Broken promises: a rights-based analysis of marginalised livelihoods and experiences of food insecurity in large-scale land investments in Tanzania

Proponents of large-scale land investments (LSLI) still promote them as a development opportunity, which can lead, among other benefits, to job creation and enhanced food security for local communities. However, there is increasing evidence that these investments often deprive affected communities of their access to land, with multiple negative impacts on livelihoods, food security and on the environment. This paper relies on empirical data to present an analysis of LSLI and food (in)security – crucially at the level of individuals in two villages in the Ruvuma region, Tanzania, over 10 years after the acquisition of village land within the Southern African Growth Corridor of Tanzania (SAGCOT). We introduce an innovative framework that permits an integration of a rights-based approach with the Sustainable Livelihoods Framework to explore smallholders’ livelihoods and experiences of food insecurity. Our paper demonstrates how this integration, along with attention we have given to the FAO’s PANTHER principles, adds the missing yet crucial dimension of accountability on the part of national governments as duty bearers. Our findings show that in the case of these two villages, the human rights principles of participation, accountability, transparency and empowerment are severely undermined, with women bearing the brunt in all these domains. This overall state of affairs is, we argue, due to inadequate monitoring and evaluation of LSLI processes themselves and low levels of commitment on the part of institutions in Tanzania to monitor the promises made by investors. This in turn demonstrates an accountability deficit on the part of duty-bearers within LSLIs, and limited capacity of affected community members to claim their rights. Individual food insecurity experience in the two communities correlates, among other characteristics, with lack of land ownership, employment and income-generating activities. The rights-based livelihoods framework applied in this study points to serious deficiencies in the LSLI model as presently endorsed in SAGCOT, and emphasises the fact that access to land in Tanzania is a precondition for the realisation of the right to adequate food and thus a critical requirement for achieving and maintaining food and nutrition security. We conclude by arguing that progressive coalitions within and beyond national states must devise policies and institutions that empower individuals and civil society actors to make demands on their governments to respect, protect and fulfil their obligations regarding the legally enforceable right to food.


Introduction
In the period following the 2007/2008 global financial crisis, there has emerged evidence of a 'global land grab' (TNI, 2013) and consequent to it, a growing body of research on large-scale land investments (LSLIs) in Sub-Saharan Africa (SSA). This literature is largely critical of the impact of LSLIs on livelihoods and food security (Alamirew et al., 2015;Cotula, 2012). For example, in a cross country analysis of the impact of 'transnational land acquisitions on local food security and dietary diversity', Müller et al. (2021, 5) found that land deals in SSA have contributed to food insecurity by shifting production away from local staples and decreasing dietary diversity among households in the region. Critics who cite a mounting number of case-studies, argue that the economic and speculative value of land in these countries are prioritised to the detriment of the rights of local communities including other social and cultural rights attached to land (Davis et al., 2015;Von Braun, 1995). For example, Bues and Theesfeld (2012) report that local communities in Ethiopia were denied their rights to water in favour of a floriculture investment because of the investor's influence on the local government. Alamirew et al. (2015) and Yengoh and Armah (2015), examining the impact of LSLIs on food security and employment in Ethiopia and Sierra Leone respectively, show that employment opportunities associated with these investments are both temporary and marginal and that there is a decline in affected households' food consumption, attributable to LSLI.
The issues can be more serious than this : Borras Jr et al. (2013, 175) state that there is a 'blurring of governance boundaries between sectors of food, energy, climate change mitigation strategies and commercial complexes', all posing significant governance challenges in the regulation of LSLIs. Significantly, the law itself, including international human rights law, has been criticised as an enabler of the commodification of land and labour, exacerbating the global order's 'poverty, dispossession and exploitation' (Cotula 2020, 475) and increasing the vulnerable status of customarily held land in many African countries (Alden-Wily, 2011;Chadwick, 2019). Indeed, governments routinely mobilise their formal ownership of land within national law to legitimise LSLIs and promote growth corridors or agricultural modernisation (Chadwick, 2019;Cotula, 2013), while private investors (domestic and foreign) maximise national law and the legal protections provided by international investment law to acquire land and shield themselves from adverse public (re)actions and legal contestations (Ferrando, 2014;Jayne et al., 2016).
Consequently, the weaknesses in global markets and financial institutions, further exposed by the 2007/2008 financial crisis and subsequent land rush and food security risks in many African countries, have variously produced and deepened a vulnerability context for ordinary citizens. This has in turn demonstrated the need for policies and institutions to mitigate and regulate the excesses of market forces (Sama, 2016;Schiavoni et al., 2018). It is for this reason that in 2009, the former UN Special Rapporteur on the right to adequate food, De Schutter (2009b), identified and recommended plausible policies focused on the linkages between land ownership, food security, sustainable development and the right to food (see Kothari, 2006;Miggiano et al., 2010). Indeed, authoritative voices like De Schutter (2011a, 505), have argued that '[g]overnments have obligations that they cannot simply ignore in the name of attracting capital' and actively advocated for rights-based approaches to LSLIs, which we use in framing this paper.
Research on and the impact of LSLIs remain highly contested due in part to the lack of accurate data (Edelman, 2013;Locher & Sulle, 2014;Oya, 2013), and due also to opposing views of different actors (Diao et al., 2018). On the one hand, advocates for LSLIs, such as governments, multi-national corporations and institutions such as the World Bank argue that these investments are a development opportunity and can boost much needed foreign direct investment (FDI), increase export revenue, and can lead to food security, technology transfer and job creation for local communities in developing countries (SAGCOT, 2011;Steffens et al., 2019;World Bank, 2017). In fact, following the 2007/ 2008 crisis, the government of Tanzania and its partners launched the Southern Agricultural Growth Corridor of Tanzania (SAGCOT) as part of the Agricultural Sector Development Strategy (ASDS) and as a means to stimulate economic development by attracting multi-national companies to invest in agriculture (Bergius et al., 2018;SAGCOT, 2011). The SAGCOT provides a platform to galvanise global investors, multi-national companies, state bureaucrats and politicians in a grand modernist vision for the transformation of agriculture (Sulle, 2020, 333). The International Fund for Agricultural Development (IFAD), the World Bank, the African Development Bank (AfDB), the European Union (EU), Irish Aid and the Japanese International Cooperation Agency (JICA) were the main donors in the ASDS (SAGCOT, 2011). These partnerships underwrote the launch of the SAGCOT in 2010 (Bergius et al., 2018;Sulle, 2020). Kaarhus (2018) provides a detailed account of the development of the SACGOT and the role of the Norwegian petrochemical company YARA International in promoting commercial agriculture as a growth model in East Africa.
In this sense, the SAGCOT is envisaged by the government of Tanzania as a means to stimulate economic development by attracting multi-national companies to invest in agriculture (Bergius et al., 2018;SAGCOT, 2011). There is no doubt that these large-scale land deals cannot avoid a certain degree of enmeshment in local dynamics and micro politics over who has the right to dispose of land, who is consulted and who benefits from sales or transfers of land within rural households and communities (Sulle, 2020). For example, political contestations around land investments in Tanzania can be traced back to the 1970s and 1980s following the state-led villagisation programme (Bryceson, 1980;Nelson et al., 2012). Bryceson (1980, 557) describes the villagisation programme as a failed top-down experiment in which the state became the coordinator in the formation of villages as commodity-producing units. Arguably, current debates around land dispossession and investments serve in part to reproduce and extend ongoing debates that form part of Tanzania's agrarian history (Schlimmer, 2018, 84). This complex history of political contestations in relation to land investments in Tanzania supports the claim by Sulle (2020, 337) that the SAGCOT did not start with a tabula rasa, but builds on existing state-led agricultural investments and local machinations around land.
LSLIs have been challenged for depriving local communities of their access to land, causing environmental degradation, and leading to human rights violations (Baumgartner et al., 2015;Kanosue, 2015;Schiavoni et al., 2018). This affects those people most severely who are already marginalised, worsening existing structural inequalities. Women, who often depend on marginal land for supplementing food supplies for their households, especially certain categories such as widows and single women with children, but also the elderly in general as well as the youth, are included here. Proponents of human rights advocate for improved accountability and transparency in designing and implementing LSLI schemes. In other words, rural communities should be better involved in land investment policy design and implementation (De Schutter, 2009a, 2011b. For this reason, civil society organisations have contributed to these debates by proposing actions in the form of voluntary guidelines, which primarily focus on improving the governance of land tenure and agricultural policy in general (see Seufert, 2013 (Bernstein & Oya, 2014;Sulle, 2020). .
As already mentioned, the lack of sound data on LSLI is exacerbated by (i) the reluctance of governments and investors to reveal information on land-based investments which are controversially deemed to be proprietary and confidential (Cotula, 2011), (ii) the varied nature and definitions of LSLIs (Anseeuw, 2013), (iii) the lack of rigour in research methodologies (Edelman, 2013;Locher & Sulle, 2014;Oya, 2013), and (iv) the risks involved in researching such a politically charged phenomenon (Cramer et al., 2015;Talleh Nkobou, 2020). These challenges and limitations can lead to what Oya (2013) describes as 'killer facts' within LSLI research, or, stated differently, 'the extent to which data reflects realities … is questionable' (Oya, 2013, 505). Further, Scoones et al. (2013) argue for the need for more research on the impact of LSLI on local communities whose livelihoods and food security are at stake. Socially marginalised groups, among others, women and the elderly, are most affected, with data on the gendered impacts of land deals largely lacking (Behrman et al., 2012;Daley & Pallas, 2014).
While this is not the first study to call for, or to integrate, human rights with the Sustainable Livelihoods Framework (SLF) (see Carney, 2003;Lemke & Bellows, 2016), this study uniquely combines the SLF with the human rights (PANTHER) principles. PANTHER stands for Participation, Accountability, Non-discrimination, Transparency, Human Dignity, Empowerment, and Rule of Law (FAO, 2005), as will be elaborated in the following section.
The paper has three objectives. First, in addressing conceptual and methodological challenges in LSLI research, this paper sets out to provide a new perspective and approach to research on LSLI. It does so by developing an innovative rights-based livelihoods framework. We draw on the right to adequate food, moving away from a needs-based to an entitlement-based perspective of food security. The SLF is used to facilitate the assessment of livelihood strategies and outcomes in the two case study communities. Second, within this broader framework, we apply the Household Food Insecurity Access Scale (HFIAS) (see Coates et al., 2007) to explore the food insecurity experience of individuals as a livelihood outcome in the context of LSLI within these case study communities. We do not perform a comparative analysis or a 'before' and 'after' evaluation of the food insecurity experiences in the community, an option which was not available to us. Such an approach will require more data, time, and resources as we have highlighted in the challenges of conducting LSLI research. We do, however, juxtapose a cross-sectional perspective of individuals' food insecurity experiences within these communities against promises made in investment contracts and policies. This provides valuable new empirical data, given that in Tanzania data on the link between LSLI and individual food insecurity experience has thus far been lacking. Third, we shed light on whether the human rights principles of accountability, transparency, participation and empowerment were adhered to in the context of the LSLIs observed here, and especially how relevant local government institutions and land management structures affect livelihood strategies in these communities.
In making these contributions, we wish to emphasise the imperative of bringing greater nuance to the complexities of LSLI as a global phenomenon, which affects local actors and communities in a wide range of ways. We recognise, however, the inherent difficulties and yet the necessity of doing so via indepth, localised and on-the-ground research into this often controversial and polarising area of enquiry, which can place researchers themselves at risk of their personal safety, as has indeed been experienced by the first author of this paper and his local hosts (see Talleh Nkobou, 2020). In the next section we introduce the rights-based livelihoods framework, which integrates the human rights-based PANTHER principles and the SLF, placing the elements of the framework in the context of LSLIs in the two case study communities observed here.

A rights-based livelihoods framework
In the debates on LSLI schemes, a rights-based perspective has increasingly been called for, both within policy and academic research (Busscher et al., 2019;Claeys & Vanloqueren, 2013;Grajales, 2015). As argued by Scoones (2009, 181), sustainable livelihood approaches have been criticised for the 'lack of engagement with processes of economic globalisation', 'lack of attention to power and politics' and its focus on local level livelihoods in isolation (also see Lemke & Bellows, 2016). Our rights-based livelihoods framework focuses on both the overarching context and conditions for the development of livelihood strategies and the resulting livelihood outcomes, such as food security. This human rights-based approach complements the SLF by introducing broader political economy debates and structures of accountability for a more nuanced understanding of LSLIs across different scales from the local, through the regional to the global level (Narula, 2013).
To do so, we draw on the widely accepted 1996 World Food Summit definition of food security as the physical and economic access to sufficient, safe and nutritious food, at all times, to meet dietary needs and food preferences for an active and healthy life (FAO, 1996). However, we extend this definition by moving from a needs-based to a rights-based approach, which includes aspects of dignity, acknowledgement of rights, transparency, accountability, and empowerment (Mechlem, 2004), as is reflected in the definition of the right to adequate food (see below). A rights-based approach introduces legally enforceable state obligations within LSLI transactions (De Schutter, 2009a;Narula, 2005). Here, ordinary citizens are rights holders, and national governments and other local institutions are the primary duty bearers, with multiple public and private sector actors also being recognised as duty bearers (Cornwall & Nyamu-Musembi, 2004).
General comment (GC) 12 1 of the UN Committee on Economic, Social and Cultural Rights (CESCR 1999) defines the right to adequate food as: '…when every man, woman and child alone or in community with others, have physical and economic access at all times to adequate food or means for its procurement'. In GC 12, para. 7-13, the term adequacy is particularly significant for the right to food since it serves to underline availability (dietary and sustainable), accessibility (economic and physical), and acceptability (consumer and cultural) of food. For its part, a rights-based approach entails focusing on those who are most vulnerable, understanding what causes the vulnerability and changing conditions to improve the situation (Chilton & Rose, 2009). This means that, as a signatory to the 1966 International Covenant on Economic Social and Cultural Rights (ICESCR), other international treaties such as the 1979 Convention on the Elimination of all Forms of Discrimination Against Women (CEDAW), the 1989 Convention on the Rights of the Child (CDC), and in line with General Comment (GC) 12 of the CESCR (1999), the Government of Tanzania has the following legally enforceable obligations: (1) to respect the right to food, meaning that it should not interfere with people's access to resources necessary for their livelihoods.
(2) To protect against non-state actors (individuals or enterprises) from interfering with the access to productive resources of individuals.
(3) To fulfil, requiring states to adopt appropriate legislative, administrative, budgetary, judicial, promotional and other measures towards the full realisation of the right to food (GC 12, 1999), para 15 2 ). The fulfil dimension includes the sub-dimensions to facilitate, where states should proactively strengthen people's access to and use of resources and services, and to provide goods and services to those, who for emergency or non-emergency reasons outside their control, are unable to obtain them (for example, food assistance).
Noteworthy, however, is Article 2 of the ICESCR, which obliges states to take steps 'individually and through international assistance and co-operation, especially economic and technical, to the maximum of its available resources, with a view to achieving progressively the full realisation of rights recognised by the covenant'. Moreover, 'progressive realisation' cannot be treated as a licence to remain passive, and GC 3 of the CESCR (1999) clearly defines the minimum core obligation of states (De Schutter, 2014, 562). In line with these normative contents, LSLIs and associated policies should be implemented in a way that progressively realises the right to adequate food for ordinary citizens (Narula, 2006(Narula, , 2013. As stated earlier, in our analysis we apply the PANTHER principles, which form part of the FAO's (2005) Voluntary Guidelines to support the progressive realisation of the right to adequate food in the context of national food security (Right to Food Guidelines). Along with legally enforceable treaties such as the ICESCR, voluntary instruments such as the Right to Food Guidelines ensure that citizens' political and social rights are embedded in political and social policy (Gready & Ensor, 2005). Authors like Das and Grant (2014) and Mohr et al. (2016) have used the PANTHER principles to assess the progressive realisation of the right to adequate food at the national and sub-national level. In line with Yeshanew (2014, 4), this research argues that these principles should govern decision-making, implementation and monitoring processes within LSLIs in countries like Tanzania. While we acknowledge that the PANTHER principles are indivisible, for the purpose of this study, we pay particular attention to the following four principles as experienced by individuals in the two case study communities, which in the context of our research are understood as follows: (also see Fig. 1): (1) Participation means ensuring the free, informed and full involvement of all segments of the population, including marginalised groups such as women, the elderly and the youth, in decision-making processes at all stages of LSLIs. This also involves engaging in meaningful consultations with relevant state and non-state actors, including impacted citizens and civil society groups.
(2) Transparency means adopting a clear and contextsensitive strategy for communication with citizens and other stakeholders impacted by LSLIs. This entails freely sharing information about the duties and responsibilities of different actors, rights and entitlements of affected citizens, and maintaining a two-way communication between investment-related actors and impacted communities.
(3) Accountability refers to duty bearers having to ensure that monitoring and reporting mechanisms, that are inclusive and context-sensitive, are put in place within LSLI schemes. Such tools should contribute to continued learning about and improvements to the LSLI scheme. Accessible and confidential grievance and recourse mechanisms, including pathways for timely responses to complaints of impacted citizens, should be a critical part of the governance system within LSLIs. (4) Empowerment of marginalised and affected citizens should play a primary role and should be provided to build resilient livelihoods. Local governments and other authorities are expected to be supported in implementing their duties to respect, protect and fulfil the right to adequate food, and building the capacity of affected citizens to claim their rights. Programmes and policies that promote LSLI should design and implement inclusive and participatory capacity-building strategies, including improving capabilities to absorb shocks and stressors and to adapt to new livelihood strategies, e.g., the protection of workers, and raising awareness of the distribution of rights and responsibilities within LSLIs.

The Sustainable Livelihoods Framework (SLF)
Developed as an intervention planning tool in development programmes (Chambers & Conway, 1992;Solesbury, 2003), the SLF has been developed further and adapted to various contexts, allowing for the assessments of the vulnerability and capability of households' assets or resources to cope with shocks and to develop livelihood strategies based on available resources (Morse et al., 2009;Scoones, 2009). Scoones (2009, 186) in a critical review of the SLF calls for a livelihood approach which 'looks simultaneously at both structure and agency and the diverse micro-and macro-political processes that define opportunities and constraints'. In the context of our research on LSLI and applying a rights-based approach, the elements of the SLF are understood as follows: (1) The vulnerability context consists of two components: external stresses and shocks people are exposed to, such as climatic and weather events, but also market shocks such as experienced during the 2007/2008 financial crisis and the subsequent increased global demand for land (Deininger et al., 2011). The vulnerability context also refers to internal stresses and shocks, which are mainly determined by the capability of individuals to cope with the loss of productive resources such as land, as well as to internal dynamics and power relations among diverse actors within communities. This includes the (in)ability of individuals to cope with shocks and stresses depending on the human, physical, natural, social, financial and political resources ('assets') available to them. Access to these resources largely determines people's capacity to diversify their livelihoods and, in line with a rights-based perspective, to hold state and non-state actors accountable for the violation of their rights.
(2) Transforming structures and processes refer to institutions, policies and regulations that impact the choices that individuals make about using their productive resources, and the types and amount of assets or resources they have entitlements to (Messer & Townsley, 2003, 4). This component of the SLF focuses on the role played by institutions and processes, such as the rule of law, in the realisation of people's right to adequate food in rural and urban settings.
(3) Livelihood strategies denote the range of activities and choices that people undertake or make to achieve certain livelihood outcomes. In the context observed here, such goals include the choice to work on large-scale farms, or to secure alternative pieces of land for farming. The capability to maintain, uphold and develop livelihood strategies can be a means to achieve alternative livelihood outcomes (Hall et al., 2015;Scoones, 2009). (4) Livelihood outcomes are the goals to which people individually and collectively aspire as a result of pursuing their livelihood strategies, for example, food and nutrition security, increased income and wellbeing (Messer & Townsley, 2003;Scoones, 2009). Here, we are specifically concerned with the food insecurity experience of individuals as one of the livelihood outcomes in the context of LSLIs. Figure 1 illustrates how all elements of the rights-based livelihoods framework are inter-linked, highlighting those aspects in red that are most relevant in the case studies observed here.

Case study location
Fieldwork for this paper was carried out in the Ruvuma region, Tanzania, during a three-week pilot study in December 2017, and from May to September 2018. A previous visit was made by the first author to Ruvuma in 2014, as part of research conducted for his M.Sc. dissertation. Selected aspects of the M.Sc. thesis (unpublished) entitled The impact of large-scale land acquisition on the right to adequate food of small-scale farmers in Tanzania, were presented at the AFHVS/ASFS Annual Conference, Chatham University, Falk School of Sustainability, Food Studies Program Pittsburgh, PA 24th -28th June 2015 (Talleh Nkobou et al., 2015). This prior knowledge, familiarity, and our existing relationship with the national network of small-scale farmers' groups in Tanzania, MVIWATA, all helped to facilitate community entry, to establish rapport and relationships of trust and to engage in more sensitive discussions related to livelihoods, food security and the right to food as addressed in this paper.
Ruvuma is located within the Southern Agricultural Growth Corridor of Tanzania (SAGCOT), an area earmarked for the development of agribusiness partnerships (SAGCOT, 2011). Ruvuma has five districts, with the regional capital being the municipality of Songea. Two villages (Village 1 and 2) were selected and were anonymised, along with the names of research participants, to protect their identity. The Ruvuma River is the primary source of water in this region. Agriculture, predominantly carried out by smallholders typically cultivating about 2-5 acres of land, accounts for over 90% of regional food production and employment. In both villages, the village government is co-led by a chairman who is elected by the village assembly and a governmentappointed Village Executive Officer (VEO). According to 2018 demographic data obtained from the VEO in both villages, there are 428 households in Village 1, with a total of 1954 inhabitants (991 male and 963 female inhabitants). Village 2 is appreciably larger and consists of 821 households and a total population of 3612 inhabitants (1793 male and 1819 female inhabitants).
A detailed account of the LSLI process in the two case study communities has been provided elsewhere (Talleh Nkobou & Ainslie, 2020). For the purposes of this paper, a brief summary of the history of the LSLIs is set out here. In the first case study (Village 1) the origin of the LSLI dates back to 1984 when it is claimed by research participants that a group of male village members acting on behalf of the village agreed to the transfer of 404 ha of village land to an investor. 3 At the time, land could be transferred by the village government, who negotiated the transaction, and provided land to anyone. It is important to understand these transactions within the context of changing legal, political and economic reforms in Tanzania (Talleh Nkobou & Ainslie, 2020). For example, following independence in 1961, The Land Acquisition Act n o 47 of 1967 repealed all colonial land law and vested all land in the president as trustee. Additionally, the Rural Lands (Planning and Utilization) Act N o 14 of 1973 declared living in villages as compulsory under the villagisation programme (URT, 1994, 42). Section 4 of the Rural Lands (Planning and Utilization) Act gave the president 'unrestricted discretionary powers to declare any part of Tanzania as a specified area'. Village held land were (is) registered as 'rights of occupancy' and public officials had (have) discretionary powers to 'revoke' village held land (URT, 1994, 51-57). Subsequent changes in land law i.e., the 1999 Land Act and the 1999 Village land Act introduced the oversight mechanism of the village-assembly which obligates the consent for any transfer of land to a potential investor (for more see Larsson, 2006).
Under the 1999 Village Land Act, a village council is tasked with making recommendations on any land transfer and is then supposed to submit a proposal to the village general assembly for approval. The same village members who had agreed to the transfer of land in 1984, some of whom were key informants in this research, were shocked when the son of the deceased investor suddenly claimed an additional 1595 ha in 2011. The total land he claimed thus constituted an area of just under 2000 ha and controversially, his claim was supported by local government officials. Several attempts by village members to contest the acquisition of this additional land failed, among others holding village assembly meetings, enlisting local law firms and writing letters to the respective district officials involved. During the course of these contestations, some village members were intimidated by the district commissioner, village executive officer and police, and some were even jailed. In 2018, while village members seem to have accepted that their land was taken, they continue to accuse local government officials of favouring the investor, to the detriment of their rights.
In the second case study (Village 2), community members agreed to the very substantial transfer of 20,000 ha of land to a local investor in 2010. The investor promised to provide farming tools and tractors, a telecommunication tower, employment opportunities, schools and a dispensary. However, village members have expressed frustration that they were tricked by a 'cunning investor' who gave them 2000Tsh (US $0.86) each for their 'voices to be silenced' (female FGD participant, Village 2, 04.08.2018). Controversially, village members have been barred from accessing this piece of land, which has not been developed since 2010, contradicting the promises made by the investor at the time of acquisition. In both case study communities, members reported an abuse of power on the part of local government representatives and the consistent disregard of their interests and rights.

Materials and methods
Research permits and ethical clearance were obtained from the Tanzania Commission for Science and Technology (COSTECH) 4 and from the Research Ethics Committee 5 at the University of Reading's School of Agriculture Policy and Development, respectively.

Data collection
The paper relies on both qualitative and quantitative research approaches. For an initial analysis of secondary data, we drew on published academic literature, policy documents, civil society and media reports as well as on community documents such as letters and village meeting minutes regarding the land deal in the two communities.
The study used a two-stage sampling technique. First, two villages were purposively selected because of their LSLI experiences (as described above). Second, 187 households were selected using a purposively stratified sample, including an average of ten households per hamlet in each village (see Table 1). All adults above 18 years of age in the sampled households were interviewed (n = 374) using a structured questionnaire. This typically took 45 min to administer to each person. In doing so, we hold that experiences regarding food security of each individual within a household must be considered, as certain voices may be marginalised when only household heads are consulted, as is still common when conducting household surveys. The questionnaires were administered with the help of two locally-recruited research assistants (one male and one female). Male and female adults were interviewed separately by the male and female research assistant, respectively. In doing so, we attempted to minimise unequal gender dynamics between interviewer and interviewee as well as possible power dynamics between household members, for e.g., husband and wife, by enabling both female and male household members to share their views. This also permitted interviewees to speak relatively freely about their food insecurity experience and livelihood strategies, with minimal influence or pressure by their partner during interviews. During the period of fieldwork, the first author also engaged in participant observation which provided additional insights into the everyday experiences of village members.

Focus group discussions (FGDs)
Participants for the FGDs were selected from each hamlet, with input from the village chairperson. To minimise bias, each hamlet was represented by one male and one female participant, who were preferably in the older generation and with sound knowledge of the village history. To encourage the participation of women, FGDs were separated into male and female groups. In Village 1, fourteen participants were split into two equal groups -seven men and seven women, allowing us to conduct two FGDs, which each lasted about two hours. The FGD in Village 2 was limited to one session with both male and female participants because of the sudden termination of the first author's research permit (Talleh Nkobou, 2020). Focus groups were specifically useful in exploring the perceptions of village members regarding institutions related to land investments; the role played by key actors, e.g., government officials in land investments; human rights and recourse mechanisms, for example the support provided by the government in case of the violations of rights; and the gendered implications of the LSLIs.

Key informant interviews
Key informants (n = 18) were selected from a variety of sectors mainly using snowball sampling (see Table 3) in the selected villages and urban areas in Ruvuma and Dar es Salaam. Initial e-mail contact with an expert in investment and human rights law at the University of Dar es Salaam (UDSM also acted as the institutional host for the first author) was established through a purposeful internet search. Key informant interviews followed a loose 'schedule of questions' but were largely (and deliberately) unstructured. Taking on average about one hour to conduct, they were especially helpful in gaining an in-depth understanding of the land acquisition process in the case study communities in particular, and the politics of LSLI in Tanzania more generally.  Anxiety and uncertainty about household food supply: (Accessibility) Did you ever worry that your household would not have enough food? Insufficient quality -includes a variety of the type of food and food preferences (Acceptability) Were you or any household member not able to eat certain kinds of foods you preferred due to lack of resources?
Did you or any household member have to eat a limited variety of foods due to lack of resources?
Did you or any household member have to eat some foods that you really did not want to eat because of lack of resources to obtain other types of food?
Insufficient food intake and its physical consequences. (Availability) Was there ever no food to eat of any kind in your household because of a lack of resources to get food?
Did you or any household member go to sleep at night hungry because there was not enough food?  2013) use the FIAS to measure food insecurity at the household level. By using the FIAS, we assigned individuals along a continuum from food secure to severely food secure (see Table 4) over four weeks during the dry season (also see Vogenthaler et al., 2013Vogenthaler et al., , 1689. Key informants also contributed to the design, adaptation and validation of the HFIAS survey instrument. For example, during the development phase of the HFIAS questionnaire, a standardised protocol developed by USAID (Coates et al., 2007) was used to operationalise and translate the questionnaire into the local language Kiswahili with the help of key informants. Translation into Swahili minimised inconsistencies, which may occur when questions are translated in an ad hoc manner during interviews. Additionally, to avoid inconsistencies raised during the piloting phase, three generic questions in the 'insufficient food intake' domain of the standard HFIAS questionnaire were excluded, while maintaining all three domains as demonstrated in Table 2. Each item was asked with a recall period of four weeks. Knueppel et al. (2010, 365) also found inconsistencies in the validation of the HFIAS in rural Tanzania, concluding that there were challenges in separating items in the HFIAS due to 'the overall high level of food insecurity in the population'. Leyna et al. (2008) document similar challenges with food insecurity measures in rural Tanzania.
It should be noted that, of the 18 key informants, only four were women (see Table 3). There is a striking gender imbalance in the academic sector and in the farmers-based Mildly food insecure (FIAS <5) Individuals expressed concerns about not having enough food sometimes or often, and/or were unable to eat preferred foods, and/or eat a more monotonous diet than desired and/or some foods considered undesirable, but only rarely. However, to fit into this category, the individuals did not express cutting back on quantity nor experience any of the most severe conditions of food insecurity, i.e., running out of food or going to bed hungry.
Moderately food insecure (FIAS <6) Respondents indicated incidences in which households sacrificed quality more frequently, by eating a monotonous diet or undesirable foods sometimes or often, and/or started to cut back on quantity by reducing the size of meals or number of meals, rarely or sometimes. However, to fit into this category, the individuals did not experience any of the most severe conditions. Severely food insecure: (FIAS ≥6) Respondents described households as cutting back on meal size or a number of meals often, and/or having experiences of some of the most severe conditions, i.e., running out of food or going to bed hungry in any instance during the four weeks. In other words, any individual that experienced one of these three conditions even once in the last four weeks (30 days) were considered severely food insecure.
*These categories can indicate tendencies or broad categories, but are more nuanced in reality, as we describe in more detail when presenting the results organisation in Tanzania. This reflects the low representation of women in the public sphere in the context of land rights more generally in Tanzania (also see Duncan & Haule, 2014).

Measuring food insecurity experience
The FIAS is calculated as the sum of the product of the item response (x i ) and the frequency-of-occurrence response (fi) across all item domains from the HFIAS for each individual in Table 2 (Coates et al., 2007). If an item response was 'yes' (coded as 1), the frequency of occurrence question was asked (often = 3, sometimes = 2, rarely = 1). Based on these responses, the food insecurity access score (FIAS) was calculated using the following formula (1): If the individual response to all six frequency-ofoccurrence questions was 'often', coded as 3, then the maximum FIAS is 18. If the individual responded 'no' to the item domain questions, frequency-of-occurrence questions were skipped, and subsequently coded as 0with a total minimum FIAS of 0. The responses on individual food insecurity experiences were then coded and categorised into four levels using the description in Table 4. Fitawek et al. (2020) use the same food security categories to explore the effect of LSLIs on household food security in Madagascar. The results were also disaggregated by gender, age and education, to investigate the prevalence in food insecurity experience between these categories (see Table 5).

Variables associated with food insecurity experience
As mentioned earlier, the FIAS was used as our dependent variable. To select associated variables linked to individual food insecurity experience, we used a backward stepwise Akaike information criterion (AIC) approach (Yamashita et al., 2007). The model with the lowest AIC was selected. Results of the regression analysis are included in Table 6 Qualitative data emerging from focus group discussions, key informant interviews and observations related to the principles of participation, accountability, non-discrimination, transparency, human dignity, empowerment, the rule of law. These and other components of the rights-based livelihoods framework were categorised through identifying emerging themes and concepts using the qualitative software NVIVO (Creswell & Poth, 2018). In the next section, we present and discuss the results of our analysis.

Socio-demographic and economic characteristics
On average, a household in Village 1 and 2 comprises five members with a standard deviation (SD) of 4. The average age of interviewees (n = 374) was 44.7 (SD 14.2). Interviewees generally fell between the age range of 40-50 years (48%). The vast majority (85%) of interviewees described their household as male-headed. 6 The daily income of interviewees was estimated by 82% of respondents themselves as below the daily minimum wage of 3815.50 Shillings (US$1.65). With regard to education, 82% of interviewees had primary level education, 4% had no education, 10% had attended secondary school, and 4% had received some form of apprenticeship (e.g. masonry). The low level of education was skewed towards female interviewees, explained by customary practices which continue to discriminate against women in Tanzania (Duncan & Haule, 2014). Since the LSLI, the average size of arable land-holdings of those interviewees who owned land (19%) in both Village 1 and 2 was reduced to 2-3 acres.

Food insecurity access score as part of livelihood outcomes
Years after the LSLIs in Village I & 2, village members experience high levels of food insecurity. A distribution of the food insecurity access scale across gender, age and education is represented in Table 5. Based on our calculations of the FIAS, 47% of interviewees were severely food insecure, 16% moderately food insecure, 4% mildly food insecure, and 33% were food secure. Even for those in the 'food secure' LCL -Lower limit of the 95% confidence interval UCL -Upper limit of the 95% confidence interval *0.05 significant level, **0.01 significant level category, some respondents expressed that they are sometimes worried about not having a full diversity of foods at all times. So, in our research context, a strict application of the definition of the right to food would result in over 90% of individuals being categorised as food insecure along the FIAS continuum, given that no individual should experience any form of food insecurity (GC 12 of the CESCR 1999).
Considering the harsh socio-economic realities in Village 1 and 2, we categorised those individuals as food secure who stated that they never ran out of food and did not have to cut back on the quantity of food (see Table 4). We argue that this minimalist categorisation in fact shows that the categories of food security applied here generally have to be regarded with a degree of caution, especially in contexts where communities experience high levels of food insecurity (Knueppel et al., 2010;Saint-Ville et al., 2019).

Determinants of individual food insecurity experience
Because the regression analysis performed in this study focuses on the individual level experiences, national-level variables such as global economic shocks, and other variables that may explain food (in)security such as climatic condition or seasonality, soil profile, GDP, and economic development profile of the country (Baumgartner et al., 2015;Müller et al., 2021) were not included in the regression model. This explains the low R-square value of 17%. The R-square value means that our variables of choice (see Table 6) explain 17% variability in the food insecurity experience of individuals. Understanding individual's food insecurity experiences like other studies on human behaviour and preferences inherently have a greater amount of unexplainable variations (Miles, 2005). Yet, the significant coefficients for the variables associated with food insecurity in this study such as land and employment still represent the mean change in the FIAS when holding other predictor variables in the model constant. The low R-square value justifies the need for a mixed research approach in LSLI research which can explain other macro level variables that impact individual food insecurity experiences and livelihoods.
So, in further explaining the variation in individual food insecurity, we compensate by triangulating via a qualitative analysis within the rights-based livelihoods framework. The framework clarifies the obligations of states at the national level and the role of local institutions as part of transforming structures and processes, as well as other components of the SLF, such as external stresses and shocks. Table 6 presents the results from the regression analysis. The coefficients of each variable are of importance here and show how various variables affect the FIAS (the measure of food insecurity experience) of individuals within the two communities. The results in Table 6 are explained along with the qualitative analysis in the next section.

A rights-based perspective of food insecurity experience in the context of LSLI
External vulnerabilities and shocks The pressures of the 2007/ 2008 financial crisis, the rapid rise in food/cereal prices in 2008/09 and the focus of EU policies on biofuels all encouraged multinational corporations to seek cheaper agricultural land in countries like Tanzania (Cotula et al., 2008;Giovannetti & Ticci, 2016). Additionally, national policies and programmes such as the SAGCOT partnership, which was launched as part of Tanzania's Agricultural Sector Development Strategy (ASDS) in 2010, created added incentives for investors to engage in the commercialisation of agriculture in Tanzania (URT, 2015). These external drivers created shocks impacted upon local communities in a variety of ways (Nelson et al., 2012;Ngoitiko et al., 2010). According to research participants in both villages, land conflicts spiked since 2010following the launch of SAGCOT and the increasing interest on the part of outside investors in land in their communities.
What this demonstrates is how 'so-called' external variables could link with 'local' policy implications to accentuate violations of the right to food for smallholder farmers. Additionally, the link between local investors and foreign capital adds to a more nuanced representation of investments in Tanzania's history (Nelson et al., 2012;Ngoitiko et al., 2010). As argued by Sulle (2020, 333), local political elites welcome large-scale investments because they create rentseeking opportunities through lease agreements and accumulating land. Within these LSLI schemes, there is a clear focus on the part of the Tanzanian government on promoting exportoriented agribusiness, rather than on adopting long-term economic policies and programmes to enhance the productive capacity and livelihoods of small-scale farmers. For example, the visit in July 2014 by the first author to Village 1 unexpectedly coincided with a visit to the village by the former president of Tanzania, Jakaya M. Kikwete. The president praised the implementation of the large-scale farm, which according to him, would improve government's efforts in alleviating poverty (observation by first author, July 2014).
More recently, however, there is recognition by the government of Tanzania that the focus on large-scale 'commercial agricultural has had little impact on poverty reduction and efforts must be made to address these challenges' (URT, 2015, 42). And indeed, researchers such as Mbunda (2013) observe that policymakers did not sufficiently integrate smallscale farmers into the conception, design and implementation of the SAGCOT initiative (also see Bergius et al., 2018). Consequently, there have been regular and worrying reports on the abuse of human rights, among them the right to adequate food, and discriminatory decisions and policy-making processes during the implementation of LSLI schemes in Tanzania (Mousseau & Mittal, 2011;Twomey et al., 2015).
According to interviewees, the LSLI process in Village 1 & 2 was done in favour of the investors with support from local government officials, and several promises made by the investors to local communities have remained unfulfilled.
Similarly, other LSLIs in Tanzania have consistently been linked to a lack of transparency, unfulfilled promises, corruption and intimidation of village community members (Mbilinyi, 2012;Ngoitiko et al., 2010;West & Haug, 2017). According to observations made by the Committee on Economic, social and Cultural Rights in the third periodic review report in 2012 (CESCR, 2012), 7 the Committee expressed concern that: 'Several vulnerable communities, including pastoralist and hunter-gatherer communities, have been forcibly evicted from their traditional lands for large-scale farming, the creation of game reserves and expansion of national parks, mining, construction of military barracks, tourism and commercial game hunting. The Committee [was] concerned that these practices have resulted in a critical reduction in their access to land and natural resources, particularly threatening their livelihoods and their right to food.' The absence of adequate legal and institutional mechanisms have contributed to the negative impacts of LSLIs on the human rights of local communities (Abebe, 2012, 878). Control mechanisms regarding LSLIs are often ineffective and transactional procedures inconclusive (Okoth-Ogendo, 1999, 7, Bélair, 2018.

Enhancing livelihood opportunities: Promises and reality
The government of Tanzania describes LSLIs as a development opportunity that will allow them to increase employment opportunities, enable technology transfer, and lead to income generation and infrastructure development in rural areas (SAGCOT, 2011). However, in both Village 1 and 2, only 8% of the 374 research participants had gained employment from the LSLIs. The low rate of employment can partly be explained by the fact that in Village 2, the investment scheme has not yet been implemented. Workers on the large-scale farm in Village 1 are mainly commuting workers from neighbouring villages, with some of them travelling daily, while others rent temporary huts from local residents and share common resources, such as food, and water with their landlords.
Job opportunities are often seasonal, and workers are employed informally, on a day-to-day basis, without formal contracts that would provide them with greater economic and social security. A representative of the civil society organisation, Business and Human Rights Tanzania, explained that investors often rely on intermediaries (known locally as middlemen 8 ) to provide the labour on the farms. Investors exploit weaknesses in the labour laws, by ensuring that workers are not employed for the statutory period of more than six days per month or a six-months-probation period, 9 which would automatically qualify workers to receive several employment benefits, such as an employment contract and a fair representation in the case of unfair termination of the contract. Hence, intermediaries and the investor/manager of the scheme resort to a day-to-day registration system for their workers in the LSLI community in Village 1.
Additionally, the lack of access to land increases concerns and anxiety about not having enough access to food within both villages, as is expressed in the following statement of a male research participant: [Before the arrival of the LSLI], I cultivated 10 acres. Now, I cultivate five acres because I was robbed of my land. Now I harvest 50 bags [of maize] while I used to harvest 200 bags. I' ve reduced the number of kids I used to live with [sic]. Two went to another village; one went to work in the garage in town. He does not want to work in the farm anymore! -Male FGD participant Village 1, 07.07.2018 The long working hours on the LSLI farm in Village 1, ten hrs per day (07:00-17:00), and the absence of adequate food consumed during working hours partly explain why community members working on the large-scale farm experience a high level of food insecurity (FIAS of 2.60), compared to those who do not work on the farm (see Table 6). The lack of adequate time to produce one's own food also explains why community members who are involved in other income generating activities have a higher food insecurity experience. However, from Table 6, those 18% of interviewees who earn above the minimum daily wage of 3815.50 Shillings ($1.65), have a lower FIAS of 2.28 compared to those who earn below the minimum daily wage (82%). Results in Table 6 also show that a 1 % increase of income spent on food reduced the FIAS by 1.54.
Participants in focus groups and key informant interviews volunteered that working conditions on the large-scale farm are not favourable for workers' health and wellbeing. For example, workers who are tasked with spraying insecticides complained about the lack of protective gear, and consequently that workers experienced burning eyes and sore hands from repeated exposure to harmful chemicals. Further, there were reports of sexual abuse by female FGD participants who stated that to gain employment, farm supervisors, who are mostly men, ask for sexual favours.
When women go to work, male supervisors ask them for sex. If they reject [these requests], they do not get the job. If you do not have sex with them, you' re not employed. So many people come here from far away to find jobs on the large farm. Some women have to agree to give sex bribes because that is the only choice to secure a job. -(Female FGD participant, Village 1, 21.07.2018) When a local government representative in Songea was asked about unfair employment conditions and violation of workers' rights in the LSLI scheme, he responded that 'village members always complain because they are lazy and are always expecting handouts from the government' (Interview, 24.07.2018). Such reasoning fits well with ideas that present welfare policies as 'paternalistic' and people attitudes as lacking entrepreneurial spirit. It highlight the regard of government officials as being responsible 'for' and not 'to' citizens (Schneider, 2003). As a response to complaints about sexual exploitation in Village 1, the investor organised sex education campaigns and distributed free condoms to farm workers, which was seen by the investor as part of the solution.

Institutionalising the PANTHER principles
Transparency Research participants argued that there is lack of transparency regarding the acquisition process surrounding the LSLI in the two villages. For example, while members in Village 1 acquiesced to the transfer of 404.ha of land to the investor in 1984, there were no subsequent consultations before the transfer of the additional 1595 ha in 2011. A letter 10 from the DC's office in fact showed that the 1595 ha piece of land was registered with an ownership certificate in the name of the investor in 1987, although the authenticity of this document could not be determined. FGD participants and key informants denied agreeing to the transfer of this parcel of village land in 1987.
These debates must, of course be situated within the broader history of Tanzania's complicated and constantly evolving land reform process. For example, oversight mechanisms for the transfer of village land at local government level were only introduced after land reforms in 1992(Alden-Wily, 2003Shivji, 2002). In other words, the village assembly (VA), comprised of all adults living in the village above eighteen years of age, is the supervisory organ, while the elected Village Council 11 should act an executive body accountable to the VA (Shivji, 2002). If the village assembly approves and recommends the transfer, the land commissioner forwards the approval to the President, who signs off on the transfer of the village land to private ownership. After the President's approval, a 14-day period is provided to allow for any aggrieved party to lodge complaints before the final transfer is made.
Tanzania's has one of the most progressive legal structures relating to land tenure in Africa, including notional gender parity in ownership of communal land (Looloitai, 2014;Nelson et al., 2012). However, while the country's land laws have provisions to protect customary rights (German et al., 2011), and the power to enact and give concrete expression to these, land laws in Tanzania are in fact still very centralised (Alden-Wily, 2003;Shivji, 1998). Indeed, the President has unilateral powers to revoke and convert village land into public land in the 'public interest' (Alden-Wily, 2012, 755): ' ... [w]here the president is minded to transfer any area of village land to general or reserved land for the public interest, he may direct the Minister to proceed in accordance with the provisions... for the purpose of public interest ' -The Village Land Act, 1999, pp. Part III, Section 4 (1),(2).
In Village 1, the village assembly was not party to the land transfer process. Expressing their discontent with the land acquisition process, FGD participants highlighted that: We have sent several letters to the district commissioner ' s office, ward secretaries, regional officers, but we 10 Ref: No. AB.81/223/02/95 letter to Village 1 from the District Commissioner's office addressing dispute between village members and large-scale investor. 11 In Tanzania, village leaders (except for the village executive officer, VEO, who is appointed by the government) are elected by the village residents or village assembly every five years. This ensures that these leaders are accountable to the village assembly (Kesale 2017, 5). Hence, the village government reports to the village assembly and village members, in theory, have the power to hire and fire village government. However, this oversight mechanism can be abused by VEOs who might regard themselves as more powerful than the village governmentas is experienced in the case of Village 1, where the VEO was described by key informants as 'more influential than the village chairman' when the land was transferred to the investor in 2011.
have received no reply. The Prime Minister came to this farm in 2018, but we did not get the chance to talk to him. In 2014, President Kikwete came here; no one was allowed to speak. These politicians only come here to speak to the investor and about coffee farming. They do not listen to the challenges we face from this investment.
The land conflict increases anxiety and concerns over food availability among village members. This very likely explains why those involved in land conflicts have a higher FIAS of 1.98 relative to those who are not involved in land conflicts (see Table 6).
Empowerment While LSLI schemes claim to contribute to the broader local economy, interviewees argue that they do not expect any benefits from the investments, as is reflected in the following statement: 'We do not have any information[from the district] about the income generated from the coffee farm. We don't benefit from the company! Last year, [we know] the company paid [some money] to the district, but where is our village share? (Male FGD participant, Village 1, 21.07.2018) It is the case, however, that the influx of migrant workers provides benefits to residents in Village 1 who resort to alternative sources of income, such as renting out accommodation facilities. Further, there is high production and consumption of a local brew (wanzuki), which is commonly consumed during social gatherings in the evenings. Additionally, some women sell sweet potatoes along the main road, while several young men are engaged in moulding mud bricks for the construction of houses in the village. It was clear that there is no strategy or programme by local government officials to provide alternative livelihood strategies to build resilient livelihoods within these communities. Instead, local government officials refer to village members as being 'lazy and expecting government handouts'. On the part of the investor, apart from distributing condoms, we did not observe any inclusive programmes that might foster capacity-building, to help village communities curb the negative impacts stemming from the LSLI.
In terms of local infrastructure, village members had expected changes in infrastructure and public services in their villages as a result of LSLIs. Research participants in Village 1 placed particular importance on road traffic signs, arguing that: 'the lack of road signs is proof that we are not important, and we fear that there are plans by the investor to expand his land-holding. We even think there are plans to remove us from this village!' -(Male key informant, Village 1, interviewed 10.07.2018). Members in this village perceive this as a sign of neglect on the part of government authorities.
And indeed, in 2017 the investor had a dispensary built in Village 1. At the time of this study, village members started advocating for the construction of school infrastructure. The investor promised to provide financial support for the purchasing of building materials, which cannot be sourced locally, such as roofing material, cement, and paint, while the local community committed to providing labour, bricks and other locally sourced building materials. These fraught and uneven clientelist type relationships between communities and investors (Banks et al., 2016) leave one to wonder what the role of the government is in the provision of these services? For their part, local government officials were clearly drawn into a complex web of relationships with the investor, with the interactions between all these parties potentially having a range of 'intended and unintended consequences' (Banks et al., 2016, 256).
Educational levels in the community remain low, and there is a need for education provision both for adults and the youth. The results in Table 6 show that those with primary education are more food secure, with a lower FIAS of 2.24, relative to those with no education, and those with secondary level education have a lower FIAS of 4.27 compared to those with no education. Women remain highly disadvantaged, exemplified by the sexual harassment experienced by them, and higher levels of food insecurity, as shown in Table 5. Further, women are not represented in positions of authority, as observed in the limited number of women who were key informants during this research.
In Village 2, FGD participants highlighted that none of the investor's promises had been fulfilled. We also observed a lack of empowerment within the LSLI process and among different actors and institutions in the affected communities. For examples, legal institutions concerned about the welfare of the local community have had to close because of the lack of financial resources needed for land dispute resolution.
Accountability According to information gathered from the FGDs in Village 1 and 2, village members made efforts to address the lack of accountability and transparency in LSLI processes. They repeatedly consulted with Civil Society Organisations (CSOs), the media and lawyers to facilitate and support these efforts: Although we have not succeeded to get back our land, journalists have helped us understand our rights to some extent. We found out about human rights by listening to the radio. Rights should be fought for. If you fight, you can get it. We have tried by going to lawyers. A local radio station came here and asked questions, what did you do after losing your land? Which problems did you go through? Then they aired the interviews and our stories over the radio! (key informant interview with the Chairman of Village 1 07.07.2018) To ensure accountability, the functions and tasks of local government institutions need to be clearly defined, regularly reviewed, and institutional provisions made for adequate monitoring and dispute resolution mechanisms. In the case studies observed here, there is a perception among village members that the state is not willing to take measures that will firstly, prevent third parties from interfering in the productive resources of local communities and secondly, institute appropriate legislative, administrative, or judicial measures to ensure fair access to these resources in local communities. The following quote from a FGD participant provides a good indication of the broader perceptions of village members: Unfortunately, the government has not addressed this issue... I am convinced that the whole government knows about the land problems in this community. Because the process of getting our land involved the regional and district commissioner' s offices, and our complaints are known from the district level to ministerial level. These people [public officials] have just decided to support the investor by staying quiet about our problems! -Key informant in Village 2, 04.08.2018.
When respondents in both Village 1 and 2 were asked if they are aware of official channels via which they could lay landrelated complaints, 73.5% responded 'no', while only 26.5% were aware of some of the procedures, as described by FGD participants: When there are land conflicts between village members, they normally bring them to the chairperson or VEO. If it is complicated, the conflict is forwarded to Village land committee and if the Village committee fails, we forward to the Ward land committee. FGD participant in Village 1, 07.07.2018.
Our finding corresponds with those of Fernandez and Schwarze (2013), who argue that there are typically no tools within land investment schemes to hold investors accountable for not fulfilling their promises within investment contracts. Additionally, the financial costs of defending the rights of local communities are high, and lawyers are not willing to invest their efforts in land disputesgiven that the villagers will struggle to pay their fees. In an interview with a representative from the Tanganyika Law Society, he explained that investigating conflicts around land is expensive, time consuming and that there is a lack of legal capacity and resources in Tanzania to dedicate to rural communities. Thus in most cases, land conflicts are investigated on a pro bono basis. According to the same representative from the Tanganyika Law Society, many civil society groups are more interested in tackling issues such as domestic violence, which can attract funding from the donor community and are less resource-consuming and intractable than land dispute resolution. These factors impact civil society interests on the continuous monitoring of LSLI deals.
Public institutions such as the Tanzania Investment Centre (TIC) have been criticised for the lack of follow-up on the promises made by investors regarding broader economic development, which in the case of TIC has been attributed to both the lack of capacity and of the will to perform monitoring and evaluation exercises in investment communities (Bélair, 2018, 379). The near-absence of a robust civil society to counter the negative consequences of LSLIs and to ensure accountable systems of governance, further weakens the rights of local communities in Tanzania (Maillard-Ardenti, 2012, 19). This often means in effect that the efforts made by local communities to defend their rights of access to productive resources can be ignored by those in authority.
Participation FGD participants were asked to share their perceptions of participation in the LSLI process. They generally defined participation as follows: 'When the majority agree, it has to be done. If few are involved and many excluded that is not participation' (Focus group, Village 1, 17.07.2018). Another participant defined participation 'like eating together.
In any discussion, I should be allowed to speak and be listened to, and what I say should be considered when taking any decision' (Focus group, 17.07.2018). These perceptions and understandings reflect the concept of participation as applied in our rights-based livelihoods framework.
Additionally, during the visit of the Prime Minister to Village 1 in 2018, the village chairman raised concerns about the lack of transparency and involvement of village members with regard to the revenue from the farm.
' During an audience with the Prime Minister, before I finished [expressing our concerns,] the microphone was taken. When we ask about the proceeds from the investor, we are silenced or told to go and read the documents [related to proceeds from the farm at the district office]. When we ask for the documents [from the district commissioner], we don' t get them' . (Interview with chairman of Village 1 07.07.2018).
These quotes illustrate that village members seek to engage with public officials concerned with LSLIs, but their concerns are often disregarded by local government officials.

Conclusion
By adopting an innovative rights-based livelihoods approach which integrates the human rights PANTHER principles and the Sustainable Livelihoods Framework (SLF), this research reflects recent calls to go beyond a needs-based perspective on food security. This approach highlights the right to food as a human right, which empowers individuals to make claims when their right to food is violated, as can be the case in the event of LSLIs. A rights-based analysis emphasises that land in rural Tanzania and elsewhere is not only a key resource for agricultural production, but a precondition for the realisation of the right to adequate food and thus a requirement for achieving and maintaining food and nutrition security. This perspective, and our approach to the issues, further enhances the critical need to understand individual experiences by giving a voice to marginalised groups in society. The rights-based livelihoods framework applied complements the limitations identified in using the SLF. It allows for a holistic understanding of the roles of various actors involved in the context of LSLI and enriches our understanding of the livelihood strategies of and outcomes for rights holders affected by these investments.
The study further highlights that rights-based approaches aren't only reactive (i.e., kicking in when people's rights have been violated) but they need to infuse rural development policies and investment agreements during their formulation and ensure that processes of free, prior and informed consent and human rights-impact assessment are carried out. Prior studies on LSLI in Tanzania have focused on food insecurity at the level of the household, without extending this to examine individual food (in)security experiences of members within these households. The anxiety levels and the experience regarding food security of each adult individual within a household must be considered, as certain voices may be marginalised when only (male) household heads are consulted. Whilst we have focused on individuals over the age of 18 years, it is arguable that future studies should cast the analytical net even wider to also include the experiences of children with households. Tellingly though, it is very much the case that information relating to livelihood strategies of individual household members could be missed or may receive little attention if such an analysis is performed without a gendered interpretation of LSLIs.
The vulnerability of impacted citizens in the case study communities is exacerbated by the lack of monitoring and evaluation of LSLI processes, and the lack of capacity and commitment of institutions in Tanzania to follow-up on the promises made by investors. This demonstrates a lack of accountability on the part of duty-bearers within the entire LSLI institutional landscape. Additionally, low educational levels and limited access to information on the part of community members restrict their ability to claim their rights. There is a lack of trust and incentives on the part of community members to work within LSLI, as they perceive the land acquisition process to be fundamentally unfair. In fact, citizens who are impacted by LSLI should be able to participate fully and play a primary role in building resilient livelihoods within the remit of these LSLI schemes.
There remains a pressing need, therefore, to understand the capability of the poor and frequently marginalised groups in dealing with external shocks and stresses as the starting point of any intervention. It is equally important to place a contextspecific, case-study-based analysis as presented here, within the broader context of cumulative macroeconomic effects and their impact at the local level. The overall and longer-term contribution of LSLIs to rural development and poverty reduction in countries like Tanzania remains questionable, given the poor integration of the concerns of local communities and the absence of mechanisms to ensure that investors keep their contractual obligations. Progressive coalitions within and beyond national states must devise policies and institutions that empower individuals and civil society actors to make demands on their governments to respect, protect and fulfil their obligations regarding the right to food. They should also be nudged to ensure the accountability and transparency of government agents and other decision-making bodies and processes in implementing such policies. made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.