Directed by
Ideas for growth
Menu
Directed by
Ideas for growth
Menu
Projects
Publications

How much do we know about the development impacts of large-scale integrated transport infrastructure?

The Infra4Dev Conference, jointly organized by the World Bank and the International Growth Centre on March 3rd-4th 2022, brought together the academic and policy-making community to exchange knowledge and insights regarding the different roles that infrastructure can play in catalyzing development. Professor Stephen Redding, of Princeton University, provided the framework presentation on our current state of knowledge regarding the development impact of large-scale integrated transport infrastructure investments. He now shares some of his key insights for a wider audience through this blog. 

Large-scale integrated transport infrastructure investments involve major changes in the transport network connecting regions and/or countries. They can involve multiple transport modes (e.g., highway, railway ports). They can change both internal transport costs within countries and external transport costs to other countries. Examples include China’s Belt and Road Initiative from 2013 onwards and the construction of Argentina’s late 19th-century railway network. 

These infrastructure investments reduce bilateral transport costs between the connected locations, which has direct effects on the spatial distribution of economic activity through three main channels, lowering the costs of: (i) trading goods; (ii) population migration; and (iii) commuting. Recent research, however, has highlighted that these infrastructure investments also have important indirect effects, as firms and workers respond endogenously to changing transport costs by switching their choice of location. This means that even a location that not directly affected by such infrastructure investments can indirectly benefit through increases in economic activity in nearby connected locations (for example, through increased accessibility of distant markets).  

A key policy implication of this recent research is that these indirect effects are shaped by propagation mechanisms and complementary investments and policies. The impact of infrastructure investments is typically magnified by agglomeration forces, that is the benefits that arise when people and firms cluster together geographically. The benefits of infrastructure are also enhanced by complementary investments in other proximate buildings and structures (such as neighboring housing or business parks), which in turn are heavily influenced and sometimes constrained by land use regulations. Therefore, in evaluating proposed infrastructure projects, it is important not only to consider current levels of economic activity along the route, but also the ways in which households and firms will endogenously respond to the presence of new infrastructure, through agglomeration forces and complementary investments. 

Importantly, traditional cost-benefit analysis of large-scale transport infrastructure has not been able to capture such indirect effects, which can be important and may significantly affect the balance of economic costs and benefits for a project. However, recent theoretical and empirical advances, including, in particular, the development of quantitative spatial models, have substantially enhanced our understanding of these wider impacts of transport infrastructure, as well as our ability to quantify them. These models are rich enough to connect directly to observed data, such as the many heterogeneous locations connected by a rich transport network. Yet these models are also sufficiently tractable as to permit conceptual analysis of their properties, structural estimation of a small number of parameters, and policy-relevant counterfactuals. 

An example is the recent quantitative analysis, in Fajgelbaum and Redding (2022), of the relationship between international trade, structural transformation and economic development in Argentina from 1869-1914. This is an attractive empirical setting to study the impact of transport infrastructure, because it involves large-scale changes in both external integration (from reductions in transatlantic freight rates driven by the invention of the steam ship) and internal integration (from the construction of Argentina’s railway network following the invention of the steam locomotive). 

External and internal integration are shown to both affect the level and composition of economic activity through a spatial Balassa-Samuelson effect: Locations with better access to world markets have higher population densities, higher urban population shares, higher relative prices of non-traded goods, and higher land prices relative to wages, as in many developing countries today. The figure below provides a graphic illustration of the expansion of the railway network and the changing distribution of population density within Argentina (Figure 1). 

Figure 1: Population Density and the Expansion of the Argentina Rail Network 

Starting from observed economic activity in 1914, the model is used to undertake counterfactuals for reversing external integration (raising transatlantic freight rates to 1869 values) and reversing internal integration (removing the railway network). Undoing external integration reduces real income per capita by 7.1%, which is broadly in line with existing estimates of the welfare gains from trade (Bernhofen and Brown 2005). Furthermore, removing the railway network reduces real income per capita by 4.8%, again comparable to existing estimates from other settings (e.g. Donaldson and Hornbeck 2016, Donaldson 2018).  

An influential approach to evaluating the impact of transport infrastructure investments is to compare the impact on land values to construction costs, as in the public finance literature following George (1879). For Argentina’s 19th-century railway network, the net present value of the increase in land income is shown to exceed historical estimates of construction costs for standard discount rates of 3 or 5 percent. Therefore, this large-scale investment in transport infrastructure can be rationalized in terms of its impact on economic activity.  

Importantly, the excess of the net present value of the increase in land income over construction costs is substantially greater at the levels of external integration in 1914 than at those in 1869. The reason is that the railway construction costs are fixed, whereas the absolute increase in the level of economic activity from the construction of the railway network is larger in the more open economy in 1914. This finding highlights an important policy complementarity between infrastructure investments to reduce internal transport costs and trade policy reforms to reduce external barriers to international trade. 

Although Argentina during the 19th-century is an attractive empirical setting, a similar pattern of results is found for China’s recent Belt and Road Initiative, as examined in Lall and Lebrand (2020). Again, there is evidence of a spatial Balassa-Samuelson effect: Population density falls with distance from trade hubs and declines more sharply for urban population density than rural population density. The Belt and Road transport investments are found to raise aggregate real income by amounts ranging from 0.4-10 percent, depending on the country considered. Furthermore, the impact of these transport investments is again magnified if they are accompanied by complementary trade facilitation policies. 

In summary, transport investments are some of the largest single investments that societies make, with transport projects accounting for about 10 percent of the World Bank’s lending portfolio in 2021. Recent developments of quantitative spatial models have enhanced our understanding of the impact of these investments, highlighting indirect effects from firm and worker relocation, and the role of propagation mechanisms and complementary investments and policies. Looking ahead, there is an exciting future research agenda applying these quantitative spatial models in new settings and making use of the proliferation of Geographic Information Systems (GIS) data on economic interactions at a fine level of spatial disaggregation. Furthermore, increasing adoption of such analytical techniques as part of the evaluation of large transport projects, promises to provide a more detailed understanding of their economic impacts. 

This blog has been co-published with the World Bank.

Comments

Leave a Reply

Comments will be held for moderation. Your email address will not be published.

Time limit is exhausted. Please reload CAPTCHA.