Beyond rationality and towards understanding

Diane Coyle has led an interesting and varied life as a professional economist. After Oxford University and a doctorate at Harvard University, she became chief economics writer for The Independent . She runs a consultancy company as well as being the author of several provocative and unconventional books on economics.

In The Soulful Science , Coyle brings to bear this rich experience to reflect on developments within the discipline of economics over the past 20 years or so. Economics has indeed been changing in exciting ways, and she describes these in a clear and authoritative manner.

Very substantial progress has been made, for example, in the apparently mundane area of data gathering. For much of the 20th century, most of the data available to economists was at a highly aggregate level - such as estimates of, say, total spending in the economy over time. Because much of economic theory consists of postulates about individual behaviour, such aggregate data is of little scientific value. Indeed, modern signal-processing techniques show that this kind of data is dominated by noise rather than by true information. Despite the best efforts of econometricians, very little progress has been made in resolving disputes within economics using such data, as Coyle points out.

But economists now have access to data sets that contain information not only about specific individuals and their behaviour, such as their detailed spending and saving patterns, but also about how these change over time.

This panel data (longitudinal data in the jargon) are capable of providing more reliable insights into behaviour.

Data are of course no more than necessary conditions for serious work to be carried out. Along with better data, economists, like everyone else, have powerful computers. And they also have the techniques. Americans such as the Nobel laureates James Heckman and Daniel McFadden have led the way in developing the difficult statistical techniques required for the analysis of such data.

The combination of data, powerful computers and technique is by no means a panacea, but genuine progress has been made in understanding a range of important problems. Perhaps the most powerful insight to emerge from this work is that individuals are heterogeneous. In other words, people are different. Hardly surprising, one might think. But much of conventional theory is built on the assumption that the economy can be described using a single so-called representative individual. Instead, differences are a fundamental feature of the real world.

Coyle is by no means starry-eyed about economics, and she argues that some of these developments have their downside. In particular, excellent econometric programmes combined with huge computing power make it easy to generate large amounts of statistical analysis with no real understanding of what it all means, a weakness that many young economists, bristling with technique, exhibit.

She also describes fascinating developments in data collection in economic history by scholars such as Angus Maddison and Nick Crafts. The massive prosperity of the West over the past 200 years is entirely without precedent in human history, and describing what happened before and during the early years of this economic take-off is scientifically important.

Despite these advances, it should be said, the teaching of economic history in pure economics degrees has declined dramatically in a regrettable way.

Coyle makes optimistic claims about the ability of economics to understand why some economies grow and others do not. A key development here is the notorious theory of post-neoclassical endogenous growth. Yet even mainstream economists express doubts that this gives a better explanation than the simpler 1956 growth model created by Nobel prizewinner Robert Solow. And both theories predict convergence in per capita income across economies over time, which seems empirically falsified by the persistence over many years of non-trivial differences even between regions of the same economy.

A fundamental area of economic research over the past 20 years has been in behavioural and experimental economics, which Coyle explores at length.

Instead of merely assuming that individuals act in a particular way, the concept of the "rational" agent, economists have tested how people actually behave.

The conclusions of this research programme are a devastating blow to the postulates of the rational, maximising decision-maker. In general, people gather limited information, reason poorly and act intuitively rather than rationally. All scientific theories - even quantum physics, which has survived the most rigorous empirical tests - are approximations to reality.

In limited circumstances, the conventional economic view of rational behaviour is a good one. But most of the time it is a poor approximation, sometimes very poor. And its use can give seriously misleading views of how the world actually operates.

Rather like Japanese soldiers emerging from the jungle, many academic economists still cling to the comforting postulate of rationality, unaware that the war is over. For over it most certainly is. As an article in the January 2007 issue of Economic Journal , the most prestigious in the UK, notes: "For most of the 20th century the axioms (of economic theory) were not tested against the evidence of real decision-making behaviour. But now behavioural economics has called the bluff." The challenge for 21st-century economics is to construct theories based on these empirical insights.

One problem that Coyle notes is that the teaching of economics has not really kept pace with these developments. The textbooks used to instruct most students have, if anything, gone backwards in recent years. These works, aimed at the mass market of US community college students, have dumbed down the subject to a terrifying degree. Even the material presented to strong students is replete with "theorems" and "lemmas" based on postulates of behaviour that have been discredited empirically within economics itself.

Coyle's style is very accessible, and this book is an excellent survey of the frontiers of economics for the general reader. Students of economics at all levels will benefit, as the work's academic credentials are strong, with more than 300 references to leading books and articles. Overall, The Soulful Science can be recommended highly.

Paul Ormerod is a director, Volterra Consulting, and author of Why Most Things Fail .