Editor's note: This column first appeared as a chapter in the Vox eBook, The Long Economic and Political Shadow of History, Volume 2, available to download here.
At the turn of the 19th century, sub-Saharan Africa was the least urbanised region in the world, with only about 50 cities of more than 10,000 inhabitants. By 2010, the number of cities had increased to almost 3,000. This column explores whether colonial railroad investments transformed Africa’s economic geography. Were railroads so instrumental that without them economic outcomes would have been different and development delayed?
It is commonly acknowledged that transportation infrastructure is needed in order for countries to develop. However, it is less clear whether transportation infrastructure has the influence to determine where and what cities develop, and whether they permanently change the economic landscape of a country. This comes down to a ‘the chicken or the egg’ problem. At the root of city development could be physical geography: the fact that there is a river, a fertile plain, a natural harbour, and so on, nearby. Thus, geography could give locations a natural advantage and railroads and roads would therefore be built as a response to this. If geography is the main driver, then there is little scope for history (i.e. man-made advantages) to determine the spatial distribution of economic activity (i.e. cities).
Davis and Weinstein (2002), Bosker et al. (2007) and Miguel and Roland (2011) studied the bombing of Japan, Germany, and Vietnam, respectively, and showed that many of the destroyed cities subsequently recovered their population and their importance within the national economy. Davis and Weinstein (2002) attributed this to the effect of their geography (which was not altered). However, there is also a literature which finds that localised historical shocks, or ‘accidents’, can have permanent spatial effects. For example, Redding et al. (2011) showed how the division of Germany, despite later reunification, permanently changed the location of its main airports. Likewise, Bleakley and Lin (2012) showed that many US cities originally formed at such obstacles to water navigation as attracted commerce and supporting services. These cities still exist today, in spite of the fact that the importance of water navigation has become marginal over time.
The African context is different: when colonial railroads were built, African countries were poor and non-urbanised. The overwhelming majority of people lived through agriculture, which has less scope for economies of scale. In Jedwab and Moradi (2016) and Jedwab et al.(2016) we studied railroads and development in Ghana, plus Africa and Kenya, respectively.
Colonial railroads in Africa – historical accidents?
Why did colonisers build railroads? Through studying the history of each railroad, we identified three motivations:
- Military domination: the line was built to exercise ‘effective control’ in the Scramble for Africa, or to dispatch troops for better control of the native population;
- Mining: the line was built into the interior to further European mining interests, and;
- Cash crops: the colonial power built the line to connect agriculturally rich areas. With limited budgets, colonisers expected the railroads to pay for themselves, hence they connected areas of high economic potential.
So, what room is there for historical accidents? Plenty.
While the origin and the endpoint of a line had some advantages, none of the localities in-between may have been considered worthwhile to connect. Railroad placement was essentially an engineering problem – figuring out the cheapest (typically the shortest) route from a point A to a point B. The Kenya-Uganda railroad is a case in point (see Figure 1).
Figure 1 Southern Kenya: The railroad followed the least-cost path from Mombasa to Kisumu
Source: Jedwab et al.(2016)
This railroad was built between 1896 and 1901 to connect landlocked, and relatively prosperous Uganda to the coast (i.e. Mombasa). Kenya was merely transit territory. We verified that this railroad indeed followed the route that minimised construction costs. Using engineering reports, we fed into the computer geographic information and historical cost parameters associated with earth formation, clearing, permanent way, and sidings and culverts. We also noted that the best locomotives at that time could only overcome gradients below 5-10%. The overlap of the cheapest-to-construct route, as identified by the computer, and the actual line is striking. Tellingly, contemporary critics called the Kenya-Uganda railroad the ‘Lunatic Express’, ‘going from nowhere to utterly nowhere’. In fact, the railroad bypassed highly populated areas en route to Kisumu (Lake Victoria) and Uganda. Hence, locations on or near the route were lucky to have received access to the railroad, but others, with similar or better geography missed out.
In Ghana, five railroad routes were proposed, but, due to random events, never materialised (Figure 2). First, the Cape Coast–Kumasi line (1873) was proposed to link the capital of Cape Coast to Kumasi, sending troops to fight the Ashante. This project was dropped because the war came to an abrupt end in 1874. Second, Governor Griffith wanted a central line from Saltpond to Kumasi (1893), in order to tap the palm oil areas and link the coast to Kumasi. When he retired in 1895, he was replaced by Governor Maxwell, who gave in to the gold-mining lobbies and instead built the Western Line, which connected European interests to the coast. For the third and fourth proposed routes, Maxwell thought that a second line was needed, and projects were suggested with two different sets of termini: Apam–Kumasi and Accra–Kumasi. A conference was to be held in London to discuss the proposals, but Maxwell died before reaching London. Following him, Governor Hodgson favoured Accra, but thought that the line should instead be built to Kpong. He retired in 1904, before work began, and was replaced by Governor Rodger, who finally built the Eastern Line.
One of our strategies was, therefore, to compare the locations actually connected to locations hypothetically connected by those alternative routes that were equally likely to have been built but – by the toss of a coin – were not.
Figure 2 Southern Ghana: More than one railroad route was surveyed
Source: Jedwab and Moradi (2016)
Economic development during the colonial era
In Kenya, railroad placement led to a curious situation, whereby lines traversed sparsely settled areas with no Kenyan freight to transport. In the hope of creating an agricultural export industry, land was alienated and offered to European settlers, with plots near the railroad the most sought after. In our analysis, we found that between 1901 and 1962 – a year before Kenya became independent – the railroad led to a concentration of both European and African populations (see Figure 3). The cultivation of coffee and tea expanded along the railroad lines. The effects decrease when moving away from the railroad, and are zero after 30 km. In addition, we found a strong effect of the railroad, within a distance band of 10 km, on skilled European, Asian (mostly traders) and urban populations, whereby the former two were concentrated in the latter.
Figure 3 Railroads, city locations and urban growth in Kenya, 1962-2009
Source: Jedwab et al. (2016)
For Ghana, the story is similar. Railroads decreased transportation costs, thereby making cocoa production for export markets profitable. Ghana became the world’s largest exporter of cocoa by 1911. The rural population increased along the railway lines because cocoa cultivation required more labour, thus creating villages. In turn, cities also emerged, as trading stations along the railroad, servicing the villages (see Figure 4).
Lastly, we expand these findings to the rest of Africa circa 1960, when most countries became independent. Using data for 39 sub-Saharan African countries, Figure 5 shows that railroads built during the colonial period strongly predicted the current location of cities.
Figure 4 Railroads, city locations and urban growth in Ghana, 1931-2000
Source: Jedwab and Moradi (2016)
Figure 5 Railroads, city locations and urban growth in sub-Saharan Africa, 1960-2000
Development in the long run: Stability despite the demise of railroads
Railroads gave an initial advantage to the cities they created. However, this advantage disappeared over time. First, railroad systems collapsed after independence, as a result of mismanagement, lack of maintenance, and the adoption of a new transportation technology, namely, motor roads. The new roads were not necessarily built along the colonial lines, which should have led to the decentralisation of economic activity, away from the railroads. Second, in the case of Kenya, the European population and a large part of the Asian population left the country after independence. In Ghana, the ageing crop of cocoa trees along the colonial railroad declined in productivity.
Despite all this, we find that for Kenya (2009, see Figure 3), Ghana (2000; see Figure 4) and most of sub-Saharan Africa (2000, see Figure 5):
- locations along the old railroad lines are still more developed and urbanised today, and
- railroad cities, i.e. cities that emerged with the construction of the railroads, are still wealthier today than non-railroad cities of similar sizes.
Sources of path dependence
What is the reason for the strong persistence in spatial economic and urban patterns? For Kenya, we contrasted four different explanations based on: (i) institutional persistence, (ii) technological change, (iii) sunk investments, and (iv) spatial coordination failures.
First, we find that institutional persistence, for example the fact that the new African political elite may have purposely tried to preserve the colonial distribution of population in order to better control the resources of the country, does not account for path dependence.
Second, we verified that path dependence is not explained by changes in transportation infrastructure, as measured by the roads that may have replaced deteriorating railroads.
Third, we find colonial railroad cities were better endowed with non-transportation infrastructure (e.g., hospitals and schools) at independence, and that these colonial investments partially explain path dependence.
Fourth, we argue that persistence is also explained by the fact that the early emergence of the railroad cities served as a mechanism to ‘coordinate’ locational decisions and spatial investments in subsequent periods. For example, cities, once they have emerged and reached a certain size, indicate to potential rural-to-urban migrants the locations to which they can move if they want to be better able to exchange goods and ideas with other individuals. We therefore contend that expectations about where people will live in the future may matter as much as history (i.e. colonial investments), which suggests that regional policies may be ineffective if they do not also change expectations.
For Ghana, we came to the same result. Sunk investments during the colonial period and the resolution of these spatial coordination failures both explain path dependence.
Lessons for infrastructure investments in Africa today
Africa still suffers a massive infrastructure deficit (Calderón & Servén 2010). During the last decade, many countries have invested in modernising their infrastructure networks. Various giant transportation projects are on their way, such as the Tanzania–Gabon railway ($33 billion), the Mombasa–Kampala–Kigali railway ($14 billion), the Trans- Kalahari railway ($9 billion), and the Abidjan–Lagos motorway ($8 billion), and there are many more ambitious plans, such as the proposed rail/road corridor connecting South Sudan to the Indian Ocean. Chinese firms are playing a leading role, with striking similarities to the European colonial endeavours at the turn of the 20th century, seeking to unlock the continent’s resource potential.
In the light of our findings, what lessons does history provide to inform policy? First, infrastructure investments can produce economic change by reducing trade costs and integrating markets. Second, the impact will, however, depend on how much these transportation investments decrease existing trade costs. The construction of the railroads in colonial Africa constituted a ‘transportation revolution’, because trade costs were extremely high before. However, the construction of roads in post-colonial Africa did not constitute a transportation revolution, since railroads had already significantly decreased trade costs and cities had already been created and grown to significant sizes. This implies that transportation investments will have smaller effects in more developed regions.
Conversely, new infrastructure could transform the economic landscape in poor, remote regions with high trade costs. Finally, the effects of new investments in transportation infrastructure in poor, remote regions will likely depend on a region’s intrinsic economic potential. This is because, at the beginning of the development process, city growth is fuelled by the agricultural activities and rents surrounding them. In both Ghana and Kenya, we found positive effects of transportation investments on agricultural development and structural change.
Bleakley, H and J Lin, (2012), “Portage and Path Dependence.” Quarterly Journal of Economics, 127(2), 587-644.
Bosker, M, S Brakman, H Garretsen, and M Schramm (2007), “Looking for multiple equilibria when geography matters: German city growth and the WWII shock”, Journal of Urban Economics, 61(1), 152-169.
Calderón, C and L Servén (2010), “Infrastructure and Economic Development in Sub-Saharan Africa”, Journal of African Economies, 19(S), 13-87.
Davis, D R and D E Weinstein (2002), “Bones, Bombs, and Break Points: The Geography of Economic Activity”, American Economic Review, 92(5), 1269-1289. doi:doi: 10.1257/000282802762024502
Duranton, G (2015), “Growing through Cities in Developing Countries”, World Bank Research Observer, 30(1), 39-73. doi:doi:10.1093/wbro/lku006
Glaeser, E L and J D Gottlieb (2009), “The Wealth of Cities: Agglomeration Economies and Spatial Equilibrium in the United States”, Journal of Economic Literature, 47(4), 983-1028.
Jedwab, R, E Kerby, and A Moradi (2016), “History, Path Dependence and Development: Evidence from Colonial Railroads, Settlers and Cities in Kenya”, Economic Journal.
Jedwab, R., and A. Moradi (2016), “The Permanent Effects of Transportation Revolutions in Poor Countries: Evidence from Africa”, Review of Economics and Statistics, 98(2), 268-284.
Lucas, R. J. (1988), “On the Mechanics of Economic Development”, Journal of Monetary Economics, 22(1), 3-42.
Miguel, E., and G. Roland (2011), “The Long-Run Impact of Bombing Vietnam”, Journal Of Development Economics, 96(1), 1-15.
Redding, S. J., D.M. Sturm, and N. Wolf (2011), “History and Industry Location: Evidence from German Airports”, Review of Economics and Statistics, 93(3), 814-831.