a funded or unfunded primary alpha source. In many of these cases the investors go for a diversified fund of hedge funds investment. If the bottom is funded, due to investment constraints or cost considerations, the alpha in the middle part is usually captured by active deviations in the asset classes of the benchmark. The top slice is the overlay part, which really consists of two layers, a passive beta management layer designed to reduce unwanted residual betas coming from the bottom and middle part of the sandwich and an active alpha management layer, eg. adding alpha through a GTAA overlay focusing on multi-asset class investing on a relative basis in a number of countries as well as to some extent timing betas between asset classes. The overlay part is unfunded and is possible to put on top of any sandwich regardless of how the other parts have been funded.

This all might sound complicated and therefore costly. However, the construction of investment sandwiches like the one outlined above is fairly straightforward and is already starting to be implemented by some larger hedge funds and banks. Implementing synthetic benchmarks through futures and swaps is becoming increasingly commoditised and cost-efficient.Overlaying fund of hedge funds investments with broad-based GTAA overlays was pioneered by Millennium Global Investments more than two years ago.

When calculating the costs of a sandwich it is important to think about opportunity costs as well as actual costs. The combination of a benchmark portfolio and active risk-taking constrained to the asset classes of the benchmark, can lead to big opportunity losses and in some cases actual losses through expensive and poor alpha creation. The proper way to evaluate the alpha creation from these more traditional long-only managers is to look at their total fees and then subtract the price of a synthetic benchmark or a tracker, which gives you the implicit fee for the alpha. This alpha should then be assessed for concentration, ie. the annualised volatility, and the quality, eg. the Sharpe ratio and Sortino ratio. In this comparison it becomes clear that the best price for a unit of high quality alpha usually is to be found within broad-based overlays such as GTAA, diversified fund of hedge funds and multi-strategy hedge funds. The best proposition is likely to be a combination of cost-efficient benchmark creation and alpha generation through GTAA overlays and fund of hedge funds. The future looks bright for the sandwich and for GTAA.

Arne Hassel is Managing Director GTAA & Fund of Hedge Funds at Millennium Global Investments.