Global Governance Meets Local Land Tenure: International Codes of Conduct for Responsible Land Investments in Uganda

Abstract Throughout the last decade, the international donor community has developed a plethora of regulatory initiatives for responsible agricultural investments. It remains unclear how such guidelines are invoked in practice in investment cases, and whether their use can prevent conflict and protect local land rights, as promoted. Uncovering how international guidelines work necessitates an understanding of the formal-legal setting and underlying land tenure regimes that shape investment projects. In Uganda, these contexts vary from region to region and investments take place on land held under various tenure regimes, including private, state-owned, and customary land. Based on 8 months of fieldwork in Uganda, I compare three cases of large-scale land investments in different settings and argue that variation in the underlying land tenure systems determines the variation, uneven applicability and effectiveness of global governance mechanisms.


Introduction
The management of conflict associated with the rise of foreign large-scale land investments in the developing world is a prominent topic on the global governance agenda. Agrarian justice activists and peasant rights organisations strongly advocate against foreign investments, pointing to violations of land rights of smallholder farmers, forced displacement and environmental pollution. In contrast, most national governments and mainstream development agencies, such as the World Bank and the United Nations organisations, emphasise the great potential of large-scale investment to contribute to poverty reduction by introducing employment opportunities, transferring technologies, and generating foreign exchange in developing countries (Deininger et al., 2011;Liu, 2014).
While acknowledging the risks and harm that have been associated with investments, these mainstream actors believe these are largely avoidable by improving state regulation and investor behaviour. In their view, if investments are done right, they can generate 'win-win' scenarios that benefit not only the investor, but also the local economy and communities in the vicinity of the investment. Efforts by the international community to mitigate negative effects of investments have thus centred around the creation of voluntary codes of conduct to discipline land deals so that they become more socially and environmentally sustainable. In this paper, I understand these international guidelines and codes of conduct as a range of 'soft laws' spanning from voluntary global governance

Sub-national Variation in African Land Tenure Regimes
This article engages with debates around the interactions between the global and the local governance sphere and, more specifically, is situated in a stream of literature emphasising domestic institutions and actors in shaping and governing large-scale land acquisitions (Boone, 2015;Fairbairn, 2013;Moreda, 2017;Peters, 2013;Widengård, 2019).
The phenomenon of land investments and the accompanying narratives and initiatives around the use of global governance mechanisms are still often perceived as an impact of 'the foreign upon the domestic' (Fairbairn, 2013, p. 36). This view, I argue, obscures the agency and relevance of and, importantly, the variation within national level politics, formal-legal frameworks and land tenure regimes in the context of land deals. The existence and significance of formal-legal frameworks governing land in African countries is often underestimated. In Uganda, numerous laws and policies exist to govern the land sector, even though these laws are not always implemented in practice. And in Uganda as in other African countries, the nature of land laws and practices governing land holdings and transfers varies across space. This means that global governance mechanisms are not introduced into a lawless void when invoked in situations of large-scale investment, but rather come into force within an existing formal-legal arena of varying land tenure regimes. This produces a varied pattern in the extent to which international guidelines gain traction as instruments to protect those using the land, even when investors are open to conforming to international norms. This variation is the central focus of this paper.
A large body of literature has recognised the importance of land tenure regimes for understanding historical and contemporary socio-economic and political dynamics in sub-Saharan Africa (Berry, 2002;Boone, 2013Boone, , 2014Lund, 2008;Mamdani, 1996;Peters, 2013;Platteau, 1996). This study builds on Boone's (2014) conceptualisation and theory of variation in land tenure regimes, defined as 'property regimes that define the manner and terms under which rights in land are granted, held, enforced, contested, and transferred ' (p. 4).
In most African countries, these institutional configurations vary across sub-national jurisdictions, meaning that different land tenure regimes exist within the same country. In Uganda, officially recognised land tenure regimes are the freehold, leasehold, Mailo, and customary regimes, as I will elaborate further below. Marking a departure from the theoretical precedent of rural Africa as an invariant and institutionless space, Boone's model of variation in land tenure regimes thus allows for systematic structured comparison and contrasting of different land tenure regimes (2014, p. 5). Land tenure regimes can be differentiated from one another according to forms of political authority, rules around land rights and citizenship rights, and territorial jurisdiction (2014, p. 6). Building on this theoretical concept, this paper zooms in on the differences in the recognition of land rights in different land tenure regimes in Uganda. Some land tenure regimes embody land rights that are more recognised and/or protected by African governments and international actors than others. For example, land rights under private property regimes, such as freehold tenure, usually enjoy de jure and de facto recognition and protection at the national level. But land rights of people living on other tenure regimes, such as customary tenure, are often not fully recognised or protected. The vulnerability of customary land rights in particular has been widely acknowledged (Alden-Wily, 2011;Kapstein, 2018;Peters, 2013). Peters (2013) argues that since the colonial era, customary land in much of sub-Saharan Africa 'has been treated as less than full property' (p. 2) and that customary landholders are particularly threatened by the rise of global large-scale land investments. Similarly, land rights of people living on state lands, such as state-owned forests or wetlands are often unrecognised by national governments and international actors. These differences, I argue, are significant for the way that global codes of conduct can gain traction in investments cases (or not).

International Governance Mechanisms
Advocates for international guidelines see these as a landmark achievement towards achieving food security and promoting sustainable development (FAO, 2012a). Many civil society organisations also praise the emergence of regulatory initiatives for their potential to provide a framework of reference to support local resistance against forms of 'land grabbing' (Milgroom, 2015). In contrast, Borras and Franco (2010) warn that the use of such guidelines could lead to a reframing of the narrative of 'land grabbing' towards capitalist and economics-based understandings of land investments as 'opportunities', and away from a social justice and 'pro-poor' perspective (p. 510). An important question coming out of these debates is how are such global governance mechanisms placed vis-à-vis national formal-legal frameworks?
Some argue that when formal-legal frameworks or human and financial capacities at the domestic level are not enough to provide desirable outcomes, the international community and global codes of conduct can step in to correct this (Deininger et al., 2011;Von Braun & Meinzen-Dick, 2009). International guidelines on responsible investment can be used to pressure governments and investors to uphold the host country's laws with regards to land rights, since these laws are not always adhered to in practice during investment cases. But international guidelines tend to go over and beyond the land laws of most countries, including principles like Free Prior Informed Consent (FPIC) for land acquisition processes and recommendations regarding human rights-based, environmental, gender, and pro-poor best practices. 'Outside pressure' from the international donor community is therefore often perceived as necessary to get states to adhere not only to their own laws but to also incorporate global governance norms for responsible investments. Sassen (2013) is sceptical about such interactions between global and national rulemaking. She sees foreign large-scale investments and associated global governance mechanisms as part of a larger transformation process in which national sovereign territory is slowly being converted to a commodity on the global market, and its governance subjugated to a form of global geopolitics rather than national laws.
In this article, I problematise the notion of the global level penetrating the local level and instead draw attention to the important role of (the variation of) local institutions, as outlined above. I perceive of international guidelines as sources of soft law that cannot replace national formal legal frameworks but are, in fact, dependent on them and on domestic authorities for the way they are invoked (or not) in structuring and regulating land investments.
The Voluntary (CFS, 2014, Art. 20). Other codes, such as the International Finance Corporation's (IFC) Performance Standards or the Equator Principles, have similar provisions. By calling for the recognition of legitimate land rights, international guidelines may recognise that some land rights are socially and historically perceived as legitimate by different groups, but these rights are sometimes not recognised in law (i.e. they are not de jure land rights) or may actively be denied. This poses several challenges. Who is going to decide if land rights are legitimate if the government has already taken the stand that they are not? What does it mean for international guidelines to protect legitimate land rights in situations of unclear, ambiguous and legally excluded land claims?

Methods
During eight months of fieldwork in Uganda in 2018, I collected information on 14 large-scale land investments from secondary and grey literature and conducted over 100 interviews and 10 focus group discussions. 2 The investment projects span different regions of the country, different systems of land tenure and governance, and different commodity sectors, including forestry, palm oil, sugarcane, coffee, tea, and other agricultural crops. Of the 14 cases, I selected three cases ( Figure 1) to illustrate the correlation between land tenure regime and the way that international guidelines work as observed in my larger sample of cases. These three cases capture the main lines of contrast across investment cases on Mailo, state, and customary land.

Land Tenure Regimes in Uganda
I group Uganda's land tenure regimes into three categories: private, state-owned and customary land. Freehold and Mailo land tenure both fall under the category of private land. I will focus here specifically on Mailo land as it is essentially a form of freehold with additional characteristics, but more prevalent than freehold tenure in Uganda. While not officially a recognised tenure regime of its own, I include state-land in my categorisation as substantial parts of Uganda's land is owned or under the control of the government, and many large-scale land investments are implemented on stateowned land. In turn, I do not include leasehold tenure as a stand-alone category in the analysis. Leasehold tenure is recognised in Uganda's Constitution but is, in practice, not a land tenure regime of its own. Leasehold titles can be issued on all other land tenure categories in Uganda. All foreign investors must acquire a leasehold title as foreign ownership of land is impossible. Customary land tenure prevails throughout much of Uganda. Land tenure regimes roughly map onto geographical regions in Uganda ( Figure 2). Mailo tenure is practiced today only in Uganda's Central Region, which comprises the Buganda Kingdom, as well as Kibaale District (Western Region), which was part of Buganda Kingdom until 1964. Customary tenure is mostly found in northern and eastern Uganda, but also features in the Western Region. Noticeably, state-land does not feature on this map, as pockets of state-land exist all over the country, not just in one specific region.

Private (Mailo) Land
Both freehold and Mailo tenure are officially recognised and protected by law, registered in a central registry, and equip the landowner with exclusive ownership rights. In contrast to Freehold, Mailo tenure is characterised by a dual landownership structure, involving the landlord, who owns the land title to a plot of land 'in perpetuity', and occupants or 'tenants', who occupy a part of this plot, and also have transferable rights to this part of the land. Legal occupancy rights on Mailo land are divided into the categories of 'lawful' and 'bona fide' tenants. A lawful tenant on private Mailo land includes a) a person occupying land by virtue of repealed earlier laws, b) anyone who settled on the land with the consent of the landlord, or c) a person who had occupied land as a customary tenant but whose tenancy was not disclosed or compensated for by the registered owner (Republic of Uganda, 1998, Art. 29). A bona fide occupant on Mailo land is someone having occupied and farmed the land uncontested for at least twelve years before the 1995 Constitution, as well as persons who had been resettled on the land by the government before that date (Republic of Uganda, 1998, Art.29;Coldham, 2000, p. 66). Legal and bona fide occupants may undertake comprehensive transactions on the land they occupy, subject to consent by the landlord. Land rights by Mailo occupants are protected and enforceable by law -the result of a long history of land conflict and land reform -and hence provide the claimants with legitimate and legible land rights in the eyes of the Ugandan government and the international community. Mailo ownership titles are registered and recorded in a central registry. Even though tenants have no outright ownership or documented protection of their claims (titles), the land rights of lawful and bona fide tenants are robust in the Ugandan political-legal context. It is difficult for the state to legally evict Mailo occupants. Conflicts are frequent but are handled in legal or legalistic ways (i.e. court cases, grievance mechanisms).

State-owned Land
There is ample room for confusion about what exactly constitutes state land in Uganda. State-owned land is often conflated with 'public land' and both concepts tend to take on different meanings at the national and local level. With the coming into force of the 1995 Constitution and its declaration that 'all land in Uganda belongs to the citizens of Uganda' (Republic of Uganda, 1995, Art. 237), stateowned land technically ceased to exist, and the concept became questionable. However, plenty of land is still under the control of the Ugandan government. Art. 238 of the Constitution gives the Uganda Land Commission the mandate to hold and manage any land in Uganda owned by the Government of Uganda. Such land refers to government-owned landholdings that are registered and titled -understood as private domain government assets -and handled by the state like freehold land, meaning it can be sold, leased, mortgaged or divided up. This also refers to government land in the public domain, which, in the Ugandan context, can be further separated into public land at the national and the local level. At the national level, public land held in trust by the government includes natural resources and nature reserves (including forest reserves, wetlands, rivers, game reserves) vested in the state. At the local level, public land refers to land that is 'not owned by any person or authority' (Republic of Global governance and local land tenure 587 Uganda, 1998, Art. 59). The management of the latter is vested in District Land Boards, who can issue leasehold titles on such land to individuals, groups or companies (Republic of Uganda, 1995Uganda, , 1998. State-owned land, especially public land at the national level, gives rise to multiple informal and ambiguous land ownership and use claims. In contrast to claims on private land, these are not protected or recognised by law. As I will show in the case studies, they are broadly ignored by the Ugandan government in cases of large-scale land acquisition.

Customary Land
Under customary tenure, land is governed and managed according to the norms and practices of a particular (customary) community. Abundant literature on African land politics perceives customary tenure in sub-Saharan Africa as a co-creation of colonial authorities and local people. Colonial rulers used state-backed local leaders (recognised as 'traditional' or 'customary' by the colonial administration) to indirectly govern rural populations (Boone, 2014(Boone, , 2015Mamdani, 1996;Peters, 2009). Most land in Uganda (and in most of sub-Saharan Africa) is not held by individuals as formal and titled private property. Instead, over 80 per cent of Ugandans hold their land under customary tenure (Obaikol, 2014, p. 55). Customary land holding structures vary from region to region, with customary land held by nucleus families mostly found in western Uganda, while extended family (clan) structures are more predominant in northern Uganda. Customary tenure can also include communal ownership of common grazing and hunting grounds. This is predominant, for example, in northeastern Uganda (Karamoja), where cattle herding predominates. In Uganda today, customary land is de jure on par with other forms of land tenure and was officially recognised for the first time in the 1995 Constitution. De facto, however, customary tenure is not treated as equal to the other tenure regimes. There is no registry for customary land. The Land Act of 1998 provides for the option of acquiring certificates of customary ownership (CCOs) that can serve as confirmation and evidence of customary ownership of land and give the owner transactional rights in accordance with customary law (Hunt, 2004, p. 177). In practice, however, these certificates are seen as inferior to a land title (i.e. freehold, Mailo, leasehold), especially by financial institutions. The 1998 Land Act provides for the option of converting a CCO into a freehold title, following official surveying procedures. This further indicates that CCOs are considered an inferior and intermediate step towards what is often deemed by states and international actors to be the ultimate goal, a freehold title. Very few such certificates have actually been applied for or issued in Uganda so far. 3

Case Study 1: A large-scale Investment on Mailo Land
Based on my larger sample of cases, the following case is typical of dynamics emerging around largescale land investments on Mailo land. While conflict emerged during the implementation of the project, the investors and the government demonstrated strong political will to avoid and make amends for potential infringements of Mailo land rights. The cementing and legibility of Mailo land rights in Uganda's formal-legal framework made the invocation of international guidelines not only possible but also straightforward. This helped steer a transparent procedure of land acquisition and land conflict redress.

The Case of the Vegetable Oil Development Project (VODP)
In 2003, the Vegetable Oil Development Project (VODP), formerly known as the Oil Palm Uganda Limited (OPUL) project, was established as a public-private-partnership between the Government of Uganda and two international palm oil companies, the Kenyan company BIDCO Ltd. and the Singaporean company Wilmar International. The Government of Uganda owns 10 per cent of the shares of the VODP project. The project was funded by the International Fund for Agricultural Development (IFAD) and supervised by the World Bank (Vorlaeufer, Schilling, Kirk, & Graefen, 2018). Located off the shores of Lake Victoria, on Bugala Island, in Kalangala District in Central Uganda, operations under Phase 1 began in 2003. The island comprises an area of approximately 270 square kilometres. 16,000 hectares were initially allocated for the development of oil palm plantations (Carmody & Taylor, 2016). In late 2017, the second phase of the VODP project included the expansion of the project to neighbouring Buvuma Island in Buvuma District, where the cultivation of another 10,000 hectares of palm oil was planned (Interview with company representative, 30 August 2018).
The government committed to identify and allocate 6,500 hectares of land to BIDCO and Wilmar, which was to be 'free of encumbrance' (that is, free of claims to land ownership or use by local communities), and suitable for agricultural production under a 99-year lease (IFAD, 2011). Approximately 3,000 of the 6,500 hectares were retrieved from formerly public land and 3,500 hectares were purchased by the government from private Mailo landowners and tenants.
As a partnership between the private sector, the government, and farmer organisations (Kalangala Outgrower's Trust), and funded by international organisations, the project was considered an 'example of innovation in development cooperation' ( Recommendations contained in international guidelines featured at many stages of the project implementation. Particular attention was given to securing local land rights. The government set up a Land Acquisition Task Force to carefully monitor the Mailo land purchases. Land acquisition was based on a 'willing buyer -willing seller' concept (Barbanente et al., 2018, p. 13), and the principle of Free Prior Informed Consent (FPIC) was applied for both landlords and tenants.
However, the Mailo land purchases posed several challenges to the project. Due to a complicated history of population displacement during the colonial era, much of the Mailo land on the islands belonged to 'absentee landlords' and has, over the years, been settled by a mix of Mailo occupants, some of which claimed legal or bona fide tenancy status, as well as those the law considers to be 'squatters'. This multi-layered structure of Mailo land claims has complicated the process of land purchases since disentangling ownership and tenancy rights, lack of consent by absent landowners, and disagreements over boundaries took time and resulted in land-related disputes in some instances (Barbanente et al., 2018, p. 13). These quickly captured the attention of donors, civil society, the government and the investors, who responded swiftly and diligently. NGOs supporting local communities referred to national legal requirements for protecting Mailo land rights as well as to international guidelines to pressure the government and the investors to conform to 'best practices' for responsible land acquisition. NGOs also invoked the use of grievance mechanisms, provided by the international financial organisations involved in the project. In 2016, the NGO Friends of the Earth, together with the Ugandan NGO National Association of Professional Environmentalists (NAPE), filed a complaint about the activities of one of the foreign investors, BIDCO, at the Social and Environmental Compliance Unit of the United Nation's Development Programme (UNDP) (Vorlaeufer et al., 2018, p. 15). This was possible since BIDCO is a member of UNDP's Business Call to Action Alliance, a global advocacy platform promoting pro-poor business models. The complaint against BIDCO accuses the company of violating guidelines by causing displacement of land users and environmental harm (Vorlaeufer et al., 2018, p.15). In early 2017, NAPE filed a further complaint to the International Finance Corporation (IFC) on behalf of the Bugala Farmers Association. This avenue for grievance claims was possible since BIDCO was partially funded by the IFC.
The provision of grievance mechanisms is part and parcel of most international institution-specific principles for responsible investment. For example, the UN Guiding Principles on Business and Human Rights state that 'business enterprises should establish or participate in effective operationallevel grievance mechanisms for individuals and communities who may be adversely impacted' (United Nations, 2011, Art. 29). I consider the use of grievance channels by local communities and NGOs to be part of a larger repertoire of strategies for pressuring governments and investors to conform to international guidelines.
The involvement of international actors in the project, the advocacy by NGOs and the use of international grievance mechanisms had substantial effects on the further development of the project. The complaint filed by NAPE to the IFC resulted in the establishment of a voluntary dispute resolution process between members of the Bugala Farmers Association, BIDCO, the Ugandan government, and aggrieved Mailo landowners, resulting in the granting of compensation payments to all Mailo tenants and people living on the land, regardless of their legal status (Barbanente et al., 2018, p. 14). The investors also changed their land acquisition strategy for the expansion phase to Buvuma Island. According to a VODP representative, '[l]and challenges are even higher in Buvuma Island. So here we decided to simply compensate everyone, regardless of the status of the tenant' (Interview with company representative, 30 August 2018). The project thus surpassed the requirements of Uganda's formal-legal framework, which only requires compensation for 'lawful' and 'bona fide' tenants, and compensated all people living on the land, regardless of their status. In line with the principle of FPIC and other international standards of inclusive deal-making, the investors further undertook 'a comprehensive, participatory stakeholder mapping (. . .), covering all land occupants prior to land acquisition (. . .) and both owner and occupants must agree on the acquisition'. (Vorlaeufer et al., 2018, p. 18).
Mailo land rights are both de jure and de facto recognised and protected by the Ugandan government, and therefore constitute graspable and legible claims for actors invoking international guidelines for responsible investment. Despite the occurrence of land conflicts in the first phase of the project, the way these were redressed demonstrates strong political will on the part of the government and the investors to avoid any potential (further) infringement of Mailo rights. This may be partially due to the nature of the project as a high-profile international public-private-partnership. More importantly, I argue, the firm recognition of Mailo land rights in Uganda's laws as well as the history of protection of these claims by the state created the conditions for a legal and transparent procedure of land acquisitions for the project and the possibility for international actors and NGOs to pressure the investors and the government to adhere to intentional principles, beyond the national legal requirements.

Case Study 2: A Large-Scale Investment on State-Owned Land
In the context of my larger sample of cases, this case is representative of dynamics emerging on stateowned land. Diametrically opposed to the case on Mailo land, the government does not recognise claims to land rights emerging from land users on state-owned forest reserves. Reference to global governance mechanisms did not lead to the protection of land rights or the resolution of land conflicts. Instead, the thrust of the international guidelines was refracted and came to focus on environmental aspects of the investment project.

The Case of the Busoga Forest Company
In 1996, the Government of Uganda, through the National Forestry Authority (NFA) granted the Norwegian Company Green Resources over 11,000 hectares of forest land within two of its stateowned Central Forest Reserves for the purpose of plantation forestry and the restoration of degraded forest land. The company now manages two pine and eucalyptus plantations in the form of 49-year lease concessions. The licenced area under the name Busoga Forest Company, which constitutes this case study, covers over 9,000 hectares of the Bukaleba Forest Reserve on the northern shores of Lake Victoria, in Mayuge District, Uganda's Eastern Region. The company's other plantation, the Kachung Forestry Project, covers 2,669 hectares in Dokolo District, Northern Uganda. The Bukaleba Forest Reserve is one of 506 state-owned Central Forest Reserves under the management of the NFA. These Reserves fall under the category of 'public land' at the national level. The central government is effectively the 'landlord' and can lease out these forests (or parts thereof) by way of concession, licence or permit (MLHUD, 2013b, p. 27). Leases to private individuals are prohibited and it is illegal to pursue subsistence farming, livestock grazing, or similar non-forestry related activities inside these Reserves (NFA 2003).
The implementation of the Busoga forestry project triggered a myriad of conflict dynamics. When Green Resources first acquired the investment licence in 1996, several communities, comprising around 15,000 people, were located inside the forest area licenced to the company. In 2000, the NFA evicted people from the Reserve claiming they were 'squatters' on state-owned land. Some communities managed to avoid evictions and remained in a cluster of villages inside the Reserve.
Not all forest dwellers claim to have legitimate land rights in the forest. Some acknowledge that it is state-owned land to which they have no formal right but argue that they have 'nowhere else to go' and no material means to resettle (Focus Group Discussion, 13 July 2018). Others are adamantly Global governance and local land tenure 591 defending the legitimacy of their land rights on the basis of two different historical claims. One group argues that they migrated into the area in the 1970s to work on a state farm inside the forest and were given land by the Idi Amin regime. They claim they have lived there for several decades, invoking a 'bona-fide' settlement status (Focus Group Discussion, 13 July 2018). Others argue they have ancestral (customary) land rights in the forest reserve from the pre-colonial era. After being forcibly evicted as part of the Tse Tse fly eradication programmes led by the colonial government in the Lake Victoria region in the first half of the 20th century, these people claim to have returned to their ancestral land in the 1980s (Focus Group Discussion, 13 July 2018).
The NFA does not consider these communities to have legitimate land rights in the forest and refutes the notion of 'ancestral land' altogether. 'They are playing the political card and it is basically all about political bargaining' (Interview with NFA officer, 6 June 2018). In 2011, after ongoing tensions over the issue of forest dwellers in the project area, President Museveni issued a directive to allocate 500 hectares of the company's licenced area to these communities. This, however, did not resolve the conflicts. Instead of land being 'allocated' to communities in a formal and legalistic manner, the situation is better described as an extra-legal 'toleration' of communities to stay in the forest. Without land titles or other documentation, their land tenure status remains unresolved and in a legal grey zone. There is no formal procedure in place to legally carry out the presidential directive to give 500 hectares of forest land to communities. This caused substantial confusion: On the one hand, communities are 'tolerated' to stay and cannot be evicted anymore. But the NFA still invokes laws about land use inside forest reserves, which strictly prohibit non-forest related activities. In many cases, people were arrested, and their cattle confiscated (Focus Group Discussion, 13 July 2018). The communities are effectively still treated as 'encroachers': The problem is that this agreement [to hand over 500ha] was made on a loose communication by the president. (. . .) There is no individual nor an institution that is assigned to follow through on this or follow up. It was merely a political move. In the meantime, we are still seeing the people living there as encroachers. We are chasing away the grazers, who bring their livestock into the forest (Interview with NFA officer, 6 June 2018).
Referring to the illegal nature of 'encroachers' in Central Forest Reserves, as outlined in Uganda's formal-legal framework, the company also does not recognise a legal basis of land rights for these communities.
If you were to use the law, if you were to use any forms of (. . .) regulation, then all those people would be gone by tomorrow. And they [the communities] would never go to any court of law to win any sort of case because they have no grounds (Interview with company representative, 6 June 2018).
Considered a matter of national jurisdiction, the company effectively deferred the responsibility to deal with land issues back to the Ugandan government. 'We do not have the tools nor the mandate to deal with [evictions]' (Interview with company representative, 6 June 2018). In the absence of government action to resolve the question of forest dwellers, the company started to apply 'soft' pressure to get people to leave the forest. By financing schools and hospitals outside of the Reserve, they aimed to incentivise people to relocate (Interview with company representative, 6 June 2018) They also continued to cultivate tree plantations in close proximity to the villages, which substantially limited the communities' farming and livelihood possibilities, thus driving them out.
In this case, international actors invoked global governance mechanisms as part of forestry certification schemes with regards to sustainable tree planting practices, climate change mitigation strategies, and biodiversity safeguarding (Interview with company representative, 6 June 2018; Green Resources, 2016, p. 7). In 2011, the Bukaleba Forest Plantation was certified by the International Forest Stewardship Council (FSC), and, in 2012, as an Afforestation and Reforestation project under the Verified Carbon Standard (VCS), a system to certify carbon emissions reductions. Despite the invocation of these guidelines and certification instruments, the protracted conflict surrounding the remaining communities inside the Reserve remained unsolved and unaddressed by the government, the company, and international and civil society actors.
The lack of mobilisation by NGOs -usually acting as 'watchdog' organisations -in response to the plight of forest dwelling communities is further striking in this case, particularly in comparison to the well-documented case of the company's sister plantation in northern Uganda. 4 I argue the lack of engagement on the part of international actors and NGOs in this case comes as the result of the de jure non-recognition of land rights of forest dwellers in Uganda's legal framework. Instead of addressing the ambiguous, illegible and historical claims to land rights of forest communities, international guidelines invoked in this case focused exclusively on environmental aspects of the project. In applying 'soft pressure' to incentivise communities to relocate, the company was seemingly forced into action by institutional paralysis, and the lack of guidance from the international instruments.

Case Study 3: A Large-scale Investment on Customary Land
The last case study is an example of a large-scale investment on customary land. Even though customary land rights are enshrined in Uganda's legal framework, they are often not recognised or protected in practice by the government due to their lack of documentation. The general call of international guidelines to protect legitimate land rights in this case was subverted by the conversion of customary land rights to a more 'legible' tenure regime and by the arbitrary decision over legitimacy of rights by the government and the investor.

The Case of the Atiak Sugar Factory
The Atiak Sugar Factory is a sugarcane investment by Horyal Investments Holding Company Ltd., owned by well-known entrepreneur Aminah Hersi Moghe. Located in Amuru District, northern Uganda, the project is based on a nucleus estate as well as substantial outgrower farming areas. The size of the farm can be estimated to be between 4-6,000 hectares (Interview with local government official, 17 August 2018). Operations started in early 2019. The government supported the project with around 26-28 billion UGX Shillings (approx. 6-7 million USD). In 2018, the government also bought a 10 per cent share in the company.
The land on which the project was implemented is customary land and the land acquisition was a two-fold process. The Horyal company negotiated a contract with an extended family that claimed ancestral customary land rights over the plot of land destined for the sugarcane project. After an agreement was made, the company helped to formalise the family's land claims by facilitating the issuance of a freehold title in the family's name at the Amuru District Land Board. Thereafter, as the landlord with a documented land title, the family granted the company a leasehold title over a large portion of this land for the sugarcane project.
The process of land demarcation and the formalisation of the family's land title sparked a dispute between the extended family and surrounding communities. According to my interviewees, some neighbours and community members argued that the family overextended their land boundaries during this process, and then leased land to the company that was not theirs to give (Focus Group Discussion, 17 August 2018). For example, part of the land that was leased to the company was a customary communal hunting grounds that they had used before the war (Focus Group Discussion, 17 August 2018). This land is not inhabited but is subject to communal ownership and could not be leased out by any individual family. The 'landlord' family is wealthy and politically well-connected. According to my interviewees, complaints over land claims violations voiced by neighbouring communities were squelched, people were arrested, and some villagers were forced to relocate or lost access to their farmland (Focus Group Discussion, 17 August 2018; Interview with local government official, 17 August 2018). The paramount chief (rwot Kweri) of Atiak, together with several community members, sued the extended (landlord) family over land rights violations. The case was taken up by a law firm in Kampala. However, before the case reached the courthouse, President Museveni himself invited the conflicted parties to his residence for a private audience and asked that the case be handled outside of court (Interview with defence lawyer, 10 September 2018). This demonstrates the level of influence of the landowning family and the involvement of political interests in this case.
While no explicit reference to international guidelines was made in this project, numerous global norms for responsible investment seemed to feature in the case. The Atiak Sugar Factory was praised by the general public and media in Uganda for a participatory and inclusive ideology, the empowerment of women groups and sugarcane-outgrower farmers, and an overall pro-poor, developmental and even philanthropic orientation (CWEIC News, 2018;Muwanga, 2020). The investor seemed to follow a 'bottom-up' approach to land acquisition by negotiating a lease contract directly with what was believed to be the rightful landowning family (Interviews 17 August 2018 and 18 November 2018). Direct engagement with local land users is seen as a 'gold standard' of responsible investment by most international guidelines (i.e. VGGT, CFS rai). By facilitating the formalisation of a freehold land title for the family, the call for securing legitimate land rights, as outlined in most international guidelines on responsible investments, was seemingly met.
But there are two problems with this. First, the formalisation of the land rights for the family effectively converted the tenure regime from customary to freehold. Adherence to international guidelines thus seemed to endorse the transformation of such land into a 'legible' and formalised tenure system, indicating the ill-fit of international guidelines with undocumented and ambiguous customary land claims as such. Next, the government and the company played a central role in identifying and solidifying the land rights of one (wealthy and well-connected) family, and with the issuance of a freehold title made the land indisputably the family's private property. Other claims to customary rights to the land by neighbouring communities and families were thus erased. It seems that in the absence of documented land rights, the government has substantial leverage to decide which rights are going to be recognised and which not. Potentially legitimate land rights of surrounding families therefore constituted a legal 'grey zone', in which there is no formal-legal basis for advocates of international guidelines to pressure the government or the investor to protect or recognise these rights.

Discussion and Conclusion
This paper has extended the scope of theorisation around land tenure regimes to a new domain, namely, the role of international guidelines in influencing large-scale land investments in African countries. The theory and concept of variation in local land tenure regimes offers good leverage in explaining why global governance mechanisms around large-scale land investments might work differently in different settings.
I argued that the adherence to global governance norms by various actors is shaped by whether and how the land rights in question are already legally recognised and protected by the state, which varies from one tenure regime to another. I presented case studies of land investments on three different land tenure regimes with varying degrees of recognition of local land rights by the state, ranging from a 'strong' recognition (Mailo), to a more ambiguous level of recognition (customary), to the absence of such recognition altogether (state land).
On land tenure regimes, where land rights are de jure recognised, legible (documented), and enforceable in a court of law, international guidelines can gain traction to force the government to honour its own laws, as well as integrate international best practice standards. In the case of the Vegetable Oil Development Project (VODP) on Mailo land, the rights of Mailo landlords and tenants were firmly enshrined in Uganda's legal framework. During the land acquisition process, disputes around overlapping Mailo occupancy rights were 'visible' conflicts that were rapidly addressed in legalistic ways by the government and the investors. Through the funding and supervision by IFAD and the World Bank/IFC, an important link to the international arena and 'outside influence' was established. NGOs took up an active 'watchdog' role to protect Mailo land rights and supported aggrieved Mailo land users in filing grievance reports to these organisations, which led to an altered strategy for the next land acquisition phase of the project: In order to avoid further potential disputes over occupancy rights, and to ensure the principle of FPIC, the company decided to compensate all land users even though national law only required the compensation of Mailo tenants with official lawful and bona fide status. This is thus an investment case in which the condition of legal recognition of land rights by the state was met, and in which 'outside pressure' by international organisations and NGOs resulted in additional adherence to international guidelines and worked to protect local land rights and resolve conflict.
In contrast, on land tenure regimes where land rights are not de jure recognised by national law, international guidelines have no legal basis to protect these. In the case of the Busoga Forest Company on a state-owned Central Forest Reserve, claims to land rights invoked by forestdwelling communities were not recognised by the Ugandan government. Local communities claiming rights to forest land on the basis of ancestral (customary) rights or having received land as gifts by previous governments were considered 'squatters' by the current government. They were sometimes allowed to remain on the land in a legal 'grey zone'. This case shows that if the government does not recognise the land claims in question as a minimum condition, then the hands of international organisations and NGOs are tied in terms of pressuring the government or the investor to adhere to national laws or global norms with regards to protecting these land claims. In this case, global governance mechanisms were not altogether absent as the project was funded by international financial institutions. However, reference to global norms seems to have been limited to addressing environmental aspects of the investment instead of the ambiguous land claims of forest dwelling communities. These remained unresolved and were treated as a matter pertaining to national (sovereign) jurisdiction.
In turn, on land tenure regimes where land rights are de jure recognised in principle but are often ambiguous and illegible due to the absence of documentation of these rights, as is the case with customary tenure, the ability of international guidelines to gain traction is dependent on the government's recognition of these rights in individual cases. Not all customary land rights are thus recognised unproblematically, in and of themselves. This means that even if land rights are legally (de jure) recognised by the Ugandan government, this does not automatically mean that the state will protect them in practice. The Atiak Sugar Factory on (originally) undocumented customary land in northern Uganda was hailed an example of a 'best practice' and pro-poor investment. The company was commended for negotiating a land deal directly with a local family and helping to secure and formalise their land claim. Yet, while this family's land claims were recognised and, in the course of the investment converted into private property, land claims of neighbouring customary land users, including communal land rights, were invisible and ignored. The apparent conformity with international guidelines by the investor and the government in terms of engaging at the local level and securing the family's land thus actually worked to delegitimize customary land claims for other people. In fact, by formalising the land rights for some already privileged local landowners, the investor contributed to a process of transformation of land tenure away from customary and communal (unregistered) land use to increased privatisation and formalisation of land, and the emergence of landlordism in northern Uganda. This case demonstrates that actors invoking or making reference to international guidelines are often unable to engage with ambiguous, undocumented, contested, and unprotected land rights. As discussed by Boone (2019), titling and formalisation of land always generates winners and losers. During investments on customary land, the already privileged benefit while the poor often lose out. The Ugandan elite families leasing land to international investors benefit from formalisation processes '[b]y laundering power as legitimate authority and by taking possession of land as property through government instruments of law and policy' (Peluso & Lund, 2011, p. 647).

Global governance and local land tenure 595
This paper provided insights and empiric evidence of the way that international guidelines work in practice in cases of large-scale land investment, which has so far been based mainly on presumptions. International guidelines tend to go over and beyond the national laws of most countries. Even if investors conform to the country's land laws, international guidelines are still needed to ensure truly responsible and 'pro-poor' investments (in the eyes of the international donor community). These cases from Uganda suggest that states will be more inclined to respect their own land laws if the land rights in question are clearly legible and recognised in the formal-legal framework. If this is the case, states and investors will be further inclined not only to conform to national laws but also to make reference to international guidelines if there is enough 'outside pressure' from the international donor community and/or domestic or international civil society to do so. In turn, when local land rights are not state-recognised, or state-recognised but not clearly 'legible' as in the case of customary land, the advocacy for international guidelines will risk irrelevancy if these are promoted as the only blueprint for addressing investment-related land conflict.
It is not the intention of this paper to discredit global governance mechanisms and their potential to guide and discipline large-scale land investments. The use of such instruments, often propelled by NGO activism, can be immensely helpful, as the case study on Mailo land shows. At the same time, these guidelines, with their intention to protect legitimate land rights, often provide little guidance in cases of ambiguous and unclear land rights and leave fundamental questions unanswered. On state-owned land, for example, how should the Ugandan government deal with local land rights that derive their legitimacy from the land allocations, de-gazettements, and forest use policies of previous regimes? Can such land rights be simply erased, and can people be evicted if the present government has not engaged with the forest dwellers in decades, thereby indicating its tolerance of these communities? Similarly, for undocumented customary land claims, how can actors such as civil society organisations invoke global governance mechanisms to protect land rights if the central government decides to ignore them, or if there are conflicting claims at the local level?
Given these shortcomings in our understanding, there is much need for further first-hand evidence of how such guidelines work (or not) in practice. Variation in the effectiveness and effects of international guidelines across the three land tenure systems in Uganda suggest that we should expect such variation in other countries.
Similarly, more research could analyse the impact of international actors' tendency to focus on land tenure regimes and land rights that are 'legible' for outsiders (i.e. titled and registered property rights), and how land rights that may be acknowledged locally, even by governments past or present, often fall off the radar of international initiatives. In this regard, there is much scope to further explore methods of securing land rights through locally appropriate forms of formalisation, including documentation of land rights through state-sanctioned protocols that could result in wider legibility of ambiguous land rights.