Rubber Plantation Land Grabs and Agrarian Change: A Political Economy Analysis of Livelihood Pathways of Ethnic Minority Groups in Northwest Vietnam

Abstract

This paper critically examines the consequences of land grabs for livelihoods and agrarian change, based on a case study of rubber plantations in ethnic minorities in the uplands in Northwest Vietnam. Building upon Scoones’ agrarian political economy of livelihood framework, an integrated conceptual framework of a ‘livelihood pathway’ is developed to analyze the impact of rubber plantation land grabs on livelihoods and the agrarian political economy. Drawing on qualitative analysis and survey data from 205 households across six villages inhabited by Thai, Hmong, and Kho Mu communities, this study finds that rubber plantation land grabs have led to differentiated livelihood strategies—ranging from subsistence farming and wage labor to commercial agriculture—shaped by each group’s socioeconomic status, political connections, and access to resources. Consequently, the land grabbing undertaken by a domestic state-owned enterprise has caused the emergence of a set of distinctive livelihood pathways within a complex web of intersections across class and ethnicity in the upland area. This study concludes by arguing that an integrated conceptual framework of a ‘livelihood pathway’ offers a useful tool for analyzing the long-term socio-political consequences of land grabbing in similar contexts across developing countries and beyond.

1. Introduction

In this era of the global land rush, there has been an explosion of scholarship on ‘transnational land deals’ or ‘global land grabs’. These terms are generally defined as large-scale, cross-border land deals through lease, concession, or outright purchase by transnational corporations or foreign governments for diverse profit-seeking activities including food/non-food production, natural resource exploitation, and land speculation [1,2]. In the last decade, there has been a surge of literature on global land grabs, revealing the causes, drivers, and dynamics of the phenomenon unfolding at the local, regional, or global scales [3,4,5]. Scholars have shed light on various spectrums of the effects of land grabbing on local configurations, especially concerning the transformation of land property relations [6,7], labor pattern changes [8,9,10], social conflicts and local reactions [11,12], and human rights concerns [13]. The impacts of land grabs on local livelihoods have also been debated [1,14]. Generally, these scholarly efforts have contributed evidence of the damage a land grab caused to the livelihoods of the affected local people, especially those who were either requested or forced to contribute the land they had owned, including a loss of access to farmland and resultant food insecurity, a decrease in productivity, displacement of their residence, and mounting indebtedness, among others [6,15,16,17,18,19,20]. However, some evidence on the positive impact of land grabs on livelihoods has also been presented, such as increased consumption and income levels, adoption of new agricultural practices, increased employment as wage workers, and improved local infrastructure [3,16,21].

Although the burgeoning literature on land grabbing has greatly enriched our understanding of the possible effects of land grabbing on the livelihoods of affected local people, two important issues have been inadequately addressed. There is a lack of studies analyzing the impacts of land grabs in relation to wider social and agrarian change. The majority of academic land grab studies focus their analysis on the short-term effect of land grabs on local livelihoods. While their findings are insightful in numerous spectrums, they are often confined to micro-level contexts related to inter- and/or intra-household dynamisms or a local community and associated groups, with wider medium- to long-term social relations and alternations left unexamined. Indeed, the academic research has overlooked how land grabs mediate the emergence of new livelihood trajectories in multiple locales. This gap leaves some critical questions about livelihood outcomes unaddressed in the existing literature, such as why certain livelihoods are possible for some groups of people, but not for others; who would win and who would lose in the local developments of the post-land loss phase, and how the local agrarian transformations via land loss are linked to the wider processes of a long-term social change. Further, this void requires a close understanding of interlinkages between the micro-level contexts of local people’s livelihoods and state-driven processes of land grabbing, whereby previously non-existent, new collective livelihoods might come into being. It is also important to understand how a set of collective livelihoods that have emerged in different locales or according to varying identities present a social differentiation whereby newly formed social classes acquire relevance to the long-term processes of broader social and agrarian change (e.g., agricultural modernization, urban industrialization, penetration of capitalism, etc.).

This paper seeks to fill those gaps in the light of a conceptual framework related to the agrarian political economy of livelihood developed by Scoones [22] for understanding the differential impact of a land grab program on the communities and livelihoods of multiple ethnic groups. This study investigates how different ethnic minority groups, who had customarily owned the land, have developed diverse livelihood strategies following land loss, and how these strategies have shaped collective livelihood pathways. To meet the objectives, we draw on empirical evidence from our case study on rubber plantation land grabs in Dien Bien province, Northwest Vietnam.

This paper is structured as follows. The next section introduces the theoretical framework that underpins the study. This is followed with a discussion about the study site and research methods. After that, the findings are presented and discussed. The final section takes a step back and address how the paper’s findings speak to the broader land grab literature.

2. Theoretical Underpinning

Land grabbing has become the main concern for land use sustainability at any level from local to global. The drivers of land rush range from global dynamics such as increasing demand for, and prices of food and non-food agricultural commodities [2,23], corporate food regime restructuring [24], energy system transition, and environmental purposes such as carbon sinks and global biodiversity [1,21,25], to national and subnational forces including national development strategies [26,27], privatization of land by land titling programs or land reforms [4,28], and various political ends such as territorialization, sovereignty, and authority [4,29,30]. A land grab process immediately impacts land users whose livelihoods depend largely upon the land and the associated natural resources that are acquired [31]. There is consensus among scholars that a land acquisition could significantly transform the livelihoods of people residing in the target area by having their land use and control altered. The impacts of these processes, positive and negative, on local livelihoods have been intensely debated [1,16,32]. The above findings contribute to offering diverse empirically grounded explanations of livelihood strategies and changes in specific contexts. However, they do not provide any accounts about the collective livelihood trajectories and associated social differentiation [22,33]. Indeed, common livelihood patterns that could be conceived of beyond individual/household contexts remain elusive. Thus, the critical questions of why certain livelihoods are possible for some social groups but not for others, as well as how new agrarian classes emerge, have not been addressed fully by those approaches.

To theorize the relationships between a micro-analysis of livelihoods and a macro-analysis of agrarian change, Scoones [22] champions a combination of the insights from agrarian political economy and rural livelihood analysis. In developing a framework of the agrarian political economy of livelihood, Scoones [22] draws on insights from ‘rural livelihood analysis’, which provides a useful framework to reveal the specifics of diverse agrarian livelihood contexts. In the framework, contexts, conditions, and trends constitute the broad spatial or temporal terrain which influences all aspects of people’s livelihoods historically, politically, economically, socially, or environmentally. The livelihood analysis emphasizes the presence of the ‘vulnerability’ context, in which exogenous or endogenous factors such as natural disasters, economic crises, or seasonality adversely affect livelihoods [34]. This approach also concerns diversified economic and social activities that individuals and households pursue, livelihood assets (also known as capital) that they deploy for forging their livelihoods, and livelihood outcomes they achieve. Institutions and organizations are considered as key vehicles relating to the structures and processes that affect the household assets deployed, intervene in the household livelihood strategies pursued, and mediate the livelihood outcomes achieved, for different individuals and households [22].

The livelihood perspective has been criticized by some critics, however, due to three reasons. First, it focuses on the micro-level politics of inter- and intra-household relations within the boundaries of a local community and its resident groups, without due consideration of wider political configurations and dynamics. Second, it rests on an instrumental mapping of livelihood assets as a basic component of rural poverty analysis, framed by an economic analysis. Third, a livelihood analysis has a very limited focus on political space by focusing mainly on ‘empowering’ the poor, without making it clear how this process takes place. The class relations are also absent from the discourse on livelihood analysis [35].

The agrarian political economy provides the opportunity to ‘bring the politics back in’ to livelihood analysis to show political and economic alliances being forged between different classes and the structuring of wider agrarian change dynamics [22]. Scoones notes that “a grounded political economy approach allows for the detailed description of a diversity of livelihood strategies and evaluation of long-term livelihood trajectories and their structural conditioning and shaping” [22]. By adding two more questions from Bernstein’s initial four questions of political economy underpinning agrarian analysis [36], Scoones presents six core questions in analyzing the agrarian political economy of livelihood: (i) ‘who owns what (or who has access to what)?’; (ii) ‘who does what?’; (iii) ‘who gets what?’; (iv) ‘what do they do with it?’; (v) ‘how do social classes and groups in society and within the state interact with each other?’; and (vi) ‘how do changes in politics get shaped by dynamic ecologies and vice versa?’ [22] They are a set of questions on livelihood that have been answered primarily on the micro-scale through a detailed livelihood analysis addressing the ownership of household livelihood assets, livelihood strategies, and their consequences. However, a profound understanding of collective livelihoods allows us to reveal the contextual and experiential specificities of accumulation and differentiation in specific locales or according to particular identities of social groups. An agrarian political economy perspective, on the other hand, helps elucidate how social relations on diverse scales shape a broad network of accumulation, determine patterns of class formation, and ultimately identify patterns of winners and losers under a changing agrarian landscape. By incorporating the strengths of livelihood analysis and agrarian political economy, we can better identify the nuances and intersections of livelihoods associated with particular locales or identities and also understand their implications for a broad social and agrarian change, in such a way that neither approach can do by itself.

There is a theoretical challenge when it comes to combining these two approaches. A conventional livelihood analysis attempts to unfold individualistic strategic behaviors. The agrarian political economy primarily focuses on analyzing social differentiation, including power relations and institutional processes at meso or macro levels. At a glance, the level of analysis in each analytical frame appears to diverge. To link complex, individualized livelihoods to the dynamics of accumulation and social differentiation, some scholars argue that a ‘livelihood pathway’ approach is useful to provide a way forward [37,38,39,40]. According to de Haan and Zoomers [40], a livelihood pathway is defined as “a pattern of livelihood activities which emerges from a co-ordination process among actors, arising from individual strategic behavior embedded both in a historical repertoire and in social differentiation, including power relations and institutional processes, both of which play a role in subsequent decision-making” [40].

Three vital issues are associated with the ‘livelihood pathway’, a concept that forges synergetic links between livelihood analysis and agrarian political economy. First, a livelihood pathway highlights the formation and development of a collective pattern of livelihoods relating to a locale, a community, or any other identity. Accordingly, its plurality (livelihood pathways) signifies a differentiation of livelihood patterns across different locales, communities, or other identities. The genealogy of such a collectivized understanding of livelihoods could be connected to the emphasis of a political economy perspective: a certain common attribute that is shared by a group of people beyond individual differences (e.g., land ownership, income level, gender, age, ethnicity, and so on) provides a context for them to collectively engage with the livelihood of a similar trait.

Second, a livelihood pathway analysis highlights the collective dimensions of interaction between agency and ‘institutional processes/organizational structures’. Accordingly, it allows us to unpack the ‘black box’ of institutions and organizations regarding how they affect and transform the livelihoods of people at a certain collective level. This point could be better conceived of by considering multiple layers of commonality and difference observed for a people. For example, a group of rural dwellers might be seen as subsistence peasants living in the same region. However, the very same people could also be considered as an amalgam of sub-groups living in different communities with varying socio-cultural backgrounds and political networks. From the viewpoint of the original commonality of livelihood (subsistence farming), it would be difficult to understand how and why any difference in the livelihood emerges across the group of people after external institutions and organizations influence them (e.g., land loss through grabbing). It is possible, however, that certain collective variations at the sub-group level (e.g., ethnicity, educational level, political affiliations, etc.), which appear rather irrelevant at first sight, have played decisive roles in the formation and development of a distinctive set of livelihood pathways.

Third, as de Haan and Zoomers [40] argue, by incorporating political economy perspectives, a livelihood pathway analysis allows us to relate the identified collective patterns of livelihoods to the wider context of a society. As discussed above, a group of rural dwellers may share or shape a common trajectory of livelihood. In doing so, they develop a particular ‘habitus’, which, according to Bourdieu [41], is a system of dispositions primarily defined by social class and to be acquired through socialization. An agrarian political economy perspective helps link this domain of social differentiation and class formation to the macro-scale structure and processes, so that a livelihood pathway as the middle level concept could be relevant to the wider social and agrarian change (e.g., agricultural modernization, urban industrialization, and penetration of capitalism). Thus, there is a need to examine the transition from micro-scale (livelihood perspectives) to macro-scale (agrarian political economy approach). In addition, inspired by the temporal perspective of political economy, a livelihood pathway analysis adds historical moments (short- to long-term) to a livelihood analysis. These spatial and temporal moments of a livelihood pathway provide a useful bridge to the structural concerns of an agrarian political economy. Such an integrated approach is recursive in that it connects inter- and intra-household relations with the structural process of social differentiation within which their livelihoods are pursued, as shown in Figure 1.

By incorporating these conceptual leverages, an extended livelihood framework is built as an integrated model involving the elements of land grabs and agrarian political economy, as shown in Figure 2. In the integrated framework, land grabbing is considered as a critical factor for agrarian changes that affect individual household livelihoods. In the light of the framework, the processes of land acquisition appear to have changed the ability of households to access a bundle of livelihood assets (especially land and land-based resources). The transformation of land use reshapes the opportunities and constraints relating to the livelihood strategies pursued, as well as the outcomes achieved, of individual households.

In addition, land grabs are a critical driver pushing affected people into vulnerable conditions such as displacement, distress from land use transformations, food insecurity, erosion of customary culture, and social conflicts, among others. In response to land grabbing, individual households, depending on various given contexts, pursue different livelihood strategies with the combination of livelihood assets at their disposal to achieve the desired livelihood outcomes.

The peasants who have lost their land could have various options according to their given circumstances and priorities in needs. They would strive, covertly or overtly, to protect their land or get access to another land for sustaining the farming practices on which they have relied. Those who have lost their farmland but do not have the power to regain it nor have access to any other available livelihood sources may be left with no choice but to become wage workers for the investors. Meanwhile, others who have some good physical and financial capital at hand might prefer earning more income by offering new services to others. Others might also end up moving away from their neighborhood to a distant city to earn and send remittances home.

While individual household responses to land loss may vary, certain common patterns of livelihood strategies and outcomes may be observed at a micro- or meso-social level, which can be discerned as a set of livelihood pathways. As mentioned above, livelihood pathways would be a useful conceptual liaison to identify patterns of collective accumulation and social differentiation occasioned by land grabs over time.

As seen in the agrarian political economy of livelihood approach, five out of the six basic questions of agrarian study, as offered by Bernstein [36] and Scoones [22], have been ingrained to link the livelihood framework to the wider patterns of agrarian change. For the purpose of our inquiry, the question of “who owns what?” (or “who has access to what?”) relates to the issues of property and ownership of land and other livelihood assets that have been changed following a land grab. The questions of “who does what?” and “what do they do with it?” relate to the concerns of individual livelihood strategies for the short term, while from a medium- and long-term macro-scale viewpoint, they pertain to the social division of labor, patterns of consumption, savings/indebtedness, and investments taking place in the post-land grab phase. The question of “who gets what?” directly refers to short-term individual livelihood outcomes, but it also relates to the medium- and long-term patterns of macro-scale accumulation and socio-economic differentiation. The final question of “how do social classes and groups in society and within the state interact with each other?” offers a useful lens to examine the social relations among key stakeholders and other actors involved in the land grabs.

By adapting the foregoing framework, this paper attempts to respond to the above core questions for an in-depth understanding of agrarian change dynamics in the post-land grab phase in Vietnam. We aim to examine how livelihood pathways have been formed after the rubber plantation grabbing by a state-owned enterprise in the mountainous area.

3. The Study Site and Research Methods

3.1. Rubber Plantation Context

With the economic reform called “Đổi Mới” launched in 1986, rubber trees became the pivotal industrial crop, their agricultural industrialization and modernization serving the national goals for a key crop for export-oriented production and promotion of supportive policies for the development of the rubber sector [42]. Accordingly, the rubber plantation area and production in Vietnam increased sharply from about 105,000 ha and 82,000 tons in 1985, to 972,000 ha (over 172,000 compared to the 2020 national planning target) and 1.296 million tons in 2023 [43].

The government has encouraged large-scale rubber plantations through support for the Vietnam Rubber Group (VRG)—the largest state-owned rubber company in Vietnam—converting ‘insufficient agricultural land’ and forestland to rubber planting [44]. A representative example of rubber development is the case of Dien Bien province in Vietnam (Figure 3). Rubber plantations were first introduced in the province in 2010. To plant rubber trees in the province, a model of collaboration between the rubber company and local people was established, under the contract that the former provides the latter all the necessary investment capital and services including seedlings, input materials, training, and access to the markets. The latter in turn contribute their land to the rubber company (while still maintaining their land use rights certificate or ‘Rubber Red Book’ (Sổ đỏ cao su) on that land)1. The duration of the contract was 27–30 years, with options for renegotiation or extension. The contract guaranteed that, when the trees become productive, 10 percent of the financial value obtained from latex harvesting and wood liquidation would be shared among the land contributors. Local people were also recruited as workers for the rubber company and were promised financial compensation for their land contribution from the provincial budget. By the end of 2023, the rubber trees aged 10 to 17 years old covered a total of 5110 ha of the terrace cultivation land that had been contributed by 4135 households in the province [45]. In 2024, the company employed 661 permanent workers and approximately 400 seasonal wage-laborers. Among the company’s labor source, local ethnic minorities accounted for 84.6 percent [46].

3.2. Site Selection

To understand the land grabbing processes, ethnic minority households’ differentiated responses to the loss of their land, and the associated changes in their livelihoods and overarching agrarian transition, we examined the cases of three ethnic minority groups, namely Hmong, Kho Mu, and Thai, residing in six villages representing the six largest plantation plots of Dien Bien Rubber Company (DBRC) in the province. Those villages were the first rubber-planted locations and provide information to examine the long-term implications of land grabbing related to the development of rubber plantations. The criteria considered for selecting study villages include the demographic distribution of different ethnic groups, size of the land contributed to the DBRC, farmers’ options of available farming practices, as well as their accessibility to the new land in the post-land grab (

Table 1. Main features of selected villages in the study site.

Village Name	Tin Toc	Huoi Chan 1	Xa Nhu	Na Sang	Ta Lao	Co Ha
Commune name	Muong Pon		Hua Thanh	Na Sang	Na Say
Total land area (ha)	374.6	394.0	120.2	411.0	146.5	151.0
Population (person)	378	502	196	760	302	234
2024 household number	73	123	45	158	58	51
Ethnicity (percent) (Hmong/Kho Mu/Thai)	0:100:0	0:44:56	100:0:0	56:23:10	0:0:100	0:0:100
Year of land contribution	2008	2008	2008–2009	2009–2013	2010–2015	2010–2015
Percentage of land-contributing households in all the village households	79.4	61.4	91.1	58.5	80.0	93.3
Arable land size (percent) (Rubber plantation/terraced land/paddy rice land)	69:27:4	49:45:6	72:24:4	27:67:6	65:30:5	85:10:5
Number of sampled households	33	36	38	51	24	23

Sources: Author’s compilation based on secondary document review.

).

3.3. Research Methods

Before starting the fieldwork, we sought permission from the DBRC, the Provincial People’s Committee (PPC), the District People’s Committee (DPC), the Commune People’s Committee (CPC), and the administrative heads of the six villages. To foster trust within the community, the first author spent one week residing in each village before engaging in any discussions involving sensitive or potentially contested issues. During this preparatory phase, regular consultations were held with village leaders to mitigate the risks of cultural misinterpretation and ensure that all interactions were informed by local norms and expectations. We conducted a household survey using semi-structured and in-depth interviews, focus group discussions, and analysis of secondary information obtained from the local government and the DBRC. Furthermore, we carried out field observations throughout to gain insights of the villagers’ daily activities.

The field research was conducted in two phases. The first phase took place from July to September 2018, during which the rubber plantations in the Northwest region of Vietnam were in the end of the nurturing stage. The second phase was carried out from December 2023 to March 2024, at a time when the plantations had been in productive operation for five years. This longitudinal approach enabled an in-depth assessment of the livelihood pathways of the ethnic minority communities under the impact of rubber plantation land acquisitions.

Prior to a household survey, data from all the households whose land had been acquired by the rubber company in the study villages were collected, and then cross-checked with a household list provided by the DBRC. To address concerns regarding potential sampling bias, this study employed a stratified random sampling method to enhance representativeness and analytical rigor. The theoretical basis for stratification was informed by the literature on agrarian differentiation, which highlights the importance of ethnicity, landholding size, and livelihood strategies in shaping responses to rubber plantation land grabs.

The total sample of 205 households was proportionally allocated across three main ethnic groups—Thai, Hmong, and Kho Mu—according to their relative representation in the study villages. In the period preceding the land grabs, the Hmong people (in the case of Na Sang village and Xa Nhu village) traditionally settled further uphill, on steeper land separately from other groups, growing maize, cassava, and upland rice by slash-and-burn cultivation practices. The Thai people (in the Huoi Chan 1, Ta Lao, and Co Ha villages) basically inhabited valley areas, with their traditional livelihoods relying mainly on the combination of wet rice production in the lowlands and slash-and-burn cultivation of upland crops2. The Kho Mu people (in the Tin Toc, Huoi Chan 1, and Na Sang villages) resided in the areas ranging from valleys to middle-hill land based on traditional slash-and-burn cultivation, with several households engaging in wet rice cultivation (Table 1).

Among each ethnic group, the respondents were categorized into three sub-groups based on landholding size at the time right after they contributed land to the rubber company. These categories—small, medium, and large landholders—are not based on standardized academic benchmarks but rather reflect locally recognized distinctions. During preliminary fieldwork and discussions with village leaders and residents, households consistently described themselves using these thresholds: ‘small landholders’ (S) were those with less than 15 sào (0.54 ha) of farmland, ‘medium landholders’ (M) with landholdings between 15 sào and 50 sào (0.54–1.8 ha), and ‘large landholders’ (L) with more than 50 sào (1.8 ha)3. These distinctions are grounded in local perceptions of land sufficiency and agricultural capacity and are closely tied to villagers’ evaluations of livelihood viability and household production potential. By incorporating emic definitions, the study aligns with ethnographic best practices and enhances the contextual relevance of our analysis.

This proportional stratification ensured that our sample captured key socio-economic variations both within and across ethnic groups, thereby enhancing the analytical rigor of our comparisons. Household heads and/or their spouses were selected as primary respondents. The distribution of sampled households by ethnicity and landholding size is presented in

Table 2. Number of surveyed households by ethnicity and landholding.

Criteria	Small Landholders		Medium Landholders		Large Landholders		Total
	n	Percent	n	Percent	n	Percent	n	Percent
Hmong	20	30.3	29	43.9	17	25.8	66	100.0
Kho Mu	32	42.7	26	34.7	17	22.7	75	100.0
Thai	44	68.8	16	25.0	4	6.3	64	100.0
Total	96	46.8	71	34.6	38	18.5	205	100.0

Sources: Author’s compilation from the household survey.

.

Following the preliminary analysis of household survey data, six households were purposively selected from each village to participate in in-depth interviews. These interviews were designed to explore, in greater detail, the livelihood strategies adopted by households in the aftermath of land dispossession and the various forms of community responses to land acquisition. In addition, two administrative village leaders from each site were interviewed to provide contextual information on the village’s historical development, shifts in farming systems, socio-economic challenges, and the evolving prospects for both agricultural and non-agricultural livelihoods. Complementary insights were also gathered through a series of interviews with a staff member from the rubber company, who had facilitated our fieldwork logistics and offered perspectives on the implementation and local reception of the plantation project in Dien Bien province. The field research with focus group discussion employed the seasonal calendar and village history for deeper understanding of the local people’s land contribution processes as well as their strategies to adapt themselves to their land loss. The research also involved 13 semi-structured interviews with commune, district, and provincial locals and seven semi-structured interviews with the rubber company staff. Hence, the subsequent analysis is predominantly based on the secondary data gathered from the DBRC and authority offices, as well as the observations made during the fieldwork.

Data from the semi-structured interviews, focus group discussions, and field observations were analyzed using qualitative thematic analysis. An inductive coding approach was adopted, where key codes and themes emerged from a close, iterative reading of transcripts and field notes. The lead researcher first conducted open coding manually in Microsoft Excel, identifying recurring ideas, narratives, and patterns across the dataset. These initial codes were grouped into broader thematic categories related to land access, livelihood assets, livelihood strategies, and perceptions of land-use change. To ensure reliability, we employed peer debriefing and collaborative cross-checking. Two trained research assistants independently reviewed a subset of transcripts and the proposed coding framework. Differences in interpretation were discussed in multiple rounds of coding meetings until consensus was reached on how codes were applied. While we did not conduct formal inter-coder reliability testing, this collaborative process aimed to enhance consistency and reduce subjective bias. Field observations were documented systematically through handwritten notes and, where appropriate, photographic evidence. These materials were reviewed and thematically analyzed alongside interview data to identify patterns and corroborate findings, while descriptive analysis of quantitative data was completed using Statistical Package for Social Sciences (SPSS, https://www.ibm.com/products/spss-statistics).

Further triangulation was employed by comparing findings across household surveys, in-depth interviews, focus group discussions, and direct observations. This allowed us to cross-validate themes and strengthen the robustness of interpretations. Field notes, including photographs and contextual annotations, were analyzed alongside interview data to deepen the understanding of patterns and to support analytical claims. All direct quotes were lightly edited for clarity, and pseudonyms were used to ensure the anonymity of informants.

4. Results

4.1. Land Acquisition for Rubber Plantation in the Study Site

4.1.1. Site-Specific Land Acquisition Processes in the Study Villages

The land acquisition process in the studied villages differed in terms of timing of the acquisition, land tenure status before and after the acquisition, and terms and conditions of the land contribution contract, as shown in

Table 3. Land acquisition in the six studied villages of Dien Bien province.

Characteristics of Land Acquisition	Tin Toc and Huoi Chan 1 Village (Muong Pon Commune)	Xa Nhu Village (Hua Thanh Commune)	Na Sang Village (Na Sang Commune)	Co Ha and Ta Lao Village (Na Say Commune)
Year of acquisition	2008	2008–2009	2009–2013	2010–2015
Land tenure in the pre-acquisition phase	Land tenure was controlled by customary regulations.	Land tenure was controlled by customary regulations.	The land was distributed to villagers by customary tenures
Land demarcation in the re-allocation and post-acquisition phase	The rubber company’s bulldozers entered the fields of land contributed by farmers to the company, and evened out the terrain, consequently erasing any pre-existing border lines for the identification of the specific plots.		The demarcation of land was conducted before the land clearance by the rubber company. The boundaries of specific plots were identified under the consensus between villagers and local authorities. Customary tenure was recognized during the process of legal demarcation of land.
Rubber tapping period	Since 2016		Since 2017 in Na Sang and since 2018 in Co Ha and Ta Lao
Status of benefit-sharing	Benefit-sharing (10 percent of latex harvesting value) was paid to villagers in October 2018.		Benefits have been shared by villagers as of the end of 2021.

Source: Authors’ compilation based on secondary document review, focus group discussion, and in-depth interviews.

. With strong support from the local government and the directions from the VRG, this company rapidly acquired agricultural land and forestland in the targeted area, although the provincial policy guidelines had not been issued yet. With the same approach to acquire land in the four communes, the company staff came to the village with the local authorities and sought to persuade villagers to contribute their land by organizing several meetings to explain the benefits that the local people would receive. In the persuasion process, a very positive picture was painted such that the villagers would move from a poor and backward to a rich and prosperous status, bolstered by new roads, schools, and health clinics to be built by the company, as well as stable and well-paid employment at the plantation. Further, compensation from the government for land contribution and other supports such as the provision of rice and salt within the first two years following their land contribution were suggested.

In the case of villagers refusing to join the rubber plantation project, the company staff forced the villagers to succumb by arguing that the project was to be implemented in line with party and government policies and the local authorities had already agreed to dispatch the company for land clearing. Then, DBRC workers brought bulldozers to clear the land so quickly that villagers did not have sufficient time to make any meaningful response. Along with force and pressure, the company also resorted to using workers hired from the households who had agreed to clear their field land in exchange for a promised high wage payment. Subsequently, rubber seedlings were planted by the company workers and local casual wage-laborers.

4.1.2. The Control of the Rubber Company over the Land in the Study Villages

The rubber project has had significant impacts on land use and control in the affected villages. Both the landscape inside the rubber field and the typology of land in the village have been transformed dramatically. Rubber monoculture has become a dominant form of land use, accounting for 69.1 percent and 27.7 percent of the arable land in Tin Toc and Na Sang, respectively (Figure 4 and Figure 5). Most of the land for rubber plantations in these two villages was used for cultivation of upland rice, maize, and cassava, rather than maintained as production forestland and ‘unused land’, which were designated by the provincial government in the official land-use planning for rubber plantation development.

It was found that the largest proportion of the cultivated land had been acquired to plant rubber trees. For example, the Kho Mu people lost the largest areas of farmland with an average of 2.61 ha per household, accounting for 74 percent of their average total farmland, followed by Thai people with 2.19 ha (84 percent), and Hmong people with 1.5 ha (53 percent) (

Table 4. The transformation of land size of the surveyed households by ethnicity (average per household). Unit: m².

Indicator	Hmong (n = 66)	Kho Mu (n = 75)	Thai (n = 64)
Total farmland before contribution to the rubber company (A) **	28,274 a,b	35,021 a	26,635 b
Farmland contributed to rubber company (B) ***	15,033 a	26,135 b	21,917 b
Rubber land redistributed to households for benefit-sharing (C) ***	6969 a	13,976 b	15,000 b
Total remaining farmland right after contributing to the rubber company (D = A − B) ***	13,241 a	8886 a	4718 b
Farmland bought/rented after contribution (E) ***	7723 a	133 b	18 b
Land sold/rented out after contribution (F) **	167 a	960 a	0 a
Farmland reclaimed after contribution (G) *	4682 a	3995 a	2908 b
Total current land area (D + E − F + G) ***	25,479 a	12,054 b	7644 c

Note: *, **, and *** represent significance at 10%, 5%, and 1%, respectively. The test was conducted with Turkey’s post hoc test (the same subscript letters in a row indicates the mean values were not significantly different). Source: Author’s household survey, 2024.

). Classified by landholding scale, small-scale households accounted for 93% of the total farmland area contributed to the DBRC, followed by medium-scale and large-scale households, with corresponding contribution rates of 65% and 41%, respectively (

Table 5. The transformation of land size of the surveyed households by landholding (average per household). Unit: m².

Indicator	Small Landholders (n = 96)	Medium Landholders (n = 71)	Large Landholders (n = 38)
Total farmland before contribution to the rubber company (A) ***	23,328 a	31,724 b	48,647 c
Farmland contributed to rubber company (B)	21,887 a	20,776 a	20,123 a
Rubber land redistributed to households for benefit-sharing (C)	12,608 a	11,384 a	11,483 a
Total remaining farmland right after contributing to the rubber company (D = A − B) ***	1441 a	10,948 b	28,524 c
Farmland bought/rented after contribution (E)	2382 a	2151 a	3729 a
Land sold/rented out after contribution (F) ***	0 a	476 a,b	1514 b
Farmland reclaimed after contribution (G)	4468 a	4113 a	1643 a
Total current land area (D + E − F + G) ***	8291 a	16,735 b	32,381 c

Note: *** represent significance at 1%. The test was conducted with Turkey’s post hoc test (the same subscript letters in a row indicates the mean values were not significantly different). Source: Author’s household survey, 2024.

).

The DBRC controls acquired lands by various strategies. At varying times, ranging from 3 to 7 years after the conversion of land to rubber plantations, villagers received a Rubber Red Book by which they were claimed to be officially certified for the land redistributed for benefit-sharing. However, the land size registered in the Rubber Red Book was often significantly smaller than the actual size of the originally contributed land. The survey results show that on average, each household of Hmong, Kho Mu, and Thai people contributed 1.51, 2.61, and 2.19 ha, whereas the land size in the Book was only 0.69, 1.39, and 1.5 ha, respectively. There are two reasons for this situation. First, in the Tin Toc, Huoi Chan 1, and Xa Nhu villages, the rubber company redistributed only the land cultivated by villagers at the time of land clearance, while they did not recognize either fallow land, communal land, or ‘unused land’ owned by the communities. Therefore, they claimed not to have a responsibility to reallocate such lands to villagers. After a series of villagers’ requests to the rubber company to return communal land to the village, the company redistributed a small portion of rubber plantation land, called the ‘rubber land fund of the village’ (Quỹ đất cao su của bản), to the village. Subsequently, a parcel of the ‘rubber land fund’ was allotted to some extremely poor households. The remaining area of the ‘rubber land fund’ would be used for benefit-sharing when rubber trees on the land were tapped. The financial return from the rubber land fund would be added to the villages’ budget.

Second, in the case of the Na Sang, Co Ha, and Ta Lao villages, the land demarcation was undertaken by the cadastral office and the DBRC before the land was contributed to the rubber company. The villagers were promised that every household would receive a Rubber Red Book which would accurately show the land area that they had contributed. However, like the aforementioned villagers, those in Na Sang, Co Ha, and Ta Lao were allotted significantly smaller land areas than what they had contributed. In the process of measuring the land size and issuing the Rubber Red Book, the rubber company negotiated with the district cadastral office to restrict the area of the land designated as ‘contributed’ only to the exact area where rubber trees have been planted. Accordingly, the areas without planted rubber trees, including abandoned land and roads inside the rubber plantations, were excluded from the Rubber Red Book.

Regarding the size of the land redistributed to the villagers for benefit-sharing, the results of survey and group discussions reveal that no villagers in any of the surveyed villages have ever seen the certificates, even though LURCs had been issued. The reason is that, after issuing the Rubber Red Book, the DBRC negotiated with the provincial government to add a specific article under the Land Contribution Contract, thereby allowing the district cadastral office to keep the book away from the villagers for 27 years. It is likely that the practice of keeping LURCs away from the villagers was intended to withhold information on rubber land allocation from the knowledge of the land contributors. The villagers stated that during the tending and tapping periods, the DBRC divided the rubber plantation plots into many small parcels that were largely inconsistent with the official recognition of land use rights derived from the LURCs. Some villagers claimed that this strategy was designed to ensure that villagers would not be able to take their land back nor to take any control over the use of the land they had owned. In addition, the tapped parcels were allocated to local workers and informally contracted tapping laborers for their plantation activities of the next 20 years. The assignment of rubber parcels to each laborer for tapping was carried out in a randomized fashion. On average, a laborer who taps on 4.0 ha of land (a planting of roughly 1400–1600 trees) is paid VND 4000 (USD 0.16) per kilogram of harvested latex. With this financial incentive, they would become dedicated guards for the rubber plantation.

In sum, the DBRC’s investment in rubber plantations in Dien Bien, backed by the government, demonstrates that the rubber company exercised full authority and power to capture the common land and dispossess the villagers’ lands for their commercial interests. Indeed, the socialist state-led land grabs reveal the specific mode of ‘control grabbing’ in which the local government and its affiliated agencies imposed rules and regulations designed to legitimize such activities, facilitating an environment in which the domestic state-owned enterprise (SOE) could take control over the villagers’ lands for the project. This was done by manipulating the original land users’ land entitlements through an official land tenure program, resulting in a significant transformation of land use and control following the land grabbing. The scrutiny into the whole process reveals not only economic but also political and social incentives of the state. The introduction of the rubber plantation project was driven not only by the upsurge of rubber prices, or the so-called ‘crop boom’ occurring like elsewhere in Southeast Asia, but also backed by the State’s notion of shoring up ethnic minority people’s loyalty. The villagers’ loss of farmland and the takeover of the rubber plantations by the DBRC compelled villagers to seek a new set of livelihoods for their survival.

4.2. Livelihood Pathways Among Local Peasants

The ethnic minorities living in the mountainous areas of Dien Bien province over generations had maintained their traditional farming practices for subsistence purposes. The rubber plantation land grabbing advancing into their areas significantly transformed their land use systems and practices, resulting in a divergence of individual and household livelihood trajectories as well as of collective livelihood pathways (Figure 6).

The first livelihood pathway comprises a group of households who pursued commercial farming as their main livelihood strategy through retention and accumulation of farmland after their land loss. In particular, Hmong farmers used multiple strategies to maintain the size of their farmland. Hmong farmers in Na Sang village were aware of the potential of commercial pineapple cultivation for sustaining their livelihoods through observation of their relatives’ success in the visit to a neighbor province in 2007. Some of them transplanted pineapple seedlings in 2009. With the first harvest, pineapple farmers increased their income by three to four times compared to before4. Then, farmers who pioneered pineapple farming shared information and encouraged others to grow pineapples on their farmland that had been used mainly for maize and rice cultivation. The growth of pineapple production in their villages helped them to retain much of their farmland when the land acquisition processes for rubber plantations were in place. Medium and large Hmong farmers were able to retain their land thanks to their close political relations with commune officials. The farmers made a plea to the officials for disallowing the rubber company to take away their land for rubber plantations. The request was firmly approved. The profits obtained from commercial pineapple cropping also helped them to accumulate wealth in order to buy new uplands from large Kho Mu and Thai landholders in the village. Consequently, most of the interviewed Hmong villagers have become large landholders. The land accumulation also contributed significantly to improving their food self-sufficiency. In addition, more income, land, and reserve capital have enabled the Hmong people of the wealthy strata to extend their farming to include paddy rice cultivation in the lowland area, further contributing to the accumulation of their wealth. Hence, the livelihood pathway of Hmong farmers generally achieved an upward trend.

The second livelihood pathway comprises those locals who sought to diversify their livelihoods by combining on- and off-farm activities. As observed in the study villages, the large Kho Mu landholders still had a sufficient amount of farmland for household food consumption after they contributed their land to the rubber company. They were actively involved in the rubber plantation in the first three years of rubber planting. Among them, there were also those who acquired a terraced paddy land either for subsistence rice farming or for selling it to wealthy Hmong people, or both. Ownership of a large area of land allowed some of them to sell or rent out their land to Hmong people, as they were more or less in need of money for spending on health care, children’s education, opening groceries, or agricultural inputs. Since the rubber trees were tapped, large Kho Mu landholders also participated actively in rubber plantation work. Consequently, the diversification of their livelihoods via increased on- and off-farm work opportunities made them better off.

The third livelihood pathway embraces a group of households who engaged in wage labor work for the rubber company as their main source of livelihood while expanding their subsistence farming through acquisition of new farmland. The small and medium Kho Mu landholders, who tended to have a low educational background and limited access to the output market, contributed the largest amount of land to the rubber company. During the long gestation period of the rubber trees, they had to endure severe economic hardship. Accordingly, the loss of their farmland had resulted in a significant reduction of food self-sufficiency, which compelled them to rely on casual or permanent wage work for the rubber company, coupled with gathering NTFPs from protected forests. Their employment opportunities at rubber plantations were limited, however. Hence, they suffered from serious indebtedness for food consumption and reclaiming of land (which entailed fees for land clearance and deployment of in-field irrigation systems). Overall, the livelihood pathway of this group was downward in the post-land grab phase.

There is yet another livelihood pathway that has been formed in the ethnic minority uplands in the post-land grab phase. It comprises the inhabitants who focused on seasonal out-migrated wage work as their main livelihood strategy. After the introduction of the rubber plantation project, Thai people retained entitlement to the smallest land after they contributed their land. While they retained some land for paddy rice cultivation for subsistence, the significant loss of their farmland forced them to look for alternative livelihoods. During the first three years of rubber tree planting, the villagers initially found opportunities for rubber plantation wage work such as clearing land for planting and tending of young rubber trees. In addition, based on their reliable relationship with their relatives, relatively high educational backgrounds, and available social networks in industrial areas, the Thai people have engaged in seasonal migrant wage work as their major livelihood source since they contributed their farmland for rubber plantations. Part of the earnings were sent back home for the purchase of housing facilities and other household assets. Some large Thai landholders diversified their livelihood activities by initiating a grocery shop business or engaging in trading. Overall, the livelihood pathway of Thai people achieved a significantly upward trend.

5. Discussion

Considering the findings above, this study solicits a renewed understanding of the consequences of land grabs for livelihoods in the light of the agrarian political economy of livelihood framework. The key discussions are outlined below.

First, right after their land loss on behalf of contribution for rubber plantations, the local peasants diversified their formerly subsistence-based livelihoods into various forms such as cash crop cultivation, acquisition of new land, rubber wage labor, and out-migrated urban labor. As observed in the study villages, the choice of livelihood strategies was made consciously based on the livelihood assets that each ethnic group used to deploy, and the socio-political relations as well as market conditions they were involved with. As we have seen, the total remaining farmland in the aftermath of the land contribution to the rubber company played an essential role in determining households’ livelihood choice, such as on-farm strategies, off-farm activities, or both in combination. Small landholders contributed almost all their land to the rubber company, with many of them being unable to acquire new land. Hence, they had no choice but to find wage labor work wherever available. Here, their relationships with their relatives, educational background, and social networks were the key factors that determined what type of wage labor work they could engage in. Thai people, who tended to have relatively high educational levels, available social networks, and reliable relationships with their relatives, were able to get more lucrative wage work in the city. Meanwhile, Kho Mu people, who tended to lack those conditions, were engaged in the relatively less lucrative wage work for the rubber company. In contrast, medium- and large-sized landholders of all the three ethnic groups were able to maintain existing farming practices while being engaged in wage work for the rubber company, or urban industrial wage work. On the other hand, political connections with local officials allowed Hmong people to retain large areas of farmland, on which they increasingly pursued commercial farming by taking advantage of the availability of output market access.

Geographical location has a profound influence on households’ livelihood choice. To minimize operational costs, the Dien Bien Rubber Company prioritized the establishment of rubber plantations in villages characterized by favorable biophysical and infrastructural conditions. In villages such as Tin Toc, Huoi Chan 1, Xa Nhu, Co Ha, and Ta Lao—where road access is relatively convenient and the terrain is less steep—much of the cultivable land previously managed by local households was converted into large-scale rubber plantations. In these villages, the residents were left with little agency over their livelihood choices, as the large-scale conversion of farmland to rubber plantations severely constrained their access to productive land. Consequently, many households were compelled to shift their livelihood strategies toward wage labor employment with the rubber company or to engage in seasonal out-migration for off-farm income opportunities. In contrast, Na Sang village is situated in a more remote and mountainous area, with limited transport infrastructure and reduced connectivity to markets. These spatial constraints, alongside the Hmong community’s strategic engagement with local political structures, enabled households not only to retain their agricultural land but also to expand their holdings. As a result, many pursued a distinct livelihood pathway centered on commercial farming, diverging from the agrarian transformations observed in the other study sites.

All these findings clearly suggest that though sharing the identical operational source of land acquisition, the affected local people with different social-economic backgrounds under varying political, cultural, and spatial contexts had significantly different responses to their land loss. Hence, it is argued that in analyzing the impacts of land grabs, communities could be seen as heterogeneous units, and that our analysis should not confine itself to short-term, immediate economic impacts and responses, but embrace the reformulated long-term patterns of social differentiation and class formation in wider contexts.

Second, in the land grab literature, there is a common argument that in the post-land grab phase, larger landholders have the potential to achieve higher incomes and assets than smaller landholders by taking advantage of their asset status to accumulate further land from desperate small farmers through an “everyday process of accumulation and dispossession” [9]. Access to land plays a key role in creating winners (large landholders) and losers (smallholders), with the losers being forced to sell their remaining land, seek wage employment locally, or migrate seasonally or permanently. This ‘everyday process’ gives rise to a differentiation among farm households in terms of their access to land, which consequentially leads to the formation of contrasting livelihood pathways. However, the development of the pathway toward forming large landholders in our case does not fit in this process. In the post-land grab stage, Hmong people retained a large area of land by transforming their farming system from slash-and-burn to intensive cash crop cultivation. As they continued commercial farming, they accumulated financial capital, and using it, they bought more land over time. However, the large landholders did not buy land from small landholders because the latter tended to retain their small plots of land for food subsistence rather than sell them to the former. Hence, to acquire new land, large landholders negotiated with other large landholders who needed more money for technical investment in intensive farming or children’s education.

Third, the rubber plantation land grabs taking place in Dien Bien point to the conscious efforts of the socialist state to involve ethnic minority populations in the ongoing process of “market integration, replacing the common property with private land use rights, pressing shifting cultivators to become settled farmers” [48]. Like numerous state policies introduced in the ethnic minority uplands, the rubber plantation project shows that the state aims to “handle highland minorities in the most effective and economic way” to avoid impeding Vietnam’s steady ‘modernization’ [49]. Ultimately, by employing these projects, the State’s goals are to “ensure that [ethnic minorities’] economic activity was legible, taxable, assessable, and confiscatable or, failing that, to replace it with forms of production” [50].

Fourth, these findings also have broader implications when viewed through the lens of ethnic inequality and agrarian justice. The differentiated livelihood trajectories among the Thai, Kho Mu, and Hmong communities are not simply the result of individual household choices or market dynamics but rather reflect deeper structural inequalities embedded in Vietnam’s upland development policies and land governance regimes. The rubber plantation expansion, though officially promoted as a modernization and poverty reduction initiative, has in practice privileged groups with greater access to land, education, political connections, and market resources—reinforcing existing hierarchies and creating new forms of exclusion [51,52]. For instance, the Hmong’s ability to retain and commercialize large landholdings was closely tied to their connections with local officials, while the Kho Mu, with fewer institutional ties and weaker educational backgrounds, were more often pushed into precarious forms of labor. This suggests that the state-led process of “legibility” and “market integration” has not affected ethnic groups equally; instead, it has reproduced patterns of marginalization under the guise of development [7,47,53,54]. By reshaping land access, labor relations, and market participation, these land grabs have contributed to a new agrarian order in which inequality is restructured along both class and ethnic lines. As such, our findings call for a more critical engagement with the justice dimensions of land-based development projects in multi-ethnic contexts, where the long-term social consequences may outweigh short-term economic gains [53,55].

Last, this study advances the agrarian political economy of livelihood framework originally articulated by Scoones [22] by proposing an integrated conceptual approach that incorporates the land grab phenomenon as a key analytical entry point. While Scoones emphasized the need to connect micro-level livelihood strategies with broader political-economic structures, our framework operationalizes this connection by tracing how land dispossession reshapes household livelihoods in the short term and contributes to longer-term processes of social differentiation. Specifically, we show how land grabs—implemented through state-corporate alliances—initiate immediate disruptions to local livelihoods, which in turn catalyze divergent livelihood pathways based on households’ differential access to assets, socio-political capital, and ethnic identity. Over time, these differentiated responses evolve into broader patterns of agrarian restructuring at the community and regional levels. Thus, the proposed framework enables a dual-level analysis: capturing both the short-term micro-level consequences of land loss and the long-term macro-level outcomes in terms of emergent agrarian transitions. In doing so, this study offers an important theoretical extension to Scoones’ framework by deepening its capacity to account for the temporal dynamics of land-based transformations and the intersectionality of ethnicity, class, and state power in shaping rural livelihoods.

6. Conclusions

In the last decades, there has been a proliferation of case studies in the land grab literature that are focused on the impacts of land grabbing on the livelihoods of households and individuals within a community. In this regard, local people in the affected community are treated as if they were a homogeneous unit, with the community members uniformly affected by the land loss. However, the impacts may not necessarily be uniform but could be significantly different among the different groups within a community, let alone across different neighboring communities. In addition, there is a lack of studies analyzing the impacts of land grabs by integrating livelihood perspectives with wider contexts of agrarian political economy. This paper has attempted to fill these gaps by highlighting the consequences of rubber plantation land grabs on the livelihoods and social stratification of ethnic minority peasants in Northwest Vietnam.

Drawing on the integrated framework of the agrarian political economy of livelihood, this paper has highlighted the consequences of land grabs by revealing distinctive patterns of livelihood pathways within a complex web of intersections across class and ethnicity in an upland area. An emphasis on the livelihood aspect of the integrated framework has allowed us to illuminate the emergence of a set of distinctive livelihood pathways. Following the land grabs, individual households, with different socioeconomic, political, ethnic, and spatial backgrounds, have drawn on a bundle of livelihood assets at their disposal. In doing so, they have pursued various livelihood strategies that might be most congenial to their needs, thereby achieving various livelihood outcomes. Accordingly, as time goes on, common livelihood patterns have emerged at the local community level, influenced to a significant degree by the given local context after land contribution. This paper has highlighted that the activities of these groups were identified, chosen, and carried out as part of their conscious livelihood strategies based on the livelihood assets at their disposal, including the socio-political relations and market conditions they were involved with.

This study further argues that, as a conceptual approach, the agrarian political economy of livelihood is a framework that has an inexhaustible potential for analyzing the impact of land grabs on livelihoods thanks to its integrative feature to link short-term micro-scale and long-term macro-scale contexts. Although our research findings drew on a limited number of villages in Northwest Vietnam, there are several areas and communities across the country that accommodate similar local contexts of land property relations, livelihood conditions, and surrounding geographical, socio-political, and cultural contingencies. For instance, in many upland communities, local people have maintained customary land rights over generations. Their livelihoods mainly rely on subsistence farming, which heavily depends on the land and land-based resources. Hence, enclosure of communal land and associated land dispossession from the original land users could be operated relatively easily because of the existing ambiguous land tenure system and the land users’ limited bargaining power to maintain the land [47,56]. Consequently, similar patterns of livelihood pathways, as we observed in our research, could be seen widely in Vietnam. Unpacking these patterns, as Scoones noted, could itself be an act of ‘bringing politics back in’, which was originally central to livelihood analysis, whose essential domain has often been downplayed and even lost through various externally driven aid deliveries in the past three decades.

The findings of this study highlight significant differences in the impacts of rubber plantation expansion in the Northwest region of Vietnam, largely influenced by local governance capacity, land tenure systems, and ethnic minority livelihood strategies. To address these disparities, several policy recommendations are proposed. First, strengthening community-based land governance is critical, especially in areas where ethnic minorities face challenges securing land rights. Implementing participatory land-use planning and formalizing customary land rights would help ensure tenure security and reduce the risk of land dispossession. Second, local authorities should design region-specific livelihood support programs, tailored to the unique socio-economic conditions of ethnic minority communities, particularly those affected by land conversion for rubber plantations. A uniform approach to livelihood assistance is insufficient, as it fails to account for the diverse needs and vulnerabilities of different groups. Third, there is a need to enhance transparency and accountability in rubber plantation investments by establishing robust monitoring mechanisms. Affected communities must be adequately informed and consulted throughout the process, and environmental and social impact assessments should be conducted transparently. This would allow for adaptive policy interventions and ensure that the development of rubber plantations is sustainable and beneficial to all stakeholders in the long run.

This study acknowledges several limitations. The research is confined to a small number of villages in Northwest Vietnam, which limits the broader applicability of the findings. While the analysis offers deep qualitative insights supported by descriptive statistics, it does not include cross-regional quantitative comparisons. To enhance understanding, future research should adopt a comparative, multi-site approach across different ethnic groups and regions. Longitudinal studies are also recommended to trace how livelihood pathways and social differentiation evolve over time. Despite these limitations, this study proposes that its findings on social differentiation and livelihood pathways may be relevant to other regions with similar characteristics, such as customary land rights, subsistence farming, and limited political influence. The core intention of this study is not to generalize its findings universally, but to offer a conceptual framework that can be applied to investigate comparable land-related transformations in other contexts.