Investors talk a lot about risk, but nobody seems to be able to define exactly what it is. We like to use metrics and terms – such as volatility, drawdown and ‘permanent impairment of capital’ – to capture it but our reliance on such measures is more because they are observable rather than right. While we don’t need to precisely calculate risk, understanding it is essential to sound investment decision making. So, how should we think about it?

Perhaps Elroy Dimson put it best, and most succinctly, when he said: “Risk means more things can happen than will happen.”

This definition gets to the core of how investors should consider risk. Risk is an absence of certainty.  Risky situations are defined by there being a range of potential future outcomes and some unknowable probability attached to those outcomes.

Imagine if I make a parachute from items I found around my house and then proceed to jump from the top of the tallest building in the world – this is an extremely high risk decision with the range of outcomes as wide as you could get (I live – unlikely, I die – probably).

Alternatively, if I decide to simply jump from the top of that same tall building (absent homemade parachute) that wouldn’t be a risky decision because the result is certain (or close enough to it).  Choices are not risky because they are bad, but because the consequence of them is uncertain.*

When investors assess risk they should always be thinking about the range of potential outcomes stemming from a decision and their likelihood. I think most people instinctively think about risk in this way – certainly our behaviour often suggests we do – but we are rarely explicit about it. This is perhaps because it is harder to think about than commonly used risk metrics, and we do love to be able to measure things – even if those measurements are deeply flawed.

Risk in Action

To bring this idea to life a little, let’s consider two types of financial market environment and what happens to our perception of risk during them.

In times of extreme market stress, such as that witnessed in recent weeks, our feeling is that risk is increasing and we tend to behave in a risk averse manner. Why is this? Two things are happening. Firstly, our sense of the range of potential outcomes is widening. Secondly, we start ascribing higher probabilities to bad outcomes.

There are behavioural explanations for both of these phenomena. Our expectations for the future are heavily influenced by what has happened recently, so when markets are volatile in the near-term that fuels a sense that future outcomes are becoming inherently more uncertain. Furthermore, we tend to judge the probability of events based on how available they are to us – that is how easy it is for us to see and imagine something. In the midst of a market sell-off we will place an increasingly high likelihood on negative future outcomes.

Something close to a reverse of this situation occurs during bubbles. In spells of market exuberance our sense of the range of future outcomes narrows around high and unrealistic return outcomes, and we think the probability of such positive outcomes is increasing.

Bubbles are about performance chasing, social proof and storytelling, but they are also about a growing complacency around the real risk of an investment. In a bubble risk is growing but we act as if it is dissipating.

Time on our side?

Investor time horizons have a huge impact on risk. When individuals extol the virtues of long-term investing and suggest it is less risky than adopting a short-term approach, what do they mean?

If we are investing in equities for one year – what does the risk look like? The range of plausible outcomes is incredibly wide (+40% and -40% is not unreasonable for diversified index exposure); furthermore, the probability of negative returns is not immaterial. While, even over just one year, equity returns are more likely to be positive than negative that is far from guaranteed.

If, however, we adopt a longer term philosophy – let’s say 30 years – the range of potential outcomes narrows and the probability of positive outcomes is materially improved. In simple terms, I would be far more confident that equities will generate positive performance over thirty years than a single year.

There are good reasons for this. Over one year equity returns are driven largely by changes in sentiment, over 30 years it is the compound impact of earnings / reinvested earnings that will dominate.

When we are comparing the risk of investments across different time horizons, it is far easier to understand the concept if we ask the question – what does changing the time horizon do to the range of outcomes? And what does it do to the probability of those outcomes?

But, there is a catch. Extending our investment horizon does not always improve the odds of good outcomes. In fact, longer horizons can become the enemy.

A long-run horizon is generally a very good idea for most investors, unless we have an investment approach which carries a meaningful risk of disaster. If a strategy has the capacity for catastrophic  or complete losses, long horizons work against us. If we are leveraged or concentrated then our risk increases as our horizon extends.

A three stock investment portfolio is a risky proposition with a very wide and uncertain range of outcomes. It is far riskier holding that portfolio for 10 years than one day.

Diversification and Risk

Although diversification is by no means a free lunch, it is an effective means of reducing and controlling risk, if done prudently. It works because by combining securities and assets with different future potential return paths it significantly constrains the range of outcomes of the combined portfolio.

If we move from a single stock holding to a diversified 50 stock portfolio we greatly lower the potential to make 10x our money, but also (nearly) entirely remove the risk of losing everything.

Diversification is a tool whereby we can (very imperfectly) create a portfolio with a range of potential outcomes that we are comfortable with. When individuals complain about over-diversification, what they typically mean is that the range of outcomes has been narrowed so that average outcomes are very likely. There is, however, no right or wrong level, it simply depends on our tolerance for risk. Or, to put it another way, our appetite for extremely good or extremely bad results.

The Stakes are High

When talking about risk as there being a range of future outcomes with some probabilities attached to them, I have missed out an important element – what are we putting at risk? What is at stake? A decision may be very high risk, but of little consequence.

If I bet one year’s salary on a horse race, that is a very different type of risk than placing a £10 bet. The risk inherent in the activity remains the same (the horse I have gambled on still has the same range of outcomes no matter how much I bet), but the risks to me are not comparable.

Risk is about understanding the range of outcomes and their potential probabilities, and then judging what the appropriate stake is given our view on this.

What is Good Risk Management?

I have tried to present what I think is the most realistic and sensible way for investors to think about risk, but what does that mean for good risk management? I think there are only three things that really matter:

– Understanding the potential range of outcomes and their probabilities.

– Reducing the probability of very bad outcomes (cutting the tail).

– Increasing the probability of good outcomes.

Of course, all of these things are difficult to do  – but then that is the point. If the range of outcomes is incredibly wide then we need to know that – it will impact the choices we make. We will never know the range of outcomes or their true likelihood, but we can make reasonably educated guesses. We also know how to protect against the inherent uncertainty of the future (through diversification and time horizons) and how to avoid things that may cause catastrophic losses (such as concentration and leverage).

Risk management often goes wrong and it does so both behaviourally and technically. Behaviourally, because we have a plethora of biases – extrapolation, overconfidence, availability, recency etc – which mean we worry about the wrong things and are complacent about issues that should matter. Technically, because we try to precisely measure things where it is impossible to do so – inevitably becoming overly reliant on inherently limited metrics. If we use a single number to measure risk, we are not thinking about risk in the right way.

We don’t know much about the future, but we know more things can happen than will happen. We should invest with this in mind. 

—-

* Risk and uncertainty are technically different things, but for most of us that distinction is not particularly useful.

–

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).