As an inveterate collector of statistical oddities, I was struck by some intriguing data on rebalancing that came out at a recent investment symposium organised by the US fund management group Vanguard.
Rebalancing is one of the aspects of portfolio management that has strong theoretical justification, but is not, so I have always suspected, as easy to implement – or to justify to clients – as it should be.
In a recently reissued research paper, three Vanguard analysts go back into history to discover how different strategies on rebalancing might, with the benefit of hindsight, have worked out.
They focus on the two key issues of frequency (how often to rebalance) and triggers (what percentage deviation from a target asset allocation should lead to a rebalancing exercise).
The analysis is based, like so many such studies, on US stock and bond market returns from 1926 to the present.
Taking an old-fashioned 60/40 equity-bond portfolio as the benchmark, the data underlines how far portfolios are capable of deviating from their target allocation without rebalancing.
In the extreme case of a portfolio that was never rebalanced at all, it would have crept up by the end of the period to a 97 per cent allocation to equities. The annualised return would have been slightly higher, at 9.1 per cent a year versus 8.4 per cent for a portfolio that was rebalanced monthly to its 60/40 target, but the volatility would have been higher too. That is no surprise. An investor only interested in return, and not at all in risk, would always have something close to a 100 per cent equity portfolio if historical US data is taken as the source. The whole point of rebalancing is to reduce the risk of a portfolio, not to maximise its returns.
There are, however, occasions when rebalancing can usefully help investors to make the most of the returns that equities have to offer, and these typically occur when markets go through severe bear phases.
The Vanguard study confirms that a formal rebalancing discipline would have proved beneficial in forcing investors to go back into equities during all the worst equity market falls of the past century.
It turns out there have only been seven occasions since 1926 when the manager of a portfolio with a 5 per cent trigger for rebalancing would have needed to take decisive corrective action to go back into the equity market following a market fall. Three of these occasions took place in the 1930s (1930, 1931 and 1937) and one in the 1970s, while the other three have all come round in the 2000s (2000, 2002 and 2008).
By definition these occasions followed periods during which equity market returns were very poor, typically worse than minus 20 per cent on a trailing 12 month basis. These are precisely the conditions in which investors are reluctant to commit new money to equities and need to be forced back into the market against their natural instincts. To have to do it twice in two consecutive years, as the study suggests was required in 1930 and 1931, needs greater fortitude, not to mention a high degree of business risk.
We don’t of course yet know whether those who were compulsorily rebalanced back into equities in 2002 and 2008 will reap the same superior rewards over the subsequent 10 and 15 years as historical experience suggests they should. All that can be said is that nothing in the market reaction since the onset of the Lehman Brothers crisis suggests that the pattern of a further period of superior equity returns over time is yet at risk.
In some ways, however, the most surprising finding of the research is how little difference changing the frequency or the trigger for rebalancing seems to make to portfolio characteristics. The annualised returns and volatility of a 60/40 portfolio turn out to have been similar historically whether it is rebalanced monthly, quarterly or annually.
Similarly the risk and return figures do not change much if the threshold for portfolio changes is reduced to a 1 per cent deviation from target, or increased to 10 per cent.
The one big change, as Vanguard is naturally only too happy to emphasise, is that the burden of management and trading costs arising from rebalancing rises sharply the more frequently it occurs. Its conclusion is that six monthly or annual rebalancing strikes the best balance between portfolio efficiency and cost minimisation, and that anything more frequent adds little additional value.
However, Vanguard also recognises two additional restraints on the potential benefits. One is that rebalancing is only practical when portfolios are composed of assets that remain liquid and tradeable during times of market stress. The second is a throwaway comment that investors’ risk tolerances are in practice too rough and ready to allow them to appreciate the good that is being done in their name.
Doing the right thing, in other words, can be a thankless task.
Jonathan Davis 