The problem with confidence is that it is hard to distinguish cause from effect. Do successful salesmen, sports teams and politicians get to the top because they radiate inner certainty, or are their winning smiles the result of victory?

Markets place great faith in the notion that consumer confidence surveys help to predict economic outcomes. Stock indices in the US show a marked tendency over time to move in line with the University of Michigan's Consumer Sentiment Index and the Conference Board's Consumer Confidence Index. Both have most recently shown consumer confidence clambering back to levels last seen in early 2008, before the financial crisis took hold. Yesterday, US consumer spending, on cue, came in at its highest annual rate of growth in two years.

Sadly, none of this should be regarded as predictive of the future. Both the main sentiment surveys, and consumer spending itself, remain at levels historically associated with recession. And these sentiment indicators may well just tell people what they already know – rather than predict improvement, they merely summarise consumers' responses to levels of joblessness, and to highly visible indicators such as stock market indices. Some academic research has suggested a small degree of predictive power, but this disappears once other widely available economic variables are taken into account.