By Ewan McAlpine
Bond investors selecting benchmark indices against which to measure performance and risk assume that those indices are good representations of the investment universe. This assumption is flawed, as bond indices are constructed on the basis of rules resulting in distortions and misrepresentations. While not new, such flaws have recently gained greater visibility.
Bond index construction methodologies are less simple than for equity indices. Index membership depends on criteria such as currency, maturity and credit quality; weightings are determined by market capitalisation, but a bond’s capitalisation is related to the amount of debt taken on by the issuer. While the asset-weighted construction of equity indices makes sense, the liability-weighting of bond indices means that the most indebted issuers are the most highly represented in the index. Thus, investors will seek to invest most heavily in the most indebted issuers, which will typically be the most liquid.
The construction of portfolios managed against benchmark indices will, by definition, be influenced by benchmark structure. However, as well as investor herding, further distortions are caused by index structure. For inclusion in the most commonly used credit indices, individual issues must satisfy various criteria, the most important being minimum size and credit quality.
Minimum size excludes smaller issues that may otherwise satisfy entry requirements and be of better credit quality than other similar but larger issues. While credit quality is a sensible criterion, setting it in terms of ratings provided by major credit ratings agencies fails to recognise that not all issues will be rated – an issue size may be too small or a rating too expensive. Small and non-rated issues are therefore excluded, so rather than being heavily sought after by investors, they will be overlooked and may become relatively illiquid. More broadly, agencies focus on default probability and ignore likely recovery rates; the true value of corporate bonds is a function of both. Thus the credit quality of market indices, and portfolios based on them, will be incompletely described.
Indices do not encourage investors to look beyond benchmark universes, typically do not present them with an ideal mix of diversity and security and, because their composition changes constantly, do not encourage investment for the long term. The flaws in a benchmark-centric investment approach are evident from the variation in the index weighting of senior unsecured bank bonds over the last ten years (from pre-crisis levels of more than 40% to around 20% now). Without an awareness of these problems, benchmark indices can potentially result in unwanted portfolio risk.
However, with inefficiencies come opportunities. By not being tied to a benchmark index and focusing on the most liquid issues, investors can take advantage of opportunities to buy bonds offering real value in terms of yield spread, with security assessed on the basis of analysis of covenants, structure and, when relevant, security, and with the only penalty being relative illiquidity. This has to be considered beside the typical requirements of a long term investor in bonds: good total returns, certainty of cash flows supporting those returns and a long term view.
When sterling credit markets were in their infancy prior to the 1990s and before the dominant role now occupied by credit rating agencies, long term finance was available only to the largest corporate entities or borrowers prepared to pledge security over assets. Whilst the former may have achieved a credit rating, most long term secured borrowing was undertaken without one. While few issues are sold nowadays without a rating, a sub-set of sterling credit markets remains unrated and such bonds are often undervalued because investors under-appreciate the security offered, favouring more transitory concepts such as ratings or liquidity. Researching those areas where most added value can be found, such as secured off-index bonds, offers real opportunities to improve the risk/reward characteristics of portfolios.
Where portfolios are managed relative to benchmark indices, investment managers must have enough flexibility to invest off-index in order to benefit from such opportunities and this freedom must be clearly expressed in portfolio investment guidelines and restrictions.
An alternative approach is to break free of the constraints presented by benchmark indices and to focus on objectives such as yield spread target, duration target, sector preferences and diversification/concentration limits. Such an approach, built around key strategic objectives, lends itself to the construction of bespoke portfolios designed to capture value created by market inefficiencies across the maturity range. Such bespoke ‘Buy & Maintain’ portfolios represent a strategic solution that enables investment managers to focus on the real issue of portfolio risk.
Ewan McAlpine is senior client portfolio manager at Royal London Asset Management
Comments