Rewarding good behaviour
Behavioural finance is at the cutting edge of investment theory. It now rivals the ‘efficient market hypothesis’ as the main arena of generalised investment research. Understanding the behavioural biases that portfolio managers are prone to is crucial for a proper understanding of any out- or under-performance. This is something that has been increasingly recognised, as studies have shown that unconscious but avoidable biases to which most fund managers are prone have a huge negative effect (typically around one or two per cent per annum) on performance. Moreover, the entire market psychology can be affected by behavioural phenomena (see box, below left ).
Attribution and style analysis have an important role in portfolio monitoring. However, complementary analyses, following entirely separate approaches, can prove even more indispensable, illuminating how the investment process can be improved. Indeed, some of the behavioural traits that affect performance not only have a large effect but also persist to a far greater extent than any effect shown up in attributions. It is this combination of substantial impact and persistence that makes managing such traits so important for running portfolios going forward.
I shall look at insights from behavioural finance, backed up by new research into neuroscience, such as the disposition and endowment effects. These are crucial to advancing the investment process because they can highlight systematic effects that are not fully appreciated by the portfolio manager. The manager knows about his or her consciously chosen positions, and incidental style biases, but does not always have a clear awareness of behavioural effects. These effects can swamp the effects of skill, in large part because they persist, year after year.
One example of how the insights of this field can be harnessed is by looking at how stock selection success varies with the direction and extent of positions taken. Such analyses have been performed and there is a fascinating and fairly consistent tale. Most portfolio managers have skill; they can choose stocks to overweight, which on average subsequently outperform. Arguably, though, an even more useful insight is that those stocks which are heavily underweighted tend to subsequently outperform as well, leading to a loss of alpha. This has been related to the well established ‘endowment effect’. Simply put, this is the tendency to pay more attention to assets you own, than to assets you do not own. A stock held in the portfolio will consequently be keenly monitored. On the other hand, a large cap stock that is not owned is unlikely to be monitored to the same extent, although the manager intellectually knows the opportunity costs of that stock’s relative return on the relative performance are as great as the influence that an overweight holding has.
This is much more than just an interesting phenomenon; research on different managers has shown the extent to which the managers fall into this behavioural trap is the primary determinant of whether the portfolio manager outperforms or not. In fact, even most of the underperforming managers successfully chose stocks to materially overweight that would go on to subsequently outperform the index.
The problem is these were more than offset by the underperformance resulting from materially underweight stocks which also outperformed. Remarkably, this conclusion applies across all sizes of portfolio manager, all mainstream investment styles and all geographical regions. Accordingly, it is by increasing the focus on underweight positions that most managers have the greatest opportunity for improvement.
Another respect where behavioural finance can prove invaluable is by looking at trades. This can enable systematic trends in trading to be identified, and where appropriate, corrected. One of the pieces of folk wisdom passed down to fund managers is they should run their winners, and cut their losers. It is advice that can prove favourable or detrimental in individual instances. However, research has shown that if applied regularly, this advice is sound. Despite this advice, fund managers do the opposite and it hurts performance, to the tune of scores of basis points a year. This is explained as the ‘disposition effect’ (see box below).
Inalytics have used their database of transaction level data to demonstrate that this bias is evident in asset managers. It also showed, reassuringly, that investors generally show skill when buying (evidenced by the subsequent outperformance of most of the stocks they buy). However, they tend to lose money when selling. This is partly a result of the disposition effect, and could also relate to the observation that the typical fund manager tends to be an optimist and feels most comfortable when looking for the opportunities that will do well. This often means less research is carried out in to which stocks to sell, and often selling simply becomes a cash raising exercise. It is also noticeable that good sellers are a rare breed. They tend to be cynical, pessimistic and always looking for the hidden problem lurking behind the next corner. As a result they do not tend to ‘fit’ and not surprisingly there are few such individuals. Yet those with the ability to sell are highly valuable (even if not highly valued) members of any investment team.
To avoid any biases that may have arisen from the self selection of portfolios by our fund manager clients, we only used trades provided by our pension fund clients. This amounted to 45,000 individual trades and represents a broad spread in terms of industry, region, and benchmark. The data was taken from December 2003 to September 2006.
On examination of the database we found 57 per cent of stocks that had been sold had outperformed over the previous 12 months. This result certainly underlines the disposition effect, namely that fund managers tend to sell their winners.
Curiously, the researchers noticed the average performance turns negative in the month prior to the sale. This could suggest short term momentum is being used as a proxy for research and to oversimplify the sell decision-making process. Perhaps, having made such a profit, the managers are nervous it will erode away.
Having found evidence supporting the thesis of the disposition effect, we then examined what impact it has on performance. We found the stocks sold impacted performance negatively by a highly significant 94bps per annum. Consistent with the view that fund managers tend to be better buyers than sellers we found that the buys added value (47bps per annum.). However, it went only half way towards offsetting the losses from sales. The net effect of trading was to lose 47bps per annum.
Behavioural analyses are complementary to conventional attribution and, in some ways, more useful to the fund manager in that they provide insight into effects that are both more persistent and more actionable than typical attribution effects.
Analyses that separate out systematic behavioural biases from skill confirm the general hypothesis that managers have skill. However, this skill often fails to translate into superior performance as it is offset by a series of predictable, observable and avoidable shortcomings. In my view, neutralising those shortcomings will be one of the main challenges of active equity management over the next few years, and performance analysts who embrace these techniques will have a key role to play.
.jpg)
