Dickon Reid examines the findings of a recent study promoting the cause of ‘active
Consider the following one-off bet. I toss an unbiased coin. Heads you pay me £1,000, tails I pay you £1,500. Game over. In technical terms, the game’s value to you is £250 and statisticians say to make it fair you should give me £250 up front. Some would take the bet, for others the prospect of losing a thousand isn’t worth the fifteen hundred.
Now consider this. I toss the coin again one hundred times. For every head you give me £10, for every tail I give you £15. In statspeak the expected returns for the two games are identical. In other words, if we play the two repeatedly, you will come out better by an average £250, irrespective.
Yet psychologists have shown the second is more popular than the first. (If you’re unconvinced, up the stakes in the first bet, the number of takers will soon wane). With a stream of small bets, you’re more likely to realise your £250 and avoid a big loss.
It’s this simple premise that Don Ezra of Frank Russell Company uses in a recent article to promote the theory of active multi-managers.* Instead of coins, he applies the theory to investment bets (for argument’s sake, deviations from an index). Start from what he calls tactical asset allocation (TAA) which in its simplest form is whether to be in or out of a market. This is the equivalent of the £1,000/£1,500 bet. It’s one-off, if you’re right the rewards are big; if you’re wrong, vice versa.
Market troughs or peaks are the best time to make these bets, but the two are often years apart and you therefore have to hold positions for a long time, perhaps a few years. Anyway, over such a period, it’s hard to choose if the manager is skilful or lucky. As Ezra says, these bets are considerable and it’s tough to be convinced the long-term odds are in the manager’s favour; hence few investors find them suitable.
The second asset allocation theory deals with positions, for or against, in different countries or in different sectors. The manager will make a few bets at a time but the potential gains/losses are smaller than the TAA. As these bets are less extreme than the TAAs it’s easier to differentiate between luck and skill but as Ezra says, its largely a matter of faith you’ve chosen a competent manager.
Now to the micro level. This includes single-security selections and here a manager typically has hundreds open at a time. Potential returns and losses tend to be smaller and holding periods tend to be measured in months opposed to years. “Because market conditions within an asset class tend to vary more than conditions across countries and sectors, one gains exposure to varied conditions more quickly. One then feels more confident in making judgements about skill, because one sees a manager’s process in action over more varied conditions,” says Ezra.
If you’ve chosen a skilled manager (something you’ll soon know), these allocations are the equivalent of the £10/£15 bets and you’re more likely to realise the expected return on your assets than by making a single placement. But there are a few preconditions to the theory – namely avoiding TAA bets and identifying superior managers – something Ezra admits is difficult.
He elaborates on Frank Russell’s selection which is thorough and hence complicated. Given the complexity Ezra says if you get the selection right on average 60% of the time, you will earn a lot of money. He adds it’s extremely humbling to discover the result of such effort is a hypothetical 40% rate of failure. If you manage to choose a hopeless single manager the outcome is dire and this is another compelling argument for selecting multiple managers to make the now proven multiple security selections.
“Even with the best research, odds tilted only 60/40 in you favour in the short term tend to be uncomfortable except for the few with exceptional fortitude. The odds on success become much better if the element of luck is reduced by selecting multi managers (as long as they are carefully assembled to neutralise only the unwanted bets.) Once again, this converts a £1,000/£1,500 proposition into a portfolio of smaller propositions with the same expected outcome; but the portfolio gives the outcome greater certainty,” he says.
And there’s another bonus. According to psychologists, multi managers are more likely to avert the hanging on to bad investments, a problem Ezra describes as so pernicious in investing. “If it’s a single decision, we look at it in a ‘right or wrong’ decision. If it’s one of several, our self-esteem is easily satisfied by the recognition that most of our decisions are right, and it’s no longer necessary to prove stubbornly that a particular one has got to come out right.”
The paper finally looks at three different approaches to active management and their relative attributes. First up is a single balanced management structure. According to Ezra, such a manager makes tactical departures from an asset allocation benchmark and makes sector, country and security bets. The result is excess return tracking errors relative to passive investing. The success of such managers relies on luck. Who hasn’t experienced choosing a balanced manager on the basis of past success only to encounter failure instead? “They trustees have not learnt the reliability lesson: exposure to bets of we have termed unreliable increases the impact of luck, and reduces the impact of skill in beating benchmarks,” he says.
Secondly, there’s passive core management with specialist satellites. Each asset class has a passive core tracking a benchmark and a specialist manager managing the rest. With all else equal, the passive core will cut tracking error and reduce the payoff for success. And, says Ezra, the structure still ignores the reliability dimension and unreliable bets still have an impact. There is no way to control the impact of manager making country and sector bets, for example. “Success still retains a large element of luck … apart from the selection of specialists, one is still playing the same game just with the positive and negative outcomes scaled down proportionately,” says Ezra.
So, on to the grandly-named multi-asset, multi-style, multi-manager approach. Each asset class has a specialist and the managers make predominantly security selection bets thereby avoiding TAA, asset classes being re-balanced to their policy weights periodically. “By eliminating or at least reducing the impact of unreliable bets, it successfully reduces tracking error. Success in adding value is never assured; but now the element of luck is reduced, and success depends more on the skill available in manager research and selection. Although tracking error is reduced, there is no proportionate sacrifice of added value as 100% of the asset are actively managed,” says Ezra.
Ezra then runs simulations on hypothetical portfolios as evidence supporting the multi manager theory. He stresses the evidence is not offered as proof yet it suggests there are benefits from selecting managers and that you can reduce risk by focusing on security selections. The theory looks good but are you convinced? Some managers would no doubt fancy their chances on the equivalent £1,500 bet.
* ‘Active versus passive: A New Perspective’ by Don Ezra, director of strategic advice at Frank Russell Company