Cassandra's clueless heirs

This book is a collection of newspaper articles, essays and lectures that John Kay has produced during the past few years. The papers are grouped into four sections. The first contains an assessment of macroeconomics and specifically of forecasting, which constitutes many noneconomists' views of what economics is really about. The second is made up of the strategic issues facing firms. The third is a collection of articles on stakeholding. The final section is a miscellaneous grouping of amusing but incisive short pieces on topics as diverse as horse-racing, the Saatchis, and global media.

The book is written in a clear and accessible style and has wide appeal, in particular to those from other disciplines who are intrigued by the business world and want to know more about how it operates.

This latter group might imagine that they would also be given an insight into the world of academic economists. But in this they would be mistaken. Indeed, the one group for whom much of the contents will be anathema is the orthodox economists. For Kay is deeply critical, in a penetrating way, of a great deal of the conventional thinking in economics.

An important theme is the increasing marginalisation of much of academic economics. Any economist with direct experience of the business world, as Kay himself has, soon learns that managers have little interest in what conventional economics has to offer. Kay states quite clearly that while microeconomics in particular ought to be important to business, "because of the ways in which the subject has evolved over the past 100 years it has conspicuously failed to play that role". Conventional microeconomics is not merely dead, but condemned to perdition for its selfimposed irrelevance.

The opening section is an attack on macro-economic forecasting. The very poor track record of forecasters is well documented. Kay argues that the economy is inherently unpredictable, concluding that "when someone predicts the growth of manufacturing output in 1997, do not listen. He does not know." As a former forecaster, I share this view. Indeed, nonlinear signal processing can be used to demonstrate the point conclusively, showing that the signal-to-noise ratio in macroeconomic data series is extremely low, a fact economic modellers and econometricians find very hard to grasp.

The implications of this extend beyond mere forecasting. An army of econometricians estimates increasingly more complicated models in the search for equations that do not fail almost as soon as they are confronted with genuinely ex ante data. But the quest is inherently doomed.

Kay recognises that economic forecasts will continue to be made, "just as astrologers and quack doctors stay in business". He is scathing of the particular community of economists - mainly in the City - who continue to peddle these wares so confidently, and blames them for encouraging the credulity of noneconomists. His strictures are correct. But self-publicists in the City can always point to the fact that the Economic and Social Research Council continues to support macroeconomic modelling on a substantial scale, and such activity might therefore be presumed to have some academic credibility.

The behaviour of companies has long been Kay's speciality. This is reflected in the five papers in the second section, which are particularly stimulating and well argued. The content of each, such as the discussion of the strategies of BMW and Honda, is of interest in its own right. But the theme that connects them all is the devastating attack Kay makes on standard microeconomic theory.

There are two premises absolutely central to orthodox microeconomics. First, that the behaviour of economic agents - firms and individual consumers in everyday English - is based upon the principle of maximising utility. Second, that the behaviour of the consumption or production sector of the economy at the aggregate, macro, level can be described by the behaviour of one agent - the so-called "representative" agent.

Kay rejects both of these postulates. His criticisms are not new, and are similar to those made in technical terms by Alan Kirman in the Economic Journal in 1989. But Kay himself is a competent mathematician and has a complete command of the orthodox theory. It is precisely these qualities that make the criticisms of people such as Kay and Kirman all the more effective.

Kay argues that the principle of maximising behaviour in economic theory is "irrelevant" and that "real people and real institutions mostly do not maximise anything".

He asserts forcefully and repeatedly that each firm is different. Indeed, it is their inherent differences that explain why it is extremely difficult to replicate success by copying or cloning the structures and practices of other companies. But once we accept the point, the validity of the concept of the representative agent vanishes.

Kay's emphasis on the differences between firms allows him to mock the fads and fashions of management theory, each of which emphasises a unique way to corporate success, be it "total quality management" or "re-engineering", or whatever. For all its faults, orthodox economic theory is a serious intellectual construct, and criticisms of it need to be made at an appropriate level of rigour. But management theory, for Kay, is simply an object for ridicule.

He draws attention to the growing literature in economics on the theory of the firm that is outside the standard paradigm. Certainly, progress has been made, not least of which is the recognition in this literature of the principle of path dependency. The eventual outcome for any firm can be very sensitive to the initial conditions from which it starts.

Valuable work has been done in this area, but there are probably limits to how far models with analytical solutions - an approach Kay strongly prefers - can take us. A key assumption in economic theory is that the behaviour of any individual agent is influenced only indirectly by the behaviour of others through the workings of the price mechanism. Direct influences, such as the sudden desire to acquire a luxury car when the next-door neighbour is seen with one, are not permitted. But in reality such influences are pervasive. The problem is that models with this feature very rapidly become intractable analytically. However, advances in computer technology now mean it is possible to examine the properties of models containing individual, interacting agents using simulation. It is here that the really exciting possibilities exist.

Undoubtedly the weakest section of the book is on the amorphous concept of stakeholding. Kay firmly rejects a number of the more naive popular ideas that have come to be associated with it. For example, British capital markets are by no means short term, as witnessed by the success of our pharmaceutical industry, which requires very long time horizons on its investment.

His papers on stakeholding, however, lack the incisiveness of the rest of the book, and the concept has a quasireligious feeling to it. Kay argues that stakeholding in continental Europe has led to a system that in practice is similar to that of the lifetime employment principle in Japan. But unemployment in Europe is now 20 million, despite the widespread existence of consultation and works' councils - a fact brought home brutally by Renault's recent decision to close its factory in the very town where the Belgian prime minister lives.

Overall, though, this is a sparkling, well-argued book.

Paul Ormerod is chair of postorthodox economics and visiting professor of economics, University of Manchester.