In the complex world of investing investors are becoming more and more aware of the need to understand and control the aggregate risk of their portfolios. A growing number of pension sponsors are addressing this by establishing clear objectives and translating them into policy frameworks. These investors are using benchmarks at both the asset class and the aggregate level to monitor their investment managers. Consequently customised benchmarks are increasingly used for institutional portfolios to define, implement and control investment strategies.
There still are investors however, who try to beat the return of the median manager. This is called the ‘horse racing’ approach, where all investors ‘race’ against the same benchmark, eg the median industry return. In this scenario investment managers with good performance will be rewarded by receiving more assets to manage and vice versa. As past performance is not a good guide for future performance and not all managers can beat the median, this approach may lead to selling low and buying high in terms of manager appointments.
So-called ‘absolute return’ investors also tend to ignore benchmarks, aiming to achieve constant positive returns unconnected with market movements. Their investment objectives might include a fixed nominal return of for example 7% or a set real return of say 4% above reported inflation as opposed to the objective to beat the universe benchmark return.
Investors who avoid using benchmarks consisting of bond and/or equity indices to control their asset strategy do so because they do not want their assets to fall in value along with the benchmark, ie with the market. Such investors might use market neutral strategies, which consist of a long-short portfolio and provide cash returns plus an additional return (alpha). However, in order to establish such strategies one needs to define a benchmark against which to neutralise longs and shorts as well as any unwanted bets. Hence there is a role for benchmarks even for absolute return investors as benchmarks help to establish market neutral strategies, which target absolute returns that are uncorrelated to equity markets.1
In many continental European countries, such as Italy, Spain, Germany and Austria, institutional investors focus on absolute return. Insurance companies and Pensionskassen2 in Germany, for example, managed to show almost constant returns over the past 10 years, and were able to distribute annual book returns (Nettoverzinsung) of 6.8–7.6% on average. This was possible because they were able to level out volatility by using hidden book reserves (Stille Reserven) as a fluctuation cushion. Currently reserves built from long-standing customers are used to subsidise new business. German insurance companies will now have to reduce their distributed book returns below the psychological level of 7% as this fixed target is no longer as meaningful.
This has shifted the focus from book related returns towards market related returns and thus to customised benchmarks, which helped to structure the investment process. In that sense, the investment process can be split into three main phases: strategy, implementation and control.
In order to derive an asset strategy, ie, the allocation of the invested capital, one has to define asset classes and recognise investment restrictions. It is also useful to determine the “optimal mix” of assets via asset liability management (ALM). In all these areas benchmarks can play an important role.
An asset class defines the type of assets (eg equities, bonds) and the region or country (eg Europe, US, Asia) of the issuer. Benchmarks consisting of different bond, equity or real estate indices thus define and separate asset classes. At the asset class level, the underlying indices can also define bond or equity universes, ie the constituents that fall within certain investment restrictions. This way a specific benchmark consisting of a composite of indices is a tool to monitor investment restrictions.
The asset strategy can be determined by an ALM study, the result of which is a strategic benchmark that can be represented by a composite of bond, equity and property indices. The customised benchmark’s return represents the target returns for the total fund while its component parts represent targets for individual investment managers. The trend where more and more plan sponsors take the liability structure and expected cash flows into account in their investment strategy has also led to an increase in use of defined customised benchmarks.
A benchmark may be a useful tool for implementation purposes, eg to determine the split between active and passive management as well as for monitoring portfolio transitions. Figure 1 illustrates how a traditional active portfolio is really a blend of two portfolios - a ‘passive’ component reflecting the amount of any stock held that is also in the benchmark and an active component reflecting the ‘active’ decisions made by the portfolio manager in under- or overweighting that stock versus the benchmark. Most strategies fall between these two poles of the passive and active management spectrum. Hence a benchmark defines the passive component of the portfolio and market risk (eg benchmark volatility) and pure active risk (eg tracking error) is represented by deviations from the benchmark. In this respect, evaluation of active risk versus the benchmark is a key component of performance measurement. In declining markets it helps to determine how much market risk or stock specific risk the asset manager could avoid. For those who do not like the concept of index tracking it demonstrates clearly the advantage of looking at benchmark relative comparisons. It is essential for any active manager to identify within any portfolio under management where the active decisions have been made. The essence of a successful active managed portfolio is to outperform benchmarks with acceptable active risk parameters, which retain exposure to specific markets.
During transition management, on the other hand, when investors want to move a portfolio from one asset manager to another, benchmarks can be valuable tools to control the transition process. One benchmark could define the starting position of the transition and another benchmark the target portfolio for the new strategy at the end of the transition period. This would help to determine any implementation shortfalls such as market impact and opportunity costs.
The last phase of the added value chain in the investment process is control where a benchmark can be used for performance measurement, performance attribution and risk measurement purposes.3 In the UK and US greater transparency and discipline in selecting asset managers is the result of the use of consultants and performance measurement statistics. This is because the benchmark concept separates investment decisions into smaller parcels relating to, for example, equities, bonds, countries, currencies, industries and individual stocks (performance attribution).
Performance is affected by many decisions, both implicit and explicit, that asset managers make concerning these risk factors. Performance attribution via specific benchmarks can examine these decisions in a variety of ways and over different time periods. For asset managers this analysis is invaluable in understanding which decisions are rewarded and which are not and thus enables them to allocate returns and understand risk.
Investment risk is often measured by tracking error, ie the return deviation from a benchmark. This also requires a benchmark. More importantly, relative performance per unit of risk (eg information ratio) is a measure, which ascertains whether the level of return achieved was justifiable in terms of the level of risk taken. Hence investors can compare risk characteristics of each asset class relative to the market. A benchmark thus provides a useful focus for manager review as the investor can gain an understanding of a manager’s current outlook by discussing the active bets or deviations in the portfolio relative to the benchmark, which represents the consensus view. It also helps the investor to understand the manager’s investment process and style. An extension of this principle has been the development of risk budgeting whereby investors who have access to sophisticated support are able to identify the risk that they were willing to take on within the different asset categories and portfolios, building up to an overall risk profile for the total fund.
The decision-making process may be summarised as in figure 2 wherein benchmarks are useful to structure, implement and control investments. The investment process essentially divides into two areas - the analysing aspect (strategy and control) and the design aspect (implementation). This picture has been adopted by an increasing number of institutional investors and consultants which has led to the trend of using customised benchmarks.
Olaf John is Geschäftsführer (managing director) at Fidelity Investment Services in Frankfurt and responsible for institutional business in Germany and Austria
1 ‘Benchmarks for absolute returns’ IPE July/August 2000, pages 70–71
2 A Pensionskasse is a captive insurance company owned by a pension plan sponsor and often imprecisely translated as ‘pension fund’
3 ‘Passive Management: Understanding the passive option’ IPE Supplement: European investment: the Euro effect, September 1999, page 16