I was speaking with a friend recently and, as they also worked in the industry, our conversation inevitably turned to investment. We were discussing the difficulty of different aspects of the job, in particular the contrast between asset allocation and security selection. They were of the view that asset allocation was inherently more challenging, whilst I held the opposite position. As I reflected upon this discussion, I realised how important this issue was. If investors are undertaking an activity to improve outcomes it is critical to understand how difficult the task is and why. If we can answer this, we can judge whether it is worthwhile and what is required to carry it out successfully.
So, what is more difficult, asset allocation or security selection?
Before considering the specifics of the question, it is important to define some terms. If we are asking what aspects of investment are more difficult, we need to understand what we mean by it. We can do this by framing it as a simple problem:
How easy is it to adopt and maintain a process that can improve the odds of better outcomes?
The issue of difficulty is really about skill – the capacity to consistently link process and outcomes in an intentional and positive fashion.
And that brings us to the first question we should consider when considering the difficulty inherent of a given task:
Is it an activity where skill can influence the results?
Before even beginning to worry about the relative difficulty of anything, it is critical to understand whether we can consistently and positively impact results at all. Some activities are just too hard, and the results of our endeavours will be dominated by randomness. We can apply two tests to identify such situations:
Reasonableness test: Is it reasonable to believe that skill can influence an activity and offer some advantage? This is a simple sense check to avoid the pitfall of seeing patterns in random outcomes. If we get enough people flipping coins some will get five heads in a row, that does not mean it is reasonable to believe it is a task where skill can exert an influence.
In a field as complex as financial markets it is easy to get immediately lost in the weeds. It can save a great deal of time to start with first principles: define what the essence of the task is and the core assumptions that must hold to believe that skill can be a factor.
Evidence test: Is there evidence that the potential to influence outcomes with skill exists? If we believe that an activity passes a reasonableness test, we need to validate this by looking at the data. Do historical outcomes support the idea?
We will have an example of an investment activity that may fail both of these tests later.
Once we have made a judgement on whether the results of an activity are likely to beholden to randomness, we can move to the next question:
What is the nature of the skill?
We can often fall into the trap of believing that investment difficulty is just an issue of technical complexity. It is simple to assume that the more technically challenging a skill is, the more reward we might enjoy from performing it well. This is a misnomer. Technical factors are one aspect, but equally important are the behavioural facets.
Technical skill is where there is inherent complexity in the task and expert knowledge is a necessary but not sufficient condition. Behavioural skill is where the task itself might be simple, but our ability to make the rational decisions required to improve outcomes is compromised. Both are types of processes that can lead to better outcomes, but the requirements to execute them successfully are entirely distinct.
If we know what type of skill is required, we can attempt to judge how difficult it is to perform. Which brings us to the final question:
How difficult is it to undertake the skill to improve results?
It is not sufficient to know whether skill can exert an influence on performance, we need to know how hard it is to successfully perform it. Mowing the lawn and flying a 747 are both tasks where skill influences outcomes, but one is far more challenging to master than the other.
Difficulty matters for investors. Just because the application of skill might improve our results, it doesn’t mean it is a good idea to try. Not only because something might be so challenging that it is a waste of our time to even attempt it, but because our attempts might lead to worse outcomes than if we had not even tried.
The difficulty of any investment activity is also context dependent. For professional investors, the working environment is critical. If technical expertise is required, then the abilities of colleagues and quality of systems may greatly influence the feasibility in exercising a skill. If the skill is behavioural then steps must be taken to neutralise problematic biases. It is impossible to capture a behavioural premium from adopting a long-run approach (for example), if everyone is obsessing over quarterly results. Simple skills can quickly become impossible to perform.
Having created a framework for assessing difficulty and feasibility, we can consider the question posed in the title of this post. To muddy the waters slightly, I will address strategic asset allocation (long-term) and tactical asset allocation (short-term) separately, as they are markedly different endeavours.
Tactical Asset Allocation
Tactical asset allocation – adjusting portfolio exposures over short time horizons based on market conditions to improve returns or reduce risk – is undoubtedly the most difficult activity of those that we will discuss. It falls at the first hurdle. It doesn’t even pass a reasonableness test.
Is it reasonable to expect any person or team to confidently and consistently predict the near-term movements of a system as staggeringly complex and adaptive as financial markets? It is hard to contend that it is.
Not only is it an unreasonable expectation. I am yet to see any compelling evidence that investors can add value by calling market directionality or asset class performance in the near term on a reliable basis.
If tactical asset allocation is so difficult and there is so little evidence of it working, why is it commonplace? It is likely a behavioural phenomenon. It is the very fact that financial markets are so chaotic and unpredictable that makes the idea of tactical asset allocation so alluring; because markets are turbulent and changeable there is an overwhelming urge to act. Something is happening – do something about it!
In an absurd fashion, it is often easy to be viewed as a negligent investor by do nothing even when it is overwhelmingly the most sensible course of action.
Tactical asset allocation is really about creating narratives, stories to talk to investors and clients about. The main advantage that stems from tactical asset allocation is the comfort clients may receive from seeing activity in their portfolio through difficult market conditions. This might help them stay the course. Despite this potential benefit, my sense is that money would be better spent educating investors on how boring and long-term investment should be, rather than using the indirect and uncertain behavioural benefit of tactical asset allocation.
In a ‘traditional’ multi-asset portfolio, tactical asset allocation is usually applied relative to a neutral allocation within some form of risk tolerance. For example, a portfolio can move underweight or overweight equities based on an assessment of market conditions. Attempting to adjust such exposures seems a futile task. Provided the time horizons / objectives of the portfolio are sufficiently long, the rational, evidence-based approach would simply be to permanently overweight equities within the portfolio’s risk tolerance and capture the long-run premium. This would get the probabilities of good outcomes on our side far more than attempting to time markets.
There are two problems with such a simple approach: 1) We cannot charge for it in the way we might for tactical asset allocation. 2) We will need to endure the difficult spells of equity performance to capture the probable return advantage. Neither seems like a compelling reason to persist with tactical asset allocation.
If short-term asset class timing is so difficult, what about if we adopt a longer-term approach?
Strategic Asset Allocation
Compared to tactical asset allocation, adding value through strategic asset allocation – the long-term (ideally 10 years+) mix of assets in a portfolio – is easy. If you don’t believe me, answer this question:
Maintaining prudent risk tolerances (say +5/-5% allocation changes or 1% tracking error budget) how confident would you be in outperforming a 60/40 portfolio over the next ten years by making strategic adjustments to the allocation?
I am not confident about many aspects of investments but my answer to this question would be somewhere around 80%. Why so high? Because I would simply overweight equities and underweight bonds. As the time horizon extends, I become more confident in receiving an additional return for the additional volatility. Risk-adjusted returns might be a little worse, but that is acceptable if you remain within the mandated risk tolerances and of course: ‘you can’t eat risk-adjusted returns’.
There are inevitably challenging scenarios where long-run equity returns from here are inferior to bonds, but that is captured in the 20%. If markets take one of those paths my modest overweight to equities is probably going to be the least of my worries.
Even if we take a more intricate approach, there is still evidence to suggest that over the long-run the returns from cheaper assets will outstrip more expensive assets. Over ten years I think that emerging market equities outperform US equities, but my confidence would be nowhere near 80%, probably closer to 65%. It is still, however, an evidence-based edge. It is also one that passes a reasonableness test – is it reasonable to expect cheaper major asset classes to outperform more expensive ones on average over the long-run? I think so. There is solid evidence supporting this type of strategic asset allocation behaviour and it is not technically difficulty, but there is a major problem – behaviour.
Although most investors time horizons are more than ten years, really none of them are. We care far too much about short-run outcomes, so our ability to capture improved returns from simple strategic asset allocation adjustments is incredibly limited. Whilst we can have a high confidence in a structural overweight to equities, can we maintain it through the inevitable bear markets – can we keep our job? Imagine attempting to hold an overweight to emerging market equities and an underweight to US equities position through years of thumping US outperformance. Although the odds may be on our side, nobody will believe they are.
It is not technically difficult to make prudent, long-run asset allocation decisions that get the probabilities in our favour, but few of us operate in an environment where we are able to exploit that advantage.
How do the behavioural challenges of top-down, strategic asset allocation compare to the bottom-up complexities of security selection?
We can consider security selection to encompass picking a collection of individual stocks or funds to improve upon the returns of a default, market cap index. As with SAA and TAA, we can split the activity into two elements: identifying companies that will improve on the returns of the market (stock picking) or structuring a combination of stocks to have attributes or biases that are distinct from a market cap approach (factor tilts). These are not mutually exclusive, but for the purposes of assessing difficulty it is helpful to treat them separately.
Stock picking is incredibly difficult – the evidence clearly attests to this – only the minority achieve exceptional long-term results. Unlike tactical asset allocation, however, it does not fail the reasonableness test. Is it reasonable to believe that it is possible to pick a collection of companies with better attributes than a selection based purely on their size? Yes, it probably is. It is just very hard. Technically and behaviourally. Even if we have a technical edge that can add value, there will be arduous periods when it will appear that we don’t.
By contrast, factor tilts are not technically difficult to achieve. They are simple and commoditised. It can be accomplished by deviating from a market cap weighting methodology with virtually any other approach. There is also ample evidence that – over a long enough horizon – established factors provide some form of return advantage. It is also reasonable to assume that buying cheaper, higher quality, smaller sized etc… companies, on average, leads to better results.
The challenge of factor tilts in security selection is not technical, it is behavioural. Even factor tilts with vast amounts of empirical evidence as to their efficacy will underperform for prolonged and painful periods. This is a requirement rather than a bug, if there wasn’t some discomfort in holding a factor-tilted approach, there likely would not be a return advantage.
What’s the answer?
The answer as to what is more difficult: asset allocation or security selection is – it depends. We need to understand the specific details of the skill we are attempting to perform before making such a judgement. Strategic asset allocation may be a more worthwhile endeavour than tactical asset allocation but comes with its own profound challenges. Whereas factor-based stock investing is theoretically easy, but practically tough.
The question is also inherently subjective. Reasonable people may disagree and the difficulty of any investment activity will be highly dependent on our environment. All investors, however, should be explicitly considering the likelihood that they can deliver positive results from the task they are undertaking.
We must remember that most activities that offer the potential to deliver improved returns will either be complicated (technically) or painful (behaviourally), sometimes they will be both. Whatever the cause of the difficulty we need to understand it if we are to have any hope of achieving better outcomes.