Most actions we take, whether we realise it or not, have unintended consequences. The more scientific among you might recall the theory behind the butterfly effect and how small decisions can magnify into large and unplanned consequences.
The less scientific might make the unfortunate decision to watch the poor and declining series of Butterfly Effect films to see an erroneous interpretation of the concept. But since this is not a film review column, I will leave my critiques aside and instead focus on an important concept when thinking about building multi-manager portfolios.
We all know, when combining fund managers in a given asset class or region, that we are looking for diversification, and we typically measure this through style or factor exposures, as well as on a more qualitative basis by their philosophy and process, which gives a given portfolio manager his or her typical ‘type’ of stock. For sure, there are managers with indefinable and/or varying ‘styles’ – the ‘rotator’ – but, by and large, it is possible to categorise fund managers into ‘buckets’ and by combining managers from different camps we are able to build a diversified portfolio.
So much for the theory, what about the practice? Well, when we build a standard multi-manager portfolio the logic is that we combine the best ideas of multiple managers into one overall portfolio, giving something that is better than the sum of its parts – the benefits of multiple experts’ stock picks, but with the lower risk created by diversification. But the problem, which is too easily overlooked, is that a manager’s portfolio is not simply their best ideas; it is instead a carefully (we hope) portfolio constructed to outperform its benchmark within risk limits. But in the vast majority of cases, there will be at least some regard for risk, and so some positions will not be real ‘alpha sources’ but will, in fact, be ‘risk reducers’.
In an individual portfolio, this is not a problem – it simply reduces portfolio risk. The problem comes when we try to ‘blend’ fund managers into a portfolio. Of course we gain exposure to the (hopefully great) ideas of portfolio managers. But we also gain exposure to these risk reducers in equal measure. Now, if each manager uses different risk reducers then this is not a problem – because inevitably, unless their portfolio construction skills are non-existent, they will have smaller positions in risk reducers than in their best ideas. But if, as could be the case, their risk reducers are the same stocks, whereas their alpha ideas are different stocks (the latter being why we picked them to ‘blend’ well), we end up with a multi-manager portfolio dominated by risk-reducing names and with only a tail of alpha generators.
How likely is this to actually happen? My experience is that it is more frequent than fund selectors would like. The problem is that, whenever you think of a sector, there is usually one ‘safe’, quality, large-cap name that springs to mind – in other words, the obvious risk reducer. For a long time in UK Financials this has been HSBC. In pharmaceuticals it has probably been GSK. And the result is, when building a multi-manager portfolio, all too often HSBC becomes the biggest position, not because any one of the managers particularly likes the stock, but because all of the managers, or a majority of them, have it as a mid-sized position with the aim of risk reduction. When added together, however, this position ‘behaves’ like an alpha generator – with its performance having a disproportionately large effect on the overall portfolio’s performance. And this happens despite the fact that none of the underlying managers intended it to be the case.
