The Jonathan Davis column
Shares are still the best investment for long-term returns, but you could be looking at a very long term if you invest at the top of a bull market

Are you feeling confident about the stock market? When shares are riding high, the rising value of their portfolios gives most investors a warm feeling inside. Yet it is one of the paradoxes of the stock market that feeling good about the market is actually the time you should be worrying most.

So much, at least, I learnt from my recent trip to the United States, when I had the chance to catch up with the latest research by Jeremy Siegel of Wharton University.

If anyone can lay claim to being the person who has provided the intellectual rationale for the recent strength of world stock markets, then Professor Siegel is that man. His immaculately researched historical analysis of stock market trends, first published in 1994, established how remarkably consistent long-term stock market returns have been.

Tracing the data back to the start of the 19th century, Professor Siegel established what smart investors knew intuitively: that while shares are highly volatile in the short term, over longer periods their real rates of return are more certain than alternatives such as bonds.

"Despite extraordinary changes in the economic, social and political environment over the past two centuries," he concludes (and don't forget this period includes two World Wars, not to mention the Great Depression and the Opec crisis of the 1970s) "stocks have yielded between 6.6 and 7.2 per cent per year after inflation in all major sub-periods". What is more, the long-run real returns on shares in the UK and Germany (6.2 and 6.6 per cent per annum respectively) have been very similar - both being within 1.0 per cent of the US average.

Stocks are also more certain to provide real returns over time than either bonds or cash, the two main alternatives. Hold a portfolio of shares for 17 years, and there has never been a period when they have not provided a real return. In other words, unlike bonds or cash, stocks are in practice index-linked.

Of course, it does all depend what you mean by long term. Professor Siegel has cranked the numbers. What they show is summarised in the table. The trend is clear: the longer you hold shares, the greater the chance that your investment will outperform a similar investment in either bonds or cash (defined here as Treasury bills). In round number terms, there is a 60 per cent chance that stocks will outperform bonds over one year. This rises to roughly 70 per cent after five years, 80 per cent after 10 years; 90 per cent after 20 years - and, in effect, 100 per cent after 30 years. The figures for cash are of a similar order of magnitude.

The conclusion, says the professor, is that shares are "clearly the asset of choice for all investors seeking long-term growth". Well, amen to that, but he doesn't stop there.

There is, he goes on, "no compelling reason" for long-term investors to cut their holdings of shares "no matter how high the market seems". So, as long as you hang on long enough, you will always come out ahead. Even those who bought the so-called Nifty Fifty at the very top of the 1972 market, when fashionable growth stocks such as Merck and Xerox were selling for an average of 41-times earnings, would have made a healthy 12 per cent real return if they had held on to them through the great bear market of the 1970s.

Mind you, they would have had to have waited until the middle 1990s to catch up with the market as a whole but, even so, the patient investor who ignored the level of the market in 1972 would not have lost money as a result.

But can the professor really be saying that it is never right to sell shares, or that it can be justifiable to buy shares when the market, as now, is more highly valued than it has ever been on conventional criteria?

He himself admits in the latest edition of his book that he was worried by the example of a distinguished academic predecessor, Professor Irving Fisher, who famously predicted in October 1929 that "stocks are on a permanently high plateau". Just two weeks later, Wall Street crashed and we entered the worst bear market of the century.

Nevertheless, Professor Siegel is sticking to his view that shares have actually been "chronically undervalued" for most of history, with investors blinded to their real long-term value by the evident short-term volatility of the markets. It is the very fact that people are now wising up to the long-term value of shares, coupled with unprecedentedly favourable recent economic conditions, which helps to explain the depth and strength of the current bull market in the professor's view .

But can it last? At the end of the day, Professor Siegel concedes, markets are made by humans rather than by facts, however impressively documented. If investors come to believe the rates of return achieved in the last 15 years are in fact the norm, rather than nearly double the long-run average, then they will be disappointed when normality finally returns.

He quotes evidence which shows how periods of positive investor sentiment are usually followed by periods of market underperformance, and vice versa - the gloomier you are, the better the chances are that shares will do well in the immediate future.

"Fear", says the professor "has a far greater grasp on human action than does the impressive weighting of historical evidence." At the moment, of course, the feelgood factor is still very much with us and bargain hunters in the stock market are on short rations.