Uncertainty: learning from the mistakes of others

In his third and final article on the subject of scheme funding, Con Keating looks at randomisation in decision-making.

In 1969, after two years of deliberations, a committee of financial experts published an extremely influential report to the Ford Foundation, “Managing Educational Endowments”. Much of it is timeless, such as the debates over unrealised profits, and bond versus equity exposure. In many ways, and remember this was long before investment consultants existed, let alone had gained primacy, it set out the template for endowment and pension fund management that is still applied today.“Trustees should not themselves attempt to manage their …portfolios.” and “In choosing between investment managers, primary emphasis should be placed on who the responsible individual will be, on what his record has been in the past, and on whether there is any reason to suppose that he will not be able to do as well in the future.” Fees are dismissed as being of “less importance”.Since this report was written our empirical knowledge and theoretical understanding has improved by leaps and bounds. The work of Adamou and Peters has shown the statistical limits to the usefulness of investment track records and further cautions abound. At the recent Chief Economists conference at the Centre for Central Banking Studies, Andrew Gelman offered this: “Lots of researchers think that if they’ve rejected a null hypothesis at a 5% level with some data, that they’ve proved the truth of their preferred alternative hypothesis. Statistically significant, so case closed, is the thinking.Then all concerns about measurements get swept aside: after all, who cares if the measurements are noisy, if you got significance? Such reasoning is wrong, wrong, wrong but lots of people don’t understand.” It is also worth noting the Ford report’s presumption of competence and skill in the professional, and their performance continuity or persistence; though this is doubtless a product of its time, when deference to authority may have been usual, today the null hypothesis would almost surely differ.Every six months, S&P Global publish their “Persistency Scoreboard” for mutual funds; the most recent states: “It is worth noting that only 0.78 per cent of large-cap funds and no mid-cap or small-cap funds managed to remain in the top quartile at the end of the five-year measurement period. This figure paints a negative picture regarding the lack of long-term persistence in mutual fund returns.Similarly, only 3.7% of large-cap funds, 5.79% of mid-cap funds, and 7.82% of small-cap funds maintained top-half performance over five consecutive 12-month periods. Random expectations would suggest a repeat rate of 6.25%.” My own recent, unpublished work on segregated mandates has similar, if depressing findings. Even performance measurement remains an issue, particularly where fund riskiness is considered, and as is demonstrated in Measurement Errors in Stock Markets, which finds, inter alia, that: “We show that the application of usual performance ratios … can lead to inappropriate findings and consequently wrong conclusions.”It is notable that Ontario recently issued a consultation on their “Solvency Funding Framework for Defined Benefit Pension Plans”, when Ontario has been something of a poster child for regulatory forbearance in this regard. This consultation does note, in passing, that: “The requirement for dual valuations may explain why DB pension plans have performed better in Canada than their counterparts in other countries.” While, in my opinion, underplaying the importance of their forbearance measures, this resonates well with the earlier proposal (part 1) for multiple perspectives on scheme sufficiency.With growing support for the view of markets as complex adaptive systems, a field of intensive research, the difficulties facing trustees are evidently awesome, and in their search for solutions many have discovered behavioural finance. Research psychologists have now documented hundreds, literally hundreds of replicable biases and departures from the standard rational agent of mainstream economics.Elsewhere than economics, such a profusion would usually be taken as strong evidence of a wrong model. Notwithstanding the ardent proselytising of such promoters as David Halpern, the CEO of the Behavioural Insights Team, with claims that these approaches have “enabled policymakers to boost savings; increase tax payments; encourage healthier choices; reduce energy consumption; boost educational attendance; reduce crime; and increase charitable giving”, the hope that this will enable regulators in “anticipating and addressing ‘animal spirits’ that drive bubbles or sentiment-driven slowdowns” seems far-fetched, wishful thinking.Though there are some interesting academic papers on some aspects, notably those of Benabou and Tirole, the simple fact is that there is no overarching, coherent body of theory that contains or explains most of these observations; there is a useful taxonomy, a dendogram due to Elster (shown below). One of the more intriguing recent suggestions, due to Philip Kotler of the Kellogg School at Northwestern(1). that: “if economists now have to study and explain how consumers actually make their choices, they need to turn to marketing” and further that “behavioural economics is… another word for marketing.”However, uniquely assigning particular behavioural findings within this tree is far from trivial. The question remains how should the trustee board go about its business of making decisions when faced by this objective and cognitive indeterminacy; when the safety of the herd, (over)confidence and much else commonly employed is questionable – not least, the continuing basic need or desire for a narrative. As Elster observes: “Human beings are… reason-seeking animals. They want to have reasons for what they do, and they create reasons where none exist.”It is clear that contrary to the assertions of the technologists, the answer does not lie with technological innovation, such as big data, though that will expand the boundaries of the feasible. The personalisation we see on the internet, book recommendations and advertising, arises from the use of predictive algorithms, but they preclude precisely the form of innovation we know as evolution. “A certain amount of randomness in our lives allows for new ideas or modes of thinking that would otherwise be missed. And, on a macro scale, it is necessary for life itself.”One potential answer lies in deliberate randomisation of the process. The Nobel laureate, John Harsanyi, has long suggested that one pragmatic approach to decisions in strategic games with indeterminate solutions is the flip of a coin. Steve Levitt, of Freakonomics fame has recently published a paper, “Heads or Tails”, which reports: “…subjects having difficulty making a decision flipped a coin to help determine their choice. For important decisions (e.g. quitting a job or ending a relationship), those who make a change (regardless of the outcome of the coin toss) report being substantially happier two months and six months later.” Before dismissing randomisation out of hand, it is as well to remember that we already use random selection in a crucial decision setting: juries.For trustees, Eleanor Roosevelt’s advice seems apposite: “Learn from the mistakes of others. You can’t live long enough to make them all yourself.” For fund managers, the definition of economics itself holds the key to a successful future: “Economics is about the efficient allocation of scarce resources, and about making resources less scarce”, and focus upon the second part of that would have fund managers delivering true added value, and make the trustee problems less demanding. For regulators and supervisors, the answer lies in recognising the value of cooperative, collaborative and collective arrangements rather than flogging the competition carthorse into its overdue grave.But that would a subject for yet another article.Con Keating is head of research at BrightonRock Group