Behavioural finance has taught us a lot about the suboptimal fashion in which investors (professional and private alike) arrive at decisions. It is 35 years since Daniel Kahneman and Amos Tversky first outlined their version of what was to become prospect theory, highlighting the high value investors accord to loss aversion relative to commensurate gains. Since then the field of behavioural analysis has expanded massively, most excitingly in recent years by aligning itself to the findings of neuroscience, which can track how different parts of the brain react to intellectual and emotional challenges.
Yet we are still a long way from fully understanding, let alone embracing, the practical implications of the findings of behavioural finance for the fund management business. The mathematical framework of expected utility, the capital asset pricing model and mean variance optimisation, for all its well-documented frailties, still remains the dominant intellectual strand in professional investor training. Fund analysis, though increasingly sophisticated, remains firmly rooted in backward-looking risk-adjusted return analysis.
However the influence of behavioural finance continues to advance, as Herman Brodie, founder of the Prospecta consultancy business, pointed out in a recent Halkin Society presentation on the potential applications of prospect theory. The most exciting recent developments, he points out, is that MRI scans and other tests by neuroscientists confirm loss aversion and other behavioural biases are indeed hard wired into the human brain, meaning that behavioural theories have a scientific foundation, unlike the assumptions behind standard financial theory.
Mr Brodie knows of what he speaks, having spent a number of years in the trading department of a global investment bank. He observed at first hand his own and management’s performance-detracting behavioural biases. It was loss aversion, he discovered, that explained both his own shortcomings as a trader and the reason, when he later switched to a job devising and managing algorithmic trading strategies instead, why the results were not that much better.
The firm’s management had an uncanny ability to increase or decrease their capital allocation to algorithmic trading at precisely the wrong moment. By cutting their exposure after poor months and increasing it after good ones, with predictably poor results, the managers in effect allowed their aversion to loss to override the supposed advantage they gained from having the trading executed by an emotionless automaton rather than a frail human being.
Prospect theory has wider implications. Mr Brodie points to a number of recent academic studies suggesting it provides a convincing contributory explanation for a number of well-documented anomalies in asset pricing, including the surprisingly persistent success of momentum as an investment strategy. We are all driven, it turns out, to cash in unrealised gains too quickly and avoid suffering losses at almost any cost. Momentum works in part because different investors – pension fund managers as much as retail investors – are working off different base reference costs.
Another interesting study, by Alexandros Kostakis, senior lecturer in Finance at Manchester Business School, builds on work that highlights the importance of “skewness” in determining asset prices and returns. The logic of prospect theory would lead you to expect that investors will favour assets whose return history has a positively skewed distribution (in other words, where the possibility of significant loss is lower than in a normal distribution) and avoid those that are negatively skewed. This in turn implies that the latter will command risk premiums that investors who are able (or incentivised) to tolerate can exploit systematically.
And this is indeed what behavioural finance analysts have been able to observe. Over long periods in both UK and US financial markets, assets with negatively skewed distributions have produced higher risk-adjusted returns than those with positive skews. Differences in skew help to explain, among other phenomena, why equities outperform bonds, why credit outperforms government and agency debt and why punters prefer buying lottery tickets to buying stocks.
In the UK equity market, according to Mr Kostakis, the risk premium associated with negative co-skewness (the degree to which adding a stock increases the negative skewness of a portfolio) was a statistically significant 2.1 per cent per annum in the period 1991 to 2005. He also found that the most successful UK equity fund managers in the period were the ones who proved most adept at capturing the risk premiums associated with negative co-skewness. (They did not need to be aware that this was what they were doing, as other strategies such as value investing can at different times be a proxy for capturing the same effect.)
The majority of UK equity managers still underperformed their benchmark after fees, however. Mr Kostakis argues that they would have underperformed by an even greater margin had they failed to reap the negative skewness risk premiums. Just as importantly, he says, investors who only look at funds’ performance on standard risk-adjusted measures, such as Sharpe ratios, will not realise that they are paying the fund’s managers to take more risk than they would be comfortable with, if the strategy’s risk was understood correctly.
The same reasoning helps to explain why absolute return and long-short hedge funds produced such disappointing results during the global financial crisis. Their inherent negative skewness came home to roost, despite the tin appearing to say something different. One of the many generalised insights that prospect theory provides is that successful professional investors will always be incentivised by standard performance measures to get paid for taking risks that investors in general would not want to take if those risks were correctly appreciated and interpreted. That is an important message and there are sure to be others as this fascinating field of study continues to develop.