Large-scale green grabbing for wind and solar PV development in Brazil

The study investigates how the rapid expansion of variable renewable energies (VRES)—wind and solar PV—since 2010 in Brazil is linked to large-scale land control and tenure transformations, conceptualized as green grabbing. Brazil’s energy plans foresee major growth in wind and solar to diversify beyond hydropower and meet demand, yet weak land governance, overlapping and insecure tenure, and historical inequalities risk reproducing territorial conflicts and dispossession of traditional and Indigenous communities. The research asks who finances and owns these projects, how international capital is involved, and which land tenure regimes are used to secure project areas. It aims to provide quantitative and spatial evidence of the scale and modes of land appropriation associated with VRES, addressing a gap in empirical, nationally consistent assessments of investment–land tenure dynamics in Brazil.

The paper situates green grabbing within broader land grabbing scholarship that surged after the 2007–08 crises, expanding from farmland investments to environmentally framed appropriations for carbon sequestration, conservation, and renewable energy. Definitions are contested, but a central theme is control grabbing over large tracts through varied, often market-based mechanisms, increasingly justified by climate imperatives. In Brazil, research has emphasized illegal deforestation and land grabbing in the Amazon, the role of undesignated public lands (terras devolutas), insecure tenure, overlapping claims, and weak governance. Studies in the Global South link VRES to land struggles, loss of common lands, and non-recognition of communal rights. Monitoring tools like the Land Matrix face data gaps for investment and tenure, complicating quantitative assessments. The review highlights how digitized cadastres (e.g., CAR) can be misused to legitimize illicit claims and how policy changes enable legalization of prior grabs, underscoring the need to examine VRES-specific investment–tenure nexuses.

Study area: National-scale Brazil focusing on implemented wind and solar PV parks in Northeast, Southeast, and South regions. VRES asset mapping: ANEEL georeferenced data (as of 04/02/2022) for wind (parks, turbines, plants) and utility-scale solar PV (≥5 MW; points). Solar park polygon areas were derived from OpenStreetMap and validated with Google Maps. Ownership and investment: Bloomberg New Energy Finance (BNEF) datasets (projects, organisations, transactions) covering 2000–2021 were merged to identify direct/parent owners and direct/parent investors and their countries. The final dataset covers 25.5 GW wind and 8.6 GW solar (BNEF), acknowledging scope differences with IRENA statistics. Data integration: ANEEL assets (03/30/2022) matched to BNEF via fuzzy string matching, spatial proximity, and manual validation using public sources. Matches: 574/602 wind parks; 117/120 solar parks (aggregated to 44 polygons where multiple projects share one area; area split evenly across linked projects). Allocation of area to actors: For owners, land area shares were proportional to reported ownership, normalized where shares did not sum to one. For investors, areas were split between equity and debt using national gearing ratios (wind 70% debt; solar 58% debt), then distributed evenly within each investor category. Land tenure datasets and processing: Used national public datasets (SIGEF/SNCI for private property titles; SICAR for CAR registrations; FUNAI for Indigenous lands; INCRA for Quilombola and rural settlements; ISA for conservation units; IBGE for boundaries). Following a prioritized hierarchy based on legal security and precision: (1) private property titles (SIGEF/SNCI) take highest priority; (2) CAR claims (SICAR) supersede public layers but are overridden by private titles; (3) public lands (conservation units, Indigenous, Quilombola, rural settlements) lowest priority; remaining areas classified as undesignated public land. Data cleaning procedures aligned with Imaflora’s ATLAS approach; SIGEF/SICAR used to retain timestamps for privatization analyses. Control areas: For each park, created control areas by randomly relocating and rotating the park polygon within the same municipality, excluding overlaps with existing parks and ensuring containment. For wind, two controls: Control random and Control match wind resource (chosen to minimize absolute difference in wind power density from the Global Wind Atlas v2.1). For solar, one random control per municipality; solar radiation matching not applied due to intra-municipal uniformity. Temporal analysis: Compared dates of private property title submission/approval (SIGEF/SNCI) to the first financial transaction closure date (BNEF) for each park to assess temporal proximity between privatization and investment. Statistical analysis: Computed shares of area per land category for each park; two-sided t-tests compared means between park areas and control areas across land categories (private with legal title, private with CAR, public, undesignated public) for wind (n=574) and solar (n=44).

- International involvement: Foreign companies are involved (as owners, investors, or both) in 78% of wind parks and 96% of solar PV parks. - Spatial scale: Wind parks in operation/construction cover 2,148 km²; solar PV parks occupy 102 km². Despite smaller total area, solar PV has higher land-use intensity and more restrictive co-use. - Wind ownership/investment: Direct ownership is predominantly Brazilian (89% of wind park area), but many direct owners are subsidiaries of foreign parent companies. About 68% of wind park area is associated with foreign parent owners; Europe accounts for 52% of area, with France (12%) notable (Engie SA), and Enel SpA (Italy) as the largest foreign parent owner. Brazil accounts for 34% of parent owner area. Direct investors are mostly Brazilian (88%); parent investors 65% Brazilian. BNDES finances parks covering 15% of wind park area; 87% of BNDES-associated parks also have some foreign ownership/investment. - Solar ownership/investment: Parent owners are 90% foreign by area; parent investors 74% foreign. Non-Brazilian direct owners account for 46% of area; direct investors are mostly Brazilian (85%), though most are subsidiaries of foreign companies. Europe dominates foreign roles (parent owners 66%, parent investors 57% of non-Brazilian shares). Italy leads parent ownership (27% of area) and parent investment (30%); Brazil ranks second in parent investment (26%). Enel subsidiaries (Enel Brasil Participações Ltda; Enel Green Power Brasil Participações) are involved in 30% of solar PV area; Banco do Nordeste do Brasil SA covers 12.5% as a public financier. Actis LLP (UK) holds 12.5% as direct owner; CGN Energy International Holding (Hong Kong) holds 10%. - Land tenure composition: Legal private property titles cover 64% of wind park areas and 96% of solar PV areas. In wind parks, 28% of area is covered only by CAR claims; 8–9% overlaps with public (≈2%) or undesignated public land (≈7%). For wind turbine point locations specifically, 55% are on legal private titles, 38% on CAR-only private claims, and 7% on public/undesignated land. - Comparison to control areas: Shares of legal private titles in park areas are significantly higher than in controls. Wind controls show 21% (random) and 28% (matched wind resource) legal private title coverage; solar controls show 47%. Two-sided t-tests indicate significant differences (p<0.001) across most land categories between parks and controls, except public land in solar PV. - Temporal link between privatization and investment: 75% of privatizations occurred within 5 years prior to the first investment closure date, with marked acceleration around the investment date. Nearly half of wind park privatizations and one-third of solar PV privatizations occurred after the first investment, indicating a strong temporal association between VRES development and privatization activity.

The findings confirm the central research premise: expansion of wind and solar PV in Brazil is closely associated with large-scale land control and tenure transformation—green grabbing—facilitated by international and domestic capital. International networks, particularly European parent firms and investors, are deeply embedded, especially in solar PV. Wind shows greater domestic direct ownership, but foreign control is substantial through parent ownership structures. The predominance of private legal titles in park areas versus controls, and the significant use of CAR-only claims and even public/undesignated land (notably for wind turbine sites), illustrate diverse mechanisms of control grabbing. The temporal clustering of privatizations around investment milestones supports that VRES development drives privatization of public/undesignated lands, restructuring access and authority over land. These dynamics align with broader patterns where climate policy and low-carbon investments legitimize tenure modifications, potentially amnestying prior illicit claims. The results underscore tensions in public finance (e.g., BNDES) that, while accelerating renewable deployment, also enable foreign-corporate capture of land-based value. The implications are significant for spatial justice and rights of traditional and Indigenous communities, given fencing, security, and restricted access within park areas that curtail common-land use and livelihoods.

This study provides a quantitative, spatially explicit assessment of green grabbing in Brazil’s VRES expansion (2000–2021), linking investment/ownership structures to land tenure transformations. It shows extensive foreign (mainly European) involvement, strong reliance on privatization to secure land, and widespread use of weaker or illicit claims (e.g., CAR) and even public/undesignated lands, particularly for wind. Privatization activity is temporally associated with investment milestones, indicating VRES development is a driver of tenure modification. Policy and practice should prioritize transparent land governance, accelerate CAR validation and removal of illicit registrations, and strengthen recognition of communal and Indigenous rights (e.g., ILO 169, UNDRIP, IFC Performance Standards). Future research should combine quantitative tracking with field-based qualitative studies to establish causal pathways, document community impacts, examine pre-investment permitting and site-control dynamics, and account for non-operational or failed land deals in overall land competition assessments.

- Causality: The first investment closure date is an imperfect proxy for project initiation; permitting and land control actions may precede investment by years, limiting causal inference. - Data gaps: CAR dataset lacks time-stamped registration histories, hindering precise temporal analyses of private claims. Some private ownership claims exist only in state cadastres or local notaries and are not digitized nationally. - Coverage differences: BNEF excludes projects <1 MW (e.g., rooftop PV) and may include financed-but-not-yet-operational assets; merging with ANEEL required manual validation and may miss or misattribute some assets. - Spatial uncertainties: Overlapping and conflicting land claims, mapping errors, and potential forged titles introduce uncertainty despite prioritization/cleaning procedures. - External validity: The analysis focuses on operational/construction parks; broader dynamics include non-operational or failed land deals not captured here. - Statistical scope: Significant differences with controls are shown, but unobserved confounders (e.g., developer siting strategies, local governance changes) may influence patterns; solar radiation matching was not applied due to assumed intra-municipal uniformity.