Behavioural finance is at an exciting juncture. There is no single unified theory, no integration with conventional economics, but real, practical application is now emerging. This is happening as the early behavioural focus on biases is questioned, as is the over-simplified distinction between emotion and thinking. However, classical economic theory - resting on its pillar of equilibrium to make the maths work - still survives and is taught as the core from which presumably irrational behaviour departs.
Behavioural finance is growing in relevance as some of the softer factors in investment get attention; ESG, incentives, corporate behaviour and the best way to integrate investment professionals and artificial intelligence. There are positives and negatives in our biases, but psychology helps us understand human strengths.
For many, the starting point in behavioural finance is set out here by its best known exponent, Daniel Kahneman. It draws distinctions between the unconscious and knowing, heuristics and hard thinking, automatic and thoughtful. I will pick-up later on his reference to stories.
But this sharp distinction does not fully cover all the aspects of psychology that are relevant to investment – we can learn much more. Quite apart from cognitive dissonance, there are also differences in beliefs – not just emotion but hard-wired differences in people’s view of how the world works, the logic of it. Rationality is hard to define – in a number of areas, the science is not settled. Sensible, thoughtful people expect entirely different outcomes to well-informed considered decisions. And, with most firms making decisions in groups, social behaviour matters too. An issue I will focus on particularly here, is the role of incentives – we expect them to shape behaviour and outcomes, yet fail to understand just how precisely they can be gamed by humans. There are often unintended consequences. I prefer to view behavioural finance through the lens of the five issues I list here. Incentives explain a lot – in finance, these usually involve money.
The starting point is that biases are usually helpful. We draw on experience and context to make sense of the new. However, we may not realise how much of previous learning we bring automatically and unconsciously to evaluate new decisions, and in general navigate the world about us.
The pareidolia slide is an example of where we try to create a pattern or narrative out of randomness. We bring a lot of previous experience and modelling to this, and these patterns are quite difficult to unsee. We can see in the George W Bush and VP Dick Cheney manipulated photos, that we jump to conclusions even though the faces here have been manipulated to be the same. We do not examine everything, if it seems we do not have to. Similarly, for me standing upside down in the Museum of Illusions in Vilnius in Lithuania. Some of these actual problems are difficult to unsee even when we are aware of the error. The Bristol Café Wall illusion has been stylised in this picture making it very difficult to accept that the blue lines are in fact horizontal. It will take some new frame, such as a ruler or grid to help us to see this as it is. For our sense of scale and perspective, we rely a lot on visual cues and assumptions about vertical, horizontal and relations between objects.
Added to this challenge - interpreting what is new without drawing on unhelpful previous experience - is the difficulty of investment itself. Even the best managers find it difficult to pick a single stock and a specific timeframe, as the Sohn London investment conference stock pitch performance shows. It is a noisy environment in which we are ill-equipped to understand what is statistically valid.
A noisy environment is compounded by headlines and comment that may be based on little fact or insight. And, we can see just how difficult it is for experts to make forecasts when we see the ECB inflation forecasts here in blue, that are all wrong with no apparent learning, so rooted are they to assumed correlation with money supply. I suspect given deflation brought by Coronavirus, the December inflation projection will also prove wrong.
And company management is difficult, too. We pay attention to the social validation of reputation, but it may not be relevant. We saw this with the confidence that was attached to the high profile executive chair of Patisserie Valerie and his large shareholding. And a lot of the numbers that we must work with as analysts are not what they seem - even things that look precise, like earnings per share or organic growth. These may not be comparable between companies, or even in the same company from year-to-year. Research points to the impact this has on pay, despite the lack of correlation with actual share price performance. That is, non-GAAP or non-IFRS measures that adjust statutory earnings (also called Alternative Performance Measures) appear not to fool analysts and markets. But they do mislead company boards who reward executives based on those adjustments. [Guest et al "High Non-GAAP Earnings Predict Abnormally High CEO Pay∗" Massachusetts Institute of Technology Sloan School of Management May 2018]
And, while we might unpick the adjusted numbers, it is difficult to unsee the attractive images and trends that appear in annual reports. Even if we only take away the impression that this is a good business or is a growth company, the graphs frame our subsequent analysis. This is, of course, intended by company management and boards. These images often come early in presentations or at the start of annual reports, and appear more salient. Images, and the patterns we make of them, appear more vivid. We can more easily construct a narrative from them, although companies usually provide that anyway.
Compounding this problem is the fact that it takes many years (22) for investment performance to be meaningful, and predictive. And even then, what it indicates is the likely longer term results, not the sort of yearly variability we actually see from good managers. Some of the expense in the investment system, and cost to investors, results from the chopping and changing after periods of underperformance by managers. Managers who were thought good enough to invest with in the first place.
Amidst this noise and confusion, the search for meaning often focuses on narratives. This is not new, as we can see from the 1920 Ponzi coverage. Inevitably, managers with periods of underperformance must construct stories and there is good research on manager behaviour with underperformance. For some, it will mean cutting risk until a record is rebuilt. In other cases, some high profile managers wish to demonstrate that their confidence is undiminished and that they will actually make bigger bets after a tough period. Whilst intended to support a narrative of confidence and consistency, this signalling is in itself a style shift. [Richmond & Byrne, “How to deal with underperforming managers”, CFA Institute Monograph, 2013.]
This concept of narratives, attempting to link events in a coherent or apparently plausible way, carries through into a lot of fund manager reporting. Often this sounds convincing, even though it may seem outrageous when dissected. Consider how a manager here is unable to see his own conflict in making a significant investment in a major distributor promoting his fund, whilst readily able to interpret what the market might be basing a small bounce from share price lows on. Here, as with the other stories, the manager aims to underline confidence in the story, despite all that has happened, by adding to the investment. In the hedge fund examples, we see that managers are responding to adversity by taking on more risk – “a smaller fund that is able to make larger bets”. Other hedge managers explain events in strange ways sometimes underlining their contrarian approach. I think there is often an intellectual arrogance in contrarianism for its own sake – something that Professor Richard Taffler of Warwick Business School highlights in his emotional finance work. A heroic struggle against adversity and unworthy opponents. [Tuckett & Taffler “Fund Management: an emotional finance perspective”, CFA Institute Research Foundation Publications, August 2012, Vol. 2012, No. 2]
Sometimes it is possible to quantify confidence, or the extent to which it might be misplaced. Woodford helpfully from launch provided a lot of detail on individual holdings and performance. I captured this in mid-2017, well before there were public concerns about these funds – at which point the information was all taken down. We can see here that the large investments intended to display confidence, being 6, 7 or 8% of the fund. This represented a huge risk budget but not adding disproportionately to performance. That is, the liquidity risk of these concentrated holdings was disproportionate - effectively there was a good case for splitting these into smaller liquid, diverse holdings. Also of interest, is the fact that all but two of the holdings over 1.5% were in positive territory, which made me question how active price discovery was in these stocks.
There are often patterns in how managers change their engagement with colleagues and clients, or parse different periods to help clients frame what is happening. A manager might say “the decade plus long record clearly shows I am an exceptional manager. The last three years do not negate that, and were atypical. That’s not me. The latest three months demonstrates the beginning of a turnaround and is evidence of my true ability, as you previously recognised”. Written down, this seems a bizarre mix of selective periods, but sadly is common. It is a heroic narrative, and many investment managers struggling with performance problems do indeed describe their battle against adversity in this way. Imagined foes are short sellers, consensus, and myopic investors. The embattled manager is visionary and contrarian. Managers know the emotional appeal of epic tales.
Kahneman notes that stories tend to be judged by our perception of coherence rather than the quality of evidence. The world – and investment, in particular – is less coherent and more ambiguous than we believe.
As Kahneman notes, “the world makes less sense than you think”. “We are too easily convinced by simple vivid stories, rather than admitting our discomfort with complex abstract and often random reality”. The investment profession needs people who can be comfortable with that ambiguity, but it may not be the way we train analysts. The intake for the profession tends to be quite deterministic and numerate in mindset.
Academic research in psychology is going through an upheaval. Already the result is an improvement in methodology that should benefit other disciplines, too. Too many early psychology studies have not been replicated, or were underpowered. Experiments in incentives were often questionable – what a student might be influenced to do for $5 does not obviously translate into a trading environment or the sort of sums involved in investment and executive bonuses.
Viewed from the outside, research in psychology seemed simply to be a growing pile of increasingly specific biases. Many of these appear to contradict – which matters more; priming or recency? It seems too much has been expected of biases – many are likely to have small effects and be very context or culture dependent. It is little wonder that this has not built a coherent substantial economic theory.
Behavioural research often points to what appears to be irrationality in peoples’ choices or alternative opportunities - often expressed simply in terms of expected value. Certainly, on a repeated basis, 50% chance of £2 million exceeds a 90% chance of £1 million, but a decisionmaker would need to know more about the number of repetitions. Learnings from proportional betting strategies, such as Kelly, point us to different conclusions about rationality. This would be an example of flawed behavioural research.[Kelly criterion, en.wikipedia.org/wiki/Kelly_criterion].
But for the biases that do seem persistent, some solutions have emerged. On hiring, for example, or manager selection, it is now considered good practice to have a process that involves different stages of systematic screening and data collection with separate people handling interviews. Equity analysis typically also works best in a systematic 7 process, although it can be challenging to gather comparative company data and to be disciplined in rejecting the spurious idiosyncratic numbers that management often push out.
In recent years, much effort and legislation has been directed to company incentives and metrics. I will not deal here with the absolute level of executive reward, but bonuses and other incentives, and consequent behaviour. Unfortunately, most companies are still obsessed with growth, despite the low growth in the economy in general and the difficulty of measurement. And it is too easy to focus on adjusted numbers, which can become an exercise for CEOs to mark their own homework. Now, much of the focus on audit is on cost rather than quality. And auditors are responsible only to shareholders (as a body) despite the broader obligations of company boards.
I will pick out a few company examples to illustrate the impact of some very specific incentives. Financial incentives can over time have a very big impact on shareholder value, often negative. We should remember that value creation is usually driven by more than the few executives at the top, and question how deeply and well-aligned incentives are within a business. And, it is also worth remembering that many public and not-for-profit organisations are driven very successfully by mission-driven individuals motivated largely non-financially.
Incentives are usually taken very seriously by management and do direct behaviour. Perhaps this is to be expected when bonus can be twice salary or more - even small components of the bonus mix get attention. With M&S we can see what happened after a number of years of adjusting earnings, and maintaining a dividend not supported by the actual statutory earnings. The dividend was cut by 40%, and share price has since fallen by the same amount. M&S has seen declining earnings over the past decade, but we can see here just how much the statutory earnings have been flattered by adjustments. Interestingly, its CEO total remuneration has averaged £2.1m per year over the last 10 years (various CEOs), varying between £1.4m and £4.4m, with little sign that any particular level or composition of bonus has made much difference to the company’s performance.
While there is evidence noted in research from Robert Pozen et al, that US earnings adjustments are increasing, it appears less dramatic in the UK. The extent to which bonuses can be driven by adjusted numbers seems to be the main factor, rather than earnings smoothing. With a tailwind from easy money, perhaps there has been little need to be concerned about funding or cash needs. Indeed, there are some general patterns on presentation of numbers for growth businesses, which I have covered in a separate recent article. Analysts looking for traditional methods of earnings manipulation may miss subtle flattering of gross margins or unit economics that investors favour, and which drive high share ratings.
While adjustments are often presented as an add back of non-cash items such as depreciation and goodwill amortisation, I would contend that collapsing statutory earnings may often be a better guide to cash and the need for capital. Behavioural finance points to a number of issues here. In terms of groupthink, it is clear that a collection of talented individuals on a board and management, often deliver a poorer result collectively than they might as individuals. And individuals are ill-equipped to score their own work. Financial incentives do not help this.
Incentives can work for a while. The point is that shareholders own the whole result including the reset parts - the only group that has to join-up these different reward packages and different CEOs.
This is seen most clearly with Tesco, where the tenure of three different chief executives is marked. I would argue that rewards over the whole period of more than 20 years have owed a lot to some very narrow incentivisation and specific presentation of numbers. The point is that each of these chief executives collected significant bonus and yet in relative share performance terms the period has been a round trip for shareholders. Over 10 years the underperformance of Tesco versus the FTSE100 is dramatic.
The second Tesco slide touches on some of the issues; the level of adjustment and performance based pay. It is simple now on Bloomberg to identify results and rewards over a particular executive’s tenure. Bloomberg shows this information from Minerva Analytics and ISS. Those who wish to delve deeper into the accounting decisions that can materially affect presented results, might look at the auditor’s report in the 2019 Tesco accounts on pensions assumptions.
The trend for claw-back provisions now goes only in a very small way to bringing accountability back into a bonus package. Generally the metrics for bonuses do not address quality of earnings or service, risk, or the capital consumed by the strategy.
As Yale’s Robert Schiller commented, an obsession with growth frames our thinking and seems to derive from central bank policy. Monetary authorities have targets above zero but inflation has been so low we have forgotten to adjust for it. With bigger numbers for everything, it is easy to believe that we live in an exceptional age. Bigger GDP, larger daily swings in stockmarket indices etc. [Robert J Shiller, Silent Inflation, Project Syndicate, 23.11.2018].
Just as we have trouble in bringing these back to real and per capita measures – such as the lack in the West of real per capita wage growth – it stops us getting to the root of growth. Analysts might decide that rolling out new units in a business like Patisserie Valerie is not organic growth. (This is the source of a lot of our misunderstanding of consumer businesses, as unit economics tend to flatten out after 18 months and subsequently there is little inherently organic in growth of those units.) But is raising prices of say, iPhones but selling far fewer, really sustainable organic growth? Or maintaining pricing in consumer items but trimming off a piece of the chocolate bar. We might think much the same of shrinkflation as it now hits the long established consumer staples businesses. I think we would agree that it is not quite as good as finding more customers or actually getting those customers to buy more goods and services. I sense that stockmarkets are beginning to work out what true growth is, but there is an urgent need for boards, remuneration committees and CEOs to get the message.
In summary, the market is beginning to think more about sustainable margin and true growth rates.
While there are signs of improvement in the way that remuneration committees are looking at this, there are still widespread weaknesses in incentive packages. Investors should put in more thought to giving feedback and voting on this. It makes a strong case for ESG. My own observation is that many of the weakest incentive packages are aligned with weakness in governance structure and both should be areas of focus for ESG investors. AIM companies, for example, do not need to publish an audit committee report. And investors should spend more time looking at auditors’ comments or thinking about the judgement involved in the numbers. Increasingly, as more goodwill piles up on balance sheets and we move to a knowledge-based economy, management projections will underpin inventory valuation, goodwill valuation/impairment and other assets that lie on the balance sheet. Any change in trading or even outlook can quickly trigger a cascade of write-downs and covenant breaches - taking a red pen through a balance sheet.
The problems with growth don’t end in the listed sector. Regulation and presumed efficiency on public markets has moved many longer term investors onto alternatives, and private equity in particular.
This is a world of its own valuation metrics and often a firm belief in network effects and the winner takes all strategy of a platform business. I do find valuation of private growth businesses to be strange at times. Often the investors in a private company will re-value their investment to the level of the latest and most optimistic buyer, without it being clear what the rights attaching are.
For those entering the profession, I think it is interesting to see how investment firms are evolving. Already, many investment management firms are finding that younger analysts are bringing with them new sources of information and search tools. This has the potential to help create alpha and bring an edge back into active management.
I will finish with some salutary research. Work by Bessembinder has pointed to how narrowly performance is based. The first, 2018, paper noted that in terms of excess returns over risk-free US Treasuries, just 4% of US stocks accounted for all the gains over a 90 year period. The later 2019 unpublished work covered global stocks over 38 years and 1.3% of stocks accounted for all the relative gains. The first paper showed that the median relative return for US stocks is -100%. Many investors fail to recognise these base rates; remarkably few stocks really perform but those winners could create great wealth, whilst over time it appears the majority of companies fail to perform at all or even to survive. Without studying stockmarket history, it can be instructive to pull out a Financial Times or WSJ from 30 and 50 years’ ago and check how many of the companies in the index are still around today.