BOOK REVIEW / Temperamental animal that bucks the trend: 'The Death of Economics' - Paul Ormerod: Faber, 14.99

DEEP in the bowels of the London School of Economics sits a machine that would not look out of place in the works of Heath Robinson. Built in the Fifties by Bill Phillips, an engineer turned economist, it is an imposing collection of tubes, valves and levers through which coloured liquids flow to represent the workings of the economy.

Bill Phillips' machine embodies the methodological approach which has come to dominate economics since the Second World War. Most practitioners of what Thomas Carlyle dismissed as 'the dismal science' regard the object of their study as a hugely complicated system of many interacting parts, but one which is essentially mechanical and predictable.

None the less there is little evidence that economists actually understand any better how the economy works than they did when Adam Smith first popularised the subject in the 18th century. After all, almost 40 million people in the industrialised world are unable to find jobs that pay a wage worth working for, while the boom and bust of the late Eighties and early Nineties went largely unpredicted and misunderstood by the economics profession.

Paul Ormerod, an economist who has done his own share of fruitless crystal-ball gazing, believes that economists should try to emulate biologists rather than engineers because the economy behaves more like a temperamental animal than a machine. But he is premature to conclude from this critique that he has dealt economics a fatal blow. Ormerod admits he is essentially arguing for a return to the 'political economy' practised by Smith and his contemporaries, in which the behaviour of the economy is seen as crucially dependent on the social, cultural and moral environment in which it operates.

He is rightly critical of the way in which real-world analysis has been supplanted by an obsession with the abstract notion of equilibrium borrowed from the physical sciences - the idea that the economy, if left alone, will settle down to a stable and probably optimal state. Ironically, advances in so-called 'chaos' and 'complexity' theory have already undermined in other sciences the assumption that complicated systems behave in a predictable and regular fashion. But economists have been slow to catch on.

The most pervasive example of equilibrium-fetishism in mainstream economics is the belief that there is a particular level of unemployment at which inflation does not change, and that if governments attempt to push unemployment below this rate, it will fall only temporarily and at the cost of permanently faster price increases.

But, in practice, this 'equilibrium' rate of unemployment is difficult to measure and, if anything, tends to track slowly the actual rate of unemployment, which renders it largely useless as a guide to policy. Also, the level of unemployment around which the economy fluctuates is inclined to jump unexpectedly, after which the economy behaves erratically for a while before settling down again.

This means that forecasting economic behaviour is probably doomed to failure in anything other than the very short term. But Ormerod's conclusions are not very different from those of mainstream economists - that the aim of policy should be to shift downwards the rate of unemployment around which the economy fluctuates. This involves promoting social consensus, investing more, training more, promoting company profitability and persuading people to share work. Ormerod may have set out on an interesting and neglected route, but he has arrived at a familiar destination.