FOOL'S EYE VIEW
The $64m Question

By Stuart Watson (TMFTiger)
March 4, 2002

Carburton Street, London -- A few weeks ago a new book was published claiming that the actual returns from shares over the last century have been lower than previously claimed. It is called Triumph Of The Optimists and a summary of its findings can be seen here. Whilst it may sound like a dull and dry subject (ok, ok, it is) the ramifications for your wealth are very important. It dictates how you might expect your ISAs and pensions to grow and how much income you might expect to have in retirement. A lower growth rate means you have to put more in to get the same amount out at the other end.

As you might guess from its title, the book claims to prove the long-held assumption of optimists (like the Motley Fool) that shares do indeed deliver better long-term returns than bonds or cash. But it also claims that the difference is not quite as wide as previously thought. The number they come up with a figure of 5.5% per annum since 1900 as the figure for the 'real global equity return' (i.e. the return from owning shares after you take off inflation). That compares with figures of between 6% to 8% from most other reports. Real returns are much more important that absolute returns because they measure the increase in spending power of your investments. A return of 12% is of little use if inflation is 15%. You end up with less spending power than you started with!

Differing Growth Rates

Like many such reports this one rubbishes the methodology used in previous studies. From our external viewpoint it's impossible to say who is right or wrong, although the summary of this report does seem to make all the right noises. The basic problem is finding accurate values for initial data or, in other words, the make-up of the stock market 100 years ago. If you only look at companies still around today then you ignore the negative impact of companies that have closed down. Throw in mergers, demergers, flotations, privatisations, nationalizations, dividends, rights issues, share buybacks, silly name changes, not to mention the different growth rate of different countries, and pretty soon the whole thing expands to nightmarish proportions. It's no wonder that there is so much disagreement about what the true figures are!

In fifty years time we'll have much better data because it is much easier to collect data as you go along rather than reconstructing it at a later date. But for those of us saving and investing now that is, of course, of little comfort. We're stuck with these best guesses.

The Wealth Of The World

The wealth of the world is a closed system. All assets have a value now and they will have an unknown, but probably higher value, in say, 10 years' time. The difference between the two values represents the increase in everyone's net wealth. The distribution of this wealth is determined by the trading of these assets and who holds which assets at various times. No new wealth is created by this trading, it merely shuffles the pack. Some level of trading is necessary as the gains that are made have to be realised. You may think your house is worth £250,000 but unless someone else thinks the same way, and is prepared to pay that amount, it remains a best guess.

Cash and bonds, in themselves, do not create wealth, they are merely payments for borrowing money for a set length of time. Shares and property create both income streams and the prospect of capital growth as their value is linked to the growth of the economy. So whilst some may think it is merely optimistic to expect shares to do better, there are, in fact, good reasons for such an assumption.

The new data doesn't really change the fact that bonds and cash make terrible long-term investments. It's possible that there could be a 20-year period where they do better. But even then the difference is likely to be small and the chances of accurately predicting it in advance are non-existent. You can only do so by luck.

So the $64m question is not whether shares and property outperform cash and bonds. It is by how much you  can expect them too, on average. Of course the value of shares and property do not go up in a nice straight line. The longer the period the more likely it is shares and property will do better. When you are investing for your retirement, you are investing over a large number of different time periods. Whilst they overlap they reduce the chance of underperforming cash and bonds still further.

A Lower Growth Rate?

So should we all assume a lower growth rate when playing with our compound interest calculators? Perhaps. More recent reports have tended to come out with lower growth rates, although the falls of the last two years have played a small part in that. It's reasonable to assume that the methodology improves as time goes on, and so each successive report is getting more accurate.

However, any study that looks at a fixed period is subject to bias. We have traditionally used 1918 as a start point but including the period from 1900 to 1918 (where equities actually fell by 30% in real terms according to one study) is obviously going to dampen the overall number. Why not include the 20 year period before that when shares almost quadrupled in real terms? Longer time periods are better as they iron out short-term distortions but, on the flipside, the data gets worse and worse the further you go back. It's also likely the stock market represents a smaller and smaller percentage of the companies in operation the further back you go. In other words, a greater percentage were privately owned, so the quoted companies could give a misleading sample of the economy as a whole.

In conclusion, it probably does make sense to lower your expectations about future growth. Being overambitious increases the risk of falling short of generating enough wonga to retire on. If it turns out that you were indeed overcautious then you might be able to reduce your rate of saving as retirement nears or console yourself by spending or distributing the extra cash as your heart desires.