The Dispossessed: Large-Scale Land Acquisitions, Elite Capture, and Dissent in Africa

The Cost of Enclosure: Why Large-Scale Land Deals Trigger Local Unrest

When large companies or wealthy locals take over big pieces of land in Africa, it often leads to much more protesting and rioting in those areas. This happens most often when local food crops are replaced by commercial farming on land that used to belong to communities. The study shows this isn't just a temporary problem but a long-lasting change in local stability.

For the last two decades, a global rush to acquire arable land has transformed millions of hectares across the developing world. Much of this movement was spurred by the 2007–2008 food price crisis. This crisis drove investors to seek "inflation-proof" assets in the form of farmland. While economists have long documented the logic behind these acquisitions, the political and social fallout has remained a subject of heated debate. Many observers feared "land grabs"—transnational seizures of territory—but empirical evidence has struggled to isolate the specific causes of the resulting instability.

A new study by Jonathan Dries seeks to resolve this ambiguity. He looks closer at who is actually doing the acquiring and what kind of land they are taking. The research suggests that the primary driver of unrest is not necessarily foreign corporations. Instead, it is a process of "elite capture," where domestic actors exploit legal loopholes to seize community land.

The Blind Spot in Land Governance

Current policy discussions often treat land acquisitions as a monolithic phenomenon. They focus almost exclusively on the risks posed by transnational agribusiness. However, this perspective overlooks a fundamental institutional friction found across much of Africa: the dual tenure system. This is a fragmented legal landscape where communities hold customary rights (informal, ancestral claims to land) while national governments simultaneously assert formal statutory ownership over any land that lacks official registration.

Because customary land is often managed through oral tradition rather than a central database, the state can treat these territories as "empty" or "idle" domain. This creates a loophole where authorities can lease community lands to private investors via administrative fiat (an official command). This effectively bypasses traditional rights without consent or compensation. Existing research has noted correlations between land deals and violence. However, as the paper notes, it has been difficult to separate the impact of the land deals themselves from the inherent characteristics of the land. Factors like proximity to water or fertile soil independently attract both investment and conflict.

Isolating the Impact of Implementation

To move beyond mere correlation, the authors employ a sophisticated econometric strategy. This is designed to handle "endogeneity"—the tendency for investors to pick high-value sites that are already prone to tension. Rather than simply comparing land-rich areas to land-poor ones, the study uses a "failed-deal" control group. The researchers compare successfully implemented projects against a group of deals that were abandoned or failed due to exogenous factors (external shocks like regulatory shifts) before any farming began.

The methodology follows three primary stages:

1. Counterfactual Imputation: Using a staggered difference-in-differences (DiD) imputation estimator, the authors estimate what would have happened in a specific location if the land deal had never occurred. This is analogous to a scientist using a placebo in a clinical trial to ensure the observed effect is caused by the drug and not the passage of time.
2. Spatial Buffering: The researchers aggregate conflict data from the ACLED database within a 50 km radius of each deal. This distance is chosen to represent the typical scale of a chiefdom or village cluster most affected by the transfer.
3. Mechanism Validation: Once the causal link to unrest is established, the study integrates three independent data streams. These trace how grievances move through society via media coverage (GDELT), community surveys (Afrobarometer), and electoral records (CLEA).

As seen in, the study maps 1,391 geocoded deals.

This allows for a precise spatial analysis of how unrest clusters around these points of implementation.

A 158% Surge in Civic Unrest

The results of this approach are stark. The paper reports that large-scale land acquisitions (LSLAs) cause a sustained increase in civic unrest of approximately 158% relative to the pre-treatment mean. Specifically, the authors find that implemented deals lead to an additional 1.48 protest or riot events per location-year. This is not a fleeting spike. As demonstrated in the event-study plot in, the increase in protest activity is durable and persists long after the initial negotiation.

The study further reveals that this unrest is highly concentrated in specific types of deals. The authors find that the escalation in unrest is most pronounced when domestic investors acquire community or state land for food-crop production. This is visualized in .

It shows that while transnational investors do trigger some response, the effect is significantly more intense and persistent among domestic actors.

The impact extends into the political arena as well. The authors report that in constituencies near implemented deals, opposition vote shares rise by 10.0 percentage points. Voter turnout also increases by 8.8 percentage points. This suggests that local communities are not just protesting in the streets. They are translating their grievances into formal political mobilization.

Limits of the Counterfactual

While the study is methodologically rigorous, it carries certain unavoidable trade-offs. The entire identification strategy rests on the assumption that "failed deals" are truly exogenous. If a deal fails specifically because a community was already planning to resist it, then the control group is "contaminated." This would mean the control group is biased by the very unrest the study aims to measure. The authors attempt to mitigate this by filtering out any failed deals that contain explicit mentions of community resistance. However, any remaining misclassification would likely bias the results toward zero. This means the actual impact could be even higher than reported.

Additionally, the study relies on the ACLED database for conflict tracking. The authors note that ACLED coverage is sparse in Africa prior to 2002. This means that for very early land deals, the "pre-treatment" baseline of unrest might be understated. Such an understatement could potentially inflate the perceived magnitude of the jump in conflict. Finally, the paper focuses on the 50 km buffer. While robustness checks show the results hold at other distances, the findings are inherently tied to this specific scale of community organization.

The Verdict: A Regulatory Gap

The evidence presented suggests that the current global focus on "land grabbing" is partially misplaced. If the goal is to promote stability and protect livelihoods, the primary target of oversight should not just be foreign agribusiness. Instead, regulators should look at the domestic political elites who exploit the ambiguity of dual-tenure systems.

The research points to a clear sequence of institutional failure. Land is seized through top-down formalization. This erodes trust in traditional leaders (the "custodial-failure" channel). It pushes grievances into the media and eventually manifests as electoral backlash. For practitioners and policymakers, the verdict is clear. International Environmental, Social, and Governance (ESG) standards must be expanded. Current frameworks focus heavily on transnational corporations. However, the data suggests that extending similar due diligence requirements to large domestic acquirers is essential to preventing localized cycles of unrest.

Figures from the paper