Tracking Underreported Financial Flows | Geospatial Impact Evaluations

Natural Resources and Poverty in Liberia

Can Liberia escape the "resource curse" by demanding more from investors in the natural resource sector?

Download the Working Paper (PDF)

Natural Resource Sector FDI and Growth in Post-Conflict Settings →

Download the Dataset from Github

Liberia Concessions Geocoded Resesarch Release Level 1 Version 1.0 →

This dataset is broadly compatable with the Open Contracting Data Standard

Explore the AidData Labs Portal

Geocoded Liberian Concession Boundaries and Covariate Data →

One third of Liberia's land is a natural resource concession.

Since 2006, the Liberia government has granted more than a third of its land to foreign investors. These companies have earned rights to extract and export iron ore, gold, palm oil, rubber, and other natural resources.

Photo Credit: Harsh Desai

Liberia's "Development Corridor" Strategy

The Government of Liberia requires these investors to build and maintain public infrastructure, such as roads, bridges, ports, and electricity grids. The central government wants to target these infrastructure investments in specific geographic areas in order to set in motion broader economic agglomeration processes and establish new "growth poles" or "development corridors."

2006: Ore Concession to MSL

$1.6 billion


25 years

Mittel Steel Liberia

Mittel Steel Liberia, a subsidiary of ArcelorMittal, agreed to a 25-year, $1.5 billion concession agreement in 2006 which included Corporate Social Responsibility (CSR) positions.

  • Spend $800 million rehabilitating a 267km railway
  • Preference Liberians in hiring decisions
  • Pay $3 million / year into a social development fund

2009: Ore Concession to China Union

$2.6 billion


25 years

China Union

In 2009, the Government of Libreria and China Union successfully negotiated the single largest investment agreement in the country's history: a 25-year, $2.6 billion iron ore investment in Bong Mines. The local employment provisions in China Union's contract are nearly identical to the provisions contained in the concession contracts held by Western investors like Mittal Steel Liberia.

Even countries with weak public sector institutions can negotiate contracts with foreign investors that require the provision of public goods.
AidData Working Paper 34

2011 IMF Survey:

"My local economy is directly benefitting from concessions agreements"

46% strongly disagree

8% agree

The general public seems to view the government's concession-led growth and development strategy with a similarly high level of skepticism. A 2011 survey of nearly 1500 rural and urban households in Liberia revealed that 46% of the population strongly diagreed with the notion that their local community was benefitting from concessions granted to investors since 2008 (IMF, 2012). Only 8% of surveyed households agreed with the statement, "[My] community is directly benefitting from concession agreements signed and ratified by the government since 2008."(IMF, 2012)

AidData Working Paper 34

Has Liberia's "Development Corredor" Strategy Worked?

Our study presents first-of-its-kind evidence on the local economic growth impacts of FDI in the natural resource sector. We first built a subnationally georeferenced dataset of all natural resource concessions granted to foreign investors and then merged it with survey- and satellite-based outcome and covariate data. We used remotely sensed data on nighttime light to measure local economic growth and quasi-experimental methods to compare growth in otherwise similar locations with and without FDI.

Our findings suggest that the Ellen Johnson-Sirleaf administration's spatial development corridor strategy is beginning to deliver real results.

Key Findings

Mining FDI — especially from Chinese companies — improves local economies, but futher research is needed to examine effects on non-economic outcomes.
  1. The evidence suggests that natural resource concessions improve local economic growth outcomes.

    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 subnational GDP.

  2. Mining concessions have a positive effect on local economic growth outcomes.

    By contrast, there is no robust evidence that agricultural concessions register consistent effects on local economic growth outcomes.

  3. Corporate social responsibilty (CSR) projects do not have not detectable effects on local economic growth outcomes.

    We found no evidence that concessions with CSR provisions economically outperform concessions without CSR provisions.

  4. U.S. concessions do not have any discernible effect on local economic growth, while Chinese concessions do.

    Chinese mining investments increase nighttime light output by 1.26% in the 20 km surrounding their concession areas, which is roughly equivalent to a .38% increase in subnational GDP.

Read the Full Working Paper

Natural Resource Sector FDI and Growth in Post-Conflict Settings: Subnational Evidence from Liberia →

Citing the Working Paper

Bunte, Jonas B., Harsh Desai, Kanio Gbala, Brad Parks, Daniel Miller Runfola. 2017. Natural Resource Sector FDI and Growth in Post-Conflict Settings: Subnational Evidence from Liberia. AidData Working Paper #34. Williamsburg, VA: AidData. Accessed at http://aiddata.org/working-papers.

How was this research conducted?

Phase 1: Tracking the universe of FDI in the natural resource sector

Compiling a first-of-its-kind dataset of natural resource concessions through an innovative, open-source methodology

With support from Humanity United and in partnership with the Concessions Working Group in Monrovia, AidData developed a first-of-its-kind dataset of all known natural resource concessions granted to investors in Liberia from 2004 to 2015. We systematically categorized 557 concessions on 43 different dimensions, including the names and nationalities of the investors, the nature of the rights granted to these investors (exploratory or extractive), and the presence or absence of contractual commitments to undertake corporate social responsibility activities (e.g. building schools and health clinics). We also developed a novel, polygon-based (rather than point-based) geocoding methodology that identifies the specific tracts of land granted to concessionaires (investors) to explore, develop, extract, or sell natural resources.

The methodology that was developed to assemble this investment-level dataset involves first standardizing and synthesizing several official sources of information on concessions in Liberia, and then supplementing these data with open-source information. The methodology is available here.

Phase 2: Testing hypotheses with credible counterfactual evidence

Quasi-experimental methods to estimate the treatment of all natural resource concessions on economic growth and compare projects by foreign investors.

Using a set of geospatial impact evaluation methods and tools developed by AidData, an interdisciplinary team of researchers from the College of William and Mary, the University of Texas at Dallas, the London School of Economics, and TrustAfrica were able to estimate the effects that different types of concessions and concessionaire have on local economic growth outcomes. They first used the polygons that correspond to the specific tracts of land granted to investors to calculate at a high-level of spatial resolution whether or not a particular location had been “treated” with FDI activity. They then merged these geocoded investment data with remotely sensed nighttime light data (a proxy for subnational economic development) at the 1km by 1km grid cell level. They then used propensity score matching methods to compare nighttime light growth (between 2006 and 2013) in otherwise similar subnational localities with and without FDI activity. To address the possibility that locations with FDI might be fundamentally different from locations that did not receive FDI and that these different might explain differences in nighttime light growth, the research team used a rich set of spatial covariates (e.g. wealth, literacy, employment conditions, proximity to population centers and markets) from satellite imagery, weather stations, household surveys, and administrative records to identify matched pairs of treated and untreated locations that were equally likely to receive the "treatment" of natural resource sector FDI. The method was used to establish a credible counterfactual — that is, to identify identify a set of comparison cases (1km by 1km grid cells) that had remarkably similar pre-treatment characteristics to the treated cases but that did not receive treatment.

This work is being generously supported by the International Growth Centre.

Unanswered Questions

Further research is needed to build on this study.
  1. How do concessions affect the very poor

    From the Working Paper...

    Follow-up studies could strengthen the robustness of our findings by analyzing outcome variables other than nighttime light growth. A potential weakness of our outcome variable is that it does a very poor job of detecting small or modest gains that accrue to very poor areas.

  2. How widely are economic benefits shared?

    From the Working Paper...

    Future research could examine the dynamics that determine the extent to which eocnomic benefits of natural resource concessions are shared. Distinguisghing between members of different ethinc or religious groups might help to uncover the political processes that determine who benefits the most and least from concessions-led growth and development strategies. It would also be useful to understand the effects of concessions on the income and employment prospects of men and women.

  3. What effects do concessions have on conflict and deforestation?

    From the Working Paper...

    Future research ought to examine the effects that concessions have on non-economic oucomes such as social protest, land conflict, violent conflict, and deforestation.... The growing availability of subnationally georeferenced investment, outcome and covariate data now makes this type of analysis possible.

For more information, contact info@aiddata.org