Reading a comment the other day that noted how the stock market is today still at the level it was seven years ago sent me back to my database of long-run returns to see how unusual that seven-year experience is. To make the analysis more accurate, I looked at total returns (that is, including dividends as well as capital growth) in both nominal and real (that is, adjusted for inflation) terms.
I also looked at how the figures changed for various combinations of equities and gilts, to give some idea of whether a diversified investor would have been better off over seven-year periods. The data I used covers the period 1899-2000.
There is nothing magical about taking seven years as the holding period, though it has some attractions on mathematical grounds. The Rule of 72,still used by some senior professional investors, tells us that a 10 per cent annual rate of return will roughly double your money every seven years.
If you believe in mean reversion as a guiding feature of long-run stock market returns (as most serious investors do), you will also be familiar with the biblical parallel of a seven-year cycle of feast and famine.
A rule of thumb that assumes that if stock market returns have exceeded their long-run average over a period of seven years, then the subsequent seven years will reverse the trend, seems a plausible hypothesis.
In any event, seven years is arguably a sensible minimum holding period to think about when looking at prospective stock market returns. Most investors would certainly be disappointed not to make a positive total return over that sort of time frame. The question is then, how justified are they, looking at historical experience, in making that kind of assumption?
According to my calculations, the answer goes like this. Out of the 95, seven calendar-year periods between 1899 and 2000, there was only one in which the stock market failed to produce a positive total return.
That period, unsurprisingly, was the seven-year period to the end of December 1974, which happened to mark the all-time nadir of the post Second World War stock market. The average gain over seven-year periods, in nominal terms, was 110 per cent, which is probably about what you would expect, as that represents (thanks once more to our old friend the rule of 72) an annual compound growth rate of just over 10 per cent.
Allowing for inflation, that boils down to 6 to 7 per cent in real terms, the long-run rate of return recorded by equities in the 20th century. The highest seven-year return was 560 per cent, and the lowest minus-12 per cent.
You will not be surprised to discover that the period with the highest return was the seven years that began in 1975, immediately after the afore-mentioned bear market low point of December 1974.
Allowing for inflation does introduce a more sober picture. In real terms, there were 18 seven-year periods, about one in five, in which the stock market failed to produce a positive real return. The average real return for seven-year periods was 58 per cent. In the best, you would have made a return of around 240 per cent, and in the worst lost half your money.
In real terms, investors holding a representative market portfolio of shares have had about a one in two chance of making a real return of more than 50 per cent over seven-year periods, and slightly less than a one in four chance of making a real return of more than 100 per cent.
This is the traditional historical basis on which shares and equity funds are sold to investors, although of course you need to allow something like 2 per cent a year for transaction costs.
What about the returns from a diversified portfolio, defined here as a mixture of gilts and equities, or equities, gilts and cash (there being no really good long-run data for other asset classes, such as property and commodities that would now feature in a diversified portfolio)?
I looked at a portfolio divided 50-50 between equities and gilts. There was again only one seven-year period in which the portfolio failed to make a positive return in nominal terms, but more occasions (23 against 18) when investors would have experienced a negative return after allowing for inflation.
These were concentrated in three main high inflation periods, the years around the First and Second World Wars, and the late 1970s.
Looking at other combinations of equities and gilts, the trend that appears is that the greater the proportion of gilts in your portfolio, the greater the chance that you would have experienced a negative return in real terms over a seven-year period.
In fact, a portfolio with 100 per cent in gilts and nothing in equities would have lost money in nearly half the seven-year periods of the 20th century. Combine the higher probability of loss with the fact that a higher holding of gilts has also historically reduced the potential returns in a diversified portfolio and you can see why a generation of investors and their advisers, scarred by the experience of inflation, turned their back on gilts as an asset class.
The average seven-year return on a portfolio split 50-50 between equities and gilts in our covered period was 80 per cent in nominal terms and just 36 per cent in real terms. Investors had only a one in six chance of making a real return of more than 100 per cent over that time frame.
But as we have learnt many times, history is a good teacher but a poor master. In periods of low or falling inflation, fixed interest securities can do very well. While the returns data for seven-year holding periods makes the case for a high proportion of equities, when inflation is falling gilts can easily provide a much superior risk-return payoff.
Those who knew the power of mean reversion and increased their gilts holdings in the latter stages of the 1990s bull market as inflation fell, have had a happy time since. They have made strong positive real returns while equity disciples have mainly suffered. It is only a mild exaggeration to say just two simple metrics - the current seven-year return from equities and the outlook for inflation - will take you a long way to finding the right weighting.Reuse content