Benchmarks: confidence game or imperfect science?
This blog originally appeared on SharingAlpha.com.
I’ve never been much of a gambler, I always expect the house will usually win and that the odds are stacked against the gambler. My one exception was in Vegas in 2013 when I turned $50 into about $500 over 2 hours at the roulette wheel playing simple Red-Black mean reversion based on counting the previous 12 turns; when most others around me lost their shirts on low probability odds. Small wins, small losses. More wins than losses. Just luck of course. True story. I think you either tend to have more of a risk-taking or risk-avoiding mindset.
It's difficult to generalise but I find more of the latter among fund buyers. Fund managers, no comment. Take a confidence game, it builds up a player’s framing and anchoring bias that a certain outcome becomes more likely through repetition. The bet (thus risk) increases as the player’s confidence grows, just before the trick. In some ways mutual fund benchmarks and indices can have a similar relationship to funds, and for investors. They set an expected outcome that may eventually disappoint.
What do we, as investors, expect a benchmark to do?
Simply it should provide some sort of understandable, transparent reference point by which to measure fund manager’s value against. Arguably this is best done either in aggregate, over various time periods or at the individual client level. The aim is to somehow compare and match the expected utility of the fund with its outcome. Where there is confusion is when discerning between benchmarks and targets. One relates to a compliance requirement; the other an aspiration or intended customer outcome. It is here when fund managers and investors need to tighten-up reviewing those expectations and outcomes. That conversation probably requires a far better view on investment horizon, risk-adjusted outcomes, absolute returns and factor based benchmarking, which is already beginning to occur within Professional Fund Investor (PFI) circles.
How are benchmarks being used today?
Benchmarks today now appear to operate as blunt measures to bludgeon active managers on a month-month basis; instead of assessing the fund manager over their target investment horizon and accurately in context to the positioning and profile of the strategy. They have also become much more readily investable and so the relationship between fund manager and benchmark has irrevocably changed. There needs to be a clearer connect between how a fund is run and what its benchmark and aims are presented as. Today there is a disconnect, partly due to the regulator, partly due to Sales and Marketing having a blasé attitude towards fact sheet benchmarks for 30 years. The one thing worse than a vague benchmark is no benchmark and so again a clear connect from portfolio manager to investor is vital. Benchmarks not relevant to the strategy should not be stated or relied upon. Peer group and sector averages are a good example. If used they need to be relevant.
If a fund manager has a 3 year horizon then the fund investor should be interested in that rolling number, versus benchmark and hit rate of the number of times a fund manager meets the aim, beats the benchmark and otherwise builds up a picture of persistency. Ditto if the fund is aimed at; 1 year, 5 years, 7 years, 10 years+. The investment horizon of the fund should also be clearly stated alongside the benchmark and assessed against. This extends even more into fixed income benchmarks; should the benchmark have a similar spread and term duration to the fund manager, refined by credit ratings, coupons or future cash flows? These are not questions being widely discussed nor addressed. In absence of properly setting fixed income benchmarks I fear we overlook structural risks in investor portfolios, especially given the proliferation of fixed income ETFs.
Does the type of fund matter?
To even discuss passive benchmarking is quickly consigned to being irrelevant. However what isn't being spoken about is how passives should be benchmarked today. Currently they are simply being benched to themselves and thus far too much focus is being put on tracking error monitoring rather than stepping back and questioning the end outcomes for the investor. Without which passives operate with complete morale hazard, free of accountability for downside or mismatch of investment horizon to an investor. Tracking error has become a proxy for good customer outcomes even it can be anything but. Likewise when it comes to target return multi-asset funds - the use of risk free rates as benchmarks has been deemed unjustifiable, assuming they tend to acquire risk above the risk free rate? Clearly not, as the funds should be default benefit from carry above that rate and I'm sure absolute return and diversified growth fund managers in the U.K. and globally will (should) be reviewing their position on this and quickly we would suggest. Otherwise benchmark the naive portfolio be that a broad 60:40 or 70:30 equity:bond allocation.
How have we got here?
Most active managers today still invest in absolute terms for their clients but compliance requirements necessitate that benchmarks be added for performance comparison. Often the benchmarks were selected on market weight grounds, being the standard or a ‘typical index' for a particular sector rather than bearing any resemblance or correlation to the strategy. The S&P500 is a great example in a stock market of around 5000 listed US companies. Conversely should some active managers be criticised for mediocrity? Absolutely. Regulators are culpable, with a degree of regulatory arbitrage over the years, it's been a shifting goal-post as one rule creates an issue to later be fixed by another rule, and the needs of investors have also changed over time. Regulation continually fails to catch up with investor needs. For example Style-Cap indices became popular in the US to help group the growing number of fund managers and apply attribution but again have been used to effectively corral active managers; then we criticise them for style 'drift' or poor active share. Is it a coincidence that the group of funds most flattened by the broad market has been the US equity sector. I don't think so. The whole approach of simply allocating on style is 30yrs out-moded and more attune multi-factor and outcomes benchmarking should make far more sense today. Factor suppliers like PureGroup.io or StyleResearch could be of immense value here to product teams.
The Professional Fund Investors view
Professional Fund Investors’ (PFI) experience of active fund manager benchmarks tends to be one of agnostic bemusement. The common but slightly galling statement we most often read being ‘unconstrained but benchmark aware’ in terms of performance comparisons. However those benchmarks have effectively become the most heavily invested strategy in the market and active managers have found themselves steered into a regulatory cul-de-sac. Meanwhile PFIs often set their own benchmarks, specific to the attributes of the fund manager, strategy, liability or honed set of expectations of what a fund should deliver over precise investment horizons. Meanwhile the industry remains overly fixated with standard time period returns, which has made some investors more short-term in their risk appetite. PFIs continue to review and monitor the situation, often assigning custom indices, swim lanes and peer groups to manage.
Confidence benchmarking?
Fund managers should clearly set out an expectation that is a fair and transparent representation of the riskiness of that fund. Any target relies on confidence, illustrating an expected range of return might be far more informative to an end investor. Best and worst and most likely. If an 'upper range' target is internally used only; should Fund managers ask themselves whether being publicly accountable would drive different behaviours? It might. Perhaps illustrating fund performance against a range of expected outcomes would make for a more informative conversation with the end investor? It's worth reminding ourselves that internal risk monitoring also seeks to capture abnormal ex ante riskiness in upside returns whereas the regulator and media remain prefixed with only ex post downside loss. There needs to be a clearer connect between expectation, outcomes and the persistency of meeting that expectation over time. Therein PFIs will seek to understand drivers if the successfulness of the fund manager, positions, factors, near misses and so on.
Risk-adjusted benchmarks evolved from rudimentary Sharpe ratios have merit. My friend Con Keating’s Omega ratio for example. Question whether it is fair to benchmark an active manager, with a defined risk budget, against an index that does not? Risk management expects the manager to control risk, why not then the reference index? Risk-adjusted benchmarks that remove more than 2 times deviation risk would change the very tone and narrative of the active-passive debate today.
Passive proponents may (perhaps rightly) baulk at such notions of rehypothecation but we should at least debate the practicalities, pros and cons. Correlation adjusted benchmarks, to effectively strip-out the impact of off-benchmark positions, is another possible approach. The days of referencing sectors that allow 20% off-core exposure must surely be numbered for they skew any transparency of alpha generation. Even examining a fund versus an index on a money weighted basis can produce quite different results to time weighted or non weighted. Index trading volumes are subject to vagaries just are fund flows. Devil can be in those details.
Benchmark related Fee structures
In terms of how performance benchmarks relate to fees, the FCA has missed an opportunity to more widely review the Annual Management Charge (AMC) business model, and performance fees, as part of its market study. Areas not openly explored include CASS' symmetrical AMC by Prof Andrew Clare and the proportionality of AMC to the performance fee. The current popular tone is that performance fees are bad and this is not in itself true. It's about how well the fee structure is aligned between fund manager and end investor. We see greatest innovation coming from many boutiques whereas large scale asset managers happily eschew performance fees for building scale on conventional AMC earnings. It is far too simplistic to argue one approach is better than the other and both can engender poor behaviours.
Better monitoring
There is a good argument to be had that investors should annually review the efficacy of the benchmark alongside the performance and remuneration of the fund manager, as happens in closed ended trusts, US mutual fund boards and some hedge funds. Target, hurdle, waterlines, fulcrums, hit rates, and cure periods all need to be considered in context to the riskiness of the fund, point in the market cycle, reference back to absolute outcome and target, the benchmark and the intended investment horizon. Simply regulating upper ranges will only push those down and not necessarily improve outcomes. A more thorough review is needed here.
What rapidly becomes apparent are that there are many areas that the FCA could expand its benchmark review, well beyond what is being discussed in the mainstream or outlined in CP17/18. For example the lack of competition among index providers is not widely recognised but is well known. The above views are therefore only initial thoughts for the Professional Fund Investor to consider; the APFI fully intends to consult, engage its members and respond to the FCA's consultation paper CP17/18 in the fullness of time.
JB Beckett is a Chartered MCSI, UK Director, APFI, Author of New Fund Order, Author for CISI, SharingAlpha Advisory Board Member, Ambassador of Transparency Task Force, Fund Gatekeeper and Consultant.