Skill is at the heart of all active investment processes, and yet identifying it up to now has largely been an art rather than a science. But this is changing. Investment skill is capable of being identified in a quantitative and measurable manner, and although the techniques may be in their relative infancy the results speak for themselves. In the past the first place investors turned to when trying to identify skill was the track record, but in reality it simply tell us how funds performed and whether they had met their targets or benchmarks. They say next to nothing about the investment process that produced the numbers or the manager's strengths and weaknesses.
No wonder track records are notoriously volatile and are now subject to the inevitable risk warnings that they are no guide to the future. They are right, they are no guide to the future.
Put simply track records are a quantitative measure of how successful a manager was over a particular period of time, but says little about the fundamental issue of skill. Or to use a sporting analogy, they tell us whether they won or lost the game, but nothing about how it was played.
Identifying how managers played the game in the past was difficult, and it didn't help that they like to portrait skill as something intangible or even mercurial.
Our firm analyses every decision a manager has taken to see what skills they possess and how they contributed to the end result - alpha. It works in sport; it is now time to apply these techniques to the world of investment.
It is our view that the best and most talented managers can be identified. They tend to have great timing skills and are able to generate alpha from a wide range of investment opportunities. On the other hand poor managers tend to fall foul of the old adage of buying late, selling too early and simply following their favourite stocks. Traits which can destroy value, and turn the best track records to dust.
The fact that quantitative measures can identify the best, and for that matter the worst fund managers, doesn't mean that there is no role for traditional approaches such as a qualitative assessment of a manager. On the contrary, in our view these two disciplines complement each other and our clients have found that they have the best of both worlds. In practice traditional approaches find out what managers say they do, whereas our quantitative assessment establishes what they actually do. Clients benefit from having both.
So what does quantitative analysis hope to achieve?
Using this framework of the best, a quantitative assessment of a managers' decisions is designed to find and identify whether they have the requisite timing skills and whether there is genuine breath to their process or whether alpha is being generated from a narrow range of opportunities. There are two specific areas of manager analysis: identifying timing skills when buying and selling stocks; conviction, that is how efficiently and consistently a manager's view leads to alpha.
Our research, conducted through the analysis of equity portfolios, has produced some very interesting observations on the typical behaviour of fund managers. Starting with our findings on timing, the evidence clearly suggests that the typical manager is a much better buyer than seller - no great shock to those who have spent any time with managers and therefore know how optimistic they are, in general.
What we have found through our consulting work is that many major European managers spend a great deal more time and effort on the buy side of switches than the sell. Sells are often made purely because the fund manager needs to raise cash to buy a stock they felt should be added to the portfolio. These represent classic examples of behavioural tendencies where certain information is under researched or glossed over, leading to a stock being sold in the face of overwhelming contradictory evidence.
I think the importance of this evidence cannot be overstated. Even the most talented people can easily fall foul of their own personalities when making important investment decisions.
Portfolio construction has come on leaps and bounds over the past 20 years with the advent of risk modelling and clearer client guidelines. But even in this much improved form, we have found that the construction process is not immune from common personality biases. These can put performance objectives at risk or simply waste opportunities to generate alpha.
Our analysis of 70 major equity portfolios found that most fund managers demonstrate two key skills: the ability to pick stocks that outperform; and the ability to allocate them efficiently across their portfolios. But the research suggests that managers struggle to identify stocks that are destined to underperform and consequently lose performance through their heavily underweight positions or decided against owning at all (the assumption being they will underperform over a certain period) then subsequently outperform.
This was true for nearly 75% of the portfolios analysed, and the negative impact was running into several hundreds of basis points. In fact only 5% of portfolios were losing performance through heavily overweight or ‘high conviction' positions.
So, while the overweights usually added to performance, the underweights offset these gains and drag performance down. Over a 12 month period we found that the average contribution from heavily positions was positive every single month, whilst the average contribution from underweight positions was negative in 10 months out of 12, leading to massive underperformance. Again, this is despite the fact that fund management houses have processes for managing these "negative bets".
This tendency tends to arise because of what is known as the ‘endowment effect', in essence a recognition that people value stocks more when they own them. When fund managers overweight a stock they spend a lot of time on analysis, scrutinising every announcement, any change in the share price, and any piece of information which they find out about the company. Conversely, when a poor performing manager underweights a particular stock the same level of analytical attention is not applied.
So why does it happen?
Managers have not invested as heavily in these stocks and therefore do not view them as being as important as their overweight stocks. We have found that fund managers have a tendency to become over confident in their ability to identify stocks that they believe will underperform. Once they have made a decision, it is very difficult to get managers to change their minds.
What has been most remarkable about our analysis of manager personality is that all these same tendencies were replicated through an analysis of mathematically programmed quant funds investing using pre-set formulae that showed exactly the same common characteristics of buying late, selling early and concentrating on their favourite stocks in any given portfolio.
By studying their personality habits, managers can cut out the biases that have a negative impact, thereby boosting their overall performance, and by identifying these particular traits, chief investment officers and heads of fund desks can ‘manage their managers' more effectively and provide an assessment of performance. The difference between poor performing managers and highly successful managers could be a series of finite decisions that could be pinpointed by objective analysis with hard data for managers to act upon and change their methodology.
Similarly, pension funds and consultants can use this data to make more realistic assessments of which managers have strengths in the required areas and make manager selection more objective through a formal qualitative and quantitative process.
Rick Di Mascio is chief executive of Inalytics, based in Croydon
No comments yet