We live in the past. Try as we may – if, indeed, we do try – we can’t escape it. The past shapes the present like the remorselessness of a glacier shapes a valley. As a 3,000-year-old struggle set on the eastern shores of the Mediterranean performs its latest cycle of blood-letting and another – slightly less ancient – stains the steppe lands of eastern Europe, we shouldn’t doubt this.
Why then should we expect much different from the world of money? Financial con tricks are older than the Mesopotamian shekel and financial crises, really just con tricks writ very large, are almost equally old. They all come from a generic blueprint, based on leverage, moral hazard and the abuse of other people’s money. And the pattern is repeated; never quite the same, but near enough. The tools John Law used to pump up the Mississippi Company in early 18th century France had much in common with those used by Bernie Cornfeld and his Investors Overseas Services in the late 1960s. Then they were finessed to horrible perfection in an unstable mix labelled ‘collateralised debt obligations’ that ripped apart the financial system in the 2008 sub-prime mortgage crisis.
So, if the present is only ever an echo of the past, we should expect nothing very different from investment returns. They, too, will generate their distinct sonar as the performance of any asset class swings through its familiar staging posts – revive, pump, surge, peak, slide, crash and crumble – and goes from negative, to positive to ludicrous and back again. In that context, investment returns are predictable. We know where they come from and to where they will go. But that’s not quite enough. Infuriatingly, we’re never quite sure where they are in the here and now, or whether this time it really will be different.
To understand why, we have three charts using data mostly for long-term returns from UK equities, which we’ll assume is the default asset class for this magazine’s readers. The charts will tell us – as well as charts can – whether we should be invested in UK equities and, if so, what outcomes we might expect. Chart 1 is based on the notion that if investment returns really are predictable – at least, to an extent – then there will be a familiar pattern of good, bad and ugly years, which will reveal themselves in an approximate symmetry around their average outcome. This is what a chart plotting the normal distribution of some variable or other is designed to show and, in this case, it’s the year-on-year returns of UK listed equities since 1900 adjusted for inflation.
In a way, normal distribution is one of the world’s wonders. It tells us that, despite its complexity, life is almost as predictable as the results of tossing a coin. Flip a coin hundreds of times to see how often it lands on, say, heads six times running and the result will be close to the theory (one chance in every 64 flips). This is the result of an unbiased binomial outcome – the coin can only land on heads or tails and past results won’t effect future flips – and it offers an uncannily good estimate of what happens in real life. Events will cluster around their average chance of occurring and the incidence of extreme events will drop according to their extremity.
The theory of normal distribution works so well it is widely applied to financial tasks. Arguably the financial world’s most influential piece of maths, the Black-Scholes options-pricing model, runs on the normal distribution of asset returns. Yet even a glance at Chart 1 tells us that, if something is not completely wrong with this notion, it’s not entirely right, either. If returns were normally distributed, they would cluster around their average in a beautiful symmetry and, if calculus were used to fill in the gaps between the bands, then the chart line would show the classic bell-curve shape. The bell could be steeply curved or flat; but the point is, it would be symmetrical.
Chart 1 offers a pale imitation of symmetry. It drops annual returns for each of the 123 years 1900 to 2022 into 21 bands, each with a width of six percentage points. The average return for all 123 years is 5.5 per cent so intuition tells us that the middle band, for returns between 3 per cent and 9 per cent, should hold the most observations, as it does – 21. But, rather than falling away symmetrically, the bands diminish messily. Returns for the band 3 per cent to minus 3 per cent feature less often than theory says they should; returns of minus 3 per cent to minus 9 per cent more often, and so on. If it’s a bell curve, it’s hardly harmonious.
Nevertheless, useful information is conveyed by analysing how returns are likely to cluster around their average. A typical inflation-adjusted return of 5.5 per cent a year seems worth having, while expected variation around the average says that only in one year out of six are losses likely to be worse than 14 per cent. Yet, more tellingly, there is a huge gap between 121 of the individual returns clustered together and two solitary outliers at either extreme. These – and other wild years – offer another lesson in securities’ investing, which is better explained in Chart 2.
In an ideal world, there are periods when investors want to be in their market of choice and some when they don’t. Put another way, just a few periods can – and do – make a huge difference to overall investment returns, for better or for worse.
Actually, this is a generic phenomenon, known as a Pareto distribution or – more intuitively helpful – the 80/20 rule of thumb. Basically, this says that, for any quantifiable variable, 80 per cent of the output will come from 20 per cent of the input. So it’s a decent bet that, say, 80 per cent of the profit made by supermarkets operator Tesco (TSCO) comes from about 20 per cent of its outlets. Similarly, and often cited, 80 per cent of a nation’s wealth is owned by 20 per cent of its households. The effect shows up in investing. Eighty per cent of a portfolio’s profit is likely to come from 20 per cent of its holdings and, as discussed here, 80 per cent of its lifetime returns will be accumulated over just 20 per cent of its investment horizon.
The Pareto effect is clear in Chart 2, where the vertical bars plot each year’s return for UK equities going, left to right, from best to worst. Then the chart line shows the cumulative proportion of overall returns, which rises steeply and tops out at 181 per cent in year 73 before absorbing the effect of loss-making years. So it’s not a conventional Pareto distribution, which confines itself to positive values. Nevertheless, the message is clear – 80 per cent of the net returns are made from the 13 best years.
The value of the exercise is to grasp the significance – both for good or for ill – of just a few years in overall investment performance. Investors should have been desperate to be in the market when UK equities made their range-topping return of 120 per cent and should have hid in fear the year the return was a 58 per cent loss (see either end of Chart 2).
But there’s the rub. It would have been tough to have had one without the other. Those two extremes were successive years. In 1974 UK shares suffered a 58 per cent drubbing only to bounce back by 120 per cent the following year.
Which prompts another basic truth – that securities’ returns are like a random walk, higgledy-piggledy and unpredictable. Sure, it would have been marvellous to have been invested in any two of the 10 best years that UK shares have posted since 1900. Yet for those investing before World War 2 that would have been impossible; only one year – 1921 – figures in the top 10. So that favours post-War investors, especially – and perhaps surprisingly – those active in the inflation-afflicted 1970s.
Meanwhile, for anyone whose investment career started after 1990, forget it. Good though the 1990s were for equity returns, they began what looks like a long-term trend in declining returns. Chart 3 compares the five-year rolling average of inflation-adjusted annual returns for the UK’s FTSE All-Share index and its US equivalent, the S&P 500. We have to go back to the late 1980s for the most recent period the All-Share’s trend was clearly beating the S&P index; and for the past eight or nine years the lag has been wide and getting wider. What does this tell us about the UK’s prospects, both for its economy and for its listed companies? To be concluded next time.
bearbull@ft.com
