CHASING THE HOLY GRAIL? SKILL, STARS, MARKETING AND BLOODY MATHS GEEKS?

As I sit here once again stuck in London City airport with a delayed return flight home my thoughts turn to a journalist query and I feel compelled to share my response in my CIO blog. I don’t often rant but then I can’t let my CEO Mr. Alexander have all the fun!

In my new Ambassador role I attended and spoke at the Transparency Task Force Symposium at the Unison Centre, an excellent event which included luminaries, MPs and the odd gobby Scot talking about FinTech and B-movie monsters! The room was shaped like a mini-UN chamber. The speaker line-up was varied, insightful and included one Mathematician who, having run lots of normal distribution data, concluded that ‘skill’ cannot be statistically observed without 600 years of Gaussian data (normal distribution or bell curve). He also quickly caveated he had no professional background in asset management. He made some nice comparisons to the rules relating to optics between two objects, the analogy being the benchmark beta and active ‘skill’. It was a well-paced delivery and populist for the mood in the room. The audience applauded and praised the presenter with an array of superlatives, as if he had cracked an unknown conundrum or the origin of black holes. How many actually understood the notation on the slides is debatable, but there were enough colourful probability density functions to guide even the most notation illiterate. I was later left surprised that my esteemed friend Con Keating retorted that the Maths was sound, closing ranks around geek assumptions over the vagaries of subjective qualitative analysis. My Transparency colleagues and passive proponents were wowed by the revelation, I sat rather less enamoured. As an ex-quant (not a very good one), but by no means a mathematician, I looked on cynically. A chance to ask a question presented itself. I put up my hand and pressed the annoy button.

‘Ahem, thanks for that, it was ‘interesting.’ But if I may note that 1) markets do NOT follow Gaussian patterns, 2) skill is not a constant and as a professional fund buyer, 3) we do not assume ‘skill’ when we observe a residual return. Any view on that?’ 

The speaker stuttered, stumbled, recoiled. Jovial cheers soon turned to gasps that someone may offer such a non-partisan view amid an otherwise anti-active manager agenda. The room fell quiet. Throughout the remaining day other TTF members showed their support for the Maths geek but it’s hard to know what they liked about it. Perhaps my question was too harsh, had I just beaten a puppy?

I’ll be blunt (I usually am). It was an interesting presentation for laypeople not involved in fund selection; it’s nothing I or any professional fund buyers hasn’t heard before. It was well presented and articulated but ultimately a naive academic view, one made by a charming observer who, by self-admission, had never analysed a fund manager up close. Any study that then starts with a hypothetical statistical data easily stating 12 times in duration of that of reliable observable data is absurd. That’s the problem with mathematics, all theory no real world. Funds are not observable assets; they are organic and run by fleshy machines we call humans. Ignoring the tiny number of fund managers still in charge of funds for more than 10 years, if the presentation at least challenges those who rely on medium-term performance to buy funds then I’m all for change.

Persistency of ratings based on 3-5 year data has been challenged academically for some time. However they make for good marketing. Likewise, GBM (Geometric Brownian Motion) approaches are largely contested in professional fund buying circles. Retail markets have been beholden to basic statistical theory for 50 years, enshrined by GIPS and Morningstar ‘Star’ ratings and equivalents. They are generally ignored by professional institutional fund buyers.

The reality is that I do not believe performance regression can ever be used in isolation to pick funds. It is used as a means to make fund screening more manageable. Neither is the benchmark (S&P500) reliable, alpha is not a constant and nor are asset markets Gaussian in nature, as 2008 demonstrated. Risk, too, is not constant or even volatility, beta, liquidity and the market factors are that will ultimately push and pull asset prices. Forget even the behavioural or heuristic aspects at play.

However, the answer is not then, by default, just hold the benchmark as subtly inferred. Buying passive at any index level will drive very different outcomes. Beta does not try to protect your capital, someone buying the S&P500 in 1999 or 2007 will be all too aware of this but any given manager can rotate to cash or Defensives when a Bear Market is in effect. Fidelity’s 2005 Magellan study ably demonstrated the importance of holding period and entry-exit point. Does that require market timing? Yes, but not forecasting. Running out of the way of a runaway train once it’s starting to roll is still better than staying on the tracks. Most long run studies don’t make any assumption for entry point like buying a value-biased manager who is 10% behind market and has a book at 20% discount.

Unlike amateur buyers; Professional fund buyers rarely assume ‘skill’ when they observe a residual return to a benchmark. They will also understand the asset make up of the benchmark as much as the fund, the difference in beta and ultimately the factors that drive the manager’s return. Professional fund buyers have superior tools, expertise and access to the fund manager by which to derive that view. There is a growing divergence between retail buyers (including advisers) and institutional buyers. There are a number of studies like Diane Del Guercio at University of Oregon that showed that US advisers were poor pickers of active funds. Poor fund buying I believe skews the wider active-passive debate just as much as the US biased very large studies like SPIVA.

Mutual Fund Performance and the Incentive to Generate Alpha Paper

‘To rationalize the well-known underperformance of the average actively managed mutual fund, we exploit the fact that retail funds in different market segments compete for different types of investors. Within the segment of funds marketed directly to retail investors, we show that flows chase risk-adjusted returns, and that funds respond by investing more in active management. Importantly, within this direct-sold segment, we find no evidence that actively managed funds underperform index funds. In contrast, we show that actively managed funds sold through brokers face a weaker incentive to generate alpha, and significantly underperform index funds.’

The implication is that poor fund selection practices existed among commission-based advisers and the flows to popular retail funds may have skewed the ‘average’ performance of active funds. In my paper ‘Core-Satellite Conundrum’, I proposed that fund buyers may be falling for game theory by taking safe ‘bets’ into the same well known funds. In other words fund buyers may be opting for ‘safe’ core options on an assumption of an inability to identify superior choices with any certainty (see below). How much was commission, how much marketing, how much bias?

Asset management is far too dynamic to pick funds based on performance alone, understanding the workings is essential. For me numbers can pose questions, but not answers. I avoid aggregated time period analysis and rather assess managers at the holding level through key non-Gaussian market events. Some of our members will employ quantitative techniques, but they are typical far removed from basic rear-view-mirror approach. I believe more real world econometric and factor based analysis has value.

Moreover what this highlights is that professionalism of fund selection and due diligence is critical and standardisation of some practices overdue. If scheme trustees, advisers or IGCs are trying to look at funds this way then they need help. Fund selection should be more transparent and I call to my peers to embrace crowd-rating initiatives, like SharingAlpha.com, to demonstrate our value. This is something we take very seriously at the Association of Professional Fund Investors and personally as an Ambassador for the TTF.

At Gemini we are committed to finding managers that can demonstrate alpha or deliver compelling cost adjusted returns and risk solutions. We do that by looking beyond outputs and understanding our managers and how they can add value to our fund buyers and investors.

Happy travels, now where’s my bloody flight?

#thinkgemini

Jon "JB" Beckett

Consulting CIO

Twitter: @JonSBeckett