Despite repeated emphasis on the links between the natural environment and human well-being and the disproportionate and direct dependence of the rural poor on natural resources, these links have not been well addressed in poverty assessments. Common poverty profiles neither reflect the contribution of nature to well-being nor the multiple values and meanings that people ascribe to nature. Building on a conceptual grounding for including environmental components in well-being measures, our work aimed to determine for which components it is legitimate to do so according to the people whose well-being is measured. We developed a focus group discussion protocol to elicit perceptions of environment-well-being relationships in rural settings in Rwanda and Malawi. The protocol included a well-being free-listing exercise, a matching exercise linking the listed items to predefined well-being dimensions, and a discussion of environment-well-being connections. We found that severe environmental degradation, hazards, and conflicts over access to land and forests in these diverse rural areas are deeply and directly linked to well-being. Environmental changes such as flooding or extended drought led to losses of income, crops, and assets, as well as prolonged periods of psychological stress, constrained freedom of choice, and in extreme cases, death. Our results suggest that some environmental components are constituent to well-being. We emphasise the importance of validating the precise environmental components that are considered relevant to well-being in different contexts. Extending poverty measurement with relevant environmental components can help in targeting action towards reducing poverty in a more legitimate, context-specific way.

It is well recognised that people living in impoverished situations are disproportionately and more directly dependent on natural resources, particularly in rural areas (Angelsen et al., 2014; Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services, 2019; Millennium Ecosystem Assessment, 2005). While access to healthy ecosystems can positively influence one’s quality of life, such dependences also make people’s well-being more vulnerable to ecosystem degradation and changing access regulations (Barett & Bevis, 2015; Kangalawe & Noe, 2012).

Incorporating locally defined values of the natural environment (hereafter, environment) could lead to different approaches in conservation and poverty alleviation (Schreckenberg et al., 2016; Sterling et al., 2017). Beyond physical dependence, the natural environment takes on a defining role in various cultures and worldviews as an entity in which humans are embedded (Satterfield et al., 2013; Schnegg et al., 2014). For example, the Maori worldview of water and its impact on their well-being is more complex than a measurement of ambient water quality—the focus of many conservation initiatives. The Maori consider some water sources as sacred (tapu) and do not distinguish between spiritual health and the ecological state of water sources. Focusing on ambient water quality is incomplete in its ability to assess Maori values (and thus influence on well-being) including the role of particular locations in creation stories and the ability for a site to be used by future generations (Tipa & Tierney, 2003).

The recognition of local livelihood dependencies on the environment has not been adequately translated into policy interventions towards poverty reduction or biodiversity conservation, jeopardizing the effectiveness of these interventions (Arias-Arévalo et al., 2018; Schleicher et al., 2018). Poverty reduction projects that fail to recognise these values and dependencies overestimate their impact if they ignore well-being losses that result from, for instance, agricultural development at the expense of forests, such as loss of spaces for cultural practices, identity, safety, and diversity in livelihood activities (Gross-Camp et al., 2019). Ignoring the role of the environment in poverty statistics can lead to misidentification of those living in extreme deprivation (Li et al., 2020), risking the implementation of misguided interventions to improve well-being (Adjei et al., 2017; Nayak et al., 2014). Yet, mainstream poverty indices, such as income-poverty lines, the Human Development Index (HDI), or the Global Multidimensional Poverty Index (MPI), do not recognise the contribution of the environment to well-being. While it is beyond the scope of this article to explore the cause of this omission in detail, it may be linked to discourses around poverty as a driver of environmental loss and vice versa (Adams et al., 2004), mainstream belief in Environmental Kuznets Curves where economic growth for poverty alleviation trumps nature conservation (Liu, 2012) and confusion between environmental hazards and degradation (Gray & Moseley, 2005; Satterthwaite, 2003).

One way of increasing the visibility of the role of the environment in well-being is to include an environmental dimension in poverty indices. Recognition of diverse local values of nature requires methods and approaches that are sensitive to local specificity. Plural valuation highlights the need to consider diverse values (and their incommensurability) related to justice, love, friendship, and knowledge, often shaped by human relationships with nature or other people (Arias-Arévalo et al., 2018; Beckerman & Pasek, 1997; Trainor, 2006). Moreover, plural valuation can inform environmental governance by providing information about values that reflects local contexts, amplifying more marginalised voices on how to manage nature (Adger et al., 2003; Liu & Opdam, 2014).

This article builds on the theoretical justification for including environmental aspects in well-being assessments from Schleicher and colleagues (2018). In this article, we explore whether including an environmental dimension into an MPI is legitimate according to the people whose well-being is being measured. Specifically, we ask whether all qualities, including environmental components, are represented with respect to how they perceive their well-being. This is a step towards defining what has been referred to as a “democratically legitimate” (Boulanger et al., 2011) or “socially acceptable” (Dolan and Metcalfe 2012) account of well-being.

We posit that poverty is a reflection of a lack of well-being. Like poverty, we define well-being as multidimensional including traditional emphasis on what a person has (i.e., assets), as well as what they can do with what they have, and what they think and feel about what they can do (Gough & McGregor, 2007). Among the environmental components, we do not conceptualise the environment as a static, separable, or benevolent entity; we include positive and negative well-being-environment relationships and acknowledge that environmental change often has social drivers.

We focus on the Global MPI (Alkire & Foster, 2011), which is used by approximately 100 countries to assess progress towards eradicating poverty and, for some, progress towards achieving SDG1 (Sustainable Development Goal; https://mppn.org/sdgs-and-mpi/).1 Building on the capabilities approach (Nussbaum, 2011; Sen, 1992, 1993), the MPI is based on an “objective list” theory of well-being (Nussbaum, 1992), a theory that is pluralistic (Rice, 2013). The MPI measures acute poverty, a condition defined as the experience of multiple deprivations in three main dimensions: health, education, and living standards (Alkire & Santos, 2014). These deprivations are identified at a global level for the global MPI and adjusted to country-specific conditions for national MPIs. Although global well-being indicators allow for comparison across countries, they often lack the local specificity necessary for local legitimacy. Moreover, the MPI does not include linkages between poverty and the environment.

Before expanding the MPI with indicators to reflect such linkages, it is important to understand the nature of environment-well-being relationships. In this article, we evaluate whether environmental change drives changes in well-being (as a determinant with an instrumental role) or is an intrinsic or integral part of well-being (a constituent element; Dasgupta, 2001; Holland, 2014). The environment can provide goods and provides conditions that enable people to use their resources; these instrumental roles reflect determinants. The environment can also be integral to a well-being concept where harmony with nature (and nonhuman entities) is inseparable from human well-being, for the importance of harmony and the relationships in themselves, or for the freedom in terms of safety from harm (disease, death).

In this article, we focused on Rwanda and Malawi based on the authors’ prior experience in these countries as well as the interests of the UNEP-UNDP (United Nations Environment Programme and United Nations Development Programme) Poverty-Environment Initiative for their Africa Programme. Existing studies in Malawi and Rwanda have empirically evidenced links between well-being and nature. Ecosystems provide important crops (Aiga et al., 2009; Nsengimana et al., 2016), water for livestock and irrigation (Coulibaly et al., 2015), and trees for cultural ceremonies (Dawson & Martin, 2015; Von Maltitz et al., 2016). Negative impacts of protected area formation include crop raiding by wildlife and reduced food security (Munanura et al., 2018). Wider ecosystem degradation has been linked to social conflict over intensified competition for access to limited natural resources (Nabahungu & Visser, 2013) and risks to life and property (Msilimba & Holmes, 2009). Lack of ecosystem access has been linked to food insecurity (Kamat, 2014) and violence (German et al., 2014). We explore to what extent these links and other human-nature relationships are present and can be classified as determinants or constituents of well-being.

Methods

Country Background

Malawi is characterised by high levels of income-poverty, with 70% of the population living below the US$1.90/day poverty line. The country ranks 174 (of 189) on the HDI 2020, and 53% of the population is qualified as poor according to the Global MPI. The economy of Malawi is largely dependent on agriculture in terms of gross domestic product and even more so in terms of employment and informal economy. Over 80% of the 19 million population is rural. Agriculture is the main source of rural livelihoods and half of the rural households are subsistence farmers. The average cultivated area per household is about 0.6 ha (National Statistics Office [NSO] 2017). Malawi became independent from colonial rule in 1964.

Rwanda has the highest population density on mainland African with 89% of the labour force involved in agriculture (McMillan & Headey, 2014). The population is growing rapidly (National Institute of Statistics of Rwanda [NISR], 2007), with land scarcity being a palpable concern. The average household owns less than a hectare of land (NISR, 2010), with approximately a quarter of the population being practically landless (Jayne et al., 2003). Some have argued that land scarcity was a major factor in the 1990–1994 war and genocide that killed 10% of the population (Andre & Platteau, 1998). Despite these obstacles, Rwanda has experienced a fall in income-based poverty (from 2006 to 2011; NISR, 2012) and is considered by some to be a shining example of social and economic development (Crisafulli & Redmond, 2012). There are others, however, that hotly debate the validity of this decline (Booth & Golooba-Mutebi, 2012; Zorbas, 2011) and in particular, local relevance of the defining variables (Clay & Zimmerer, 2020; Ingelaere, 2010).

Site Selection

We focused on forested areas and wetland ecosystems based on the predominance of these ecosystems and their importance to peoples’ livelihoods (see table 1). Eight villages were selected: two villages in each habitat type and country based on consultation with local partners, including the Rwandan Environmental Management Authority, WorldFish (Malawi), and the UNEP-UNDP Poverty Environment Initiative (both countries). Villages were selected in areas identified as “high” and “low” quality. “Low”-quality areas had a higher proportion of environmental risks or degradation to that of “high”-quality areas.

Table 1.

Brief Description of Case Study Sites.

Location HabitatQuality
Malawi 
Mangochi Forest “High” recent establishment of a protected area 
Dedza Forest “Low” heavy deforestation with recent establishment of a newly planted village forest and strictly, state-protected forest reserve 
Salima Wetland “High” major flood led to evacuations and negative impacts on local livelihoods 
Zomba Wetland “Low” adjacent to Lake Chilwa where severe droughts have negatively impacted local livelihoods 
Rwanda 
Gishwati A Forest “High” households were closer to the National Park, some having been involuntarily shifted 
Gishwati B Forest “Low” further from the National park in a previously forested area 
Bugesera A Wetland “High” located closer to Lake Rweru with participants from a government-supported tomato cooperative with more secure livelihoods and irrigation infrastructure 
Bugesera B Wetland “Low” further from Lake Rweru and more broadly representative of the population including individuals living in extreme poverty, some from the Twa ethnic group or historically marginalised people 
Location HabitatQuality
Malawi 
Mangochi Forest “High” recent establishment of a protected area 
Dedza Forest “Low” heavy deforestation with recent establishment of a newly planted village forest and strictly, state-protected forest reserve 
Salima Wetland “High” major flood led to evacuations and negative impacts on local livelihoods 
Zomba Wetland “Low” adjacent to Lake Chilwa where severe droughts have negatively impacted local livelihoods 
Rwanda 
Gishwati A Forest “High” households were closer to the National Park, some having been involuntarily shifted 
Gishwati B Forest “Low” further from the National park in a previously forested area 
Bugesera A Wetland “High” located closer to Lake Rweru with participants from a government-supported tomato cooperative with more secure livelihoods and irrigation infrastructure 
Bugesera B Wetland “Low” further from Lake Rweru and more broadly representative of the population including individuals living in extreme poverty, some from the Twa ethnic group or historically marginalised people 

Focus Group Discussions (FGDs)

We used FGDs and iterative questioning to understand local conceptions of well-being and its connection to the environment. FGDs were selected based on logistical constraints and their ability to capture a diversity of views (Bryman, 2004). Two focal groups in each of the eight selected villages were completed, one all male and one all female of mixed sociodemographics to facilitate a greater understanding of potential differences in well-being due to gender, wealth, age, duration of residency in the village, and education level. Participants were selected through culturally appropriate norms and in conjunction with local government representatives in the respective countries and villages. We discussed the nature of our work and interest in capturing individuals across gender, wealth, age, and so on, prior to participant selection, and all respondents were asked for oral consent prior to participation. FGD involved four parts including (1) a free-listing exercise generating a list of locally relevant dimensions of well-being, (2) a matching exercise relating the dimensions listed under part 1 to a set of a priori defined well-being dimensions, (3) a discussion about the connections between the dimensions and the environment, and (4) a ranking exercise exploring the relative importance of each well-being dimension. All fieldwork was carried out over a period of 2 weeks in 2016. FGD with a single group was completed in approximately a half-day (3 h). Field assistants and native speakers of the local language (Kinyarwanda in Rwanda, Chichewa in Malawi) were involved from the inception of the project to pilot the translated protocol and ensure it was being interpreted as intended.

Following an ice breaker, FGD participants were prompted to free-list components of well-being through a series of questions. The first question, what do you need/what does it mean to live a good life here in this village? served to frame follow-up questions that challenged participants to differentiate well-being within and between villages (see Annex I for protocol).

FGD participants were then asked to match their components to an a priori list of nine dimensions (table 2) developed through a review of existing well-being frameworks (Alkire & Santos, 2010; Millennium Ecosystem Assessment, 2005; Nussbaum, 2011; Schleicher et al., 2018) as well as prior work in Rwanda (e.g., Gross-Camp et al., 2015; Martin et al., 2015) and Malawi (e.g., Schaafsma et al., 2018). If a free-listed well-being component did not fit into one of these predetermined areas, the creation of an additional dimension was allowed. We utilised pictograms to represent each dimension and assist participants in recalling what dimensions stood for. Pictograms were shown to country partners to determine cultural appropriateness.

Table 2.

Well-Being Dimensions Based on the Review of Literature.

Well-Being DimensionsDescription and ExamplesPictograma
Food and nutrition The ability to provide in your personal and your households food and nutritional needs throughout the year, including food that you buy, produce yourself, or collect in the area in and around your village  
Health (physical) Feeling strong and well, able-bodied, and your ability to maintain your health, for example, through acquiring medication or doctor assistance  
Education The ability to obtain the schooling you want personally, to send your children to school, including the required materials (e.g., books, uniforms, materials, fees)  
Living standards Shelter (adequate flooring, roofing and walls, sanitation, and electricity), motorbikes or bicycles, mobile phones, farming/fishing equipment, livestock, safe drinking water, and fuel  
Social relations Your ability to have meaningful relationships with your family and friends; to have family cohesion and respect within families, communities, and external actors; your ability to help or rely on others in times of need. This includes, for example, your ability to care for, raise, marry, and settle children, and to participate fully in society and social events such as celebrations, weddings, and festivities  
Security, safety from other people Safety and confidence in the future; peace and harmony—free from harm inflicted by other people, such as crime, mugging, physical violence (including rape), lack of protection from police, and lack of justice  
Living in safety from risk inflicted by nature, and in a clean, healthy environment Your ability to feel that your life is safe from droughts, floods, heatwaves, mudslides, storms, tsunamis, earthquakes, etc.
Your ability to live surrounded by clean water in rivers and lakes, breathe clean air, that is, live in a safe and healthy environment free from pollution
Your ability to live without suffering crop losses and killings (by elephants, hippos, lions, etc.) 
 
Relations with nature (nonconsumptive) Your ability to be part of nature, relate to your natural environment, visit and enjoy it, without doing damage to it/extracting resources. This includes your spiritual engagement with nature and ancestral lands  
Freedom of choice and action Your ability to live the life you want, with a sense of power to control and agency over your own life, according to your values and norms, being independent of the goodwill of others, the ability to choose and achieve your goals in life, and your ability to influence decisions that are made by others in your community and beyond that affect your life; to be empowered. And the freedom to conduct traditional, cultural, tribal, and religious practices and a life without discrimination (race, gender, etc.), including your livelihood such as a self-sustaining farmer/fisherman  
Well-Being DimensionsDescription and ExamplesPictograma
Food and nutrition The ability to provide in your personal and your households food and nutritional needs throughout the year, including food that you buy, produce yourself, or collect in the area in and around your village  
Health (physical) Feeling strong and well, able-bodied, and your ability to maintain your health, for example, through acquiring medication or doctor assistance  
Education The ability to obtain the schooling you want personally, to send your children to school, including the required materials (e.g., books, uniforms, materials, fees)  
Living standards Shelter (adequate flooring, roofing and walls, sanitation, and electricity), motorbikes or bicycles, mobile phones, farming/fishing equipment, livestock, safe drinking water, and fuel  
Social relations Your ability to have meaningful relationships with your family and friends; to have family cohesion and respect within families, communities, and external actors; your ability to help or rely on others in times of need. This includes, for example, your ability to care for, raise, marry, and settle children, and to participate fully in society and social events such as celebrations, weddings, and festivities  
Security, safety from other people Safety and confidence in the future; peace and harmony—free from harm inflicted by other people, such as crime, mugging, physical violence (including rape), lack of protection from police, and lack of justice  
Living in safety from risk inflicted by nature, and in a clean, healthy environment Your ability to feel that your life is safe from droughts, floods, heatwaves, mudslides, storms, tsunamis, earthquakes, etc.
Your ability to live surrounded by clean water in rivers and lakes, breathe clean air, that is, live in a safe and healthy environment free from pollution
Your ability to live without suffering crop losses and killings (by elephants, hippos, lions, etc.) 
 
Relations with nature (nonconsumptive) Your ability to be part of nature, relate to your natural environment, visit and enjoy it, without doing damage to it/extracting resources. This includes your spiritual engagement with nature and ancestral lands  
Freedom of choice and action Your ability to live the life you want, with a sense of power to control and agency over your own life, according to your values and norms, being independent of the goodwill of others, the ability to choose and achieve your goals in life, and your ability to influence decisions that are made by others in your community and beyond that affect your life; to be empowered. And the freedom to conduct traditional, cultural, tribal, and religious practices and a life without discrimination (race, gender, etc.), including your livelihood such as a self-sustaining farmer/fisherman  

aPictograms were added to assist in recall and an understanding of what the dimension stood for.

Next, participants were asked to explore each of the nine (or more) dimensions and their connections to the environment. We used prompts, such as Is the environment relevant to food and nutrition, or education, or security? If so, how? Participants were encouraged to provide specific examples wherever possible. The nature of these well-being-environment relationships and the reasons why nature was important, as stated by participants was used to infer whether the role of the environment was constituent or determinant for well-being. Although prolonged deliberations would have enabled participants to self-define the type of role, the required resources for such an approach were prohibitive for our study.

The final exercise involved ranking the well-being dimensions in the order of importance, performed first at the individual level and then as a group. Dimensions could be ranked in parallel, that is, two dimensions could be considered equally important. Rankings were used to propose weighting values for environmental aspects of well-being as additional dimensions to a poverty index. A series of nonparametric tests were conducted to compare ranking across different groups (i.e., by country and gender).

Questionnaire and MPI Calculation

Participants’ sociodemographic characteristics including MPI dimensions were collected (Annex II). This provided basic information about the participants and an indication of sample representativeness, comparing the MPI statistics of our sample populations to the national averages. The MPI, which is based on household-level data, was calculated using the Alkire and Santos (2011) components of the Global MPI. The Global MPI consists of three dimensions (education, health, and living standards) that are measured with two or more indicators (see also table 3). We omitted body metrics due to limited time; we used only child mortality for the health dimension. Our list of nine dimensions encompassed the MPI dimensions of education, health, and living standards.

Table 3.

Descriptive Statistics (Means) of Samples in Malawi and Rwanda.

SampleAge (Years)No. of People in HouseholdYears of Education of RespondentYears in VillagePlot Size (in ha)Total Assets (Q 21:8 Possible)
Malawi (N = 65) Mean 39 5.5 30 0.9 3.1 
Rwanda (N = 66) Mean 43 6.7 21 0.8 2.9 
SampleAge (Years)No. of People in HouseholdYears of Education of RespondentYears in VillagePlot Size (in ha)Total Assets (Q 21:8 Possible)
Malawi (N = 65) Mean 39 5.5 30 0.9 3.1 
Rwanda (N = 66) Mean 43 6.7 21 0.8 2.9 

We conducted 15 FGDs (6–9 individuals per group, N = 131 individuals), two per village, with the exception of one village in Rwanda where attendance of a cultural gathering the following day limited us to a single, mixed-gender discussion (figures 1 and 2).

Figure 1.

Female focus group discussion in Rwanda discusses how their free-listed well-being components connect to the environment.

Figure 1.

Female focus group discussion in Rwanda discusses how their free-listed well-being components connect to the environment.

Figure 2.

Male focus group discussion in Rwanda matches their free-listed well-being components to those of the nine a priori.

Figure 2.

Male focus group discussion in Rwanda matches their free-listed well-being components to those of the nine a priori.

In both countries, individuals similarly identified agriculture as the most important livelihood followed by casual labour and fisheries (table 3). Land and household size in our study villages were slightly different from national averages based on the most recently available statistics. In Malawi, plots were larger than the Southern Region (0.5 ha) and similar to those of the Central Region (0.8 ha), with slightly larger households (4.2–4.4 individuals per household on average; NSO, 2017). In Rwanda, plot sizes were higher than the 2008 national average of less than 0.7 ha for 60% of households (Republic of Rwanda, 2008). Similarly, the mean household size for our Rwandan site was higher than the national average, 6.7 and 4.3 individuals per household, respectively (NISR, 2016).

Our study villages scored similarly for most MPI dimensions, although Rwanda had a higher overall percentage of people described as multidimensionally poor (table 4). Fewer Malawi participants had no electricity or dirt flooring than their Rwandan counterparts, whereas Rwandan participants had more than double the rate of child mortality (consistent with NISR, 2016). These values are similar to the national averages in Rwanda (53.8%; Oxford Poverty and Human Development Initiative [OPHI], 2017b) but considerably lower in Malawi (53.7%; OPHI, 2017a).

Table 4.

Scores of Samples on Global Multidimensional Poverty Index (MPI) Dimensions.

DimensionIndicatorDeprived If…MalawiRwanda
Education Years of schooling No household member has completed 5 years of schooling 11% 11% 
Child school attendance Any school-aged child is not attending school up to class 8 3% 6% 
Health Child mortality Any child has died in the family 23% 50% 
Nutrition Any adult or child for whom there is nutritional information is malnourished Not calculated 
Living standards Electricity The household has no electricity 100% 71% 
Improved sanitation The household’s sanitation facility is not improved (according to MDG guidelines) or it is improved but shared with other households 95% 92% 
Safe drinking water The household does not have access to safe drinking water (according to MDG guidelines) or safe drinking water is more than a 30-min walk from home, round trip 54% 55% 
Flooring The household has a dirt, sand, or dung floor 94% 88% 
Cooking fuel The household cooks with dung, wood, or charcoal 100% 100% 
Assets The household does not own more than one radio, TV, telephone, bike, motorbike, or refrigerator and does not own a car or truck 60% 68% 
Total Deprivation score Weighted sum of deprivations in all dimensions (no deprivations = 0, fully deprived = 1 0.379 0.459 
 Multidimensional headcount ratio (H) Proportion of sample who are multidimensionally poor (in percentage of individuals) 34% 59% 
 Poverty intensity (A) The average deprivation score of all multidimensionally poor individuals 0.585 0.596 
 MPI MPI = H × A 0.199 0.271 
DimensionIndicatorDeprived If…MalawiRwanda
Education Years of schooling No household member has completed 5 years of schooling 11% 11% 
Child school attendance Any school-aged child is not attending school up to class 8 3% 6% 
Health Child mortality Any child has died in the family 23% 50% 
Nutrition Any adult or child for whom there is nutritional information is malnourished Not calculated 
Living standards Electricity The household has no electricity 100% 71% 
Improved sanitation The household’s sanitation facility is not improved (according to MDG guidelines) or it is improved but shared with other households 95% 92% 
Safe drinking water The household does not have access to safe drinking water (according to MDG guidelines) or safe drinking water is more than a 30-min walk from home, round trip 54% 55% 
Flooring The household has a dirt, sand, or dung floor 94% 88% 
Cooking fuel The household cooks with dung, wood, or charcoal 100% 100% 
Assets The household does not own more than one radio, TV, telephone, bike, motorbike, or refrigerator and does not own a car or truck 60% 68% 
Total Deprivation score Weighted sum of deprivations in all dimensions (no deprivations = 0, fully deprived = 1 0.379 0.459 
 Multidimensional headcount ratio (H) Proportion of sample who are multidimensionally poor (in percentage of individuals) 34% 59% 
 Poverty intensity (A) The average deprivation score of all multidimensionally poor individuals 0.585 0.596 
 MPI MPI = H × A 0.199 0.271 

Notes: Based on data collected in Rwanda and Malawi in 2017 (N = 65 in both countries). The Global MPI consists of three dimensions (education, health, and living standards) that each has a weight of 1/3. Indicators within each dimension have the same weight (i.e., 1/6 for education and health indicators and 1/18 for living standards indicators). MDG = Millennium Development Goal.

Results regarding the first and second parts of our FGD are included in Annex III.

Linking the Environment to Well-Being Dimensions

Participants commonly described direct and indirect well-being connections to the environment during part 3 of the FGDs. Here, we summarise the expressed perceptions of participants under the nine well-being dimensions highlighting similarities and differences between the two countries.

Food and Nutrition

Participants expressed that they had experienced a reduction in the quantity and quality of land and, subsequently, availability of certain foods and resources important for food production, that is, firewood. Rwandan communities described conditions of decreasing land availability, particularly for food production. Whereas previously children leaving home purchased or simply obtained new land by clearing it, government restrictions have resulted in families subdividing plots that have low productivity. In Malawi, having land was deemed similarly important and perceived to be increasingly limited due to population growth. Participants also described agricultural land to be increasingly less fertile, mandating an increase of chemical and animal-based inputs (both countries), placing a strain on women who are responsible for various farming tasks (Malawi only).

Participants recognised the contributions that “healthy” forests and water sources make to their food production and connected current experiences of food insecurity to a decline in ecosystem health and, subsequently, their well-being. In Malawi, many FGDs described deforestation as leading to lower rainfall. Participants mentioned that trees improve soil quality, provide soil cover, and reduce erosion and flooding. Forests also provide food, an important free resource for those without sufficient money. Rivers and lakes were identified as sources of fish as well as of irrigation water. One respondent stated, “We depend on irrigation agriculture and when there is no enough water for irrigation we don’t harvest enough.”

Finally, participants in both countries expressed concerns over an inability to predict the onset or intensity of the rains. These changes were perceived as particularly problematic for crop production because they led to extended dry periods followed by flooding.

Health

Participants connected physical health to natural phenomena that were sometimes listed under “living in a clean and safe environment.” For example, a reduction in water quality due to natural disasters was linked to gastrointestinal problems and waterborne diseases. Ecosystems (particularly, forests) were said to provide traditional medicine as well as clean water and air: “…which implies that there is a link between a human body and the environment.”

Education

Although education was certainly important to well-being across all participants, its connection to the environment was less direct and primarily instrumental. Participants in both countries described the importance of trees for the provision of shade, water, and fresh air for children studying in schools. Rwandan participants further included the use of timber for the construction of classroom tables and chairs. Only one participant highlighted that it is important to educate children about the forests and wildlife and preserve historical knowledge.

Living Standards

Material components, in particular a house and land, were viewed as key to well-being. Most often being constructed from locally extracted materials, houses were often directly associated with ecosystems.2 Land was associated with food production and the physical placement of a house. Some groups raised the issue of a lack of secure tenure rights and land access. In Malawi, respondents mentioned that deforestation had led to lower quality housing, as timber was replaced by thatch and reed, materials less able to withstand heavy rains. They also raised concerns over a reduction in living standards due to droughts and forest degradation resulting in longer times for water and firewood collection for women and scarcity of fuelwood.

Social Relations

Relationships with others, particularly the immediate family, were deemed very important: “a country without harmony, people cannot care for each other.” However, social relations were said to be strained by environmental stresses like drought and deforestation. In Malawi, participants described that a reduction in forest and firewood resulted in social tensions. Deforestation meant that forests could no longer provide sufficient forest income, which was linked to the emigration of men and subsequent divorces. Similarly, droughts resulted in greater intra- and interhousehold tensions, due to women having to travel further to search for water and spending more of their time retrieving water or neighbours with wells denying others access.

Living in Security and Safety From Other People

Environmental connections to well-being under this dimension pertained largely to reducing resources and land use regulations. In Malawi, the deforestation and forest degradation of village forests meant having to collect firewood in a strictly protected area located far from people’s homes. If caught by a park guard, one became subject to public humiliation, beatings, and sexual violence (women only). Participants also mentioned that reduced environmental quality had an indirect effect on safety. For example, when the environmental quality is poor, food security and income decrease, leading to increased theft of food and firewood from houses and fields.

Living in Safety From Environmental Risk and in a Clean, Healthy Environment

People readily recognised the effects of a clean and healthy environment on their well-being and described that crop raiding by wild animals, floods, and droughts is particularly challenging to their quality of life.

Crop Raiding

Crop raiding wildlife was a notable concern of all villages. Crop losses appeared pervasive, with significant time invested in reducing such loss. In Malawi, one forest-adjacent community reported crop losses of up to 66% and explained, “To control the wildlife from destroying our fields we spend days at the field to chase them away, sometimes even at night. This is done by both men and women.” In Rwanda, Gishwati communities have responded to crop raiding by planting Irish potatoes, which are less palatable to crop raiding animals. Participants suggested that this reduction in crop diversity might be a factor resulting in lower yields and exacerbated declines in soil fertility, leading to hunger. They indicated the use of chemical fertilizers in response to decreasing soil fertility. They perceived crop raiding to have increased since the establishment of the national park in 2015 and the exclusion of human activity in the park.

Floods and Droughts

In the water-adjacent villages, issues of droughts and floods were discussed. The risk of flooding affected participants psychologically causing considerable mental stress. Participants indicated that they spent time worrying about the potential of flooding and their feeling helpless to mitigate such. This was particularly acute in Malawi where one of the communities had been badly affected by floods in 2017, resulting in two casualties, loss of assets and crops, experiencing long evacuation times, and living in temporary camps where disease outbreaks occurred. In one forest-adjacent community in Malawi, participants reported that droughts had led to extreme food insecurity with fatalities.

In both countries, participants associated flooding events with postdrought periods when the land was unable to absorb rainfall. The rain moved down hillsides, destroying crops and sometimes homes. For example, in 2015 in Bugesera village, the rains were so severe that 18 homes were destroyed. One group in Malawi linked flood risk to a lack of arable land. Having enough land would reduce flood risk: Lack of land in lowland areas drives farmers to cultivate on slopes, increasing deforestation and thus flood risk. Similarly, respondents associated droughts and dry spells with deforestation, describing negative feedback: Dry spells led to dry farmland, lower yields, and higher forest extraction to compensate for income loss and lack of food due to these lower yields. Participants in the Salima village identified a relevant threshold for floods: When floods were only ankle-high, floods were perceived as positive or desirable events with fishermen arguing that this increases fish stocks.

In one Malawi village, drought was perceived to have caused a reduction in lake water levels, forcing farmers to seek alternative livelihoods. During our field study, we noticed that waters in Lake Chilwa had receded extremely and that fishing activities were non-existent. This was affecting not only on fishing households directly but also indirectly on farmers and traders who depend on fisher families to buy their produce. Rwandan communities indicated that changes in fish stock have at times resulted in short-term migration, particularly of men.

Relations With Nature

Participants described an enjoyment of nature through various activities and sensations. Being in nature, surrounded by forest, and breathing fresh air were a few of the descriptives used. In Malawi, swimming in Lake Chilwa was mentioned as a recreational activity that was impossible to do in years of drought, due to algae presence and water quality. In both countries, participants mentioned that forests make the landscape more attractive not only for local communities but also for tourists.

In Rwanda, nature was also seen as a backdrop for social gatherings including religious meetings, weddings, political, and cultural activities. Critically, people acknowledged these non-consumptive uses of nature as contributing to their well-being, indicating a sense of remorse at the thought of losing these connections.

In Malawi, responses on this topic needed probing. However, questions about initiation ceremonies demonstrated the impact that deforestation has had on people’s ability to conduct these activities and (gendered) differences in the availability of substitutes. “If we didn’t have the trees and the forest, it would be a problem to us because we wouldn’t have space for initiations as they require confidentiality.” Some participants indicated that their village had stopped performing initiation rites because there was no forest left in which to carry out the ceremonies; in other places, men had shifted to using forests reserved for burial or using grass shelters. Participants noted that while substitutions to alternative ceremonial grounds were possible for men, it was not an option for women as cultural taboos prevented them from accessing burial sites. They had either stopped holding the ceremonies or had moved them indoors.

Sense of place or place attachment was described by participants in at least one Malawi village. Settlers among the participants—“amtsatamadzi” (literally meaning “water followers”)—feel strongly attached to the land and lake. They said, “leaving fishing would mean trouble in our hearts and bodies; having no land and livelihoods is a hopeless situation.”

Freedom of Choice and Action

Participants from Gishwati District described considerable anxiety over the future availability of land to cultivate due to the recent establishment of a National Park and creation of a buffer zone.3 Participants were concerned about where they will be shifted and whether or not this will include land or only financial compensation. One woman described that she was intensively cultivating her land with food crops without concern for the future due to her lack of freedom of choice to live where she wanted to live and uncertainty of when she might be moved by the Rwandan government.

In the two forest-adjacent villages in Malawi, participants associated restrictions imposed on forest access with a loss of security and freedom of choice. In the other village, forest access restrictions had caused not only a decrease in forest incomes but also a loss of freedom of choice and agency over ancestral lands. Importantly, respondents reported a lack of political voice on the matter and a lack of information on new regulations.

Ranking Well-Being Dimensions

Participants indicated that all of the well-being components listed in the initial stage of FGDs were able to fit into the nine a priori categories. Participants ranked the nine well-being dimensions in the order of importance, first as individuals and then as a group, where one was the highest rank and nine was the lowest. In the individual ranking, all except three dimensions were significantly different by country. The dimensions social relations, relation to nature, and freedom of choice and action were viewed as having similar importance (table 5 and figure 3).

Table 5.

Mean and Median Individual Ranks by Country With Kruskal-Wallis Comparison.

OverallMalawiRwanda
DimensionMean (SD)MedianMeanMedianMeanMedianKruskal-Wallis
Food and nutrition 2.28 (1.57) 1.41 (0.63) 3.13 (1.75) 57.0*** 
Health (physical) 3.28 (2.27) 4.66 (1.90) 1.92 (1.74) 60.1*** 
Living standards 3.42 (2.16) 2.57 (1.91) 4.26 (2.08) 29.0*** 
Education 4.94 (2.04) 4.34 (2.09) 5.53 (1.82) 13.3*** 
Social relations 5.53 (2.03) 5.54 (1.96) 5.52 (2.11) 0.0 
Living in safety from other people 5.54 (2.27) 6.55 (1.68) 4.54 (2.34) 24.7*** 
Freedom of choice and action 5.70 (2.06) 5.86 (1.85) 5.55 (2.24) 0.3 
Living in a safe environment 6.89 (1.68) 6.55 (1.71) 7.23 (1.60) 5.5* 
Relations with nature 7.45 (2.00) 7.51 (1.99) 7.39 (2.04) 0.5 
OverallMalawiRwanda
DimensionMean (SD)MedianMeanMedianMeanMedianKruskal-Wallis
Food and nutrition 2.28 (1.57) 1.41 (0.63) 3.13 (1.75) 57.0*** 
Health (physical) 3.28 (2.27) 4.66 (1.90) 1.92 (1.74) 60.1*** 
Living standards 3.42 (2.16) 2.57 (1.91) 4.26 (2.08) 29.0*** 
Education 4.94 (2.04) 4.34 (2.09) 5.53 (1.82) 13.3*** 
Social relations 5.53 (2.03) 5.54 (1.96) 5.52 (2.11) 0.0 
Living in safety from other people 5.54 (2.27) 6.55 (1.68) 4.54 (2.34) 24.7*** 
Freedom of choice and action 5.70 (2.06) 5.86 (1.85) 5.55 (2.24) 0.3 
Living in a safe environment 6.89 (1.68) 6.55 (1.71) 7.23 (1.60) 5.5* 
Relations with nature 7.45 (2.00) 7.51 (1.99) 7.39 (2.04) 0.5 

Notes: Sample sizes: N = 65 in both countries (N = 130 in total). Figures in brackets represent the standard deviations of the means.

P levels * = 0.05, ** = 0.01, *** = <0.0001.

Figure 3.

Spider plots of group and individual means of ranks of different well-being dimensions within each country. (A) Malawi. (B) Rwanda. Note: Sample sizes: N = 65 individuals in both countries (N = 130 in total); eight groups in Malawi and seven groups in Rwanda.

Figure 3.

Spider plots of group and individual means of ranks of different well-being dimensions within each country. (A) Malawi. (B) Rwanda. Note: Sample sizes: N = 65 individuals in both countries (N = 130 in total); eight groups in Malawi and seven groups in Rwanda.

In the group ranking exercise, groups came to an agreement with relative ease, which is reflected in the comparison of the individual and group ranks (figure 3). In Rwanda, individuals ranked two dimensions, social relations and freedom of choice and action, higher than the group. However, our sample size was too small to test the statistical significance of these group differences.

Weights

From the ranking exercises, inferences were made about the relative importance of the environmental dimensions (table 6). When the respondents’ mean ranks were taken as weights of the dimensions of an MPI that included environmental dimensions, then the weights of the “living in a clean and safe environment” and “enjoying nature” were approximately 5% each, compared to 20% for food, 15% each for physical health and living standards, and 10% each for education, social relations, living in security, and freedom of choice.

Table 6.

Weights for Dimensions in Multidimensional Poverty Index (MPI).

DimensionMean RankMax-Mean RankaDifference/SumbSuggested WeightWeight in MPI
Food and nutrition 2.28 6.72 0.19 0.20 0.167 
Health (physical) 3.28 5.72 0.16 0.15 0.167 
Living standards 3.42 5.58 0.16 0.15 0.333 
Education 4.94 4.06 0.11 0.10 0.333 
Social relations 5.53 3.47 0.10 0.10 
Living in safety from other people 5.54 3.46 0.10 0.10 
Freedom of choice and action 5.70 3.30 0.09 0.10 
Living in a safe environment 6.89 2.11 0.06 0.05 
Relations with nature 7.45 1.55 0.04 0.05 
Sum 45.03 35.97  
DimensionMean RankMax-Mean RankaDifference/SumbSuggested WeightWeight in MPI
Food and nutrition 2.28 6.72 0.19 0.20 0.167 
Health (physical) 3.28 5.72 0.16 0.15 0.167 
Living standards 3.42 5.58 0.16 0.15 0.333 
Education 4.94 4.06 0.11 0.10 0.333 
Social relations 5.53 3.47 0.10 0.10 
Living in safety from other people 5.54 3.46 0.10 0.10 
Freedom of choice and action 5.70 3.30 0.09 0.10 
Living in a safe environment 6.89 2.11 0.06 0.05 
Relations with nature 7.45 1.55 0.04 0.05 
Sum 45.03 35.97  

Notes: Here, the mean individual ranks from table 5 are used. Finally, as the weights have to add up to 1, the suggested weights were rounded to multiples of 1/20. Results of the calculations of the suggested weights using means are similar to those using medians.

aIn order to convert these to weights that would sum to 1 (see Santos & Alkire, 2011), the difference between the maximum score (9) and the mean is calculated.

bThe difference is divided by the sum of differences.

Environment-Well-Being Relationships

The aim of our article was to explore which components of the environment can legitimately be included in an MPI. Our results showed that participants in different rural areas of two countries readily recognised and articulated the role of the environment to their well-being, including both its constituent and determinant role in all well-being dimensions and in various ways.

The most prominent environmental-human well-being dependencies expressed by our participants that are currently excluded from poverty metrics include the notion that (a) environmental degradation leads to a range of well-being changes, usually negative, (b) where degradation is severe or, in the case of natural hazards, psychological well-being is negatively affected, and (c) having safe and secure access to and control over environmental assets improves people’s freedom of choice.

Environmental degradation was often seen as a determinant of negative impacts on well-being for most well-being dimensions. For example, deforestation was linked to droughts that were associated with lower agricultural productivity. Other problems included floods, declining soil fertility, water quantity and quality, reduced land availability, and crop raiding by wild animals. When natural hazards or environmental degradation threatened the stability and provision of these ecosystem services and resources, participants often reported a decline in multidimensional well-being.

We posit that other relationships demonstrate the constituent role of nature in well-being. Beyond certain thresholds of severity, persistent environmental degradation and natural hazards become disastrous, leading to prolonged stress and psychological health effects that continue beyond the duration of the disaster event, as well as loss of lives. This stress and anxiety associated with the natural environment created an emotional unfreedom or incapability for people severely restricting their quality of life. Being safe from severe natural hazards and environmental degradation corresponds to the “security” constituents of well-being in the Millennium Ecosystem Assessment framework and is included in SDG1 (target 1.5) but not in poverty metrics like the MPI.

Restricting access and unclear property rights were also associated with a loss of freedom of choice, such as a loss of self-confidence, place attachment, and autonomy. We argue that this would call for considering these factors as a constituent dimension of well-being. The importance of access to the environment is acknowledged in target 1.4 of SDG1 and corresponds to “control over one’s environment,” one of Nussbaum’s 10 central capabilities. Rights and access to land could be included in an MPI as one of the multiple subdimensions of “freedom of choice and action.”

Despite participants’ recognition of the environment’s role in well-being, the existing global MPI dimensions are ranked highest, with the new environmental dimensions being ranked low. First, this suggests that the MPI captures the main deprivations. Even in communities where environmental risks were severe, rankings of “living in a safe environment” remained low. Second, our results provided evidence that the determinant role of the environment in well-being is important for nearly all dimensions. Inputs related to provisioning and regulating ecosystem services should clearly maintain a prominent feature in ecosystem services assessments. Third, our FGD participants did not rank cultural or relational values (Chan et al., 2018) of nature as a highly important constituent to well-being. This may reflect the fact that we sampled in areas where histories of (forced) migration and colonisation have affected human-environment relationships, including ways of working the land (Glasson et al., 2010). It may also be explained by the short duration of our study, limiting researchers’ ability to understand deeper notions of environment-well-being relationships, embodied in other expressions than FGD.

Our protocol provides a relatively low-resource tool to identify and characterise environment-well-being links that are locally relevant. We could not include a more quantitative assessment of environmental determinant and constituent factors in our survey. An initial next step would be to define indicators that are legitimate in terms of being socially acceptable as well as deemed relevant to those using and being affected by the index (Dolan & Metcalfe, 2012). We refrained from formulating indicators because of the context-specific character of well-being. A subsequent step for such quantification would be to develop methods to measure those indicators in a way that allows for their statistical integration into an extended MPI. Quantitative thresholds (poverty cutoffs, below which one is considered poor in an indicator) for the environmental indicators would need to be established. For example, one could make the distinction between ankle-high floods (which might be considered partly positive) and meter-high floods (which can be highly destructive and life-threatening).

We focused on extending the MPI with constituent factors related to the environment. We suggest that questionnaires provided the opportunity to assess individual- or household-level data for relevant indicators of constituent factors, including aspects of access or environment-related anxiety. Questionnaires may be used to establish thresholds and be preferred where existing data on access are not reliable or sufficiently detailed.

In addition, the evident relevance of the determinant role of the environment in well-being, in combination with the low ranking of the constituent factors, call for careful consideration of ways to reflect this determinant role in poverty profiles (Thiry et al., 2018). One option would be to analyse the impact of change in the natural environment on existing dimensions of the MPI: levels of food security and health, education, and living standards. If the direct dependence of well-being dimensions on nature were assessed, what-if scenarios could be explored: How much would poverty statistics change if those environmental inputs were no longer available? The increasing availability of remote sensing and other environmental data may provide opportunities for such quantitative analyses.

A related option would be to complement existing multidimensional poverty indices with environment-sensitive indicators. For example, the MPI health dimension could be expanded by including an indicator reflecting the proportion of food that is collected or cultivated or freedom from vector-borne diseases. Asset ownership could be expanded to reflect land tenure and other environmental resources. A challenge for achieving this is that land ownership, natural resources, and asset losses due to natural disasters often relate to public goods where ownership is either at community- or government-level, yet poverty statistics operate at the household- or individual level. Ground-truthing would be critical in ensuring accuracy when using existing data sets, especially for ownership statistics.

As an intermediate approach between a determinant and constituent factor, one could define “high-risk households” based on, for example, proximity to the crop-raiding wildlife habitat or exposure to floods/droughts. This risk factor could then be included as a single additional dimension, reflecting potential vulnerability. The advantages of such an approach are that secondary data to quantify such risk may already be available. To calculate poverty statistics adjusted for this high risk would require setting a weight and threshold risk level beyond which people are considered “poor.”

Limitations

Our study is presented with some limitations. Firstly, in focusing on a small number of villages, the study is comprehensive in neither ecosystem nor sociocultural diversity. Even within the two countries in our research, a wider diversity of human-nature relationships is likely to exist related to different worldviews and therefore to notions of well-being (Apgar et al., 2011). We sampled in areas subjected to a conflictual historical past with violent repercussions for cultural identities. There may have been more diversity within communities than we were able to capture with our FGDs. Nonetheless, discussions were consistent across villages and countries in participants’ recognition of multiple ways in which the environment contributes to well-being.

In our aim to include the environment in a globally defined and widely used poverty index, our study foregoes a critique of the definition and composition of the global MPI. We found that a number of other well-being dimensions besides health, living standards, and education (included in the global MPI) were perceived to be relevant components of well-being, in line with previous research (Dawson & Martin, 2015).

Existing dimensions of well-being and human-environment relationships may be conceptualised and defined differently when taking biocultural, indigenous, or decolonial approaches (Biedenweg & Gross-Camp, 2018; McLennon & Woods, 2018; Sterling et al., 2017). Alternative conceptualisations would have applied different approaches than those we used in our study. For example, they might not have used the FGD protocol and its elements such as the pictograms used in table 2.

Our results provide locally specific evidence for the determinant role of the natural environment to well-being, as well as the link between severe environmental degradation and natural hazards and psychological well-being—reflecting nature’s constituent role in well-being. Furthermore, in both countries, participants repeatedly emphasised the importance of having access to and control over environmental resources, forming part of their “freedom of choice” and well-being.

Our FGD protocol supported in-country assessment of the social legitimacy of including environmental dimensions in multidimensional poverty measures in terms of defining locally relevant dimensions. A next step would be to define indicators that are deemed legitimate by those using and affected by a new index, for example, through deliberative processes. This is a crucial step before embarking on the inclusion of environmental indicators in poverty measures in any country, as multidimensionally “poor” people are not the only potential beneficiaries but also the agents of change if the SDGs are to be achieved and sustainable management of ecosystems and biodiversity is to support that.

Including both the determinant and constituent aspects of nature in poverty metrics could help to reconcile objectives of poverty alleviation and ecologically sustainable ecosystem management or make trade-offs apparent. In this article, we have outlined a number of steps to further develop MPI assessments that are inclusive of the determinant and constituent role of the environment. Ideally, these steps would be undertaken in collaboration with NSOs to ensure relevance, ownership, and capacity building and with relevant ministries and departments for uptake in policy design and evaluation.

  1. What is the difference between constituent and determinant factors of well-being?

  2. What are the advantages and disadvantages of including environmental factors in well-being statistics?

  3. What are key connections between well-being and environment as found in this study?

MS & NGC: conceptualisation, methodology, data collection, analysis, manuscript writing and reviewing, and approval of final version.

We are very grateful for the profound support, input, and feedback from David Smith (PEI Africa), Michael Mmangisa and James Mbata (PEI Malawi), and Jan Rijpma, Janet Umugwaneza, and Fred Sabiti (PEI Rwanda). We would like to thank colleagues at UNEP-WCMC and the University of Cambridge, especially Dr. Judith Schleicher, for the various intellectual conversations and direct inputs to the project. This work would not have been possible without the professional assistance of Mr. Owen Makaka (Dept. of Economic Planning and Development, Malawi) and Mr. Chakhumbira Khaila (Research Assistant Malawi). We would also like to thank Mr. Asafu Chijere (WorldFish Malawi) for his help in sampling. We also acknowledge feedback from the participants representing relevant Ministries, Universities, and nongovernmental organisations during workshops held in both countries. Finally, the time and input from community members were crucial to this project. We hope that the results respect and reflect their perspectives as closely as possible and will be beneficial to them.

The authors declare no conflict of interest.

This project was funded by grant IAF-2017-18-002 of the Ecosystem Services for Poverty Alleviation (ESPA) programme. The ESPA programme is funded by the Department for International Development, the Economic and Social Research Council, and the Natural Environment Research Council. Additional funding was received from the UNEP-UNDP Poverty-Environment Initiative Malawi.

Annex I Focus group discussion and weighing exercise (.pdf)

Annex II Socio-demographic survey (.pdf)

Annex III Linking bottom-up well-being indicators to a priori dimensions (.pdf)

1.

We focus on the Multidimensional Poverty Index as a baseline concept without being prescriptive: Other multidimensional poverty indicators based on objective list theory would require similar considerations as discussed in this article.

2.

In Rwanda, a government programme replaced thatch roofing with metal sheets; this programme was a response to reduce the ecological pressure of thatch collection and also houses’ susceptibility to collapse due to saturation.

3.

The Park is a strict conservation area, and upon its formation in 2016, several households were physically shifted. Other households are in the buffer zone, where human residence is permitted, but natural resource collection and other livelihood activities are restricted. The buffer zone is expected to affect more than 2,000 households who will (eventually) only be able to cultivate timber on their lands in joint management agreements with the government. Additional households in government-defined ecologically sensitive areas will be moved.

Adams
,
W. M.
,
Aveling
,
R.
,
Brockington
,
D.
,
Dickson
,
B.
,
Elliott
,
J.
,
Hutton
,
J.
,
Roe
,
D.
,
Vira
,
B.
, &
Wolmer
,
W.
(
2004
).
Biodiversity conservation and the eradication of poverty
.
Science
,
306
(
5699
),
1146
1149
.
Adger
,
W. N.
,
Brown
,
K.
,
Fairbrass
,
J.
,
Jordan
,
A.
,
Paavola
,
J.
,
Rosendo
,
S.
, &
Seyfang
,
G.
(
2003
).
Governance for sustainability: Towards a “thick” analysis of environmental decision making
.
Environment and Planning A
,
35
(
6
),
1095
1110
.
Adjei
,
P. O.-W
,
Buor
,
D.
, &
Addrah
,
P.
(
2017
).
Ecological health effects of rural livelihood and poverty reduction strategies in the Lake Bosomtwe Basin of Ghana
.
GeoJournal
,
82
(
3
),
609
25
.
Aiga
,
H.
,
Matsuoka
,
S.
,
Kuroiwa
,
C.
,
Yamamoto
,
S.
,
Chak
,
S.
,
Daun Penh
,
K.
, &
Penh
,
P.
(
2009
).
Malnutrition among children in rural Malawian fish-farming households
.
Transactions of the Royal Society of Tropical Medicine and Hygiene
,
103
,
827
833
.
Alkire
,
S.
, &
Foster
,
J.
(
2011
).
Counting and multidimensional poverty measurement
.
Journal of Public Economics
,
95
(
7–8
),
476
487
.
Alkire
,
S.
, &
Santos
,
M. E.
(
2010
).
Acute multidimensional poverty: A new index for developing countries
.
Working Paper No. 38
,
Oxford Poverty & Human Development Initiative
,
Oxford, United Kingdom
. http://doi.org/10.2139/ssrn.1815243
Alkire
,
S.
, &
Santos
,
M. E.
(
2011
).
Training material for producing national human development reports. MPI: Construction and analysis
.
Oxford Poverty and Human Development Initiative
.
Alkire
,
S.
, &
Santos
,
M. E.
(
2014
).
Measuring acute poverty in the developing world: Robustness and scope of the multidimensional poverty index
.
World Development
,
59
,
251
274
.
André
,
C.
, &
Platteau
,
J. P.
(
1998
).
Land relations under unbearable stress: Rwanda caught in the Malthusian trap
.
Journal of Economic Behavior & Organization
,
34
(
1
),
1
47
.
Angelsen
,
A.
,
Jagger
,
P.
,
Babigumira
,
R.
,
Belcher
,
B.
,
Hogarth
,
N. J.
,
Bauch
,
S.
,
Börner
,
J.
,
Smith-Hall
,
C.
, &
Wunder
,
S.
(
2014
).
Environmental income and rural livelihoods: A global-comparative analysis
.
World Development
,
64
,
S12
S28
.
Apgar
,
J. M
,
Ataria
,
J. M.
, &
Allen
,
W. J.
(
2011
).
Managing beyond designations: Supporting endogenous processes for nurturing biocultural development
.
International Journal of Heritage Studies
,
17
(
6
),
555
570
.
Arias-Arévalo
,
P.
,
Gómez-Baggethun
,
E.
,
Martín-López
,
B.
, &
Pérez-Rincón
,
M.
(
2018
).
Widening the evaluative space for ecosystem services: A taxonomy of plural values and valuation methods
.
Environmental Values
,
27
(
1
),
29
53
.
Barrett
,
C. B.
, &
Bevis
,
L. E.
(
2015
).
The self-reinforcing feedback between low soil fertility and chronic poverty
.
Nature Geoscience
,
8
(
12
),
907
912
.
Beckerman
,
W.
, &
Pasek
,
J.
(
1997
).
Plural values and environmental valuation
.
Environmental Values
,
6
(
1
),
65
86
.
Biedenweg
,
K.
, &
Gross-Camp
,
N.
(
2018
).
A brave new world: Integrating well-being and conservation
.
Ecology and Society
,
23
(
2
),
32
.
Booth
,
D.
, &
Golooba-Mutebi
,
F.
(
2012
).
Developmental patrimonialism? The case of Rwanda
.
African Affairs
,
111
(
444
),
379
403
.
Boulanger
,
P. M.
,
Bauler
,
T.
,
Van Ootegem
,
L.
,
Lefin
,
A.-L.
,
Ruwet
,
C.
,
Gérard
,
A.
, &
Spillemaekers
,
S.
Towards theoretically sound and democratically legitimate indicators of wellbeing for Belgium.
Final Report
.
Brussels
:
Belgian Science Policy
2011
121
p. (
Research Programme Science for a Sustainable Development
).
Bryman
,
A.
(
2004
).
Social research methods
.
Oxford University Press
.
Chan
,
K. M.
,
Gould
,
R. K.
, &
Pascual
,
U.
(
2018
).
Editorial overview: Relational values: What are they, and what’s the fuss about?
Current Opinion in Environmental Sustainability
,
35
,
A1
A7
.
Clay
,
N.
, &
Zimmerer
,
K. S.
(
2020
).
Who is resilient in Africa’s green revolution? Sustainable intensification and climate smart agriculture in Rwanda
.
Land Use Policy
,
97
,
104558
.
Coulibaly
,
J. Y.
,
Gbetibouo
,
G. A.
,
Kundhlande
,
G.
,
Sileshi
,
G. W.
, &
Beedy
,
T. L.
(
2015
).
Responding to crop failure: Understanding farmers’ coping strategies in Southern Malawi
.
Sustainability
,
7
(
2
),
1620
1636
.
Crisafulli
,
P.
, &
Redmond
,
A.
(
2012
).
Rwanda, Inc.: How a devastated nation became an economic model for the developing world
.
Palgrave Macmillan
.
Dasgupta
,
P.
(
2001
).
Constituents and determinants of wellbeing. Human wellbeing and the natural environment
.
Oxford University Press
.
Dawson
,
N.
, &
Martin
,
A.
(
2015
).
Assessing the contribution of ecosystem services to human wellbeing: A disaggregated study in western Rwanda
.
Ecological Economics
,
117
,
62
72
.
Dolan
,
P.
, &
Metcalfe
,
R.
(
2012
).
Measuring subjective wellbeing: Recommendations on measures for use by national governments
.
Journal of Social Policy
,
41
(
2
),
409
427
.
German
,
L.
,
Mandondo
,
A.
,
Paumgarten
,
F.
, &
Mwitwa
,
J.
(
2014
).
Shifting rights, property and authority in the forest frontier: “Stakes” for local land users and citizens
.
Journal of Peasant Studies
,
41
(
1
),
51
78
.
Glasson, G. E.,
Mhango
,
N.
,
Phiri
,
A.
, &
Lanier
,
M.
(
2010
).
Sustainability science education in Africa: Negotiating indigenous ways of living with nature in the third space
.
International Journal of Science Education
,
32
(
1
),
125
141
.
Gough
,
I.
, &
McGregor
,
J. A.
(
2007
).
Wellbeing in developing countries: From theory to research
.
Cambridge University Press
.
Gray
,
L. C.
, &
Moseley
,
W. G.
(
2005
).
A geographical perspective on poverty–environment interactions
.
Geographical Journal
,
171
(
1
),
9
23
.
Gross-Camp
,
N. D.
,
Martin
,
A.
,
McGuire
,
S.
, &
Kebede
,
B.
(
2015
).
The privatization of the Nyungwe National Park buffer zone and implications for adjacent communities
.
Society & Natural Resources
,
28
(
3
),
296
311
.
Gross-Camp
,
N. D.
,
Rodriguez
,
I.
,
Martin
,
A.
,
Inturias
,
M.
, &
Massao
,
G.
(
2019
).
The type of land we want: Exploring the limits of community forestry in Tanzania and Bolivia
.
Sustainability
,
11
(
6
),
1643
.
Holland
,
B.
(
2014
).
Allocating the Earth: A distributional framework for protecting capabilities in environmental law and policy
.
Oxford University Press
.
Ingelaere
,
B.
(
2010
).
Peasants, power and ethnicity: A bottom-up perspective on Rwanda’s political transition
.
African Affairs
,
109
(
435
),
273
292
.
Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services
. (
2019
). Summary for policymakers of the global assessment report on biodiversity and ecosystem services of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services. In
S. M.
Díaz
,
J.
Settele
,
E.
Brondízio
,
H.
Ngo
,
M.
Guèze
,
J.
Agard
,
A.
Arneth
,
P.
Balvanera
,
K. A.
Brauman
,
S. H. M.
Butchart
,
K. M. A.
Chan
,
L. A.
Garibaldi
,
K.
Ichii
,
J.
Liu
,
S. M.
Subramanian
,
G. F.
Midgley
,
P.
Miloslavich
,
Z.
Molnár
,
D.
Obura
,
A.
Pfaff
,
S.
Polasky
,
A.
Purvis
,
J.
Razzaque
,
B.
Reyers
,
R.
Roy Chowdhury
,
Y. J.
Shin
,
I. J.
Visseren-Hamakers
,
K. J.
Willis
, &
C.
Zayas
(Eds.),
IPBES secretariat
(p.
56
).
Bonn
.
Jayne
,
T. S.
,
Yamano
,
T.
,
Weber
,
M. T.
,
Tschirley
,
D.
,
Benfica
,
R.
,
Chapoto
,
A.
, &
Zulu
,
B.
(
2003
).
Smallholder income and land distribution in Africa: Implications for poverty reduction strategies
.
Food Policy
,
28
(
3
),
253
275
.
Kamat
,
V. R.
(
2014
).
The ocean is our farm: Marine conservation, food insecurity, and social suffering in Southeastern Tanzania
.
Human Organization
,
73
(
3
),
289
298
.
Kangalawe
,
R. Y.
, &
Noe
,
C.
(
2012
).
Biodiversity conservation and poverty alleviation in Namtumbo District, Tanzania
.
Agriculture, Ecosystems & Environment
,
162
,
90
100
.
Li
,
C.
,
Yang
,
W.
,
Tang
,
Q.
,
Tang
,
X.
,
Lei
,
J.
,
Wu
,
M.
, &
Qiu
,
S.
(
2020
).
Detection of multidimensional poverty using Luojia 1-01 nighttime light imagery
.
Journal of the Indian Society of Remote Sensing
,
48
(
7
),
963
977
.
Liu
,
J.
, &
Opdam
,
P.
(
2014
).
Valuing ecosystem services in community-based landscape planning: Introducing a wellbeing-based approach
.
Landscape Ecology
,
29
(
8
),
1347
1360
.
Liu
,
L.
(
2012
).
Environmental poverty, a decomposed environmental Kuznets curve, and alternatives: Sustainability lessons from China
.
Ecological Economics
,
73
,
86
92
.
Martin
,
A.
,
Akol
,
A.
, &
N.
Gross-Camp
. (
2015
).
Towards an explicit justice framing of the social impacts of conservation
.
Conservation and Society
,
13
(
2
),
166
178
.
McLennan
,
V.
, &
Woods
,
G.
(
2018
).
Learning from mistakes and moving forward in intercultural research with Aboriginal and Torres Strait Islander peoples
.
Higher Education Research & Development
,
37
(
1
),
88
100
.
McMillan
,
M.
, &
Headey
,
D.
(
2014
).
Introduction–understanding structural transformation in Africa
.
World Development
,
63
,
1
10
.
Millennium Ecosystem Assessment
. (
2005
).
Ecosystems and human well-being: Synthesis
.
Island Press
.
Msilimba
,
G.
, &
Holmes
,
P.
(
2009
).
Landslides in Northern and Central Malawi: Awareness, perceptions and coping strategies
.
South African Geographical Journal
,
91
(
1
),
38
45
.
Munanura
,
I. E.
,
Backman
,
K. F.
,
Hallo
,
J. C.
,
Powell
,
R. B.
, &
Sabuhoro
,
E.
(
2018
).
Understanding the relationship between livelihood constraints of poor forest-adjacent residents, and illegal forest use, at Volcanoes National Park, Rwanda
.
Conservation and Society
,
16
(
3
),
3291
3304
.
Nabahungu
,
N. L.
, &
Visser
,
S. M.
(
2013
).
Farmers’ knowledge and perception of agricultural wetland in Rwanda
.
Land Degradation & Development
,
71
,
363
374
.
National Institute of Statistics of Rwanda
. (
2007
).
National population projection 2007–2022
(p.
123
).
National Institute of Statistics of Rwanda
.
National Institute of Statistics of Rwanda
. (
2010
).
National agricultural survey 2008
(p.
249
).
National Institute of Statistics of Rwanda
.
National Institute of Statistics of Rwanda
. (
2012
).
The evolution of poverty in Rwanda from 2000 to 2011: Results from the household surveys
(p.
40
).
National Institute of Statistics of Rwanda
.
National Institute of Statistics of Rwanda
. (
2016
,
November
).
Statistical yearbook
(
SYB2016 ed.
). https://www.statistics.gov.rw/publication/statistical-yearbook-2016
National Statistics Office
. (
2017
).
Integrated household survey 2016–2017. Household socio-economic characteristics report
.
Republic of Malawi
,
November 2017
,
Zomba, Malawi
.
Nayak
,
P.
,
Oliveira
,
L.
, &
Berkes
,
F.
(
2014
).
Resource degradation, marginalization, and poverty in small-scale fisheries: Threats to social-ecological resilience in India and Brazil
.
Ecology and Society
,
19
(
2
),
73
.
Nsengimana
,
V.
,
Weihler
,
S.
, &
Kaplin
,
B. A.
(
2016
).
Perceptions of local people on the use of Nyabarongo River Wetland and its conservation in Rwanda
.
Society & Natural Resources
,
30
(
1
),
3
15
.
Nussbaum
,
M. C.
(
1992
).
Human functioning and social justice in defence of Aristotelian essentialism
.
Political Theory
,
20
(
2
),
202
246
.
Nussbaum
,
M. C.
(
2011
).
Creating capabilities
.
Harvard University Press
.
Oxford Poverty and Human Development Initiative
. (
2017
a).
“Malawi country briefing,” multidimensional poverty index data bank
.
Oxford Poverty and Human Development Initiative, University of Oxford
. http://ophi.org.uk/multidimensional-poverty-index/mpi-2015/mpi-country-briefings/
Oxford Poverty and Human Development Initiative
. (
2017
b).
“Rwanda Country Briefing,” Multidimensional Poverty Index Data Bank
.
Oxford Poverty and Human Development Initiative, University of Oxford
.
Republic of Rwanda
. (
2008
). Strategic plan for the transformation of agriculture in Rwanda—Phase II (PSTA II).
Final Report
.
Ministry of Agriculture and Animal Resources, Republic of Rwanda (ROR)
,
Kigali
(p.
22
).
Rice
,
C. M.
(
2013
).
Defending the objective list theory of well-being
.
Ratio
,
26
(
2
),
196
211
.
Santos
,
M. E.
, &
Alkire
,
S.
(
2011
).
Training material for producing national human development reports. MPI: Construction and analysis
.
Oxford Poverty and Human Development Initiative
.
Satterfield
,
T.
,
Gregory
,
R.
,
Klain
,
S.
,
Roberts
,
M.
, &
Chan
,
K. M.
(
2013
).
Culture, intangibles and metrics in environmental management
.
Journal of Environmental Management
,
117
,
103
114
.
Satterthwaite
,
D.
(
2003
).
The links between poverty and the environment in urban areas of Africa, Asia, and Latin America
.
The Annals of the American Academy of Political and Social Science
,
590
(
1
),
73
92
.
Schaafsma
,
M.
,
Utila
,
H.
, &
Hirons
,
M. A.
(
2018
).
Understanding trade-offs in upscaling and integrating climate-smart agriculture and sustainable river basin management in Malawi
.
Environmental Science & Policy
,
80
,
117
124
.
Schleicher
,
J.
,
Schaafsma
,
M.
,
Burgess
,
N. D.
,
Sandbrook
,
C.
,
Danks
,
F.
,
Cowie
,
C.
, &
Vira
,
B.
(
2018
).
Poorer without It? The neglected role of the natural environment in poverty and Wellbeing
.
Sustainable Development
,
26
(
1
),
83
98
.
Schnegg
,
M.
,
Rieprich
,
R.
, &
Pröpper
,
M.
(
2014
).
Culture, nature, and the valuation of ecosystem services in Northern Namibia
.
Ecology and Society
,
19
(
4
). https://www.jstor.org/stable/26269669
Schreckenberg
,
K.
,
Franks
,
P.
,
Martin
,
A.
, &
Lang
,
B.
(
2016
).
Unpacking equity for protected area conservation
.
Parks
,
22
(
2
),
11
26
.
Sen
,
A.
(
1992
).
Inequality re-examined
.
Clarendon Press
.
Sen
,
A.
(
1993
). Capability and wellbeing. In
M.
Nussbaum
&
A.
Sen
(Eds.),
The quality of life
(pp.
30
53
).
Clarendon Press
.
Sterling
,
E. J.
,
Filardi
,
C.
,
Toomey
,
A.
,
Sigouin
,
A.
,
Betley
,
E.
,
Gazit
,
N.
,
Newell
,
J.
,
Albert
,
S.
,
Alvira
,
D.
,
Bergamini
,
N.
, &
Blair
,
M.
(
2017
).
Biocultural approaches to well-being and sustainability indicators across scales
.
Nature Ecology & Evolution
,
1
(
12
),
1798
1806
.
Thiry
,
G.
,
Alkire
,
S.
, &
Schleicher
,
J.
(
2018
). Incorporating environmental and natural resources within analyses of multidimensional poverty.
OPHI Research in Progress 50a, University of Oxford
.
Tipa
,
G.
, &
Teirney
,
L. D.
(
2003
). A Cultural Health Index for streams and waterways: Indicators for recognising and expressing Maori values.
58
(
Ministry for the Environment, 2003
).
Trainor
,
S. F.
(
2006
).
Realms of value: Conflicting natural resource values and incommensurability
.
Environmental Values
,
15
,
3
29
.
Von Maltitz
,
G. P.
,
Gasparatos
,
A.
,
Fabricius
,
C.
,
Morris
,
A.
, &
Willis
,
K. J.
(
2016
).
Jatropha cultivation in Malawi and Mozambique impact on ecosystem services, local human wellbeing, and poverty alleviation
.
Ecology and Society
,
21
(
3
),
3
.
Zorbas
,
E.
(
2011
). Aid dependence and policy independence: Explaining the Rwandan paradox. In
S.
Straus
&
L.
Waldorf
(Eds.),
Remaking Rwanda: State building and human rights after mass violence
(pp.
103
-
117
).
Sage
.

Supplementary data