
==== Front
Environ Res Commun
Environ Res Commun
erc
Environmental Research Communications
2515-7620
IOP Publishing

ercad76ff
10.1088/2515-7620/ad76ff
ad76ff
ERC-101662.R3
Letter
Policy, Socioeconomics, and Law
From consumption to context: assessing poverty and inequality across diverse socio-ecological systems in Ghana
https://orcid.org/0000-0003-4966-8397
Cavanaugh Alicia C 1alicia.cavanaugh@mcgill.ca

https://orcid.org/0000-0002-0513-5292
Bixby Honor R 2honor.bixby@essex.ac.uk

https://orcid.org/0009-0005-5436-471X
Mangwani Saeesh 3saeesh.mangwani@imbrsea.eu

https://orcid.org/0000-0002-6418-393X
Agyei-Mensah Samuel 4sagyei-mensah@ug.edu.gh

https://orcid.org/0000-0001-7346-535X
Awuni Cynthia Azochiman 4cawuni@gmail.com

https://orcid.org/0000-0001-8397-6380
Baumgartner Jill C 56jill.baumgartner@mcgill.ca

https://orcid.org/0000-0003-3859-5540
Owusu George 78gowusu@ug.edu.gh

https://orcid.org/0000-0002-8972-8318
Robinson Brian E 1∗brian.e.robinson@mcgill.ca

1 Department of Geography, McGill University, Montreal, Canada
2 Institute of Public Health and Wellbeing, University of Essex, Colchester, United Kingdom
3 IMBERSea, University of Ghent, Ghent, Belgium
4 Department of Geography and Resource Development, University of Ghana, Accra, Ghana
5 Department of Ethics, Equity and Policy, McGill University, Montreal, Canada
6 Department of Epidemiology, Biostatistics & Occupational Health, McGill University, Montreal, Canada
7 Institute of Statistical, Social & Economic Research (ISSER), University of Ghana, Accra, Ghana
8 Centre for Migration Studies, University of Ghana, Accra, Ghana
9 Author to whom any correspondence should be addressed.
1 9 2024
17 9 2024
6 9 09100912 10 2023
30 3 2024
03 9 2024
11 12 2023
17 9 2024
© 2024 The Author(s). Published by IOP Publishing Ltd
2024
https://creativecommons.org/licenses/by/4.0/ Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

Abstract

Local social and ecological contexts influence the experience of poverty and inequality in a number of ways that include shaping livelihood opportunities and determining the available infrastructure, services and environmental resources, as well as people’s capacity to use them. The metrics used to define poverty and inequality function to guide local and international development policy but how these interact with the local ecological contexts is not well explored. We use a social-ecological systems (SES) lens to empirically examine how context relates to various measures of human well-being at a national scale in Ghana. Using a novel dataset constructed from the 100% Ghanian Census, we examine poverty and inequality at a fine population level across and within multiple dimensions of well-being. First, we describe how well-being varies within different Ghanian SES contexts. Second, we ask whether monetary consumption acts a good indicator for well-being across these contexts. Third, we examine measures of inequality in various metrics across SES types. We find consumption distributions differ across SES types and are markedly distinct from regional distributions based on political boundaries. Rates of improved well-being are positively correlated with consumption levels in all SES types, but correlations are weaker in less-developed contexts like, rangelands and wildlands. Finally, while consumption inequality is quite consistent across SES types, inequality in other measures of living standards (housing, water, sanitation, etc) increases dramatically in SES types as population density and infrastructural development decreases. We advocate that SES types should be recognized as distinct contexts in which actions to mitigate poverty and inequality should better incorporate the challenges unique to each.

social-ecological systems
multiple dimensions of poverty
multiple dimensions of inequality
sub-Saharan Africa
Wellcome Trust 10.13039/100010269 209376/Z/17/Z ccc2515-7620/24/091009+14$33.00
crossmarkyes
==== Body
pmc1. Introduction

The experience of poverty and inequality are relative. How it feels to be poor depends on how different you are from the people around you and your ability to fully participate in society [1, 2]. In many cases, how one participates in society must be understood within local socio-cultural and ecological contexts [3]. Well-being—which we refer to as quality of life and the ability to meaningfully participate in society—depends on local infrastructure, economic context, climatic conditions, culture, and ways in which populations access and use environmental resources [4, 5]. Absolute poverty lines identify the level of income needed to achieve basic needs across all societies [6]. While the World Bank’s absolute poverty line of $2.15 a day aids in international comparison, it is less useful for understanding poverty in a particular place, as what constitutes minimum standards of living and nutritional needs can be highly location-specific [7]. Context affects the types and diversity of livelihood opportunities available, as well as the ability or need to use local resources [8, 9]. A social-ecological systems (SES) lens can help us understand and mitigate poverty and inequality, especially as the need for climate adaptation grows [10–13]. Thus, we use relative poverty to understand how it feels to be poor and how living standards differ in various SES contexts.

Consider, for example, the livelihood options and constraints of a financial consultant in Dakar, Senegal versus a subsistence groundnut farmer in eastern Mali that is only weakly integrated into the economic market system [14]. These two livelihood contexts have different opportunities and constraints, in large part due to interactions and dependencies in local environments. Yet we typically assess well-being in these two cases in the same way, namely through measures of consumption or income [15]. While monetary-based metrics enable cross-context comparisons, they provide little insight into different mechanisms for addressing poverty or livelihood issues at the local level.

Metrics used to define poverty and inequality are foundational for development policy. How these interact with the local ecological contexts are not well explored yet are critical for understanding local human well-being and strategies for community support [16]. For example, in some rural contexts, such as pastoral systems, livestock are often considered a better proxy for wealth compared to income or consumption [17–19]. Complications with consumption-based metrics have given rise to widespread adoption of multidimensional and capability approaches to measuring poverty, inequality, and well-being [20–30]. While multidimensional and locally-relevant indictors are being recognized [4, 11], there are no empirical and systematic examinations of how diverse SES context relate to human well-being, especially at a national scale.

In this paper, we use a multidimensional poverty and inequality approach to evaluate the appropriateness of consumption (monetary expenditures on goods and services used by households) as a proxy for human well-being within SES contexts in Ghana. To do so, we use a novel and unprecedentedly detailed dataset to explore differences in how a variety of well-being metrics relate to consumption among five anthropogenic biomes (‘anthromes’) [31]. Anthromes are a land system classification that incorporate the ecological context of anthropogenic systems [32] and provide a resource-based grounding for considering how poverty and inequality manifest in a particular location. This paper contributes to the literature in three ways. We (i) provide granular evidence of the distribution of demographic and housing indicators related to quality of life, (ii) provide evidence that populations are different (and should be considered differently) in various SES contexts, and (iii) show how SES context can guide policy priorities for poverty and inequality reduction.

2. Ghanaian context

Ghana has one of the fastest growing economies in Sub Saharan Africa [33, 34] yet in 2016 over 23% of the population still remained below the World Bank defined poverty line [35]. While there is a steady reduction in national poverty it is often a more rural phenomenon and is highly concentrated in the North [36]. Inequality has also steadily increased [36] as the benefits of growth have favored certain places over others [37]. Northern Ghana has seen the greatest rates of poverty reduction over the past two decades, though development in the rural north still lags behind the more urbanized southern regions [36, 38, 39]. Inequalities are increasing nationally due to disparities within districts and regions, rather than inequality between these areas [36, 38, 40].

Current patterns of poverty and inequality in Ghana have colonial roots related to investments directed toward resource-rich regions that had goods for the export market, ideal climates for cash crops (e.g. cocoa, coffee, rubber), and access to coastal trade [41]. Investments in infrastructure, hydroelectric projects, services, and housing were concentrated in the south [42]. Southern regions were connected by modernized transportation networks to increase agricultural and mineral exports [41] which spurred growth in the Accra, Kumasi, and Sekondi-Takoradi triangle [43]. Formerly a 19th century trading and food production center, economic activity in Northern Ghana slowed as a result of Southern bias in colonial expenditure [44, 45]. Northern Ghana’s arid climate and distance from the coast meant it received less investment, instead becoming an effective labor reserve for colonial production in the south [44]. However, the northern cities of Tamale, Wa, and Bolgatanga have experienced more recent steady growth [46, 47].

Ghana is also ecologically diverse, with distinct agroecological zones characterized by variations in climate, vegetation, and soil types, resulting in a wide variety of habitats and ecosystems [48]. Northern Ghana is relatively dry and characterized by savannah made up of open grasslands, scattered trees, and shrubs. Guinean savannah and a transitional zone occupy the middle part of the country, characterized by greater density of wooded areas and a diversity of plant and animal species. The transitional zone is a mix of forest, savannah, and grasslands habitats, lying between the savannah and more forested ecosystems to the south. The coastal regions of Ghana feature the coastal savannah to the east, including Accra, and coastal forest to the west. Southern inland Ghana around the Volta Basin is home to moist semi-deciduous forest [49]. This combination of dynamic and changing economic conditions across a range of ecological contexts makes Ghana an excellent candidate for exploring dynamics between well-being and SES contexts.

3. Data and methods

To assess how living standards vary across SES contexts, we pool multiple data sources available for Ghana. A challenge is synthesizing SES and measures of well-being at the same spatial scale. Below we describe methods for classifying SES types across Ghana, estimating consumption in small geographic units, defining measures of living standards (as proxies for household well-being), and assessing inequality.

3.1. Developing social-ecological system types

While some literature examining poverty and inequality has examined urban and rural differences [36–41], deeper examination of ecological context has been rare. Yet, especially from a development perspective, ‘rural’ is far too coarse to meaningfully capture characteristics of a social-ecological context to provide concrete insight for how living conditions or livelihood dynamics interact with poverty and inequality [50]. Agricultural households are quite market-oriented and integrated, while residents in largely undeveloped ‘wild’ lands may depend on local environmental resources for most of their daily needs. Poverty may be better understood when considering the systematically different resource and livelihood realities of people in, say, urban versus agricultural versus pastoral versus undeveloped contexts.

Here we use the major categories developed by the Anthromes project [31] to classify different social-ecological system (SES) types. While there is still a great deal of heterogeneity in these classes, it gives us a better sense of the ecological context in which a community dwells. The major anthrome classes [31] that exist in Ghana are: urban areas, mixed settlements & villages, croplands, rangelands, and wildlands. Urban can comprise business districts, slums, planned residential neighborhoods, and peri-urban areas. Villages and settlements (hereafter referred to as just settlements) are areas with mid-levels of population density and are a mix of more developed and rural environments which can also include towns, hamlets, and denser agricultural settlements. Croplands are areas with annual crops mixed with other land uses and land covers. Rangelands are areas dominated by livestock grazing, with few crops and forests. Wildlands are places with very limited development and low population densities but can represent areas with national parks as well as subsistence communities that rely on forest products.

The Anthromes database provides a global gridded classification at 5-arc minute resolution, resulting in some anthrome heterogeneity within districts. To account for variation between urban and rural contexts, we (a) masked out built-up areas [51] and (b) assigned each district a single population-weighted [51] anthrome (SES) class. EAs outside of urban areas were assigned the SES where the most people live in their district. We describe these two steps below.

3.1.1. Identifying built-up areas

To identify the spatial extent of urban areas, we used the ESA Worldcover dataset [51]. ‘Built-up area’ indicates the spatial extent of urban infrastructure at a 10 m resolution. We used modal aggregation to and aggregate urban pixels at a 250 m2 resolution and identify large built-up areas in Ghana. The result was smoothed and polygonized to create contiguous units (figure 1(a)). Identified built-up areas aligned well with urban areas in the anthromes dataset, but also included smaller areas identified in the Worldcover data.

Figure 1 Classifying Social-Environmental System types. We (a) identify built-up areas, (b) overlay original anthrome land use classifications, and (c) assign districts a non-urban population-weighted anthromes type, resulting in (d) final resulting district-level social-ecological system types.

3.1.2. Identifying district-level SES types using population-weights

Our Ghanian census data is comprised of 10 regions subdivided into 216 districts and 36,593 enumeration areas (EAs)—the smallest census administrative units in Ghana with an average population of ∼650 people. Shapefiles are only available for districts, but EAs are always labeled as rural or urban. We classified districts containing solely urban EAs (as identified in the Census) as ‘urban’ districts. For all other districts, we assigned an SES type based on the environment where most people live. We applied gridded population data (from the GPWv4 dataset [52]) to the masked anthrome data and assigned districts SES types based on the most common population-weighted anthrome pixels outside of urban areas in that district. In these districts, census-identified EAs were considered urban, and EAs identified as rural were assigned the district SES type. Figure 1 summarizes the steps taken to assign households in Ghana to one of five SES types and the number of districts and EAs in each category.

3.2. Developing consumption estimates for all households in Ghana

Since consumption data are not available in the census, we estimated consumption for 100% of households enumerated in the 2010 Ghanian Population and Housing Census [53], in collaboration with the Ghana Statistical Service (GSS), using small area estimation (SAE) methods ubiquitous in the poverty mapping literature [54, 55]. SAE borrows strength from a detailed, but less representative dataset to then predict an outcome for a representative but less detailed dataset. In our context, we first estimate predictors of equivalized consumption (log consumption in Ghana cedis (GH ) divided by the square root of the household size) using the 6th Ghana Living Standards Survey [56] in a linear mixed model that includes district-level random intercepts (SI table 1). Independent variables were selected from household characteristics measured in both the GLSS6 survey and the census using a LASSO regression model that assessed the top 20–25 most relevant correlates with GLSS consumption (SI table 1). Following procedures established at the GSS to account for regional-level heterogeneities, we estimated separate models for each of the 10 regions [57]. Then we applied the parameters and district-level random intercepts to predict consumption for Census households, and summarized poverty and inequality metrics at the EA level. Cavanaugh et al [58] provide more details on these procedures, but our dataset for analysis has EAs as the units of observation, with consumption and well-being metrics summarized at that level. We present results on relative poverty using consumption deciles based on the nationally-defined and regionally-defined distribution of consumption for two reasons. First, absolute poverty lines are set based on the ability to meet food and non-food needs at a national scale, creating a tendency to underestimate urban poverty [59–61]. Second, since Ghana’s absolute levels of poverty have recently risen, we want to understand the ability of those at the bottom to meet their basic needs relative to their neighbors. Deciles based on the national level show the inequalities throughout the country, while regionally-based deciles help account for variable living standards and aspects of relative poverty across regions.

While a sample of the latest Ghanian 2021 census was recently released, our current approach utilizes the full 100% sample of the 2010 census and the spatial distribution of the data, neither of which are available for analysis for the 2021 Census. The consumption estimates used in this study have also been used to examine spatial inequalities in air pollution [62], noise levels [63], child mortality [64], and drinking water [65] for the city of Accra.

3.3. Developing households well-being metrics

Well-being is a complex construct for which we do not have a singular ‘gold-standard’ measure [66, 67]. Accordingly, we use multiple indices that relate to well-being. Using data from the 100% Ghanian Census data, we chose variables that relate to education/access to information and household living standards in order to characterize two of the three dimensions of the Multidimensional Poverty Index [22] and the Human Development Index [68], as shown in table 1. We use the household head’s education attainment, mobile phone ownership, and internet access as measures of education and access to information as they can have a positive effect on the use of health services [69]. While this may be a relatively high standard as an average for the whole population, it is a reasonable comparative measure for the head of the household. Water sources, lighting, sanitation, and fuel sources are related to defense mechanisms that protect habitants from pollutants, disease, and environmental and social risks [70]. We categorize each household as having ‘improved’ versus ‘unimproved’ levels of these variables following UN Sustainable Development Goal guidelines and WHO/UNICEF’s Joint Monitoring Programme nomenclature [71–73].

Table 1. Household improved/unimproved living standards definitions.

	Improved	Unimproved	
Education & access to information	
Education	Obtained high school degree	Not graduated high school	
Mobile Phone	Owns a mobile phone	Does not own a mobile phone	
Internet Access	Accesses the internet	Does not access the internet	
Living standards	
Water	Pipe-borne inside dwelling	Dugout/Pond/Lake/Dam/Canal	
	Pipe-borne outside dwelling	Other	
	Bottled water	Rainwater	
	Sachet water	River/Stream	
	Bore-hole/Pump/Tube well	Tanker supply/Vendor provided	
	Protected spring	Unprotected spring	
	Protected well	Unprotected well	
	Public tap/Standpipe		
Lighting	Electricity (mains)	Candle	
	Electricity (private generator)	Crop residue	
	Solar energy	Firewood	
		Flashlight/Torch	
		Gas lamp	
		Kerosene lamp	
		Other	
Sanitation	KVIP	Bucket/Pan	
	Pit latrine	No facilities (bush/beach/field)	
	W.C.	Other	
		Public toilet (WC, KVIP, Pit, Pan, etc)	
Fuel	Electricity	Animal waste	
	Gas	Charcoal	
		Crop residue	
		Kerosene	
		None, no cooking	
		Other	
		Saw dust	
		Wood	

3.4. Assessing consumption versus well-being measures

We assess how ‘appropriate’ consumption appears to be as a measure of well-being and inequality, in terms of its agreement with other metrics within distinct SES classes. We judge this appropriateness of consumption in two ways. First, we examine the pair-wise joint distribution of our well-being indicators by deciles of consumption using a multidimensional inequality approach that captures the differences between distributions but also the correlations between well-being and consumption [26, 27]. To compare to other well-being measures, deciles of consumption are defined based on regional distributions to reflect potential local differences in consumption. Each household is assigned to a decile group based on their position in the distribution, with the first decile representing the 10% of households with the lowest levels of consumption, and 10th decile representing the greatest. A common relative poverty line is defined as 60% of the median consumption, which here aligns well with the 30th percentile threshold [74, 75]. Importantly, none of the well-being measures (section 3.3) were used as predictors of consumption in the SAE estimation procedure (section 3.2). Thus, the well-being measures and our estimates of consumption are statistically independent and valid for comparison [74]. While even broad overlap in variables (e.g., a well-being measure constructed in part from a variable included in estimation) could result in statistical circularity, all our well-being measures are broad aggregates of the percent of households in an EA with an ‘improved’ measure, thus is unlikely to be related to any of the ‘raw’ data used in the initial SAE procedure.

Second, we examine inequalities in consumption and other measures of well-being by calculating Gini coefficients to compare EA-level distributions for each indicator within each SES type [27, 29, 30]. Gini values range from perfect equality (zero) to perfect inequality (one). Gini coefficients are generally categorized as: < 0.3 is low inequality, 0.31–0.40 is medium inequality, 0.41–0.50 is high inequality, and > 0.5 is extremely high [75–77]. Our Gini coefficients are based on the percent (%) of households within an EA that have ‘improved’ measures (table 1) or, for consumption, EA-level average consumption measured in GH . Statistical offices do not generally report Gini values on EA-level summary statistics, but this gives us a common comparative metric by which to evaluation inequalities within SES types.

Since consumption is measured as continuous in Ghana cedis (GH ), we can apply additional procedures to estimate the distribution of household-level consumption with each SES type to better allow for comparison with commonly reported estimates of poverty and inequality by national statistics offices. To estimate household-level inequality, however, we first note that Gini coefficients are sensitive to the skewness of underlying distribution of the data. Our raw modeled consumption estimates do not capture long tails in the true empirical distribution (predictions from models do not predict outliers or long tails of distributions by design). To allow our data to match the true empirical distribution of income, we applied ‘mean-constrained integration over brackets’ methods [78] to our estimates of EA-level average consumption. This effectively proportionally rescales our modeled estimates to match the range, variance, and non-parametric shape of the true empirical distribution as measured in the GLSS6. We then use this estimated household distribution to estimate a household-level Gini that represents consumption inequality.

4. Results

4.1. Consumption distributions differ across SES types

Figure 2 shows how the national distribution of consumption varies across regions (figure 2(a)), which is the way we might typically and intuitively think of poverty and inequality in a country. Rural and poor areas have low levels of consumption that skew toward lower consumption deciles, such as the regions of Northern, Upper East, and Upper West. These distributions are colored by the SES types represented within, and we see those regions have far fewer urban inhabitants with much greater representation of croplands (yellow), rangelands (light green), and settlements (orange).

Figure 2. National distribution of consumption across regions and SES types. Count of households in each consumption group (defined by nationally-determined deciles) by administrative regions (A) and socio-environmental systems (B). National consumption deciles, ordered from left to right, from poorest to most affluent.

The distribution of consumption by SES type (figure 2(b)) reveals very different patterns. Over SES types, the positive (urban) and negative (settlements, croplands, rangelands, and wildlands) relationships between population numbers and consumption are striking. Despite significant variability in living environments within settlements and urban systems, they have the highest number of households in the lower 30th percentile of the consumption distribution, i.e., living in relative poverty (585,132 and 524,140, respectively; figure 2(b)). While the poor make up 44% of households in the settlements category, only 17% are poor in urban systems. In croplands, rangelands, and wildlands land, over 50% of households fall below the relative poverty threshold. In rangelands, over 25% of households fall into the bottom consumption decile. While levels of poverty and extreme poverty tend to increase as the intensity of infrastructure decreases, consumption-based poverty levels are slightly lower in wildlands than rangelands.

On the upper end, over 87% of households in the top three deciles are in urban systems. Affluent households make up a larger percentage of the population in more developed systems. About 43% of the urban population are among the highest 30% of consumers, while 15% are in the highest 10%. However, using a national distribution to compare living standards would exclude many poor households in higher cost of living areas like Accra and Kumasi. Thus, in the following sections we use regional distributions to help adjust for differences in relative poverty across locations.

4.2. Consumption deciles and measures of well-being

Figure 3 shows the distribution of each well-being indicator by regionally defined deciles of consumption within SES types. We use regionally defined deciles to account for differences in poverty that may be felt differently across different regions. SES types are ordered from left to right by intensity of land use. Urban areas have the highest proportion of households with improved measures of well-being across all indicators (on average, 48.5%), especially in access to improved water and lighting (93% and 80%, respectively). This demonstrates a distinct material advantage of living in urban areas over the next most developed SES (settlements at 35%). Rates of improved well-being metrics are lowest in rangeland (on average 20%) and wildland (on average 20.9%) areas, showing very different general standards of living in these areas. These findings hold regardless of where the urban areas are located. Disaggregating urban areas based on the SES type of their surrounding district, shows that well-being is highest when the built environment is surrounded by urban districts and decreases in less developed districts. However, urban areas in more rural SES are better off than their rural counterparts, suggesting urban processes that affect well-being are common throughout the country (SI figure 1).

Figure 3. Measures of well-being. Box plots show the median and interquartile range of EA-level rates of improved well-being measures for households by consumption decile and SES type. Consumption deciles were defined for each region to reflect differences in relative poverty. Average rates of improved status (μ®x) for each row describe the overall rate of improvement for that SES type. Percentages for each cell show the average improvement rate across consumption deciles for that SES type. Consumption deciles are represented from lowest to highest, represented by shaded and tinted hues, respectively.

Across all well-being indicators, the proportion of households with improved metrics are positively correlated with consumption levels in all SES types. Well-being measures and consumption are significantly positively associated in most cases. The strongest correlations with consumption for all well-being metrics are in urban areas (r®u = 0.38), with settlements and croplands showing similar patterns (r®s = 0.33 and r®c = 0.30). Rangelands and wildland areas, however, show weaker associations across the board (r®r = 0.19 and r®w = 0.29) and are notably different than the more market-integrated SES types (urban, settlement, and cropland areas) in improved outcomes for internet, sanitation, and fuel. For example, in rangelands, the correlation between consumption and sanitation is only 0.11, and only marginally stronger in wildlands at 0.26. The strongest associations in rangelands and wildlands are with rates of mobile phone ownership, as is the case across all SES types, but improved water sources are weakly negatively correlated with consumption in wildlands. Overall, rangelands, and in some cases wildlands, show weaker and more inconsistent relationships with living standard outcomes than the more market-integrated SES types.

4.3. Inequality in consumption and well-being measures across SES types

Figure 4 shows Gini coefficients for consumption and well-being measures for each SES type. We have included two measures of consumption inequality: (1) EA-level consumption inequality, and (2) household-level consumption inequality denoted by *. EA-level consumption inequalities are low, around 0.20 for all SES types. Household consumption inequalities are higher (greater than 0.30), reaching medium levels of inequality. Regardless of measure, consumption inequalities are relatively similar across SES types. Other measures of inequality are not only typically much higher, but they are also far more varied between SES types. Only in urban and village systems are there inequalities lower than consumption (i.e. water and mobile in both, lighting in urban SES), indicating relatively equal access to these services and goods in more developed areas. Unlike consumption, inequality in most other measures of well-being qualifies as high or extremely high in less developed SES types. In settlements, five of eight metrics are high. In croplands and rangelands this is the case for six out of eight, and in wildlands seven out of eight measures. Typically, fuel, internet, lighting, and sanitation inequality are the highest, followed by mobiles, water, and consumption inequality. In wildland areas, only consumption inequality is considered low. Ultimately, places that are more ecologically dependent and less developed, have the highest levels of inequality for almost all well-being measures.

Figure 4. Inequality across SES types. Gini coefficients for consumption and well-being measures for each SES type. The color of the symbols corresponds to a well-being metric. Solid circles represent the Gini with respect to EA-level percent of households with improved metrics or average consumption. The hollow circle (*) represents the Gini calculated from estimates of household-level consumption.

5. Discussion

In developing evidence from microdata linked to the enumeration area-level, we are able to explore inequalities in Ghana with a spatial lens. Not only do we demonstrate that levels of well-being differ between SES contexts, but the spatial granularity of our data also allows us to show that places within a particular SES experience vast differences in material well-being. We first discuss our findings as they relate to well-being across SES types, then inequalities within SES, and finally overall development policy. We make suggestions on how to tailor policy to different SES contexts.

5.1. Well-being across SES

Our first general finding is that our measures of well-being are, for most of the population, positively related to consumption. In the three most developed systems (urban, settlements, and croplands), there is a strong association between consumption and improved well-being. The association with consumption is notably weaker or inconsistent in rangelands and wildlands.

Education and living standards measures have similar relationships with consumption in urban areas, settlements, and cropland systems. Livelihoods in these three SES types are largely market-oriented, albeit with different types of economic activity, levels of infrastructure, and market integration. Rangelands and wildlands, in contrast, have smaller populations and environmental and economic activities are often less market-oriented. While some rangeland production is commercial, much is for local or individual consumption, especially in northern Ghana [79, 80]. Rangelands have the lowest average level of consumption of any SES type, with particularly low levels for some well-being metrics, even among the highest consumption deciles. Physical access to goods and services may be limited in pastoral-based systems, or preferences are simply different. As is well documented, rangeland and pastoral systems often prioritize different sets of material goods and patterns of consumption, and store wealth in ‘productive assets’ (i.e., livestock) [19, 81]. Our data support the idea that to understand well-being in rangelands, one must consider metrics beyond consumption and standard measures of well-being.

Wildland systems are the least populous and mostly forested. The northern-most district is part of the Guinean savannah, which is dominated by short trees, grass, and scrubland. The two central wildland regions are located in the mixed transitional zone, while the southern-most district is characterized by wet evergreen forest. Large portions of these districts are state protected national parks (i.e. Mole, Digya, and Bui National Parks & Ankasa Conservation Area). Local communities are often highly dependent on the local environment, and access to resources within the parks is limited, and thus can have significant impacts on local livelihoods [82]. Wildland SES have the lowest levels of infrastructure development, but do not always have the lowest levels of well-being.

In summary, consumption is a fair proxy for well-being in more developed SES types. However, in SES types that are not as integrated into the market economy and more ecologically embedded, different internal logics govern strategies to ensure quality of life and meet one’s basic needs such that consumption is a weaker indicator of well-being.

5.2. Inequalities across SES

Our data show that consumption inequality is a poor predictor of inequality across a broad range of well-being metrics. Consumption inequality is low, and varies little, across all SES types, in contrast to inequalities in other well-being metrics that increase with decreasing intensity of development. In fact, consumption inequality is highest in the urban SES, the opposite of all other well-being measures, where above we note that consumption tracks most closely to well-being.

Thus, we see there are wide differences in inequality across our well-being metrics. Consumption and income are justified as a global proxy for well-being based on the idea that fungible wealth can be traded for goods or materials that improve quality of life [83]. Yet we see here that while consumption may be a broad indicator of well-being in most systems, it does not provide a good measure of the material inequalities that can exist, especially for SES that are more ecologically embedded.

High inequalities in living standards, but low inequalities in consumption, could also relate to higher consumption of goods or materials that were not measured in the standard set of goods in the census. This may be the case, especially in rangeland and wildland systems where livelihood strategies and thus consumption patterns can be quite different compared to the more market-integrated urban, settlement, and cropland systems.

With respect to both poverty and inequality metrics, future work could extend this analysis to include comparable data from the 2021 Census (when spatially resolved micro data become available). This will allow for a more dynamic picture of how poverty and inequality have changed over time, and how patterns may relate to different social-ecological processes.

5.3. Policy implications

Our findings highlight how development and poverty alleviation programs should recognize both social and ecological contexts to better incorporate the challenges unique to local environments. For instance, in urban areas, where there are large numbers of poor but better conditions than elsewhere, it may be preferable to focus on improving specific material dimensions where the poor lag behind (e.g., fuel, education, ICT). However, basing decisions on a rural-urban divide would be misleading. Our evidence points more closely to differences between what we see as more market-oriented (urban, settlement, cropland) versus more environmentally dependent (rangeland, wildland) SES types. In more market-oriented rural SES of settlements and cropland areas, patterns in well-being and inequality are more similar to urban areas than rangeland and wildland rural SES. Market-oriented areas may be better served by focusing on improving conditions for the poor while improving market access in these areas. In more ecologically oriented SES types, where more households struggle to meet basic needs, development efforts could aim to improve basic services in a locally- and contextually-relevant ways. For example, policies that work in settled and agricultural communities often do not reflect the ecological or social needs of pastoral communities [84–86]. Those inhabiting ‘wildlands’ are also supported by a range of livelihoods—from indigenous forest-dwelling communities, to tourism, to park and wildlife management—that may necessitate different policy approaches to help support local communities and reduce inequality [87–89]. Many of these challenges involve complex issues around land rights [90], political and power dynamics [91, 92], and private versus public benefits from land [93, 94]. Co-production of viable policies and continuous engagement with local communities is necessary for understand development needs, wants, and modes of delivery [95–97].

While co-produced solutions will likely lead to the best outcomes, we have some ideas about policy interventions that might work in various areas [96]. For example, unconditional cash transfer programs can improve health and educational outcomes [98] but are likely more impactful for households in market integrated systems (urban, settlements, croplands) than rangelands and wildland residents. Considering place and SES when designing programs to improve access to infrastructure and services, could help policy have an impact beyond the poorest and raise well-being for the whole population.

In SES such as croplands, rangelands, and wildlands, where households are more dependent on farming, pastoral, or subsistence livelihoods, assisting in developing resilient local food systems may be relevant [99]. Land-dependent communities are more susceptible to the impacts of climate change, which places pastoralists and farmers at risk of deeper deprivation, making it important to secure household livelihoods and improve food security. Co-producing possible policy mechanisms with communities will best lead to robust solutions. One flexible initiative in Ghana is the Ghana School Feeding Programme, implemented at the school-level, which has a long-term goal of reducing poverty and malnutrition through strengthening local food production and consumption systems [100]. Sourcing foods that are locally produced for school meal programming not only improves nutrition for school aged children, but also provides a local market for small producers.

Households with limited financial resources must optimize livelihoods and their own well-being, often constrained by the local ecological or available environmental resources, services, and infrastructure. A standardized set of living standard measures may not appropriately capture these conditions, nor the preferences or cultural norms of local populations particularly in rangelands and wildland contexts. These considerations could be addressed by working with local communities to understand what they view as quality of life/well-being metrics through co-produced or participatory poverty and wealth assessments. Recent literature provides guidance on developing local indicators of well-being [101–103], including the Basic Necessities Survey [104]. These provide promise in developing locally produced reflections of community needs, but further work is needed to see how well these measures might empirically correlate with consumption or income. A paradox seems to exist in developing indicators that are simultaneously locally salient and relevant but also comparable over time and across communities.

6. Conclusion

Socio-ecological systems shape the experience of poverty and inequality. This adds nuance to our understanding of inequality between (and within) urban and rural contexts, and allows us to identify the challenges that households face in different socio-ecological contexts across Ghana. Here, we find that SES types that are more highly integrated into the market economy have higher rates of consumption (in monetary terms) and lower levels of material inequality. Unsurprisingly, more rural and less developed socio-ecological systems face the deepest deprivation along more dimensions. We find, on the one hand, that consumption is a relatively good proxy for levels of well-being, however, it is only weakly associated in less developed contexts. On the other hand, monetary-based inequality is a poor predictor of well-being inequality. Thus, it is appropriate to consider how context might shape well-being and evaluate how well measures of economic capacity, like consumption, relate to material and living standards, and inequality, in interrogating and designing policies to reduce deprivation. Using microdata from Ghana to evaluate multiple dimensions of deprivation allows us to see that in the least developed contexts there are many indicators where poor and even middle-income households are deprived, and only a small number of households have access to improved conditions. In urban areas, the poor are deprived among particular dimensions, while higher consumption groups are better able to meet their basic needs. Our approach provides evidence on where particular concerns overshadow others, which issues are most acute, and who lacks access to basic services in Ghana. Applying this framework to other ecologically diverse contexts can help policy makers understand the most pressing challenges and how to develop programs with communities to help meet their needs.

Data availability statement

The data that support the findings of this study are openly available at the following URL/DOI: https://doi.org/10.5281/zenodo.10669077.
==== Refs
References

1 Townsend P 1979 Poverty in the United Kingdom: A Survey of Household Resources and Standards of Living University of California Press
2 Sen A 1983 Poor, relatively speaking Oxford Economic Papers 35 153 169 153–69 10.1093/oxfordjournals.oep.a041587
3 Garroway C De Laiglesia J R 2012 On the Relevance of Relative Poverty for Developing Countries OECD Development Centre Working Papers No. 314 ( 10.1787/5k92n2x6pts3-en )
4 King M F Renó V F Novo E M 2014 The concept, dimensions and methods of assessment of human well-being within a socioecological context: a literature review Social indicators research 116 681 698 681–98 10.1007/s11205-013-0320-0
5 WHO 2021 Health Promotion Glossary of Terms 2021 World Health Organization (https://iris.who.int/bitstream/handle/10665/350161/9789240038349-eng.pdf)
6 Decerf B 2023 Absolute and relative income poverty measurement: a survey Research Handbook on Poverty and Inequality Edward Elgar Publishing 36 51 pp 36–51
7 Kakwani N 2003 Issues in Setting Absolute Poverty Lines. Poverty and Social Development Papers No. 3, Asian Development Bank
8 Scoones I 1998 Sustainable rural livelihoods: a framework for analysis University of Sussex Institute of Development Studies Working Paper 92
9 Chambers R Conway G 1992 Sustainable Rural Livelihoods: Practical Concepts for the 21st Century Institute of Development Studies (UK)
10 Carpenter S R 2012 Program on ecosystem change and society: an international research strategy for integrated social–ecological systems Current Opinion in Environmental Sustainability 4 134 138 134–8 10.1016/j.cosust.2012.01.001
11 Hamann M 2018 Inequality and the biosphere Annual Review of Environment and Resources 43 61 83 61–83 10.1146/annurev-environ-102017-025949
12 Berkes F Colding J Folke C 2000 Rediscovery of traditional ecological knowledge as adaptive management Ecological Applications 10 1251 1262 1251–62 10.1890/1051-0761(2000)010[1251:ROTEKA]2.0.CO;2
13 Olsson P Folke C Berkes F 2004 Adaptive comanagement for building resilience in social–ecological systems Environmental Management 34 75 90 75–90 10.1007/s00267-003-0101-7 15383875
14 Majola N G Gerrano A S Shimelis H 2021 Bambara groundnut (Vigna subterranea [L.] Verdc.) production, utilisation and genetic improvement in Sub-saharan Africa Agronomy 11 1345 10.3390/agronomy11071345
15 McGregor T Smith B Wills S 2019 Measuring inequality Oxford Review of Economic Policy 35 368 395 368–95 10.1093/oxrep/grz015
16 Sterling E J 2020 Creating a space for place and multidimensional well-being: lessons learned from localizing the SDGs Sustainability Sci. 15 1129 1147 1129–47 10.1007/s11625-020-00822-w
17 Doran M Low A R Kemp R L 1979 Cattle as a store of wealth in Swaziland: implications for livestock development and overgrazing in Eastern and Southern Africa American Journal of Agricultural Economics 61 41 47 41–7 10.2307/1239498
18 Turner M D 1999 Merging local and regional analyses of land-use change: the case of livestock in the sahel Annals of the Association of American Geographers 89 192 219 192–219 10.1111/1467-8306.00142
19 Nkedianye D K 2019 Livestock-wealth inequalities and uptake of crop cultivation among the maasai of Kenya and Tanzania World Development Perspectives 14 100106 10.1016/j.wdp.2019.02.017
20 Mayer A L 2008 Strengths and weaknesses of common sustainability indices for multidimensional systems Environ. Int. 34 277 291 277–91 10.1016/j.envint.2007.09.004 17949813
21 Lamb W F 2014 Transitions in pathways of human development and carbon emissions Environ. Res. Lett. 9 014011 10.1088/1748-9326/9/1/014011
22 Initiative O P H D 2018 Global Multidimensional Poverty Index 2018: The most Detailed Picture to Date of the World’s Poorest People University of Oxford
23 Ciommi M 2017 A new class of composite indicators for measuring well-being at the local level: an application to the equitable and sustainable well-being (BES) of the Italian Provinces Ecol. Indic. 76 281 296 281–96 10.1016/j.ecolind.2016.12.050
24 Mori K Christodoulou A 2012 Review of sustainability indices and indicators: towards a new city sustainability index (CSI) Environ. Impact Assess. Rev. 32 94 106 94–106 10.1016/j.eiar.2011.06.001
25 Ruggeri K 2020 Well-being is more than happiness and life satisfaction: a multidimensional analysis of 21 countries Health and Quality of Life Outcomes 18 1 16 1–16 10.1186/s12955-020-01423-y 31898546
26 Duclos J Y Sahn D E Younger S D 2006 Robust multidimensional poverty comparisons The Economic Journal 116 943 968 943–68 10.1111/j.1468-0297.2006.01118.x
27 Justino P Litchefield J Niimi Y 2005 1–45 1–45 Empirical applications of multidimensional inequality analysis, PRUS Working Paper No. 23, Poverty Research Unit at Sussex, University of Sussex
28 Alkire S 2005 Why the capability approach? Journal of Human Development 6 115 135 115–35 10.1080/146498805200034275
29 Saith R 2001 Capabilities: the Concept and its Operationalisation Queen Elizabeth House Oxford
30 Laderchi C R Saith R Stewart F 2003 Does it matter that we do not agree on the definition of poverty? a comparison of four approaches Oxford Development Studies 31 243 274 243–74 10.1080/1360081032000111698
31 Ellis E C 2021 People have shaped most of terrestrial nature for at least 12,000 years Proc. Natl Acad. Sci. 118 e2023483118 10.1073/pnas.2023483118 33875599
32 Ellis E C Ramankutty N 2008 Putting people in the map: anthropogenic biomes of the world Frontiers in Ecology and the Environment 6 439 447 439–47 10.1890/070062
33 World Bank 2021 Ghana Rising: Accelerating Economic Transformation and Creating Jobs World Bank
34 Coulombe H Wodon Q 2007 Poverty, Livelihoods, and Access to Basic Services in Ghana, in Ghana CEM: Meeting the Challenge of Accelerated and Shared Growth World Bank 1 86 pp 1–86
35 Dang H-A H 2020 Ghana Poverty Assessment World Bank
36 Annim S K Mariwah S Sebu J 2012 Spatial inequality and household poverty in Ghana Economic Systems 36 487 505 487–505 10.1016/j.ecosys.2012.05.002
37 Kanbur R Venables A J 2003 Spatial Inequality and Development Oxford University Press 10.1093/0199278636.001.0001
38 Cooke E Hague S McKay A 2016 The Ghana Poverty and Inequality Report: Using the 6th Ghana Living Standards Survey University of Sussex 1 43 pp 1–43
39 Bondzie E A Fosu G O Obu-Cann E 2013 Poverty and regional inequality in Ghana: a review Journal of Social Economics 1 32–36 32–36
40 Mensah F 2020 Multidimensional Poverty-Ghana Ghana Statistical Service
41 Aryeetey E Owusu G Mensah E J 2009 An Analysis of Poverty and Regional Inequalities in Ghana GDN
42 Acheampong R A 2018 Spatial Planning in Ghana Springer
43 Cobbinah P B Erdiaw-Kwasie M O 2018 Urbanization in Ghana: insights and implications for urban governance E-Planning and Collaboration: Concepts, Methodologies, Tools, and Applications IGI Global 256 278 pp 256–78
44 Plange N K 1979 Underdevelopment in northern Ghana: natural causes or colonial capitalism? Review of African Political Economy 6 4 14 4–14 10.1080/03056247908703393
45 Songsore J 2011 Regional development in Ghana: the Theory and the Reality new edn Woeli
46 Fuseini I Kemp J 2016 Characterising urban growth in tamale, Ghana: an analysis of urban governance response in infrastructure and service provision Habitat International 56 109 123 109–23 10.1016/j.habitatint.2016.05.002
47 Fuseini I Yaro J A Yiran G A 2017 City profile: tamale, Ghana Cities 60 64 74 64–74 10.1016/j.cities.2016.07.010
48 Codjoe S N A 2006 Population growth and agricultural land use in two agro-ecological zones of Ghana, 1960–2010 Int. J. Environ. Stud. 63 645 661 645–61 10.1080/00207230600963445
49 Yamba E I 2023 Revisiting the agro-climatic zones of Ghana: a re-classification in conformity with climate change and variability PLOS Climate 2 e0000023 10.1371/journal.pclm.0000023
50 Satterthwaite D Tacoli C 2014 Seeking an understanding of poverty that recognizes rural–urban differences and rural–urban linkages Urban livelihoods Routledge 52 70 pp 52–70
51 Zanaga D 2022 European Space Agency ESA WorldCover 10 m 2021 v200
52 Center for International Earth Science Information Network (CIESIN) 2016 Socioeconomic Data and Applications Center (SEDAC) Gridded population of the World, Version 4 (GPWv4): Population Count Adjusted to Match 2015 Revision of UN WPP Country Totals228
53 Ghana Statistical Service 2010 Population and Housing Census Ghana Statistical Service
54 Elbers C Lanjouw J O Lanjouw P 2003 Micro-level estimation of poverty and inequality Econometrica 71 355 364 355–64 10.1111/1468-0262.00399
55 Molina I Rao J 2010 Small area estimation of poverty indicators Canadian Journal of Statistics 38 369 385 369–85 10.1002/cjs.10051
56 Ghana Statistical Service 2013 Ghana Living Standards Survey Round 6 GLSS6
57 Ghana Statistical Service 2015 Ghana Poverty Mapping Report Ghana Statistical Servcie Report
58 Cavanaugh A C 2023 Locating poverty and inequality: an application of small area estimation methods using survey and census data from Ghana. Pathways to Equitable Healthy Cities Working Paper
59 Songsore J 2008 Environmental and structural inequalities in Greater Accra Journal of the International Institute 16 8–13 8–13
60 Owusu G Wrigley-Asante C 2020 Urban inequality and poverty in the global south Equilibri 24 79 90 79–90
61 Owusu G Yankson P W 2007 ‘Poverty in Ghana is basically a rural phenomenon’: are we underestimating urban poverty? Ghana Journal of Development Studies 4 87 105 87–105 10.4314/gjds.v4i1.35055
62 Alli A S 2023 High-resolution patterns and inequalities in ambient fine particle mass (PM2. 5) and black carbon (BC) in the Greater Accra Metropolis, Ghana Sci. Total Environ. 875 162582 10.1016/j.scitotenv.2023.162582 36870487
63 Clark S N 2022 Spatial modelling and inequalities of environmental noise in Accra, Ghana Environ. Res. 214 113932 10.1016/j.envres.2022.113932 35868576
64 Bixby H 2022 Quantifying within-city inequalities in child mortality across neighbourhoods in Accra, Ghana: a bayesian spatial analysis BMJ open 12 e054030 10.1136/bmjopen-2021-054030
65 Tetteh J D 2022 Spatial heterogeneity in drinking water sources in the Greater Accra metropolitan area (GAMA), Ghana Population Environ. 44 46 76 46–76 10.1007/s11111-022-00407-y
66 Dasgupta P 2001 Human Well-Being and the Natural Environment OUP Oxford
67 Raudsepp-Hearne C Peterson G D Bennett E M 2010 Ecosystem service bundles for analyzing tradeoffs in diverse landscapes Proc. Natl Acad. Sci. 107 5242 5247 5242–7 10.1073/pnas.0907284107 20194739
68 UNDP 2022 Human Development Report 2021-22: Uncertain Times, Unsettled Lives: Shaping our Future in a Transforming World United Nations
69 Abekah-Nkrumah G Guerriero M Purohit P 2014 ICTs and maternal healthcare utilization. evidence from Ghana International Journal of Social Economics 41 518 541 518–41 10.1108/IJSE-11-2012-0218
70 Songsore J McGranahan G 1993 Environment, wealth and health: towards an analysis of intra-urban differentials within the greater accra metropolitan area, Ghana Environment and Urbanization 5 10 34 10–34 10.1177/095624789300500203
71 WHO/UNICEF 2021 Joint Monitoring Programme for Water Supply and Sanitation (JMP) Progress on household drinking water, sanitation and hygiene 2000-2020: five years into the SDGs World Health Organization (WHO) and the United Nations Children's Fund (UNICEF)
72 UNESCO 2021 Global education monitoring report 2021/2: Non-state actors in education: Who chooses? Who loses? United Nations Educational, Scientific and Cultural Organization (UNESCO)
73 Union I T 2023 Measuring Digital Development, Facts and Figures: Focus on Least Developed Countries ITU
74 Jenkins S P 2020 Perspectives on Poverty in Europe. Following in Tony Atkinson’s Footsteps Italian Economic Journal 6 129 155 129–55 10.1007/s40797-019-00112-0
75 Database W G H O 2019 Impoverishing Out-of-Pocket Health Spending: Pushed below a Poverty Line by Household Expenditures on Health World Health Organization
76 Bourguignon F 2015 The Globalization of Inequality Princeton University Press
77 Lin Y 2017 The social income inequality, social integration and health status of internal migrants in China International Journal for Equity in Health 16 1 11 1–11 10.1186/s12939-017-0640-9 28057005
78 Srivastava S 2023 Spatial heterogeneity in intimate partner violence across the 640 districts of India: a secondary analysis of a cross-sectional, population-based survey by use of model-based small-area estimation The Lancet Global Health 11 e1587 e1597 e1587–97 10.1016/S2214-109X(23)00377-7 37734802
79 Adzitey F 2013 Animal and meat production in Ghana-An overview Journal of World's Poultry Research 3 1–4 1–4
80 Quaye A K Hall C A Luzadis V A 2010 Agricultural land use efficiency and food crop production in Ghana Environment, Development and Sustainability 12 967 983 967–83 10.1007/s10668-010-9234-z
81 Tache B Sjaastad E 2010 Pastoralists’ conceptions of poverty: an analysis of traditional and conventional indicators from borana, Ethiopia World Development 38 1168 1178 1168–78 10.1016/j.worlddev.2010.01.001
82 Abukari H Mwalyosi R B 2020 Local communities’ perceptions about the impact of protected areas on livelihoods and community development Global Ecology and Conservation 22 e00909 10.1016/j.gecco.2020.e00909
83 Killingsworth M A 2021 Experienced well-being rises with income, even above $75,000 per year Proc. Natl Acad. Sci. 118 e2016976118 10.1073/pnas.2016976118 33468644
84 Notenbaert A M 2012 Policies in support of pastoralism and biodiversity in the heterogeneous drylands of East Africa Pastoralism: Research, Policy and Practice 2 1 17 1–17 10.1186/2041-7136-2-14
85 Rota A 2018 How to do Note: Engaging with Pastoralists—A Holistic Development Approach IFAD
86 Hruska T 2017 Rangelands as Social–Ecological systems Rangeland Systems: Processes, Management and Challenges Springer 263 302 pp 263–302
87 Razafindratsima O H 2021 Reviewing the evidence on the roles of forests and tree-based systems in poverty dynamics Forest Policy and Economics 131 102576 10.1016/j.forpol.2021.102576
88 Okech R Haghiri M George B P 2012 Rural tourism as a sustainable development alternative: an analysis with special reference to Luanda, Kenya CULTUR-Revista de Cultura e Turismo 6 36 54 36–54
89 Akyeampong O A 2011 Pro-poor tourism: residents’ expectations, experiences and perceptions in the kakum national park area of Ghana Journal of Sustainable Tourism 19 197 213 197–213 10.1080/09669582.2010.509508
90 Holland M B Masuda Y J Robinson B E 2022 Land Tenure Security and Sustainable Development Springer
91 Vallet A 2020 Power asymmetries in social networks of ecosystem services governance Environ. Sci. Policy 114 329 340 329–40 10.1016/j.envsci.2020.08.020
92 Berbés-Blázquez M González J A Pascual U 2016 Towards an ecosystem services approach that addresses social power relations Current Opinion in Environmental Sustainability 19 134 143 134–43 10.1016/j.cosust.2016.02.003
93 Primdahl J 2018 Rural landscape governance and expertise: on landscape agents and democracy Defining Landscape Democracy. A Path to Spatial Justice Edward Elgar Publishing 153 164 pp 153–64
94 Penker M 2009 Landscape governance for or by the local population? a property rights analysis in Austria Land Use Policy 26 947 953 947–53 10.1016/j.landusepol.2008.11.007
95 Reid R S 2016 Evolution of models to support community and policy action with science: balancing pastoral livelihoods and wildlife conservation in savannas of East Africa Proc. Natl Acad. Sci. 113 4579 4584 4579–84 10.1073/pnas.0900313106 19887640
96 Audia C 2021 Loops and building blocks: a knowledge co-Production framework for equitable urban health Journal of Urban Health 98 394 403 394–403 10.1007/s11524-021-00531-4 33738655
97 Petzold J 2020 Indigenous knowledge on climate change adaptation: a global evidence map of academic literature Environ. Res. Lett. 15 113007 10.1088/1748-9326/abb330
98 Pega F 2022 Unconditional cash transfers for reducing poverty and vulnerabilities: effect on use of health services and health outcomes in low-and middle-income countries Cochrane Database of Systematic Reviews 3 1–198 1–198 10.1002/14651858.CD011135.pub3
99 Xu Y Abena A A N Bakpa E K 2019 Poverty and health issues in the northern part of Ghana J. Res. Humanit. Soc. Sci. 9 14 21 14–21
100 Awojobi O 2019 A systematic review of the impact of Ghana’s school feeding programme on educational and nutritional outcomes. Agro-Science 18 42 50 42–50 10.4314/as.v18i2.8
101 Weeranakin P Promphakping B 2018 Local meanings of wellbeing and the construction of wellbeing indicators Social Indicators Research 138 689 703 689–703 10.1007/s11205-017-1692-3
102 Patrick R 2019 Human wellbeing and the health of the environment: local indicators that balance the scales Social Indicators Research 146 651 667 651–67 10.1007/s11205-019-02140-w
103 Loveridge R 2020 Measuring human wellbeing: a protocol for selecting local indicators Environ. Sci. Policy 114 461 469 461–9 10.1016/j.envsci.2020.09.002
104 Detoeuf D Wieland M Wilkie D 2020 Guide 2.0 to the Modified Basic Necessities Survey: Why and How to Conduct Digital-Based BNS in Conservation Landscapes Wildlife Conservation Society 10.19121/2020.Report.38385
