Published on Let's Talk Development

Pursuing higher economic returns to extractive investment

This page in:

Across large swaths of the developing world, a new trend is taking hold: governments are targeting public and private investments in specific geographic areas in the hopes of creating spatial “development corridors.” These strategies are guided by the belief that concentrating and co-locating infrastructure investments in specific locations can create clusters of interconnected firms, nurture the development of value chains, reduce unemployment, and improve the provision of basic public services.  

By way of illustration, consider Mozambique. Since the late 1990s, its government has made a series of coordinated infrastructure investments to establish a development corridor between the country’s central provinces near eastern Zimbabwe and the seaside port of Beira. With high-resolution satellite measurements of nighttime light, one can visually track the creation and growth of this development corridor over time. In Figure 1, it is easily observable: it is the horizontal, orange-and-red “line” that bisects the middle of the country, bending southeast to the coast.


Figure 1: The Development of the “Beira Corridor” in Mozambique, seen through Nighttime Light Levels in 1992, 2001, and 2012


Some cash-strapped governments in resource-rich countries are going a step further, demanding that incoming foreign investors build and maintain local infrastructure in exchange for contractual rights to extract and export natural and nonrenewable resources like timber, rubber, oil, gold, and iron ore. 
Liberia is a case in point. The Ellen Johnson-Sirleaf administration has made foreign direct investment (FDI) the centerpiece of its growth and development strategy. However, unlike previous administrations that adopted a laissez-faire approach, it has required that private investors build and rehabilitate roads, ports, bridges, power plants, and health and education systems in and around the communities where their investments are physically sited.
The Government of Liberia has also prioritized investments in three specific geographic areas (see Figure 2) near population centers and existing markets where it sees the greatest potential to establish new spatial development corridors. Its goal in developing these corridors has been to maximize the economic multiplier effects that can be generated by infrastructure financed and supplied by concessionaires – in particular, mining concessions.

Figure 2: The Johnson-Sirleaf Administration’s Development Corridor Priorities

In partnership with the International Growth Center and Humanity United, we recently completed a geospatial impact evaluation to determine whether this investment management and coordination strategy is working. To do so, we first assembled a dataset of all natural resource concessions granted to investors in Liberia from 2004 to 2015.  We also constructed polygons that correspond to the specific tracts of land granted to investors, which made it possible to calculate at a high-level of spatial resolution if particular locations had been “treated” with FDI activity. We then merged these geocoded investment data with a remotely sensed outcome measure of nighttime light growth at the 1km x 1km grid cell level.
To establish a credible counterfactual, we used data from satellite imagery, weather stations, household surveys, and administrative records and propensity score matching methods to identify pairs of treated and untreated locations that were equally likely to receive natural resource sector FDI.  We then calculated the difference between nighttime light growth (between 2006 and 2013) in the treatment group and the control group.
Here’s what we found:
(1) Natural resource concessions improve local economic growth. On average, natural resource concessions increase nighttime light by .58% in the 25 km surrounding concession areas, which roughly corresponds to a .17% increase in local GDP.
(2) Mining concessions have a positive effect on local economic growth. By contrast, there is no robust evidence that agricultural concessions register consistent effects on economic outcomes.
(3) U.S. concessions do not have a discernible effect on local economic growth, but Chinese concessions do. In particular, Chinese mining investments seem to have a particularly large impact. We recovered evidence that that they increase nighttime light output by 1.26% in the 20 km surrounding their concession areas, which is roughly equivalent to a .38% increase in local GDP.
These results suggest that targeting and co-locating FDI projects and public goods investments in specific geographic areas can indeed set in motion local economic agglomeration processes. It is particularly notable that the natural resource concessions which were subject to more demanding public good provision requirements (i.e. mining investment projects) produced higher levels of local economic growth than those that faced less demanding public good provision requirements (i.e., agricultural investment projects). The fact that Chinese investors, who have generally been more willing and able to satisfy the Johnson-Sirleaf administration’s public good requirements than U.S. investors, are delivering stronger growth larger economic growth impacts is also consistent with the theory of change underpinning the government’s spatial development corridor strategy.
Going forward, future research ought to evaluate whether and how natural resources concessions impact non-economic outcomes, such as social protest, land conflict, violence, and deforestation. The growing availability of subnational geo-referenced investment, outcome, and covariate data has put this type of analysis well within reach.

This post draws upon findings from a new working paper, entitled “Government Policy and Natural Resource Sector FDI: Quasi-Experimental Evidence from Liberia”, that will be presented at the Annual World Bank Conference on Land and Poverty.  An interactive web portal with the subnationally georeferenced investment and outcome data referenced in this post can be accessed via  


Brad Parks

Executive Director, AidData

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000