Investors tend to use passive investment to allocate to certain asset classes, rather than as a total solution. The problem is that while there are some true benchmarks such as the MSCI World Index for stocks and the Barclays Aggregate Index for bonds, there is no such undisputed benchmark in the multi-asset space.
So, what would a true passive benchmark for multi-asset investors be like, and would it have the potential to achieve true benchmark status in the future? In Historical Returns of the Market Portfolio, former Robeco chief strategist Ronald Doeswijk, Robeco researcher Laurens Swinkels and Rabobank quant analyst Trevin Lam simulated the effects of one.
Their starting point was to invent a Global Market Portfolio (GMP) that aggregated all investors worldwide, with weights that reflect the constitution of the average portfolio today. This can be seen in the chart below.
Then they used a database that gives a historical record of returns going back to 1960, based on the work of Roger Ibbotson and Laurence Siegel, who were the first researchers to carry out a rigorous study on a theoretical GMP. During the 56-year sample period (1959-2015), this portfolio delivered a compounded annual return of 8.4%.
For the four main asset classes, equities realized the highest compounded annual return of 9.5%, followed by real estate’s 9.2%, non-government bonds’ (investment grade and high yield) 7.4% and government bonds’ 7.0%. When adjusted for the inflation that was prevalent from 1960 to 1979, the average annual real return of the market portfolio was 2.8%, while in the disinflationary period from 1980 to 2015, the GMP had an average return of 6.2%.
However, following such a passive portfolio may not be optimal, as certain markets can move away from what can be seen as equilibrium, leading to false valuations over time. This was seen during the dotcom bubble, when technology stocks traded at hefty multiples that later turned out to be wrong. In terms of asset supply, stocks and bonds are often issued when market conditions are good, and these timing effects are hidden when calculating true returns.
To get around these problems, the authors constructed three portfolios with fixed weights that are annually rebalanced; as their allocation never changes, it makes them easier to compare over time. The first was equally weighted in the four main asset classes. For the second, the size of each asset class’s market capitalization was taken into account, creating a portfolio with a 40% weight in equities, 30% in government bonds, 20% in non-government bonds and 10% in real estate. The third alternative portfolio contained 50% equities and 50% government bonds.
And the result? All three generated a higher average compounded return than the market, amounting to 31% for the equally weighted portfolio, 25% for the rank weighted portfolio and 15% for the 50/50 portfolio.
In other words, the GMP is probably not the optimal portfolio. This, in conjunction with the fact that the multi-asset market approach varies significantly when it comes to risk profiles, seems to indicate that there is limited scope for a successful future for the true passive approach, no matter how popular investing this way becomes.
This article is a summary of one of the five special topics in our new five-year outlook.