The question of what is an appropriate benchmark for global bond investors has received increased attention recently. Two main factors have been responsible for this: first the US government is running a budget surplus and the treasury is buying back bonds causing the amount of outstanding US government debt to fall. Second, Japan is pursuing an expansionary fiscal policy in an attempt to stimulate its ailing economy leading to increased issuance of Japanese Government Bonds (JGBs). These factors are having the effect of increasing the weight of Japan in the global government bond benchmarks that most global bond managers have traditionally been measured against. Often when a country gains a larger index weight from issuing more bonds, investors are compensated by higher yields, as was the case with Italy in the mid 1990s, but Japanese yields are near all time lows. Investment managers, who view risk relative to published indices, are in the uncomfortable position of having to buy Japanese bonds at historically low yields or carry even larger positions relative to benchmarks against which they are judged.
Plan sponsors have traditionally used either the Salomon Smith Barney World Government Bond Index or the J P Morgan Global Bond Index to measure manager performance. Although they use slightly different inclusion criteria, both are essentially capitalisation weighted indices. At first glance this is intuitively neat and fits with the idea of the market portfolio in the Capital Asset Pricing Model (CAPM), although this model was originally developed with equities in mind. The validity of assumptions underlying CAPM and its limitations are well documented, however its survival indicates that it is a robust concept. Indeed, it can provide a useful framework to question the efficiency of bond benchmarks and how government bonds fit into the CAPM idea.
Under CAPM the world market portfolio dominates all others in risk return space and rational investors will only hold combinations of the market portfolio and risk free asset. While the world may not be perfectly efficient, investors need to have very good reasons for moving away from the ideas behind a capitalisation weighted world portfolio.
In the 1980s quite a number of investors moved their international equity portfolios away from capitalisation weights to GDP weights in order to make respectable the decision to underweight Japan. Any such moves should be considered active management decisions and not disguised as anything else.
Global bond investors are considering similar schemes in order to reduce the weight of Japan in their benchmarks. One idea is to move to global government ex Japan benchmark. Another, slightly more popular idea, is to move to a global broad investment grade or aggregate index containing corporate and mortgage backed securities.
Given the size and depth of the US capital markets the weightings of these two sectors are comparatively larger in the US than Japan or Europe. This has the seemingly attractive effect of decreasing Japan’s weighting from 26.83% to 15.93% and increasing the US from 26.06% to 47.98%.1 However, it achieves this by substituting credit or spread risk for sovereign risk. More fundamentally, is there any better justification for bond investors changing the weighting schemes of their benchmarks than there is for equity investors?
To answer this question it is worthwhile to consider exactly what assets should be included in the market portfolio under CAPM. Theoretically it should include all investable assets regardless of whether they are publicly traded or not, though for practical application often only publicly traded assets or those of a single market are considered.

All investable return generating securities are associated with one of three economic agents that interact in the economy, namely: firms, factors of production and governments. Returns on assets of these agents are correspondingly driven by the profit motive, value in the production process and the ability to tax. It can be argued that only assets associated with firms, that is equity and corporate bonds, should be included in the market portfolio as only firms are true wealth generators. Factors of production only derive their value when used by firms in the production process, and governments in their taxing and spending capacity are redistributors of wealth rather than generators.
The conclusion that can be drawn from this is that there is strong argument that only equity and corporate debt should make up the market portfolio, but government debt should not be included.
Another argument for government debt not forming part of the market portfolio stems from the premise that rational investors will only hold combinations of the market and risk free asset. For investors with a three month investment horizon, a three month treasury bill will be the risk free asset, likewise for investors with a ten year investment horizon a ten year government bond will be the risk free asset. Although CAPM assumes that all investors have the same investment horizon, which may be questionable, it is clear that to a greater or lesser degree government bonds do constitute a risk free asset for many investors. This is evidenced by the fact that many bond investors use government bonds for asset liability matching rather than creating the most mean variant efficient portfolio as CAPM suggests.
The use of government bonds as risk free assets in liability matching would seem to be a major reason for the local market biases observed in government bond markets. Using Japan as an example, only 6.4% of the JGB market is owned by foreigners.2 Japan is a large net saver and given the poor performance of the economy and volatile stock market, JGBs are one of the few assets available to domestic investors where they can be reasonably sure of the returns over their investment horizon.
This is pushing yields ever lower despite a spiralling debt to GDP ratio and downgrades to the sovereign credit rating from the rating agencies. The alternative of investing in foreign bonds for Japanese investors exposes them to currency and interest rate mismatch risk, which again does not fulfil the aim of liability matching.
Additionally, investing in foreign bonds on a currency hedged basis over the investment horizon will not give a significantly different return to investing in a JGB given the no arbitrage condition. If Japan’s debt continues to spiral it will eventually create inflation and/ or a devaluation of the currency, since the mere issuance of a bond does not create long term wealth if a wealth generating asset is not created with the proceeds. However, domestic investors will still need to hold JGBs in spite of this to match domestic liabilities.
Although government bonds should not be part of the market portfolio, they do have a role to play in overall asset allocation. As the risk free asset in the CAPM context they should be held in a mix reflecting the investor’s consumption basket. Unless foreign investors have Japanese liabilities in a similar proportion to Japan’s weight in a global capitalisation weighted index they should consider a customised government benchmark which more accurately reflects their consumption basket or liabilities.
Corporate bonds should form part of the market portfolio and a capitalisation weighted benchmark as part of overall asset allocation is more justifiable. Lastly, if global bond managers continue to be measured against capitalisation weighted global government bond indices, and we are witnessing an asset price bubble in JGBs similar to that in Japanese equities in the late 1980s, there is a strong case for active management.
Mark Talbot is head of international bonds at State Street Global Advisors in London
1 Source: Salomon Smith Barney (comparison of World Government Bond Index and World Broad Investment Grade Index at 28/02/01)
2 Source: Ministry of Finance, Japan