For economists, I am essentially interested in how to model and test agent behaviour when maximisation is not possible, because of the cognitive difficulties involved.

But complexity has implications not just for economics, but for business as well as the other social sciences.

Most of the articles available below are pretty technical, but those marked [GEN] are suitable for general readers.

The business cycle

Ever wondered why economic forecasts are often wrong? Here is an article published in Physica A, the world’s leading statistical physics journal, which explains why..

Would you like to be on the Bank of England’s Monetary Policy Committee. This article, published by the Manchester Statistical Society, tells you why membership is a sinecure [GEN]

Conventional economics lacks a satisfactory theory of the business cycle. Booms and recessions are caused by external shocks. The theory in this paper, published in Physica A, builds on the behaviour of individual firms and generates cycles which look like the real thing. But the cycle is generated by individual behaviour within the system. The individual agents behave in a very Keynesian way.

Has the UK converged sufficiently to justify joining the Euro? This paper, published in Physica A, uses some powerful maths to look at the evidence.

Is an efficient flow of information always a good thing?  Negative sentiments might cascade across firms to produce recessions.  This paper examines the severity of recessions in the US economy in the 20th century.  I show that it is precisely the fact that firms find it easy to obtain information about the views of others which leads to the particular distribution we observe.

Economic growth

Mainstream models of economic growth predict that living standards will converge across all countries in the long run.  This clearly has not happened.  The modern refinement is to postulate that they will converge amongst economies with similar cultural and historical backgrounds.  But how long is the long run?  The countries of Western Europe are the most culturally homogenous in the world, sharing the Greco-Roman heritage of 2,000 years ago.  But even amongst this group, the long run appears to be around 150 years.

Crime

Why do crime rates vary so much across time and place? Here is a long paper, due to be published by the British Home Office, which looks at crime as an epidemic. The maths used in epidemiology is very useful in understanding patterns of crime

Two major data bases record the criminal activities of particular individuals over long periods of time.  The Cambridge study is based in the UK, and the Pittsburgh study in the US.  The first records the number of criminal convictions, the second the number of self-reported offences.  This paper shows that both follow an identical mathematical relationship.  The key link in the potential criminal career of an individual is the commission of the first offence.  Once this is done, individuals can and do commit crime on all scales.

In this paper, we develop a simplified model for how the internal structure of a network evolves over time. We set up gangs with an initial internal network of connections, and explore theoretical network structures cited in recent crime and terrorism literature. In terms of the efficiency of the gang, we find that the most successful criminal gangs are those which have higher levels of communication and two way communication between criminals in the gang rather than a strict hierarchy. Success tends to be greater when new criminals are introduced at the bottom of the network hierarchy

Power laws

These are being discovered everywhere – the extinction patterns of biological species, the structure of the Internet, the social patterns of sexual contact.

Here is an article from the Financial Times in February 2001 which explains what power laws are and why they are important.[GEN]

The next two articles are published in Physica A.

The relationship which describes the connection between the frequency and size of the extinctions of firms looks very similar to that which describes the connection for biological species (though obviously the time scales differ!).

The duration of recessions in the Western market economies follows a subtle form of power law behaviour

How firms evolve and why they (eventually) fail

This article, based on a paper given at the World Econophysics Conference in Bali August 2002 and forthcoming in Physica A, explains the extinction patterns of capitalism’s largest firms during the 20th century

Here is a Powerpoint presentation, given at a UCLA conference on complexity at Lake Arrowhead May 2000, which summarises much of my work in this area.

This paper, presented at a conference in Santa Fe in October 2002, shows why external shocks to the economy are not the real reason why firms fail

Competition and market structure

Most economic regulators assume that if a firm has a large market share, there is prima facie evidence that the market is not competitive. This article shows why this is not the case

This next paper, presented at a conference in Florida in 2002, formalises the Schumpeterian concept of creative destruction. It shows why we need to make a distinction between competition between firms within an industry, and competition between industries - the competitive environment in which the industry operates. This distinction is often not properly distinguished either in the literature or in regulatory practice.

Economic methodology

Economic theory needs to be much more empirically based, rather than relying on a priori reasoning about agent behaviour.

The number of students studying economics has fallen dramatically, both the in the UK and the US.  Economists are supposed to know about markets, but do not seem to be able to provide a product in the market for their subject which consumers (students) find satisfying.  Here is an article in the new journal International Review of Economics Education which sets out some ideas about how the subject can be made more attractive.

This paper was given at a conference in Budapest in 2001, and will be appearing in the conference book of papers.

Here is a paper given at a conference of the Brisbane Group last year, and will appear in the conference volume of papers

Karl Popper has been enormously influential in the social science methodology literature. This paper was written for a conference on Popper held in Vienna, and has been published in …..

Econometric methodology

Econometrics is a particular application of statistical theory to economic (and social) data. There are basically two different types. Cross-sectional refers to the analysis of a data base at a point in time, perhaps of different individuals, firms, areas or whatever.

Time-series refers to the analysis of data over time. It is much favoured by macro-economic modellers and forecasters. Here are two papers which show why the research programme in macro-economic modelling using time-series econometrics has not got very far. One appeared in…, and the other in…. [.pdf, .pdf]

Property and financial market papers

Recent work by econophysicists has overturned a lot of standard portfolio theory.  Here is an illustration of the key techniques, using exchange rates as an example, given at a conference in Santa Fe in 2000.

Agglomerations and local economic activity