Including infrastructure systems in a new economic geography model

Robert Carlsson


Infrastructure performs a fundamental role in today’s developed economies by facilitating spatial movements of economic resources, products, services and people. The result is that previously localised economic agents are given an increased freedom to interact with each other in a multitude of ways. Ultimately, this enables agents to specialise more and make use of their comparative advantages while decreasing constraints due to local conditions. Therefore infrastructure reduces the dependence on local endowments and aids in creating a more complex economy. More specifically, infrastructure provides the agents with access to resources, information, and markets they couldn’t otherwise partake in.

However, infrastructure is frequently treated directly within non-spatial economic theories. This is often done by either aggregating infrastructure capital within the general capital stock or by creating a separate infrastructure capital stock. This approach is problematic as these models cannot, fundamentally, account for the network effects that infrastructure exhibits and are fundamentally associated with. Furthermore, such an approach assumes direct impacts of infrastructure on the economy and not the de facto indirect effects it has on agents by facilitating access to resources, products, services and people. Therefore an economic framework that includes at least some notion of space is required to properly begin to study the economic impacts of infrastructure. The New Economic Geography (NEG) is well suited to deal with this network nature of infrastructure as the economy in the NEG model consists of a network of cities or regions that are interconnected.

The NEG framework used in this model has been extended, beyond its most common formulation, in order to better capture the specific effects of infrastructure as well as increasing its sophistication by relaxing some overly simplistic assumptions. In this model the economy consists of two sectors, both reliant upon labour and capital as primary inputs, with one sector being more labour dependent and the other one is more capital dependent. Effectively this means that there is one sector that represents a service sector while the other one represents a manufacturing sector. Moreover, the production process is affected by additional parameters such as total factor productivity, human capital, total factor productivity, and energy efficiency. Both sectors produce goods that are specific to each region and are trade-able between the them. As is usually the case in NEG models, the cost of trading is expressed in terms of iceberg costs. However, in this model there are two independent connection levels as services and manufactured goods are differently difficult to trade across space. Finally, population movements, between regions, are determined using an attractiveness index that is calculated using factors that influence households decisions to relocate.

Including infrastructure systems in a new economic geography model (pdf, 231 KB). The future of national infrastructure systems & economic prosperity conference, Cambridge, UK, March 27–28, 2014.


Carlsson, R., Otto, A. and Hall, J.W.