Most readers of Investors Chronicle, I suspect, have more than half of their financial wealth in shares or equity funds.
For years, this was reasonable. Some research back in 1969 by Robert Merton, a Nobel Prize-winning economist at MIT, showed why. He proposed a simple rule for asset allocation. The proportion of your wealth you invest in equities, he said, should be equal to the expected equity premium (the amount by which shares are expected to outperform safe assets) divided by the product of the variance of equity returns and a measure of your risk aversion.
According to Credit Suisse, equities have outperformed bonds by 3.7 percentage points per year since 1900. Plugging this number into Merton's equation gives us equity exposure of around 50 per cent for reasonable assumptions about volatility and moderate levels of risk aversion. This is the logic for big equity exposure.
There's just one problem with it: you have to be very old or very lucky to have enjoyed an equity premium as high as 3.7 percentage points per year.
My chart shows the point. It shows the gap between annualised returns on the All-Share index and gilts for people who had bought in each month since 1985. It shows that you would only have enjoyed an equity premium as high as 3.7 percentage points if you had bought at troughs in the market such as in 2003, 2008-09 or 2011. More likely, you'd have seen only a small equity premium, or even none at all. Since the late 1980s, the premium has only been around one percentage point.
UK equity premium
This makes a huge difference. Plugging an equity premium of one percentage point into Merton's equation generates equity weightings of only around 20 per cent.
If the past 30 years are any guide, therefore, many of us are hugely overexposed to equities.
Worse still, there are powerful reasons why the equity premium should remain low.
Back in 1985, Rajnish Mehra and Ed Prescott published a famous paper in which they argued that, in theory, the equity premium should be less than one percentage point - essentially, because shares weren't risky enough to warrant much higher returns. From this perspective, it is the past 30 years that have been normal, while the high equity returns of the 20th century were freakish.
And it's easy to see why returns were so high back then. For much of the 20th century, there were existential threats to shares, including the possibility of war, revolution (even in the mid-1970s there were serious doubts about whether capitalism could survive) or nuclear annihilation. These risks caused shares to be underpriced for much of the 20th century. As the risks receded in the 1980s, shares soared - to levels from which subsequent returns have been in line with messers Mehra and Prescott's theory. The 20th century was, as in the title of a book by Paul Marsh and Elroy Dimson, the triumph of the optimists.
However, although the big risks to equities did not materialise in the 20th century, one big risk to gilts did; from the 1960s to the 1990s these were ravaged by unexpected inflation.
It might be, therefore, that the equity premium our fathers and grandfathers enjoyed in much of the 20th century was due to one-off events: the pricing out of existential risks caused equities to do better than they should have, while unanticipated inflation hurt bonds. If this is the case, then perhaps it is the post-1980s experience of a low equity premium that is normal and which should be the basis of our asset allocation. Which means that many of us are holding far too many equities.
So, what could be wrong with this thinking?
It's not that shares are now pricing in a big risk and so would rise a lot over the long term if this risk recedes. The dividend yield on the All-Share index is now 3.7 per cent, which is only a smidgen above its post-1985 average of 3.6 per cent. This suggests shares aren't greatly underpriced.
There is, though, another possibility. From another perspective, shares should deliver decent long-run returns. If we assume that the economy will grow by 2 per cent per year - a little below its postwar average - that dividends grow in line with GDP and that the dividend yield doesn't change, then we should expect annual equity returns of 5.7 per cent - two percentage points of capital growth plus 3.6 percentage points of yield. With long-term gilts yielding 2.6 per cent, this implies an equity premium of 3.1 percentage points.
From this perspective, a big equity weighting is justifiable for the more risk-tolerant investor.
This gives us a dilemma. Two reasonable perspectives give us very different long-run outlooks for equity returns and therefore very different implications for asset allocation.
Personally, I have a simple solution to nasty dilemmas: I try to avoid them. I have a high equity weighting in winter (over 50 per cent) but a low one in summer (under 20 per cent); this reflects the fact that equity returns are seasonal. If this pattern continues, I don't need to think about the long term. Which is just as well.
