Carry that weight
A new approach to bond indices
Jul 8th 2010
THINK of all your old classmates from school. Which of them would you lend to? There were probably a few you would not trust to return a pencil. So the idea of lending to all of them, in proportion to the size of their accumulated debts, sounds like a stupid proposition.
But that is what bond indices do. Just as equity indices are weighted by the market value of the constituent companies, fixed-income benchmarks are weighted by the value of the debt. In the case of government-bond indices, the more a country borrows, the bigger its weighting.
PIMCO, a fund-management giant, thinks it has come up with a better version, which weights issuers on the basis of GDP. This raises a bunch of practical questions: what to do about countries, like China, with capital controls and how to choose between local-currency and dollar-denominated debt? The simplicity of a value-weighted index has to be replaced by a set of complex rules. Global regions are weighted by the size of their GDP. Then securities are selected based on factors such as maturity and liquidity.
But PIMCO claims this more complex system produces significant advantages. First, the portfolio is much less exposed to the global deadbeats with most debt. An investor who tracks an index weighted by market value ends up with 29% of his portfolio in countries where the debt-to-GDP ratio is more than 120%, and only 1% in countries where the ratio is under 40%. In contrast, the PIMCO index has just 9% of its weighting in the high-debt countries and 16% in the least-indebted nations.
Second, it has a greater focus on emerging markets. This produces a higher yield than a value-weighted index (3.47% versus 2.04% for global government-bond indices). Third, Ramin Toloui, the economist who helped devise the index, says the benchmark has a mild countercyclical element. Bond prices tend to fall when GDP growth is strong and rise when GDP is sluggish or falling, so the weighting approach buys low and sells high.
PIMCO launched one GDP-weighted index in January last year and another two were due to be unveiled on July 8th. They enter a world which is increasingly crowded with financial benchmarks.
Indices began as a way of gauging the general health of the stockmarket. The original Dow Jones Industrial Average had just 12 stocks and was weighted by the share prices of the component companies. Only later were indices used to measure the performance of professional fund managers. This development had one benign side-effect (the creation of low-cost tracking funds) and one malign impact: a change in investor behaviour. Fund managers started to focus on “tracking error” rather than buying the best stocks. The definition of risk changed from losing money to underperforming the benchmark.
In the dotcom boom old-fashioned companies that actually made profits and paid dividends fell out of benchmarks to be replaced by loss-making internet stocks. Those tracking the index were forced to buy some of these stocks at the top of the boom in March 2000.
Partly to avoid such problems Robert Arnott of Research Affiliates has developed fundamental indices which weight companies by accounting measures such as sales, cashflow and dividends. In the ten years to end-May, a “fundamentally weighted” American portfolio returned 4.8% a year, against an annual loss of 0.8% from the S&P500.
In a recent paper*, Mr Arnott and others showed that similar fundamental factors, when applied to corporate bonds, produced a portfolio that outperformed a value-weighted index over the period January 1997 to December 2009. The outperformance is probably down to the “value” bias of the fundamental approach—buying stocks (or bonds) that look cheap on traditional valuation criteria and underweighting the most expensive. The advantage is that weightings cannot be affected by investor behaviour. A company’s dividends and cashflow remain the same whether the business is in or out of fashion.
The disadvantage is that costs are higher than with the most competitive tracking funds. For example, the total expense ratio of PIMCO’s GDP-weighted fund is 0.7%, compared with 0.22% for Vanguard’s total bond-market index fund. And the risk for investors is that the past outperformance of this strategy is not repeated. It is much easier to beat the market through the rear-view mirror.
* “Valuation-Indifferent Weighting for Bonds” by Robert Arnott, Jason Hsu, Feifei Li and Shane Shepherd. Journal of Portfolio Management, Spring 2010
0 comments:
Publicar un comentario