Big data and complexity: evidence for the structural approach to economic development

This post summarises some of the ideas in an interesting article from the May issue of the Cambridge Journal of Economics. The piece shows that an analysis of ‘complex networks’ using ‘big data’ lends support to structuralist arguments about growth and development. I briefly discuss the implications for industrial policies intended to promote the ‘catching up’ of poor countries with richer ones.

The Cambridge Journal of Economics (CJE) is an influential heterodox journal published six times a year. It includes as one of its patrons nobel prize-winner Amartya Sen and as associate editors Ha-Joon Chang, Mushtaq Khan and Anwar Shaikh, whose ideas I have sometimes discussed in previous posts.

As CJE articles are usually behind a paywall, I thought it would be helpful to summarise and comment on one or two when they are interesting and relevant to this blog.

Structuralist economics emphasises the importance of economic structures (sic) as shaping individual behaviour, in contrast to the methodological individualism of neoclassical theory, which dominates the mainstream.

In social theory, the relationship between human agency and social structure is subject to much debate, but this debate is absent from neoclassical economics, with its focus on the primacy of ‘microfoundations’ in the form of individual agency as determining outcomes.

Structuralist approaches tend to start with determining elements in economic performance such as dualism in the structure of production (for example agriculture versus industry), poverty traps, core versus peripheral economies, and productive linkages between firms and sectors.

What is sometimes termed ‘classical’ development economics adopted a structuralist approach, which described the importance of industrialisation in promoting growth and economic transformation in developing countries. It argued that without this transformation in the structure of production, growth in output, productivity, employment, wages and overall living standards would not be sustained.

Structuralist development economics also argued that poorer countries generally produce and export products of low sophistication, often primary commodities or non-manufactured goods, and have a low level of diversification in their productive structure.

By contrast, rich countries tend to export much more sophisticated products and have a more highly diversified productive structure.

Manufacturing, in contrast to subsistence sectors, is often subject to increasing returns to scale: as the scale of production increases, productivity increases via an increasing division of labour, as tasks become more specialised. Evolving production processes lead to technological progress. Manufacturing also tends to involve more sophisticated activities such as research and development.

The arguments of the structuralists imply that it is hard for poor countries to catch up with rich ones in terms of living standards. To do so, their economies must industrialise and shift production towards a more highly diversified range of more sophisticated and internationally competitive products with a high income elasticity of demand, capable of being exported onto the world market.

The structuralist approach challenges neoclassical development economics, which argues that market efficiency is necessary and sufficient for structural change in poor countries to take place. This should lead to a convergence of incomes with richer countries, in the absence of state intervention.

The authors of the article in the CJE use techniques drawn from what is known as ‘econophysics’, employing ‘big data’ and ‘complex networks’ to examine the case for such structuralist arguments.

Big data, at a basic level, is pretty self-explanatory. Its use has been made possible by enormous advances in computing power which have eased the collection and analysis of vast amounts of data.

The authors examine networks of international trade links from 101 countries and 762 products that are exported, generating 1,756,224 trade links in 2013.

Their analysis does indeed suggest the validity of structuralist arguments about the differences between rich and poor countries. Poorer countries have far fewer international trade links, and export less sophisticated products. Rich countries have many more international trade links and export more sophisticated products.

The international trade networks analysed also provide an example of what is known in statistics as a power law: a few hubs in the network (rich countries in this case) have many trade links, while many hubs (poorer countries) have far fewer trade links. There is thus substantial inequality in the distribution of links between countries.

As a further example of increasing returns to scale, the hubs with a large number of trade links may find it easier to create new links, with the opposite applying to hubs with few links. This provides some evidence of a divergence in per capita income levels between rich and poor countries, rather than convergence as argued by the neoclassicals.

The authors also examine economic complexity in a country, defined by the presence of exports of non-ubiquitous products plus diversified exports. They find a positive relation between the number of trade links and both economic complexity and income per capita. In short, richer countries have more complex economies, and catching up with them may be hard.

These results provide some support for structuralist arguments. However the article does not discuss any of the policy implications, which are beyond its scope. Although both structuralism and history suggest that the convergence or catching-up of poor countries with rich ones is indeed a difficult task, it is still possible, as shown for example by Japan, South Korea, Taiwan and Singapore and, to a lesser extent, economies in South East Asia such as Malaysia and Thailand.

The lessons of successful late developers such as these are that some form of industrial policy is necessary to accelerate economic transformation, which involves the emergence and growth of internationally competitive capitalist firms and sectors across the economy.

Countries such as China and Malaysia have encouraged foreign direct investment and have tried to promote learning by domestic firms via joint ventures and linkages with foreign investing firms, with some success.

By contrast, Japan and South Korea’s growth miracle phases relied largely on the development of domestic firms, which initially imported technology for learning, and eventually learned to innovate themselves.

The process of successful catch up has therefore varied among those few countries that have managed it. However, the political and institutional context remains extremely important in the design of effective industrial policies. Structuralist ideas should continue to form a part of this kind of development economics.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.