If you had immediately paused the game after England had taken the lead in the 55th minute of their World Cup Semi-Final match against Argentina, and asked the Argentine players and coaches how they would like their opposition to play for the remainder of the match, they would have probably said something like this:

• Give up on offering any attacking threat.
• Have no desire to keep possession.
• Defend incredibly deep and narrow.
• Allow Messi to operate unencumbered on the right, and Fernandez on the edge of the penalty area.

Conveniently, for Argentina, this is exactly what England decided to do, with inevitable consequences.

Why would a high-quality, talented team with a lauded coach opt to play in the exact fashion that suited their opponents, at the worst possible time?

Plenty has been said and written about this, and will continue to be so. I wanted to take a slightly different angle and focus on the psychological drivers. What would make a group behave in this way, when they are on top in a match and on the cusp of a long-awaited World Cup Final appearance?

There were two recognisable behaviours at play – outcome bias and loss aversion.

Outcome bias is our propensity to judge the quality of a process or decision by the results it delivers. If the outcomes are positive, so must have been the reasoning behind it.

England’s decision to play ultra-defensively when taking the lead against Argentina was no doubt driven by the view that it had worked well previously in the tournament – particularly against Mexico – so it made sense to repeat the trick.

There are two major problems with outcome bias, however. The first is the role of luck: sometimes our results are good even when the decisions that led to them were terrible – we were just on the right side of chance. The second is context – just because the outcome was good in one context, it doesn’t mean the same approach will prove a success when the environment is different.

In England’s case, both elements were at play. They were undoubtedly a little fortunate to hang on against Mexico, but crucially the context was entirely different. Against Mexico they were playing with ten men against a solid but limited team. They suffered no such numeric disadvantage against Argentina, and the current World Cup holders have a far greater attacking threat.

If England were going to judge the likelihood of adopting an extremely cautious approach, they would have perhaps been better off looking at their experience in recent major tournament semi-finals and finals, rather than matches in this World Cup alone.*

The other evident psychological phenomenon was loss aversion. We feel the pain of loss far more acutely than the pleasure of equivalent gains, which can have a profound impact on our behaviour. As soon as England took the lead, they went from a team trying to win something to a team aiming not to lose what they had. This was signalled clearly to the players by coach Thomas Tuchel’s defensive substitutions, which seemed to scream at his charges: “whatever you do, do not throw this away”.

When we view something as a potential loss, it can cause fear and anxiety – our responses become emotional rather than thoughtful. It is clearly possible to play in a conservative fashion in an intelligent and considered way, but this is not what England did. Instead, they defended like a junior football team tasked with maintaining a lead – put lots of defenders on the pitch, sit as close to your own goal as possible and boot the ball away when the chance arises.

The lack of evident thought about how to manage a game or hold a lead suggests that choices were driven by an emotional fear of loss, compounded by coaching decisions.

There were many factors that led to England’s depressingly predictable exit, but falling victim to some powerful behavioural biases and being unable to make clear-headed decisions certainly contributed.

—

* Outcome bias also means that if England’s approach against Argentina had worked out, it would have been regarded as a tactical masterclass.

—

===

My first book has been published. The Intelligent Fund Investor explores the beliefs and behaviours that lead investors astray, and shows how we can make better decisions. You can get a copy here (UK) or here (US).

I have a question for you. If you take all of the hearts from a standard deck of 52 playing cards, and then lay them out one by one, in how many different orders can the 13 cards be dealt?

For example:

J, 7, 5, A, 9, K, 3, 10, 6, 2, Q, 8, 4.

K, 4, 2, 9, Q, 6, A, 8, J, 3, 10, 5, 7.

The answer is…

6,227,020,800. That is over 6 billion different permutations of the 13 cards.

If you did the same with the full deck, the number of possible arrangements is so large that it is not worth even trying to write it down. It is not a number you would ever encounter in daily life. When you shuffle a deck of playing cards and deal them out, it is overwhelmingly likely that the order you reveal has never been seen before and never will be again.*

Trying to predict how the cards will land is as close to impossible as you might wish to get, but it is much easier than making accurate short-term financial market predictions.

Dealing out playing cards has a number of useful features from a prediction perspective:

• The number of cards is fixed.

• The rules of the game are stable and simple.

• You know all possible outcomes.

• When you draw one card, it has a known impact on the next step.

It is a stable, closed and well-ordered system. Financial markets are nothing like this. When investors are trying to forecast how the equity market might perform over the next year, or how a certain geopolitical issue will unfold, they are faced with a complex, chaotic and changeable environment.

It has all the features of a system where accurate predictions are not possible:

• It has an unfathomable number of inputs.

• There is significant path dependency, where each step profoundly influences what occurs next.

• The environment is adaptive – meaning the variables in the system react to what is happening.

• It is impossible to even comprehend the range of potential outcomes.

Despite this, a huge amount of time is spent espousing views and making investment decisions based on precise forecasts of an inherently unknowable future. Even in a very simple system, once the number of steps reaches even a modest level, the chances of making accurate predictions evaporate. Financial markets are anything but simple.

—

* I came across this idea in the excellent More or Less podcast, which you should absolutely be listening to if you enjoy numbers and statistics.

–

–

My first book has been published. The Intelligent Fund Investor explores the beliefs and behaviours that lead investors astray, and shows how we can make better decisions. You can get a copy here (UK) or here (US).

In a world where investors are increasingly being encouraged to react and trade (typically not for their benefit), doing nothing can often prove to be sage investment advice. In a recent paper, Hendrik Bessembinder analysed the performance of “do nothing” portfolios, and shows that even if we are doing very little, the results we achieve will be heavily influenced by the small choices we make.

To run his analysis, Bessembinder created a series of “do nothing” portfolios from the constituents of the S&P 500 at the end of each year from 1970.  The portfolios are “do nothing” as there are no trades after the portfolio is formed, apart from dividend reinvestment. If a stock is delisted, the proceeds are held in cash for the remainder of the period.  The underlying idea is to understand what would happen if we simply bought a selection of stocks on a certain day and then ‘left them in the drawer’. It is not quite a “do nothing” approach, but it gets pretty close.

I don’t want to simply repeat the findings of Bessembinder’s paper, and I would recommend reading it directly, but I thought I would draw out what I thought were the most interesting aspects of the results:

Rebalancing Matters

Over the full 55 year sample, an equally weighted “do nothing” portfolio (holding all available stocks at the same position size) grew from $1 to $678, while a value weighted approach (the typical index market cap structure) grew to only $361. This is a huge advantage for an equal weighted over a market cap strategy. Bessembinder suggests that this could be due to a ‘small cap effect’ boosting returns for the equal weighted approach. While this might explain some of the difference, I am doubtful that this is responsible for the entire gap – the underlying constituents are S&P 500 companies, so nothing is genuinely ‘small’. I think the differential is more likely due to rebalancing.

In the example above, the portfolios are not quite “do nothing” as they are rebalanced back to target weights at the end of each year. The impact of rebalancing is far more significant for an equal weighted approach where target allocations are fixed, unlike market cap where they move with performance. The paper also shows the impact of extending the rebalancing window to only every ten years. Here the gap between the returns of the equal weighted portfolio and the market cap portfolio collapses – the end value of equal weight is $387 and for market cap $342.

It appears that a substantial portion of the return advantage for an equal weighted “do nothing” approach came from a premium attributable to harvesting the volatility of the underlying assets over time. That is selling down stocks after periods of outperformance and vice-versa. Of course, a rebalancing effect does not only stem from an equally weighted portfolio, it comes from any strategy where the weightings of the holdings are not linked to the underlying price of the assets – equal weighting is simply one example.

Even when we wish to take a hands off approach to our investment strategy, there are seemingly innocuous decisions that can have a profound impact on outcomes.

The Two Types of Market Timing

I write frequently about the futility of market timing. I don’t think that people can predict the short-term movements of financial markets, and I wish they wouldn’t try. There is, however, another type of market timing that is far less deliberate, but can be even more influential – our starting point.

Although not a direct angle of the paper, Bessembinder’s work does show how the apparent efficacy of any given strategy is heavily dependent on the period over which we observe its returns.  One of the approaches analysed is simply to buy the largest stock in the index and then “do nothing”.  The fortunes of this admittedly extreme method depend entirely on when you start.

If you did nothing but owned Microsoft from 1999 you would have lost 33% over the subsequent decade, but if you had applied the same strategy in 2016 you would have held Apple and gained 1,046% over the next ten years.  

There are two important takeaways from this (aside from – please don’t invest your entire portfolio in a single stock). First, is that our investment outcomes will be influenced by the point in time when we invest – this will apply to a single stock or a highly diversified 60/40 portfolio. A great deal of this will be down to chance. Second, the more concentrated your investment strategy, the more beholden you become to the point in time you invest, because as concentration increases, so does the range of outcomes. Being diversified mitigates, but does not remove, timing risk.

The Concentration Conundrum

You may recall Hendrik Bessembinder from his previous paper which looked at how concentrated equity market returns were over time – just 46 firms accounted for over half of the $91trn in net wealth created over the course of a century, and the median buy and hold return across individual stocks was negative. When contrasted with Bessembinder’s new paper, this creates something of a puzzle.

His research on concentrated stock market returns implicitly supports taking a diversified, market cap-based approach to equity investing – this ensures that you have an increasing exposure to the companies that matter over time. This new paper, however, highlights the positive impact of an equal weighted approach, which benefits from harvesting a rebalancing premium from volatile equity markets, explicitly cutting the winners and bringing them back to their initial size.  

So, which is better?

It is important to note that the samples used for the two papers are different – his work on concentration incorporates a far greater number of companies assessed over a longer horizon. Although this means that the results of the two studies are not directly comparable, that distinction also brings us to the answer – it depends. The consequence of running winners rather than rebalancing is dependent on the prevailing environment. If performance over an extended spell is skewed towards a select group of large winners allowing concentration to build can pay-off, if this is not the case, the rebalancing effect can win out.

We cannot, however, know in advance if future returns will be concentrated and which companies will be responsible for that concentration. What we do know is that rebalancing and concentration are about trade-offs: rebalancing works because it prevents concentration from building, while a market cap approach works because it allows concentration to develop. The “right” amount of concentration is unknowable in advance, but the more concentrated your approach, the wider the range of outcomes you need to be prepared for.

—

Although not its express purpose, Bessembinder’s research is a timely reminder that there is no investing strategy that really constitutes “doing nothing”. Whether it is the precise approach to rebalancing we adopt, when we start investing or the particular index we select there are always choices that we need to make, and these will be consequential.

It is important to be deliberate when doing nothing.  

–

–

My first book has been published. The Intelligent Fund Investor explores the beliefs and behaviours that lead investors astray, and shows how we can make better decisions. You can get a copy here (UK) or here (US).