It might seem a strange time to be making that claim, but despite its apparent absurdity I am now convinced that it is by far the most sensible way for an investor to look at the world. It may even be broadly true.

The efficient market hypothesis states that historical information provides no help in forecasting share prices. That would mean that examining graphs of a share's performance, even reading this morning's FT, would not produce a reliable strategy for judging the price of a share tomorrow or next year. That is because all useful information would already have been assimilated in today's price. Paul Samuelson, perhaps the most influential economist of the 20th century, summed it up in 1965 in the title of his article: "Proof that Properly Anticipated Prices Fluctuate Randomly." Since all available information is already reflected in the price, future prices will move only as news arrives. News itself arrives unpredictably, otherwise it is not news.

If the efficient markets hypothesis is true, then sensible economists will admit that they simply do not know what the outlook is for the stock market. How dull! It is much more fun to have somebody predict the future.