Without the right analytical tools, asset owners are forced to make decisions blindly with respect to underperforming managers
Pension funds face a daunting challenge when selecting managers to run mandates. The process of screening and choosing the best manager for the job is made exponentially harder by the lack of meaningful data available to allocators, leaving them ill-equipped to make decisions that could have major ramifications for the scheme and its members.
Today, asset owners conducting due diligence on managers are largely forced to rely on limited information like past performance, investment process specifics, and some form of claimed competitive edge. It is thin gruel for an investment committee seeking to make the most informed decision possible.
The most glaring problem is that past performance, the starting point for most searches, tells asset owners almost nothing about how skilled an asset manager actually is ‒ a major concern when you consider that track records are often driven more by luck than ability. Nor do asset owners have at their disposal the means to examine claims of competitive edges. Indeed, given the dearth of useful analytics, they often have no choice but to simply decide if they believe a manager’s assertions.
Other industries take a much more progressive approach to ensuring their decision-makers are appropriately resourced. Sports franchises, for instance, are typically better placed to identify and award contracts to players because they can leverage modern analytical tools to inform their decisions. The investment industry, in contrast, forces asset owners analyzing managers to use outdated models designed decades ago, when portfolios were managed in a very different way.
Measuring What Doesn’t Matter
When the first performance attribution model was designed in 1986, it perfectly reflected the way money was then managed. Typically, portfolios held between 140 and 180 positions and the investment process focused on deciding which sec- tors or regions to overweight, then choosing the best stocks within them. That simply does not happen now.
Today, the average equity portfolio holds less than 40 stocks and managers choose companies purely because they want to own them, not because they fit within an allocation framework. Consequently, using an attribution process which captures the allocation and the selection effects when these decisions are no longer being taken renders the model redundant. Asset owners are accordingly not able to measure what matters in mod- ern investment portfolios, which means they lack the help they need to identify and select managers with genuine investment skill. To an extent, they are obliged to make educated guesses, like a radiologist forced to make a diagnosis without the help of an x-ray.
The problem is not, of course, only limited to searches: it also means asset owners lack critical information about their existing managers. That might be less problematic when managers are per- forming well, but it becomes a pressing concern when they underperform. How long do you give an underperforming manager before terminating them? That is not an uncomplicated decision to make when you do not know why they are underperforming. Optically, it may appear obvious: a factor rotation, say, that works against their style. But managers with similar styles can perform very differently in the same circumstances. Working out the reason why can only be done by looking at what really drives returns.
Finding Skill
It is an unfortunate fact that the average portfolio manager underperforms their benchmark after fees over most periods. Happily, though, not all managers are average. There is a long tail of those who demonstrate significant skill. According to our recent analysis1 of 752 institutional equity portfolios with track records of at least three years, 84 per cent outperformed their respective benchmarks. On average, they did so by 397 basis points per annum: a margin that demonstrates that elite managers, like elite athletes, are simply much better at what they do than anyone else.
This knowledge alone, however, will not help asset owners select the right manager for a mandate. To do that they need to look at how and where returns were generated. Once they understand those critical factors, they know where to devote most of their attention when investigating an investment process.
Generating Alpha
In today’s bottom-up, highly concentrated portfolios, asset managers make only a few core decisions that have the potential to generate alpha.
By far the most important of these is research – the process of investigating stock opportunities and adding them to the portfolio. So dominant is this that our analysis shows that 88 per cent of portfolios have demonstrable research skills, generating 383 bps (basis points) per annum of alpha on average. Only 12 per cent have poor research skills, generating negative research alpha of -139 bps. Looking at the distribution of results the message becomes even clearer: stock picking is the primary area in which managers demonstrate skill.
The second most important decision managers make in managing concentrated portfolios is position sizing – how much of a stock to own. We find that this makes an insignificant contribution to overall returns. Only 46 per cent of port- folios generated alpha from sizing and the amount is less than 100 bps on average. More than half of managers have poor sizing skills, albeit the loss of alpha is small (-93 bps) compared to the alpha generated from stock picking.
Given the emphasis asset managers typically place on their approach to portfolio construction and conviction levels, this may come as a surprise. But it is clear asset managers simply have not found a way to add alpha through a discipline often assumed to be critical to portfolio management. Interestingly, the reasons for its ineffectiveness vary across managers; there is no common denominator we can identify for its failure to have more than a marginal influence on returns.
There is, however, a much clearer and more universal explanation as to why stock picking so dominates alpha: asset managers have invested significant capital in building up the teams and systems that enable them to find and select the best ideas. Its centrality implies managers have long realized the importance of getting stock selection right. The fact that little else in their investment processes adds value – be it risk budgeting, portfolio construction, or conviction levels – suggests they overestimate their ability to make those decisions to any meaningful effect.
Armed with the right information, asset owners can see whether a track record has been generated by luck or skill. That is crucial because awarding a mandate to a manager whose past performance owes much to good fortune – through, for instance, simply having not owned stocks in sectors that performed poorly in a given timeframe for unforeseeable reasons – can be a costly mistake to make.
As elite fund managers primarily add value though stock picking, asset owners ought to spend an amount of time and effort understanding the efficacy of the research process commensurate to its importance when carrying out their due diligence on prospective managers. How are stock ideas generated, how are they scrutinized, and how do stocks make their way into the portfolio? Who are the key people involved? Given there is no prospect of the other elements of the investment process offsetting the loss of alpha if the research element starts to misfire, asset owners must have a thorough understanding of the answers to these questions.
The same principle, of course, applies to manager evaluation, not least when asset managers underperform. It happens: few managers, no matter how talented, can maintain their edge over extended periods of time. The analysis shows that, if this should occur, there is a strong chance that something is going awry in the research process. Knowing where to embark on an investigation should save an investment committee time and effort, and potentially help accelerate the decision on whether to persist with the manager or terminate them. If a manager is still adding value where it counts, then forbearance might be warranted. But there is no point hanging onto a manager who is no longer demonstrating skill, as much as no-one desires the disruption and costs associated with turnover.
Rick Di Mascio is the CEO and Founder of Inalytics.
