The large compositional shifts which have occurred within Australian fixed income benchmarks over the last decade have highlighted the limitations in using these indices. Such compositional changes are a reminder that investors shouldn’t take for granted that traditional benchmarks accurately represent the desired risk/return profile for their fixed income portfolios.

Limitations of traditional fixed income benchmarks

The central issue for investors when using traditional fixed income benchmarks is the complexity of fixed income markets and the resulting implications for the construction of the associated benchmarks. This is especially true when compared to equity markets, where the underlying securities are quite vanilla and traded on central exchanges; i.e. the nature of the issuer will vary but not the nature of the actual security. In contrast the fixed income universe consists of a wide range of securities of varying characteristics which are traded over the counter; i.e. traded directly between investors/issuers without the use of a central exchange.

Given the range of securities which reside within the fixed income universe traditional fixed income benchmark providers need to establish specific criteria for determining which fixed income securities will be included in a benchmark and how the weightings will be established. These requirements create two major issues. Firstly, there can be counterintuitive compositional impacts when determining what securities are included (and excluded) solely by the criteria set by the index provider. For example, the Bloomberg Australian Bond Composite 0+ Index, can include a large proportion of offshore issuers as the criteria for security inclusion is the legal domicile of the bond issue itself being Australia not the issuer being legally domiciled in Australia. Secondly, the approach to setting security weight creates a further issue as the default is to use issuance outstanding; i.e. the benchmark’s characteristics tilt towards the actions of the largest creditors. The by-product of these issues is that the evolution of the benchmark’s characteristics will be driven by the actions of issuers as opposed to the needs of investors.

The evolving risk characteristics of benchmarks creates a potential dilemma for investors as they may not coincide with an investor’s desired risk characteristics at a point in time or, continue to match over time. This has important implications for investors as the definition of the appropriate benchmark is essential to ensuring that the desired risk/return characteristics are reflected within a portfolio over the medium term. Failure to properly reflect this within a benchmark increases the risk that the longer-term investment objectives will not be realised.

Taking control of your benchmark!

Investors need to ensure that their fixed income benchmark represents the desired risk/return characteristics over a full cycle. Owning a benchmark requires the investor to construct a benchmark where the risk/return characteristics have been, as far as is practical, clearly established and more importantly stabilised. Stabilising the benchmark refers to maintaining that (as far as possible) the benchmark does not drift from the desired medium-term risk/return characteristics which are defined by the investors medium term investment objectives.

The ideal approach to achieving this end is to define the medium-term risk/return objectives as a range of factors which can then be modelled using bottom up security level data. Doing this is ideal as it facilitates the complete separation of risk factors from the choice of securities used to create those risk factors. The only issue with such an approach is the amount of technology infrastructure and data required to undertake such a process. A next best solution is to recognise that any broad-based fixed interest benchmark can be decomposed into a series of sub-indices based upon differing sectoral and duration exposures. These sub-indices may be referred to as ‘beta-benchmarks’ as their purpose is not to represent an investible universe but represent specific fixed income beta characteristics. For example, a broad-based index could be broken down into corporate, government and semi government beta benchmarks representing different credit and/or interest rate exposures. It is these beta-benchmarks which can be recombined to create more customised benchmarks representing the investors desired risk/return exposures.

Using beta-benchmarks to create a broad-based benchmark provides investors with an everyday approach to taking ownership of their benchmark. The initial step in the process entails initially identifying the set of desired underlying exposures that can then be quantified and represented by beta-benchmarks. The beta-benchmarks can then be combined to create a more stabilised overall broad-based benchmark. The benefit of such an approach is that the composition of the broad-based benchmark can now be more easily adjusted over time by the investor to ensure that it remains consistent with the desired medium risk/return pattern.

The use of beta-benchmarks to represent beta exposures also has benefits with respect to implementing and communicating decisions regarding the management of fixed income portfolios. Though there are still limitations with these beta-benchmarks they do facilitate the use of a 'common language' when communicating between parties. For example, if management is outsourced to an external manager then all that is required is that the allocations between the beta-benchmarks be periodically reviewed and reweighted and this is then communicated to the sub manager. As the beta-benchmarks are independently calculated and publicly available both the investor and sub manager share a ‘common language’ so the risk of miscommunication is reduced.

The fundamental nature of fixed income benchmarks means that if simply accepted without question there is the heightened risk that over time an investor can unwittingly drift away from their desired risk profile. To prevent such a drift it is important for investors to take ownership of their benchmarks to ensure that the actual risk profile remains consistent with their longer-term investment outcomes. Though this may sound complex it can be done relatively simply by using a series of beta-benchmarks which can be periodically reviewed and reweighted to remain consistent with the investors medium-term risk/reward characteristics. Taking control of one’s benchmark increases the potential for investors to realise the investment outcomes they are expecting over the longer term.