Evidence on the effectiveness of momentum investing is overwhelming – it has delivered a return premium across markets and asset classes, and there is over two centuries’ worth of data (see Asness et al. 2014).  Whilst the performance advantage delivered is difficult to refute, it does present something of a puzzle – why should a strategy that is both technically simple to apply and behaviourally attractive deliver excess returns? Before exploring this question, it is worth clarifying these two claims:

Momentum is Easy:  A standard time series, price momentum strategy is relatively easy to construct and operate. Although there are a vast range of iterations and nuances (this is the investment industry, after all) – the basic premise is to go long assets rising in value, and short those falling. The technical barriers to employing this strategy in some form are limited.

Momentum is Behaviourally Attractive: Most of us are wired to be momentum investors. Being part of the trend plays to our desire to be ‘right’, to be a member of the herd and to conform to the compelling narratives that price momentum inevitably generates.  Participating in a momentum trade is psychologically comfortable.

Yet whilst most of us are momentum investors of some variety (often indirectly), rather than obtain the documented return premium – we help to create it.  Our behavioural limitations mean that our attempts at discretionary momentum investing (driven by human decision making) are deeply flawed and incur a cost, which can be exploited by systematic approaches.  It is important to remember that the evidence upon which the case for excess returns to momentum is built necessarily relates to systematic / rules based strategies.

In his book ‘Following the Trend’ Andreas Clenow described a traditional managed futures approach – a basic trend following strategy – as such:

“A statistical game with a slight tilt in your favour and that you just have to keep throwing the dice long enough to get the law of big numbers on your side”

This description bodes ill for the ability of humans to capture a momentum return.  Victor Haghani and Richard Dewey produced a study where financially literate individuals (including some investment professionals) had to bet on the result of a coin flip, which was loaded in their favour 60/40. They started with $25, could bet stakes as small as $0.01 and were given 30 minutes. Despite this favourable setup, only 21% reached the maximum payout of $250 and 28% went bust. Even more curiously, 48% of players bet on tails (which had the 40% probability) more than 5 times. If our behaviour can be so irrational and erratic in such a dispassionate setting, it is no wonder that investors make consistently poor decisions in stimulating and emotion-laden financial markets.

To understand how momentum investing driven by human judgement serves to create a fecund environment for rules based momentum approaches, it is helpful to consider the central tenets of a typical systematic process and understand how this compares to the features of discretionary (or human) decisions:

Momentum	Systematic	Discretionary
Winners Purchased	Earlier	Later
Winners Sold	Later	Earlier
Losers Sold	Earlier	Later
Diversification	High	Low
Rules	Clear	Vague
Persistence	High	Low

The characteristic approach of a discretionary, ‘human’ investor described above is driven and shaped by a range of behavioural factors:

Buying Winners Late:  Investors tend to underreact to news that fundamentally alters the value of a security.  Whilst this phenomenon may be caused simply by the pace of information dissemination; it is more likely a result of commitment bias (our reticence to recant prior views) and also the gradual development of a new narrative – one piece of newsflow or data is unlikely to shift the prevailing market story, but if this persists the story surrounding it will build strength, drawing in investors.  Also, as highlighted by Mark Granovetter (1978), we all have a different threshold for joining the riot (or herd) simply based on how many other people are participating. Momentum begets momentum.

Selling Winners Early:  As detailed by Shefrin (2010), the disposition effect – the tendency to cut winners and run losers – has a material impact on asset prices and the momentum effect.  In the case of relinquishing winning positions, investor willingness to sell on good news (and capture gains) serves to slow the adjustment of an asset price to its new fundamental value, thereby helping to foster momentum.

Selling Losers Late:  Whilst arriving conspicuously late to the party, discretionary investors often overstay their welcome in loss making positions.  Investors are committed to their viewpoint and suffer from confirmation bias, so may simply ignore new information that runs contrary to the view implicit in their positioning. This situation is exacerbated by cognitive dissonance, which makes them unwilling to crystallise losses and acknowledge error.

Lack of Diversification: Spreading risk across a variety of positions and asset classes is crucial to all sound investment approaches, but is particularly relevant (essential) for momentum-driven investing.  The hit rate for individual positons can be low and predicting where momentum will arise (and be sustained) before the event is difficult. There is also the constant threat of sharp reversals, which can rapidly destroy gains made in previously successful positions.  The danger for discretionary investors is that they become overconfident in their ability to identify the best trades and also overweight the most recent market activity.  This can lead to exceptionally narrow portfolios focused on securities that have the strongest current momentum, rather than holding a diverse range of positions with varying levels of price strength.  Whilst systematic momentum approaches can become concentrated, they consistently maintain a broad opportunity set and (should) have prudent risk controls.

Vague Rules: The best defence against the impulses of emotion or narrative-led decision making is a set of decision rules, which dictate investment behaviour.  Systematic momentum approaches are founded on such principles, and actively divorce decision making from all intrusive elements except the key variable – which is typically price, but can include fundamentals, such as earnings.  Discretionary human decisions are rarely driven by a binding set of rules and, absent strict guidelines, our actions are impacted by a range of factors, most of which have limited relevance to the judgement at hand. These might include: how we feel at the time of making a decision (our emotional state), our prior experience with a particular investment, the most recent news we have read, our belief in a particular narrative, or even how hungry we are (Danziger,  Levav,  & Avnaim-Pesso, 2011), and the weather (Hirshleifer & Shumway, 2003).

Being behaviourally consistent without a set of rules to adhere to is close to impossible; this was highlighted by Daniel Kahnemann (alongside Rosenfield, Gandhi and Blaser) in an article published in Harvard Business Review in 2016.  The authors discussed how the influence of irrelevant factors can lead to huge variability in decision making; unlike biases, which tend to be consistent and persistent, the impact of noise is random and erratic – and therefore more difficult to mitigate.  Even with the best intentions, the ability of a human decision maker to remain disciplined is severely limited.

Low Persistence: Perseverance is undoubtedly a key requirement for successfully capturing the momentum premium, and its return patterns make this an exacting challenge. The hit rate is unlikely to be high – therefore you will be ‘wrong’ frequently; furthermore, there will be many false dawns where momentum appears and rapidly evaporates.  There will also be prolonged fallow periods in choppy markets where returns to momentum are poor, and short-term shifts in markets that see prevailing trends whipsaw and sharp losses incurred.  Whilst a systematic, rules based strategy can be agnostic on such a performance profile (and indeed specifically designed to withstand it); the foibles of momentum are hugely problematic for discretionary investors, for whom it would be a herculean task to remain sufficiently controlled amidst the welter of behavioural impediments.

Whilst most of us are invariably attracted to momentum investing and carry it out in some form (even if it is merely implicit in our actions, rather than an express choice), few of us do it well.  Our decision making is blighted by our behavioural shortcomings and the substantial influence of irrelevant ‘noise’.  Human decision makers are willing but inferior momentum investors, creating the opportunity for systematic approaches to capture a premium.

—

Key reading:   

Asness, C. S., Frazzini, A., Israel, R., & Moskowitz, T. J. (2014). Fact, fiction and momentum investing.

Clenow, A. F. (2012). Following the trend: diversified managed futures trading. John Wiley & Sons.

Danziger, S., Levav, J., & Avnaim-Pesso, L. (2011). Extraneous factors in judicial decisions. Proceedings of the National Academy of Sciences, 108(17), 6889-6892.

Granovetter, M. (1978). Threshold models of collective behavior. American journal of sociology, 83(6), 1420-1443.

Haghani, V., & Dewey, R. (2016). Rational Decision-Making Under Uncertainty: Observed Betting Patterns on a Biased Coin.

Hirshleifer, D., & Shumway, T. (2003). Good day sunshine: Stock returns and the weather. The Journal of Finance, 58(3), 1009-1032.

Hurst, B., Ooi, Y. H., & Pedersen, L. H. (2017). A century of evidence on trend-following investing.

Kahneman, D., Rosenfield, A. M., Gandhi, L., & Blaser, T. (2016). Noise: How to overcome the high, hidden cost of inconsistent decision making. Harvard business review, 94(10), 38-46.

Shefrin, H. (2010). How the disposition effect and momentum impact investment professionals.

Heuristics, or what we might call ‘rules of thumb’, have become somewhat maligned as a method of decision making in recent years. They are often erroneously conflated with cognitive biases and suffer their negative connotations.  Furthermore, the scepticism regarding the effectiveness of heuristics has been exacerbated by the ubiquitous system one (fast) / system two (slow) thinking construct (popularised by Daniel Kahneman), which can frame system one as the domain of rash, unthinking errors and system two the home of cool, rational and correct calculation.  System one and system two is frequently used as shorthand for ‘bad’ and ‘good’ thinking respectively, with the mental shortcuts that define heuristics firmly in the former grouping.

Even without exploring the limitations of the ‘two systems’ theory (see Osman, 2004); it is clear that heuristics do not fit within such a binary framework.  They can be either instinctive or deliberate, and often deliver results that are more robust than more complex methods of judgement.  For example, see Serwe and Frings (2006) on predicting the result of Wimbledon tennis matches, and McCammon and Hageli (2007) on avalanche risk.  Andrew Haldane’s speech around this topic is also excellent.

Unsurprisingly, the best definition of heuristics comes from Gerd Gigerenzer, who has been the pre-eminent voice on this subject, and the most vocal critic of the work of Kahneman.  He defines the heuristic approach as such:

“A strategy that ignores part of the information, with the goal of making decisions more quickly, frugally, and/or accurately than more complex methods” (Gigerenzer & Gaissmaier, 2011).

Although heuristics do feature in the world of investment there is an undoubted preference for complexity.  This is understandable – financial markets are intricate, unpredictable and often unfathomable; therefore we struggle to believe that simple solutions can be effective – a complex problem must be untangled by a complex process.  Furthermore, it is difficult to justify fees (and our own existence in the industry) by recommending quick and easy solutions.

This is not to suggest that everything in investment can be improved through simplification; rather we should not pour scorn on approaches that seek to streamline decision making in the midst of high uncertainty or difficulty. Sensible heuristics should be viewed as a valuable tool for investors, and below I have suggested five simple rules.

– Jam Yesterday Heuristic:  If an investment has produced strong returns in the recent past, reduce your expectations for future performance.

Whilst this is self-explanatory, it often seems to run contrary to the behaviour of most investors. The gains registered by an investment before you decided to purchase, are returns you will not be enjoying in the future!

– Upside Down Heuristic:  If an investment has outperformed its benchmark by a significant margin on the upside, it can also do so on the downside by a similar magnitude. If this is not acceptable, do not invest.

The lure of funds that have generated material excess returns can be irresistible, but if they have taken sufficient active risk for their performance to diverge markedly on the upside, they could generate similarly divergent returns in a less favourable fashion. Whilst index +20% is attractive over three years, only invest if you are willing to withstand the reverse 

– Keep it Simple Heuristic:  If presented with similar investment choices that differ in terms of complexity, select the simpler option.  

Simple investments are certainly no panacea, nor should they always be the preferred choice.  However, in situations where it is difficult to differentiate between alternatives in terms of quality, the preference should be towards that which is most easily understood.  Having clear and well-calibrated expectations is crucial for any investor.  If an investment strategy is straightforward, it is less likely to deliver negative surprises and you are more likely to stay invested for the long-term.

– Hedge Your Bets Heuristic: If you have any uncertainty over the timing of an investment, phase it into the market in equal tranches.

Phasing investment decisions is an effective behavioural trick, not only does it reduce the potential reference point impact of a single price on entry or exit; it also allows for the positive framing of your decision. If the price of your targeted asset rises as you stagger your investment you can feel content that you initiated the purchase when you did, whereas if it falls you can be glad that you decided not to invest in one hit.

– Double for Drawdown Heuristic:  Your absolute minimum expectation for drawdown in any diversified investment (outside of cash) should be double its expected long-term volatility.  If this is not palatable, do no further work.

This rule is not a comment on the appropriateness on volatility as a measure of risk, nor the expected distribution of asset class returns (which are often skewed); rather it seeks to translate the most commonly employed investment risk metric (volatility) into a simple yes / no decision based on a willingness to bear (temporary) losses.  Of course, volatility is a woefully inadequate measure of risk for illiquid investments with artificially low volatility, and both undiversified or ‘penny in front of steamroller’ strategies, which suffer from sharp tail risks.

None of these heuristics is likely to be revelatory, but amidst the myriad options that investors face as they seek to meet their financial objectives, it is easy to lose sight of crucial principles that can lead to improved investment behaviour and better long-term outcomes.  Although lacking the intellectual cachet of more sophisticated approaches; easy to remember, simple to apply rules of thumb can be remarkably effective way to navigate a complex investment landscape.

—

Key Reading:

Gigerenzer, G., Todd, P. M., & ABC Research Group, T. (1999). Simple heuristics that make us smart. Oxford University Press.

Gigerenzer, G., & Gaissmaier, W. (2011). Heuristic decision making. Annual review of psychology, 62, 451-482.

McCammon, I., & Hägeli, P. (2007). An evaluation of rule-based decision tools for travel in avalanche terrain. Cold Regions Science and Technology, 47(1-2), 193-206.

Osman, M. (2004). An evaluation of dual-process theories of reasoning. Psychonomic bulletin & review, 11(6), 988-1010.

Serwe, S., & Frings, C. (2006). Who will win Wimbledon? The recognition heuristic in predicting sports events. Journal of Behavioral Decision Making, 19(4), 321-332.