Dear New Fund Order followers and fellow colleagues, salutations from your humble suitcase fund analyst. Apologies for the lack of recent activity but I have been again on my travels far and wide discussing best practice in fund selection, technology and ESG.
Most recently I sat on an advisory panel (in my own time I may add) to hear Stirling University students pitch their latest stock ideas for the Student Managed Fund. Having prepared before the session I was reminded what it is to be a fund selector (and indeed what it is not). 2018 marks my 8th year at my current employer (not including my first tour back in 1996-1998) and my eighteenth year in fund selection/research. In my time I have met about 2500 fund managers over 2 decades; to put that into some context that also means I have probably heard the best part of 25,000 stock, bond, property, biotech, infrastructure, high yield, gilt, loan and derivative ideas.
In my time it has been great to have worked alongside experienced multi-managers and fund managers like Bill Bulloch and Jake Moeller before him. Between us we have used our real world experience of managing client portfolios (actually meeting investors, scary thought), selecting funds, researching stocks, applying macro analysis and we take that experience and apply it to assessing fund managers: both prospective and existing.
Fund selection then is all about experience and requires fund analysis/selection subject matter experts (SMEs), getting under the bonnet to understand a fund manager beyond the marketing literature and headline figures. Only in this way can we select fund managers with confidence. As such we are able to help various parts of the business proposition and support colleagues.
As we head towards MiFID2, fund governance, selection and value for money will become hot topics and how we approach this more critical than ever.
What is Fund Selection?
Well, firstly it is most certainly NOT about being the most senior guy in the room that approves the trade. Of course we do and should have controlled functions, delegated authorities and senior manager systems but selecting a manager that has already been recommended is not itself selection. That's like saying Donald Trump is a military commander because he has his Air Force 1 jacket or can press the big red button (as one might imagine it). True, many senior colleagues and executives may trade funds having never researched them or producing a buy note. They rely on fund selectors to fish the market and identify best in class managers. Selection is about manager discovery and researching, advising and implementing that manager for a particular need. To then be a fund selector is to be able to adapt to any sector, to be a proxy property manager one day, an equity the other, a biotech manager the next. Part psychologist, economist, accountant, risk manager, philosopher.. to understand enough to gauge the capability of the managers with confidence and apply critical thinking.
Buying a fund manager (not an index) is to underwrite an individual or collective of people. That at times appears unclear to some. It then struck me that many colleagues may not fully understand the origins or indeed what fund research/selection is.
So, ta da! Where does fund research and selection originate from? Not a straightforward question. Well, as well as train and mentor fund analysts over the last 2 decades; one of my other roles is as an external specialist for the Chartered Institute for Securities and Investments (CISI.org) where I write/teach on wealth management and collectives research.
Firstly in part, some of fund research can trace roots back to actuarial science however the application of quantitative techniques for fund selection are now very different to those applied by actuaries. The other part is qualitative analysis that heralds from the original Harvard marketing mix and progenitors like Russell. Fund selection is (statistically) very difficult - we only need to review the academic studies to tell us this. Funds are not easily observable from a mathematics point of view because they involve these fleshy squishy things we call humans (or at least until the advent of AI). The amount of data needed to make a statistical case for a fund manager is improbable because most fund managers rotate too often. Citywire notes: "Of the 192 managers in the sector, 74 (38.5%) have a 10-year track record. However, of the funds they run between them, 130 out of 189 (68.8%) have been in existence for at least a decade." Therefore if you chase long term persistency and tenure you tend to become funnelled into narrower and narrower samples. The likes of Citywire Discovery and FE Research can offer you manager performance statistics rather than fund level. These are sometimes useful but ultimately naïve to think that a fund manager will behave in one firm exactly as he/she has in another. See: http://citywireselector.com/news/discovery/list
The regular SPIVA report by Standard & Poor's reports that the % of fund managers that can beat the market consistently over time is negligible (well below 50%). This backs up the earlier works of William Sharpe that stated that all things being equal in an efficient market then 50% of managers will outperform and 50% of managers will underperform the market. All subsequent studies point that it is difficult to do this consistently over time. Then you have studies like those by Jonathan Reuter and Diane Del Guercio that noted the poor ability of financial advisers in the US (a sample of about 2000) to select good active managers; as they were far too prone to flashy marketing and commission. https:// papers.ssrn.com/sol3/papers.cfm?abstract_id=1935109
You also have Active Share [studies](https://www.cfapubs.org/doi/pdf/10.2469/ faj.v69.n4.7) by Paul Cremmer and Antti Patijesto that pour cold water on quasi-passive managers that do not deviate sufficiently from a benchmark which has gone on to be widely [critiqued](https://www.fidelity.com/bin- public/060_www_fidelity_com/documents/leadership-series_active-share.pdf) and challenged.
Lastly statistical studies like MathLab note that in order to make a selection decision based on performance data would require 600 years of data to optically and reliable identify alpha in a kernel of returns. Apparently the degree of identifiable excess returns (alpha) is less discernible than between the two parts of a quasar star x million light years away! Yikes. So quant analysis I am afraid is not a tenable basis to select fund managers; it gives us comfort much in the same way as looking in a rear view mirror and noting you hadn't hit a tree. I teach the origins and different models of quant fund analysis but the statistical support for such approaches is (ironically) pretty poor. The reality is that capital markets, funds and human decisions do not follow Gaussian distribution patterns or follow actuarial Wilkie 'Brownian motions' - they are unpredictable.
No, fund selection may infer from the past but only so much as it is concerned about the future. Quant continues to advance from Engle GARCH models, levy jumps, entropic, statistical arbitrage through to organic and complex network analysis; in my experience trying to identify and then manage fund managers through quantitative techniques alone is an unachievable holy grail. Quant can help identify risks arising but not the reasons; it can ask questions but can never provide definitive answers. People just don't compute well into numbers.
There are other models however and I critiqued some of them here: https://blog.sharingalpha.com/sharingalpha-the-mechanics-of-volatility-factoring-the-nuts-and-bolts-of-risk/
For example a Fund Selector can use a factor-based approach, both in terms of selection and when constructing portfolios. A factor-based approach is one that measures a fund's risk through sensitivity to different economic premia. It can trace its roots back to the simple Fama and French 3-factor model in 1993. Factor analysis has evolved considerably and includes tools from Style Research and Risk-Lab. For example PureGroup.io offers factor analysis of funds based on: Beta, Market, Size, Style, Momentum, Macro-factor, Default Spread, Term Spread, Interest Rates, Dividend Yield, Inflation, USD Trade Weight Index, Volatility Risk Premium. Alternatively;
• Use complexity analysis to identify points of risk. Complexity analysis assumes that the more complexity that exists, the more fragile the system and hence can infer rising risk within a fund. Ontonix has developed Universal Ratings and requires a step-change in thinking.
• Focussing on scenario, drawdown and stress test analysis that moves away from standard time period measures and observes how funds behave in different real and hypothetical conditions.
• Behavioural and technical approaches are widely used that seek to identify shifts in trading behaviour. Such approaches could potentially be automated into AI programs.
Not Art; Not Science?
Unfortunately that leaves us with the fuzzy logic of qualitative analysis which is prone to human biases. Applying the wisdom of the crowd is one solution, aggregating many biases ultimately identifies insight, the SharingAlpha model applies rigour to tracking the effectiveness of views via the ‘hit rate’. This is a game-changer for fund Ratings that are often assessed by one or two analysts and then approved by a committee no one meets. Crowd ratings are ultimately transparent in both terms of information and delivery. You know whom just as much as what. I have championed transparent fund ratings for over 5 years and we are finally beginning to see more transparency come through.
Qual analysis is still largely the dominion of investment consultants, the king pins of the fund research world. This is also where investment consultants have run aground in the latest FCA Asset Management study.
So, where did qual' fund research ratings and consultants come from?
Like Supersonic flight, the beginnings of fund selection can be traced back about 70 years. At the Association of Professional Fund Investors (APFI) we aim to represent the views of Professional Fund Investors (‘PFIs’), innovate and recognise best practice. Professional fund investing and research is a community, one I have been proudly part of for a little under two decades. However that pride has faced many tough lessons and challenges over the years. Today, as the industry digests the 200 pages that make up the FCA Interim Market Study Report, questions as to the effectiveness of investment consulting; selecting active managers and negotiating costs have arisen.
Buried deep in its report, in chapter 8 (from page 180), the FCA critiques the Multi-P (‘X-P’) approach used by many fund rating agencies and investment consultants. Why do fund buyers, consultants and agencies use a X- P approach and how does embedding Price into that process impact the findings of the FCA?
The X-P model began in the 1940s with 4-P: Product, Price, Promotion and Place.
According to Wiki (the de facto font of all information and disinformation) the origins of the four Ps can be traced back to Professor of Marketing at Harvard University, Prof. James Culliton. In 1948, Culliton published a paper ‘The Management of Marketing Costs’. Culliton described marketers as ‘mixers of ingredients’. Some years later, Culliton’s colleague, Professor Neil Borden, published a retrospective article and credited himself as popularising the concept including his presidential address to the American Marketing Association in 1953. The 4-Ps, in today’s form, was first proposed in 1960 by E. Jerome McCarthy in his text-book, ‘Basic Marketing: A Managerial Approach’. McCarthy used the 4-Ps as a framework for the entire work with chapters devoted to each of the elements, dedicated to analysis, consumer behaviour, marketing research, market segmentation and planning to round out the managerial approach. Author Phillip Kotler likewise popularised the approach and, and with it, spread the concept of the 4-Ps.
As a generation of Regan and Thatcherite Ivy graduates, entered consultancy, the rapid adoption of 4-Ps accelerated through the 1980s, and the phenomenon of the ‘business consultant’ took hold. This popularised the concept into commerce and any new consultant or business at that time was expected to demonstrate awareness of latest management techniques.
In 1981, Booms and Bitner proposed a model of 7-Ps, comprising the original 4-Ps plus process, people and physical evidence. Subsequently a number of different proposals for a service marketing mix (ranging 6-Ps, 7-Ps, 8-Ps, 9-Ps and more) have emerged through leading big management consultancies like McKinsey. Launching into this environment it comes as no surprise that the early progenitors of multi-manager and asset consultancy adopted a common language with, management consultants, whether that was the optimal approach or not.
The History of Consultants and Ratings Agencies
The rise of asset consultancies post ‘Big Bang’ in October 1986 and and the previous Employee Retirement Income Security Act ‘ERISA’ (1974 in US) led to the slow Americanisation and gradual de-institutionalisation of the UK asset management market. An influx of not only large US banks but also consultants coincided with the gradual demise of U.K. insurance With Profits and Defined Benefit schemes. In the FCA report a shadow has been cast on the dominance of the big consultants, which now have a grip on many facets of our industry. Given the power of investment consultants, as gatekeepers, then the FCA was right it its report to question whether this not only impacts the returns for their clients but also impacts the wider market for the rest of us.
“The point missing from these considerations is any strong emphasis on investment consultants’ performance. This may be because, as we found, there is limited information available to institutional investors on the quality or performance of advice when they are selecting a consultant.”
Big consultants include;
Russell Investments was founded in 1936, employs 1700 employees and manages $244bn assets under management, opening its first office in London in 1979, launching its multi-manager service in 1980 and launching the now famous Russell indices in 1984. Russell has become synonymous with equity style and capitalisation investing and for building one of the first X-P models, which it has since evolved. Russell is credited for introducing the X-P model into fund research. Russell is generally regarded as one of the most thorough multi-managers, at least from the point of view of conducting due diligence on a fund manager.
Owned by Marsh & McLennan, Mercers was founded in 1945, headquartered in New York, employs over 20,000 employees in 40 countries, and serving over 130 countries. Originally called William M. Mercer after the Canadian firm was acquired by Marsh & McLennan. The history of Mercers has been one of acquisition. Mercers moved into DC investment consulting in the U.K. around 2005, starting with 60 different funds managed by over 20 investment managers. In 2012 Mercers moved into investment management in the U.K. Mercers global team numbers 1,200 professionals and 120 manager researchers provides in-depth expertise in research, advice, and solutions. Mercers assert their advantage based around their Global Investment Manager Database (‘GIMD’) which uses data capture, analyst inputs and big data to support more conventional fund research.
Aon Hewitt was founded in 1940 in Illinois, as Hewitt Associates, has approximately 29,000 employees worldwide, operating in 500 offices in 120 countries. By the beginning of the 1990s Hewitt had ventured abroad and offered benefit programs to corporations in the United Kingdom. By 1997 more than 100 large companies outsourced their benefit programs to Hewitt, covering about nine million worldwide employees.
Morningstar influences/manages over $200bn assets under management in 27 different countries and and is the quoted largest fund ratings platform with around 180 fund analysts worldwide. Launching into the UK with its innocuous website service the business has grown exponentially through its fund analysis and reporting software ‘Direct’ and asset tool ‘Encore’. Morningstar, Inc. has been providing qualitative analyst research on funds since the 1980s. Morningstar has been one of the leading X-P rating agencies, which is a 6 factor approach (originally 5-P) that includes Price. It remains easily the largest agency in the UK retail market but also competes in the Institutional market. It is worth recognising that Morningstar’s model differs to that of investment consultancies albeit does supply consultancy and investment management, thus blurring lines. Others include JLT, Willis Towers Watson, Capita, Hymans Robertson, Fitch, Punter Southall.
What Fund Selection is NOT
Fund selection often gets confused with risk management. It is not operational due diligence (ODD) which is focussed on ensuring that tail risks and poor compliance are minimised. It is designed to ensure that a firm is able to meet its commitments. A good firm from an ODD point of view could be mediocre or even incompetent from a fund selection or investor point of view. I am reminded from when I was still in Compliance almost 20 years ago that my most compliant adviser from a regulation point of view was also the least competent in the recommendations. So too can operational, regulation and risk can frankly miss the wood from the trees. They are valuable and arguably necessary but not necessarily for fund selection. Many fund selectors will consider some operational factors but only in so much as how it enables or restricts the fund manager to perform. They may also consider from a reputational point if view (ergo 'sniff test').
It is also NOT Asset Allocation, which may come as a surprise to some. Asset allocation is all about making calls on different asset classes and blending them together to derive an expected outcome. It is about weighing up the risks against the likely cash flows, correlation and benefits. It assumes only market returns (beta) and not the effect of human decisions. By contrast fund selection is all about how humans add (or detract) value from an asset class. That margin can be significant in real world even if Ibbotson Kaplan studies indicate that 90% of a fund's returns are derived from its market returns. True but even in an age of low nominal rates and poor excess returns from active management, 10% margin can still be a lot.
Fund Selection and Asset Allocation are two distinct disciplines; often they may overlap on the job (for example when running a Fund of Funds) but generally the demands differ greatly. The value of asset allocation can be judged on how well it performs against say a standard optimisation model you can Google in 5 minutes or the long-run 60:40 allocation reported by the Barclays Gilt study. 90% contribution to return does not mean 90% of the value added. Fund selection is more easily exposed in terms of value, it can be measured by the excess risk-adjusted return of the manager against a benchmark over time. However firms offering multi-manager products have confused these disciplines to their peril. A few years ago [I critiqued](https:// www.investmentweek.co.uk/investment-week/analysis/2438797/in-the-face-of-new-multi-asset-solutions-are- multi-managers-losing-their-edge ) the demise of multi-manager and this dilemma with Stephen Walker, a top decile multi asset, asset allocator in his own right.
In short the disciplines of a multi-manager should not be overlooked; to focus solely on asset allocation exposes this approach as mutual funds are not the most economically efficient way to deploy an allocation. You buy a fund manager because either; a) an asset class relies on market timing, b) the valuation dynamics are complex, c) you lack expertise in the asset class, d) you cannot trade directly or require a fund to secure sufficient diversification e) you need alpha to justify the investment. Otherwise you deploy beta either through Exchange Traded Funds, index funds or derivatives. It is that simple. This is why managing asset allocation and fund selection as we do today ensures that the right asset decision is married to the right fund manager. One is not (nor cannot) simply be a product of the other.
Applying a multi-factor model, one that aims to balance all of the above challenges is one solution. In understanding the origins of fund selection may help debunk some of the misconceptions and confusions that can arise with other areas such as asset allocation. For those who want to pursue a career in fund analysis or selection then go to: http://www.profundinvestors.com