One of the most common statements since the outbreak of the coronavirus pandemic is that the economic and market outlook has become more uncertain.  Given that the future is inherently difficult to predict with any level of confidence and we are generally terrible forecasters can it really be true that the world is now more uncertain?

The environment unquestionably feels more unstable.  There are a range of imponderables around the development of the virus itself that none of us can hope to anticipate.  From the potential of a ‘second wave’ to the production of a viable vaccine.  We have also experienced an unprecedented economic ‘stop’, and credible cases can be made for an inflationary or deflationary future. 

Yet before accepting the ‘greater uncertainty’ idea at face value, it is worth deconstructing the meaning of such a claim.  What we are really saying is:

“I was much more confident in predicting the future of the economy and financial markets before that unexpected event occurred”.

Which can be expanded to:

“I was much more confident in predicting the future of the economy and financial markets before that unexpected event occurred, and showed that my level of certainty was exaggerated.”

It seems paradoxical to suggest that things were more certain before something we hadn’t expected happened and upended our prior beliefs. 

When we say things are now more uncertain we are usually implying that something has altered which has meant that the future has become increasingly difficult to predict.  It has not.  It was unpredictable before and it remains unpredictable now.  Rather than making the world more uncertain, unanticipated events serve to show us that we were understating how uncertain things were before. When claiming greater uncertainty in the future, we actually get it backwards.

Although we should not conflate greater uncertainty with changes in the forecastability of the world, this does not mean that the notion is meaningless; it just needs to be considered in a different way.  What we perceive to be an increase in uncertainty is really a change in the frame through which we view the world.  We have taken a particular path and now face a different distribution of possible outcomes. 

Let’s take a simple example.  My range of potential future personal outcomes involves me losing my job tomorrow.  This is (hopefully) a tail risk, yet it is a possibility.  If I am made redundant tomorrow then my life has taken that specific course.  My future is still unknowable, but the routes and their likelihood from this starting point have changed.  Although it feels like my life is more uncertain now I am without a job or salary, it is nonsensical to suggest that my life was more certain before it was struck by a remote risk that I failed to predict. 

If my life following a job loss hasn’t become more uncertain, then what has it become?  More fragile. When we discuss greater uncertainty following a particular event,  we often focus on how the shock has left us more vulnerable to future negative events.  Since losing my job I am now more susceptible to failing to pay my mortgage and losing my house.  There are a range of unpleasant scenarios that are now more likely.  This is the same for many businesses through the current recession.  The probability of ruin has changed.  Our certainty in predicting the future hasn’t.

The perception of greater uncertainty is not just about fragility, but how we make sense of the world.  When we perceive  a sharp rise in uncertainty it is a case of the narratives we use to explain the world becoming untethered.  The cause and effect stories we weave help give us some comfort navigating the randomness and chaotic nature of life.  If there is an occurrence that dramatically alters the environment then the threads that hold together our own explanatory narratives quickly break apart.  

The world has not become any harder to predict, but rather the stories we previously used to explain it are no longer valid.  If we cannot construct a coherent narrative things begin to feel distinctly uncomfortable.  To address this we will create new stories, or simply adjust our previous ones.  Restoring our prior (misplaced) sense of confidence and control

The very occurrence of the coronavirus pandemic means that any certainty we may have held before was unfounded.  We never know what will happen tomorrow.  We should act accordingly.

Performance chasing is endemic in mutual fund selection.   Despite the randomness and luck involved,  the decisions that we make and narratives we weave are inextricably influenced by historic returns.  Although we may only spend a fraction of our time considering performance, it has an overwhelming impact on the judgements we make.

Supporting the results of previous studies in this area, a new piece of research by Leonardo Weiss-Cohen, Philip W.S. Newall and Peter Ayton shows individuals using  strong past performance to inform their fund choices rather than lower fees.

The researchers devised a task where investors had to select between one high fee (0.7%) and one low fee fund (0.1%) over 60 periods (months).  Both funds generated returns which were index plus random noise.  The mean of the random noise was set at zero, so the only long-term performance differential was the fee load.  For the first selection participants were shown a 12 month historic return for each fund and asked to make an initial choice.  At the end of each subsequent period they were shown the performance of both funds and again selected between the two options.  The compensation participants received for taking part in the study was in part based on the wealth they managed to accumulate through the game.

There were two separate experiments.  The first featured subjects (400) who had previous investment experience, and two conditions – one where they received a standard disclaimer: “Past performance does not guarantee future results” and one without this disclaimer.  The second experiment recruited subjects (596) without investing experience.  Here there was a third condition – a ‘social disclaimer’ stating:  “Some people invest based on past performance, but funds with low fees have the highest future results.”.  All participants also took a test to grade their level of financial literacy.

The were a number of informative results:

– Performance chasing behaviour was evident across all trials, in both initial and subsequent fund selections. The fund with the highest returns in the last trial was selected with greater regularity.  Recent performance was more influential than fees.

– The standard performance disclaimer had limited impact on reducing performance chasing behaviour and led to worse outcomes for those with low financial literacy and no experience.

– In the second experiment participants chose the low fee option more often when they had been shown the ‘social disclaimer’.

– Participants with lower financial literacy and no investment experience actually chose the expensive fund more frequently when they received the standard disclosure: “Past performance does not guarantee future returns”.

– Participants were more likely to choose the low fee option as the trials progressed, but the overall effect of this was modest.

Of course such ‘lab’ studies are imperfect and never a substitute for natural experiments. Furthermore, the paper does not provide full details on the exact scenarios faced by the participants or a granular breakdown of the results achieved .  These issues notwithstanding there are a range of insights that are worthy of greater research and consideration:

-It is fascinating that participants tended to favour highly uncertain (indeed random) future performance at the expense of the certain performance advantage of a cheaper fund.  Their decisions were made absent any information about the forthcoming returns of the funds meaning choices were guided only by fees or previous results.  This could be an example of our struggle to accept randomness and tendency to see information in noise (the outperforming fund must have some advantage).  Or, as the authors suggests, higher fees can be associated with quality.

– In the real world, performance chasing behaviour is typically intertwined by some form of narrative.  A story that validates why past performance will indeed be a prelude to future excess returns – it could be the exceptional skill of a certain fund manager or the wonderful prospects of a particular economy.  In this experiment there was no story to tell just the numbers, and individuals still tended to be led by them.

– As suggested by the authors, one of the potential issues with recurring fees is that because they can be small relative to the volatility in performance, investors may neglect to attend to them, or fail to understand the huge influence they can have when compounded over the long-term.

– Those with low financial literacy performed worse in the task, which goes to reaffirm that it is individuals in such groups who require the most protection and direction.

– Perhaps the most fascinating aspect of the study is the role of the disclaimers. The “Past performance does not guarantee future results” wording has always seemed a woefully ineffective message because if offers no meaningful guidance.  The social disclaimer provides a clear nudge by directly stating that lower fee funds achieve better outcomes than higher fee funds. Whilst this wording is not without its own problems – fees alone are not sufficient to make an investment decision –  it does suggest that such disclaimers could be re-worded in a manner which genuinely helps investors, particularly those with less experience and knowledge.

Although it is no surprise to observe performance chasing behaviour;  it is particularly interesting to see it play out in a sterile, artificial environment absent the narratives, newsflow and incentives that contribute greatly to performance chasing in the real world.  The structure of the research also allows for a clear distinction to be made between past performance and fee levels as a decision driver, which are otherwise impossible to isolate.  The study provides further support to the idea that past performance and neglect of costs are real problems for investors, and behavioural interventions – such as changing disclaimer wordings – might be part of a solution.

—

https://www.thinkforwardinitiative.com/research/persistent-chasing-of-past-performance-when-selecting-mutual-funds

At the start of the last decade everyone was bullish on emerging market equities.  It is hard to recall now, but it was the consensus trade.  You couldn’t attend a meeting or conference without being told how many people there were in China or how underwhelming the growth prospects were for developed economies compared to their emerging counterparts.  The irresistible optimism was to prove unfounded as the asset class delivered a prolonged spell of underwhelming performance.  This disappointment is merely one prominent example of perhaps the most common (and often damaging) investor behaviour; we take past performance, combine it with a persuasive story and extrapolate into the future.

Dealing with uncertainty

 Our desire to extrapolate is an understandable reaction to the uncertainty of financial markets.  If we cannot predict the future then our best approach may be to assume a continuation of previous trends.  It is not, however, simply a case of presuming performance patterns will persist; the power of extrapolation comes from the narrative that we use to explain what has happened in the past.  The stories we tell are what make us believe that things will continue.

When themes or narratives are discussed regarding investment decisions; it is often done in such a way that suggests that the story is driving the price.  This is misleading.  On many occasions the reverse is true – our inability to explain complex, chaotic markets means that we wrap a compelling narrative around the performance after it has occurred.  Extrapolation in investment is driven by a circular relationship between these two elements.  Performance creates story creates performance.

These post-hoc rationalisations are hugely problematic however.  Whilst the futility of market forecasts and predictions about future market movements is often discussed, this doesn’t go far enough.  On most occasions it is difficult to provide robust and complete explanations for market movements even after the event.  If markets are largely chaotic and random then simply observing outcomes does not mean you can then draw a straight line of cause and effect.

We cannot confidently explain the past in financial markets, let alone predict the future.  Given this, how do we construct narratives that support our desire to extrapolate?  We simplify.  Let’s take the argument that reinforced the extrapolation of emerging market outperformance in 2010; fundamentally, it can be distilled to this: economic growth in emerging markets will be higher than in developed markets in the future.

Although a naive and fragile argument; from an extrapolation perspective it is incredibly powerful because it is easy and ‘feels’ right – of course you want to invest in higher growth markets.  Also, crucially, there was evidence to support it.  By evidence, I mean that in the prior decade emerging market equities hugely outperformed and emerging market economic growth was higher than developed markets.  Whether the relationship between these two variables is strong or not (it’s not) is immaterial – what matters is whether the story is a convincing explanation of performance.  Once the link is forged between narrative and price, it becomes hard to break.  Any attempts to refute it are met with puzzlement – “of course it is right, haven’t you seen the performance?”

Creating a simple, powerful story to describe historic performance is ineffective and illusory, but it is also relatively harmless if all we are doing is explaining the past.  Unfortunately, that is not all we are doing.  The narratives we create don’t just help us decipher yesterday; they set our expectations for tomorrow.  We believe that if our story still holds, then performance will persist.

Extrapolation in three easy steps

 Extrapolation is an exercise in simplifying things that are uncertain or unknowable in order to make some form of prediction.  The process for extrapolation in investment contains three crucial steps:

1)      Take a significant performance pattern: Performance needs to be meaningful and sustained – a convincing narrative is hard to forge if returns are noisy or absent a convincing trend.

2)      Construct simple explanatory narrative:  Markets are too complex to understand or define the precise drivers of previous performance.  With a great deal of effort we can make educated guesses, but it is far easier to simplify the task by creating a straightforward explanation that is understandable, convincing and difficult to refute.

3)      Consider whether the narrative holds into the future:  Our desire to weave simple, intuitive and optically powerful explanations often means that we assume the performance trend will continue provided the narrative explanation holds.  If emerging markets grow faster than developed markets their equity markets will continue to outperform, if rates stay low quality will outperform value etc…

We create a compelling narrative to justify why past performance has been particularly strong or weak.  We then use that same narrative to predict its continuation.

The problem with extrapolation

Investors can and do make money from the propensity of other investors to extrapolate, but whilst it creates opportunities; it also causes a host of problems:

–         Extrapolation is an exercise in simple forecasting:    Our explanation for what happened in the past quickly becomes our prediction for the future.  This is dangerous because we cannot accurately explain the past let alone predict the future.  Even in the event that we are correct in our diagnosis of the drivers of previous returns; it is unlikely that they will persist unchecked.

–          We often imply structural changes when extrapolating:  Although much market behaviour is inherently cyclical, when we extrapolate it often tends toward being a structural pronouncement.  Things are either in a new paradigm or dead, rather than be in some form of natural cyclical decline or boom.  This is because when we extrapolate we do so in perpetual terms – this will continue indefinitely.

–          Stories are often wrong and missing key variables:  The narratives we utilise are almost always too simplistic to be meaningful or even close to comprehensive.  Sometimes they are entirely wrong.  Yet when it comes to stories it is not the validity that matters but the coherence.  There will be some occasions in financial markets when Occam’s razor does apply – and a simple explanation does fit – but it has to be supported by robust evidence, not just be a convincing yarn.

–          Extrapolation can leave us concentrated and overconfident: The longer a performance trend persists and the more investors participate, the more the underpinning narrative is bolstered.  Our growing assurance can leave us unable to see any other environment prevailing; indeed, doing anything but following the dominant trend probably comes at a cost.  It is at these times when we are most vulnerable to abandoning sensible investment disciplines such as diversification.  After all, the more certain you are about the future, the less you need to diversify.

Extrapolation is simple.  If something has worked well in the past you don’t have to expend any energy explaining why it will work in the future – the evidence is in front of you.  It is easy to convince yourself and it is easy to sell to other people (try marketing an investment strategy that hasn’t worked for the last five years).  As investors we use it persistently; whether it is thinking about asset class returns, building a portfolio or selecting active managers.

Extrapolation can be effective.  Performance trends in markets can and do persist for prolonged periods.  Yet exploiting these trends is more about understanding the behaviour of other investors (the Keynesian beauty contest), than it is ‘knowing’ the fundamental drivers of market performance (past or future) or telling an accurate story.

The central problem with our tendency to extrapolate is that when we do so we are implicitly making unjustifiable assumptions about financial markets: that they are predictable, that there are simple explanations and that things won’t change.  Holding such beliefs rarely ends well.   Yet despite the pitfalls when performance and narrative align it is hard to imagine that the prevailing trend will ever cease.