This week's article examines the long-term performance of numerous asset allocation strategies. The article also explores the devastating effect of high fees on all long-term strategies.
Investment strategy study confirms importance of cost control
Asset allocation: why bother? Investors have long grasped that diversification between different sectors and asset classes is far more important to their ultimate returns than stockpicking — a choice of individual stocks. Many have offered new and varied asset allocation plans, involving extra asset classes to improve the balance of risk and return.
But new research in a book by Cambria Asset Management’s Mebane Faber shows clearly that the gap between the best and worst asset allocation schemes is narrower than the gap between the highest and lowest fees.
In other words, the precise asset allocation model you use is less important than keeping control of fees.
The research involves taking some of the most popular asset allocation models that have been proposed, and testing how they would have performed from 1973 to 2013.
It becomes clear that a detailed and disciplined spread of assets can reduce the sharpness of drawdowns over time, and so give an investor greater confidence that they will have a decent sum of money when they need it. That is the idea. Over many periods, these allocations would have beaten stocks — but the point was more to provide a smoother ride.
All the strategies were modelled using basic and popular benchmarks. The classic allocation of 60 per cent stocks and 40 per cent bonds, all in the US, would have averaged a real return of 5.13 per cent, with a maximum drawdown of 39.4 per cent. Replacing this with global indices raised the return to 5.54 per cent, while reducing slightly the worst drawdown.
Now Mr. Faber started testing models that also involve real assets, and that stipulate different categories of stocks and bonds.
Most successful was the “El-Erian portfolio”, modelled on an allocation suggested by the regular Financial Times columnist Mohamed El-Erian in one of his books.
More aggressive than some others, this had 51 per cent in various classes of stocks, 17 per cent in bonds, and the remainder distributed between index-linked bonds, commodities and real estate.
This portfolio managed a lofty average real return of 5.96 per cent, although on the flipside its worst drawdown was significantly worse than most others, at 46.5 per cent.
Weakest was the “permanent portfolio”, advocated by the famous investor Harry Browne.
This is hugely conservative, advocating a quarter each for stocks, bonds, treasury bills and gold. Its return was only 4.12 per cent, but those returns were consistent, and its worst fall was 23.6 per cent.
An array of other strategies all fell between these poles. The “all weather” portfolio is modelled on the approach made famous by the Bridgewater Associates hedge fund, and is meant to have separate allocations to prosper whether inflation rises or falls, and whether growth rises or falls, compared to expectations.
This has only 30 per cent in stocks, 15 per cent in 10-year Treasuries, 40 per cent in 30-year Treasuries, and 16 per cent in commodities — half of which is in gold. This managed a return of 5.04 per cent, while reducing the worst drawdown to 28.7 per cent.
The “Rob Arnott portfolio,” named for the founder of Research Affiliates and prolific academic financial researcher, offers another twist.
It has 20 per cent stocks, 20 per cent corporate bonds, 30 per cent government bonds, 10 per cent inflation-linked bonds, and 10 per cent each in commodities and real estate. This had a return of 5.04 per cent, and a maximum drawdown of 26 per cent.
And the “Marc Faber portfolio”, named for the famously bearish commentator and blogger, put 25 per cent each in various stocks, bonds, gold and real estate.
Although apparently very conservative, the results were similar to many of the others — a 5.26 per cent return with a worst drawdown of 28.7 per cent.
The remarkable fact jumping out of this exercise is that all these apparently diverse strategies ended up delivering startlingly clustered results by the end of 30 years.
Excluding the highly conservative permanent portfolio (which rewarded its investors with the smoothest ride), all of the others were clustered within 1 percentage point of each other per year.
All of these models succeeded in providing a smooth ride that virtually matched the final result achieved much more bumpily by investing solely in the S&P 500 (which would have returned 5.71 per cent per year over the period, but with a huge post-Lehman drawdown of 54 per cent). What matters about your asset allocation is less the precise balance between different asset classes, it would appear, and more the discipline to keep to it consistently, rebalancing each year.
More importantly, fees plainly matter more than asset allocation — providing that asset allocation is within the spectrum of sensible ideas surveyed by Mr Faber. Imposing a fee of 2.25 per cent on the winning “El-Erian portfolio” — not unreasonable if advisers charge a consulting fee and use relatively expensive mutual funds, would render it worse than the losing “permanent portfolio”, assuming it was implemented without paying for advice and using the cheapest available exchange traded funds.
So this welter of research ended up delivering a conclusion that surprises few of us. In the long term, nothing is more important than keeping fees under control.
Global Asset Allocation: A Survey of the World’s Top Asset Allocation Strategies by Mebane Faber. $2.99.
