The Importance of Context.
It’s a simple rule of wisdom that is surprisingly easy to forget or ignore. Its value in bringing long-term solutions to problems is unquestioned. A few examples in no particular order:
– Greek crisis: Tree: debt rescheduling conditions; Forest: do we want a pan-European “state”?
– Beppe Grillo’s advent: Tree: successful politician; Forest: Italy’s political system needs fixing.
– Iran/contra affair: Tree: a US president gone mad; Forest: US international policy must be long-term.
– Financial crisis: Tree: “fix” the banks and improve regulation; Forest: human nature causes crash-prone imbalances.
Recently I spent time with a few specialists in the field of manager selection, whose task is to select the best active portfolio managers for clients. They are usually referred to as “managers-of-managers” (MOMs in short). The intent was to bring myself up-to-date on the science behind this highly-valued activity.
Much importance is attached to the investment process the active manager employs and to solid due-diligence on the part of the MOMs. Clients for the most part don’t pay enough attention to these two elements, preferring to focus on the performance record instead.
Agreement exists on the fact a good active manager needs time to demonstrate skill (and can be off the benchmark for several consecutive years). Yet some MOMs will claim the ability to profitably implement rotational portfolio strategies aimed at selecting the right manager at the right time. They accomplish this notoriously difficult task by identifying in advance market “environments” or “themes” which statistically favor certain managers’ styles. This is akin to hiring and switching around the ideal human models by predicting what fashion trends are coming up next season. If it sounds crazy it’s because it is.
Market environments in these situations are characterized by factors or variables that in the past have had some association with them. While factors can essentially be anything with enough historical data, it is logically useful and reassuring to pick the ones with at least a feeble theoretical connection with the environments. Once you identify the relevant factors, you build a mathematical model that takes as input the factors and yields a forecast environment as output.
If you are not asleep yet, you might think: “This is all very fine in theory but it does not obviate the problem of accurately predicting something: either managers’ performance or the factors.” And you would be absolutely right; in fact you would appear a lot more alert than a good bunch of professional MOMs out there. Getting back to our tree/forest analogy:
– Manager selection: Tree: select the right managers for the right environment; Forest: “It’s tough to make predictions, especially about the future.” (Yogi Berra).
Final example: Tree: a common factor used in some models is “Cash versus S&P500 index”, which I pragmatically translate as “absolute performance of the S&P500;” Forest: really: if you know where the market is going next, why bother to worry about active managers at all?
Photo source: http://forestpolicypub.com/2011/03/10/not-seeing-the-forest-for-the-trees/.