Land Transport Infrastructure

Two main types of transportation have been studied in the interregional transportation literature: rail, an important transportation mode that was an especially popular investment during the early twentieth century, and highways, which gained importance in the second half of the twentieth century as motor vehicles became widely available.

Particularly in developing countries, many large rail investments were originally constructed by colonial governments with the motivation of primary resource and agricultural product extraction, and thus these projects tend to connect a productive periphery to a port. This contrasts with highways, which tend to be more recent investments and have more varied purposes, such as connecting major cities to each other or to ports.

Rail

Both for connecting cities to each other or peripheral areas to major cities or ports, the literature has broadly found that rail has a large long-run positive impact on incomes and economic growth. However, results also suggest that impacts hinge on the ability of factors, i.e. capital and labour, to move in response to changes in the transport network. It is also worth noting that the policy-relevance of these findings for modern times is unclear as these studies are mostly of historic or colonial investments in rail. Empirical tests for spillovers tend to find limited evidence that production is diverted from nearby untreated areas, suggesting that much of the estimated effects are not due to diversion from nearby areas, but rather represent growth effects.

Within the set of papers studying rail lines connecting peripheral areas to a major exporting port or main economic hub are Jedwab and Moradi (2016) and Banerjee et al. (2020), who study the consequences of rail investments in Ghana and China respectively. Both use straight-line paths to instrument for the endogenous placement of railroads, and they both examine the long-run impact of investments from the late 18th to early 19th century period. Jedwab and Moradi (2016) study cocoa production in Ghana, and find long run development gains in cocoa areas connected to the coast that persist even after the rail system subsequently collapsed. In contrast, Banerjee et al. (2020) find only a small level effect of rail on the GDP of Chinese cities connected to major cities, and no effect on later GDP growth. One potential explanation for the stark difference between these findings is the fact that labour movement was limited in China during the period of study, highlighting the importance of movement of capital and labour for spurring growth.

In order to empirically test for relocation effects, these studies look for treatment effects in areas that were not directly connected to rail. Jedwab and Moradi (2016) do this by comparing the effect of rail in different bands, e.g. 10-20 km. distance to rail, to further areas, e.g. > 20 km. away. In a similar vein, Banerjee et al. (2020) run their specification excluding the districts nearest to the rail line. Both of these tests are based on the idea that if economic activity is being relocated from nearby areas then the effect of rail should be negative for these less proximate areas that are losing firms and residents. Neither find evidence that nearby areas are negatively affected by rail development. While in the Chinese context this may be because the policy context did not permit migration or other relocation of factors, the Ghanaian case is puzzling since rail areas did see large increases in both rural and urban populations. Therefore migration must have come from other areas not nearby rail. This finding highlights the limitation of empirical tests that look for relocation effects in areas nearby treated ones, as relocation may not be limited to nearby places.

Alternatively, some rail systems connect broad areas to each other and are not focused on a connection to a particular port. Studies from the United States have found large gains from 19th century rail investments (Donaldson and Hornbeck 2016, Hornbeck and Rotemberg 2022), and while evidence from developing countries is more sparse, a couple of studies suggest that gains have also been large in India. Donaldson (2018) uses a general equilibrium (Ricardian) model together with a difference-in-difference specification to study railroads in India. From the model he shows that the own trade share is a sufficient statistic for welfare in this setting, and finds that rail spurred large increases in trade between regions and caused an average increase in income of more than 16%. Fenske et al. (2023) construct a least-cost instrument to study the effect of colonial rail expansion on city size in India, and find that rail-driven increases in market access significantly increase city size, especially among the initially most isolated cities. In the case of Kenya, Jedwab et al. (2017) study the effect of colonial railroads and find strong evidence of urban path dependence.

Overall there is less evidence on the impact of intercity high-speed passenger rail on economic outcomes, and we see this as an interesting area for future research. In a recent paper, Tian and Yu (2023) study the impact of the expansion of high-speed rail in China and document increases in export volumes, suggesting that passenger rail may promote firm productivity gains through labour productivity spillovers across locations. Similarly, Barwick et al. (2022) documents that China's rapid expansion of high-speed railways (HSR) facilitates the use of intercity travel as an effective adaptation strategy to climate change since it reduces the exposure to extreme pollution and high temperatures, improving health outcomes. 

Highways

In the case of highways, the empirical evidence on the effect of transportation improvements on incomes is fairly mixed, painting a more nuanced picture than that for rail. While many studies find that highways which connect regions to each other or to major cities or ports have positive welfare effects, the literature has also found that for rural areas connected to major cities specialisation can mean becoming more agricultural and less industrial. The effect of rural or last-mile roads that connect isolated villages to the road network on growth is particularly mixed. Despite the fact that last-mile roads appear to have strong effects on shifting labour out of agriculture in village economies, little impact has been found of roads on income or consumption for those who remain in rural villages.

Studies that focus on highways that connect medium- and large-sized cities to each other have found that these investments have the potential to generate aggregate gains, but also highlight the fact that these gains are not equal across space and may be dampened by poor location choice. For example, Bird et al. (2020), Reed and Trubetskoy (2021), and Lall and Lebrand (2020) use quantitative spatial models to study highway networks in Central Asia. Bird et al. (2020) and Lall and Lebrand (2020) study the potential impact of the Belt and Road Initiative (BRI), a massive infrastructure project that seeks to connect China with the rest of Central Asia and the Middle East. The model-driven approach in Bird et al. (2020) finds that the BRI would increase welfare by 2-3%, and that there is a large amount of heterogeneity in gains between regions. While some would double in size and experience income growth of up to 12%, other areas could see declines, a finding reiterated by Lall and Lebrand (2020) who find that gains concentrate in areas near border crossings. Coşar (2022) studies a series of road capacity upgrades in Turkey using an economic geography model with endogenous labour supply, and finds significant dispersion in welfare gains, with the median gain being 2.9% and a maximum gain of 12.4%. Sotelo (2020) develops a quantitative trade model to study how road paving would affect Peruvian farmers, finding that while gains are heterogeneous, road improvements would lead to a median improvement in farmer welfare of 2.7%, whereas a policy of developing new roads in remote areas would lead to only a 0.2% median gain in welfare. Finally, Fan et al. (2023) use customs data and a spatial equilibrium model to evaluate the benefits of expressway development in China, finding that expressway expansion led to large welfare gains and a 150% return on investment. They highlight how models that ignore general equilibrium effects, the ability of trucks to reroute, and changes in the terms of trade would severely understate the benefits from infrastructure investments.   

The economic geography literature has also made progress on the issue of where infrastructure should be placed when future changes to the climate are taken into account. Balboni (2023) extends the economic geography framework to include dynamic impacts, and finds that Vietnam’s recent coastal bias in infrastructure investments is dynamically inefficient, as future sea level rises imply that land and infrastructure in coastal districts will become submerged. Other biases in infrastructure placement can also dampen its effectiveness. Bonfatti and Poelhekke (2017) find that mines bias land transportation infrastructure and therefore trade in African countries towards overseas exports as opposed to interior trade or trade with neighbouring countries. Finally, Alder (2023) compares India’s Golden Quadrilateral, which connects large cities to each other, with China’s strategy of connecting intermediate cities. Using a general equilibrium framework, his results suggest that the income-maximising network in India would be larger and benefit initially lower income districts more than the current network.

The evidence for highways that connect peripheral areas to major cities or ports is also mixed, and an important distinction seems to be whether connected cities are primarily engaged in agriculture or manufacturing sectors, as cities tend to specialise as they open to more trade. Some of the existing research that studies highways connecting peripheral regions to major hubs uses instrumental variable approaches (least-cost, straight lines, historic networks) to isolate exogenous variation in exposure to new highways. The findings from this literature run the spectrum from negative growth effects in connected areas to large positive effects. While Ghani et al. (2016) find increases in manufacturing output of around 49% over the decade after construction began in districts connected to the Golden Quadrilateral connecting Delhi, Mumbai, Chennai and Kolkata, Martincus et al. (2017) finds that highways in Peru that connect regions to the coast increase exports in connected areas by only around 5.6%. Faber (2014) studies peripheral-to-nodal highways in China and finds that local GDP growth, and in particular industrial GDP growth, actually slowed in connected areas.

How can this disparity in findings be explained? One point that distinguishes the context of Faber (2014) from the other studies is that peripheral areas in Faber (2014) were strongly agricultural, and the evidence suggests that being connected to the highway system caused these areas to further specialise in agriculture, whereas the connected municipalities in Ghani et al. (2016) and Martincus et al. (2017) were already manufacturing centres. This suggests that the composition of industry in connected places and the relationship of these areas to the nodal cities they are connected to matter.

Finally, are these growth or relocation effects? Each of these papers design and implement empirical tests for relocation effects by imposing assumptions on the way that spillovers operate, and find limited evidence for relocation. Faber (2014) and Ghani et al. (2016) perform proximity-based tests for relocation, and find limited evidence that economic activity shifted between nearby districts, but they are not able to test if relocation came from other non-proximate areas. Because Martincus et al. (2017) use transaction-level data with firm identifiers, they are able to implement both a distance-based test and more sophisticated tests for relocation. They compare their main results to estimates that compare connected firms to unconnected firms in different industries and municipalities, with the idea that these estimates should be similar to their main specification if there are no within-municipality or within-sector spillovers. This test slightly relaxes the proximity-based parameterisation of spillovers, but is not assumption-free in the form that spillovers take. Jedwab and Storeygard (2022) more flexibly incorporate spillover and relocation effects through an empirical market access approach in studying the impact of transport investments in Africa using roads and cities data spanning 50 years in 39 countries. They build market access measures from population and travel time data to estimate the impact of roads on population and night lights. They find that market access increases population and has a large effect on night lights, and these effects are larger for small and remote cities.

While the above provide suggestive empirical evidence that gains are aggregate and involve more than just a spatial reshuffling of the economy, theoretical frameworks are better equipped to quantify aggregate gains. In this vein, Asturias et al. (2019) study the Golden Quadrilateral project using an economic geography model with variable markups, and find that aggregate gains to the manufacturing sector in India are around 2.7%, and that an important share of these gains stem from improvements in allocative efficiency. In another example, Morten and Oliviera (2023) develop a quantitative spatial model with trade frictions in order to estimate the welfare gains from the radial highway network in Brazil. They employ differences in differences and IV strategies and find that areas that were connected to the Brazilian coast via the radial highway network saw increased trade and migration. Through model estimation, they find that the radial highway network increased welfare by 2.8% with the primary channel being goods market integration between cities as opposed to labour market integration. Pellegrina and Sotelo (2023) study how a period of migration west in Brazil, driven in part by road expansion, reshaped Brazil’s aggregate and regional comparative advantage. Using a quantitative spatial model, they find that the decrease in migration costs played a pivotal role in altering Brazil’s competitive position, leading to its emergence as a prominent exporter of commodities. According to their estimates, the reduction in migration costs accounts for 25% of the observed changes in specialisation. Baldomero-Quintana (2022) finds a similar effect on specialisation studying a major road investment in Colombia.

Rural roads

In theory, roads that connect isolated rural areas to the larger transportation network could promote development through structural transformation, as previously isolated areas can now produce and trade with more industrial areas. While many studies do find sectoral change in response to roads, the evidence that this leads to increased incomes is mixed, and the literature has found that the impact of roads depends on the initial size of the village and the presence of complementary infrastructure. Perhaps because migration between isolated regions is less of a concern than migration between cities, this literature does not test for spillovers as frequently as the papers studying roads between cities.

Rural areas dominated by agriculture may benefit directly from connectivity if access to more markets and technology increases incomes or productivity, but studies so far suggest that while roads may help moderately developed places grow, they are not sufficient to cause increases in incomes in the most isolated places. For example, Alder et al. (2022) and Mitnik et al. (2018) use night lights data and a difference-in-differences approach to estimate the effect of road improvements on regional growth in Ethiopia and Haiti respectively, and both find that road improvements increase luminosity but only for relatively developed areas whereas the least developed places saw negligible or even negative effects. In India, Asher and Novosad (2020) use a regression discontinuity design based on the rule used to develop roads in different districts to study a policy that provided all-weather feeder roads to unconnected villages, and find no effect of roads on income, assets, or agricultural production. Shamdasani (2021) also studies rural road improvements in India, and finds no effect on agricultural wages, though farm households do increase their use of productivity-enhancing inputs. Gebresilasse (2023) studies the effect of a programme in Ethiopia that connected villages to an all-weather road using a differences-in-differences design and finds that road access led to increased agricultural income only when paired with access to agricultural extension services.

In contrast to these muted empirical findings on the effect of roads on incomes, some studies do find positive effects of roads in rural areas. Gertler et al. (2019) exploit the budgetary allocation process in Indonesia to isolate exogenous variation in highway funding and find that highway improvements led to increases in nominal consumption and income as well as movement of workers from informal to formal employment, implying a 0.45% welfare gain. In their study of river bridges in Nicaragua that reduce uncertainty in market access caused by flooding, Brooks and Donovan (2020) find that farm profits increase by 75% in connected areas, and that connected villages see gains in labour market incomes, particularly outside the village.

Even when roads do not increase incomes, empirical evidence suggests that connecting rural areas to the main road network does lead to shifts of labour out of agriculture and into manufacturing or service sectors. This is the case in Asher and Novosad (2020), who find that despite the null effect on incomes, roads lead to a 9 percentage point shift of workers out of agriculture and into wage labour.[1]1. The same programme was also shown to increase investments in children's education (Adukia, Asher and Novosad 2020).  Similarly, Shamdasani (2021) finds that roads lead to a 40 percentage point reduction in cultivation, but only in areas that are close to towns. In Ethiopia, Gebresilasse (2023) also finds that villages that do not receive access to extension services see a 22% shift of workers out of agriculture and towards crafts and trade occupations.

Discussion and open questions

Overall, historic rail investments seem to have large effects on local GDP when factors are able to move. Evidence on highways is more mixed but generally finds positive welfare effects particularly for larger and manufacturing-oriented areas. Though connecting rural areas tends to lead to shifts out of agriculture, whether these shifts lead to gains in income is unclear. However, this body of evidence still leaves important policy questions unanswered. First, do modern investments in railroads deliver the same gains as historic investments? The existing studies are generally limited to colonial rail investments, which were made under very different institutional, economic and political systems than today, and it is unclear whether new rail investments would deliver the same benefits when competing against modern motorised transport options such as trailer trucks. Furthermore, modern interregional passenger rail is relatively unstudied, and we know little about its effect on growth and welfare in developing countries today. For highways and rural roads, what are the factors that influence whether a road will impact economic growth and welfare? We know that the initial size and sectoral composition matter, but how important are determinants of road quality for its ability to drive trade? Finally, how at risk are transportation networks to climate change and natural hazard risk? Balboni (2023) provides an example of how highways in Vietnam are at risk for future sea level rise, but we still know little about how exposed transit infrastructure is to heat, hurricanes, and other natural disasters that will become more frequent with climate change, and what the welfare consequences of these risks will be. There is also a pressing need to learn more about the effects of transportation on environmental impacts. For example, while Asher et al. (2020) show that new rural roads in India had a precisely zero effect on local deforestation, trade models suggest that general equilibrium impacts on deforestation may be large (e.g. Araujo et al. 2023).

References

Adukia, A, S Asher, and P Novosad (2020), “Educational Investment Responses to Economic Opportunity: Evidence from Indian Road Construction.” American Economic Journal: Applied Economics 12(1): 348-376.

Alder, S (2023), “Chinese Roads in India: The Effect of Transport Infrastructure on Economic Development.” Working Paper.

Alder, S, K Croke, A Duhaut, R Marty, and A Vaisey (2022), “The Impact of Ethiopia’s Road Investment Program on Economic Development and Land Use: Evidence from Satellite Data.” World Bank Policy Research Working Paper 10000.

Allen, T and C Arkolakis (2022), “The Welfare Effects of Transportation Infrastructure Improvements.” Review of Economic Studies 89(6): 2911–2957.

Asher, S and P Novosad (2020), “Rural Roads and Local Economic Development.” American Economic Review 110(3): 797–823.

Asher, S, T Garg, and P Novosad (2020), “The Ecological Impact of Transportation Infrastructure.” Economic Journal 130(629): 1173–1199.

Asturias, J, M García-Santana, and R Ramos (2019), “Competition and the Welfare Gains from Transportation Infrastructure: Evidence from the Golden Quadrilateral of India.” Journal of the European Economic Association 17(6): 1881–1940.

Araujo, R, J Assunção, and A Bragança (2023), “The Effects of Transportation Infrastructure on Deforestation in the Amazon: A General Equilibrium Approach.” World Bank Policy Research Working Paper 10415.

Balboni, C (2023), “In Harm’s Way? Infrastructure Investments and the Persistence of Coastal Cities.” Working Paper.

Banerjee, A, E Duflo, and N Qian (2020), “On the Road: Access to Transportation Infrastructure and Economic Growth in China.” Journal of Development Economics 145: 102442.

Barwick, P, D Donaldson, S Li, Y Lin, and D Rao (2022), "Improved Transportation Networks Facilitate Adaptation to Pollution and Temperature Extremes." NBER Working Paper 30462.

Bird, J, M Lebrand, and A Venables (2020), “The Belt and Road Initiative: Reshaping Economic Geography in Central Asia?” Journal of Development Economics 144: 102441.

Bonfatti, R and S Poelhekke (2017), “From Mine to Coast: Transport Infrastructure and the Direction of Trade in Developing Countries.” Journal of Development Economics 127: 91–108.

Brooks, W and K Donovan (2020), “Eliminating Uncertainty in Market Access: The Impact of New Bridges in Rural Nicaragua.” Econometrica 88(5): 1965–1997.

Coşar, K (2022), “Overland Transport Costs, a Review.” World Bank Policy Research Working Paper 10156.

Donaldson, D (2018), “Railroads of the Raj: Estimating the Impact of Transportation Infrastructure.” American Economic Review 108(4-5): 899–934.

Donaldson, D and R Hornbeck (2016), “Railroads and American Economic Growth: A ‘Market Access’ Approach.” Quarterly Journal of Economics 131(2): 799–858.

Fan, J, Y Lu, and W Luo (2023), “Valuing Domestic Transport Infrastructure: A View from the Route Choice of Exporters.” Review of Economics and Statistics 105(6): 1562–1579.

Faber, B (2014), “Trade Integration, Market Size, and Industrialization: Evidence from China’s National Trunk Highway System.” Review of Economic Studies 81(3): 1046–1070.

Fenske, J, N Kala, and J Wei (2023), “Railways and Cities in India.” Journal of Development Economics 161.

Gertler, P, M Gonzalez-Navarro, T Gračner, and A Rothenberg (2019), “Road Quality, Local Economic Activity, and Welfare: Evidence from Indonesia’s Highways.” NBER Working Paper 30454.

Ghani, E, A Grover Goswami, and W Kerr (2016), “Highway to Success: The Impact of the Golden Quadrilateral Project for the Location and Performance of Indian Manufacturing.” Economic Journal 126(591): 317–357.

Hornbeck, R and M Rotemberg (2022), “Growth Off the Rails: Aggregate Productivity Growth in Distorted Economies.” Working Paper.

Jedwab, R, E Kerby, and A Moradi (2017), “History, Path Dependence and Development: Evidence from Colonial Railroads, Settlers and Cities in Kenya.” Economic Journal 127: 1467-1494.

Jedwab, R and A Moradi (2016), “The Permanent Effects of Transportation Revolutions in Poor Countries: Evidence from Africa.” Review of Economics and Statistics 98: 268–284.

Jedwab, R and A Storeygard (2022), “The Average and Heterogeneous Effects of Transportation Investments: Evidence from Sub-Saharan Africa 1960–2010.” Journal of the European Economic Association 20(1): 1–38.

Lall, S and M Lebrand (2020), “Who Wins, Who Loses? Understanding the Spatially Differentiated Effects of the Belt and Road Initiative.” Journal of Development Economics 146: 102496.

Martincus, C, J Carballo, and A Cusolito (2017), “Roads, Exports and Employment: Evidence from a Developing Country.” Journal of Development Economics 125: 21–39.

Mitnik, O, R Sanchez, and P Yañez (2018), “Bright Investments: Measuring the Impact of Transport Infrastructure Using Luminosity Data in Haiti.” Inter-American Development Bank Working Paper 935.

Morten, M and J Oliviera (2023), “The Effects of Roads on Trade and Migration: Evidence from a Planned Capital City.” Working Paper.

Pellegrina, H and S Sotelo (2023), “Migration, Specialization, and Trade: Evidence from Brazil’s March to the West.” Working Paper.

Reed, T and A Trubetskoy (2021), “Assessing the Value of Market Access from Belt and Road Projects.” World Bank Policy Research Working Paper 8815.

Shamdasani, Y (2021), “Rural Road Infrastructure & Agricultural Production: Evidence from India.” Journal of Development Economics 152: 102686.

Sotelo, S (2020), “Domestic Trade Frictions and Agriculture.” Journal of Political Economy 128(7): 2690–2738.

Suzuki, H, J Murakami, Y Hong, and B Tamayose (2015), Financing Transit-Oriented Development with Land Values: Adapting Land Value Capture in Developing Countries. Urban Development Series, World Bank.

Tian, L and Y Yu (2023), “Geographic Integration and Firm Exports: Evidence from China.” Working Paper.

Zarate, R (2023), “Spatial Misallocation, Informality, and Transit Improvements: Evidence from Mexico City.” World Bank Policy Research Working Paper 9990.