Introduction

Policymakers often tout economic growth and transformation as a golden goose to promote prosperity and the goals of the Sustainable Development Agenda. However, more than two-thirds of the world’s population lived in countries where inequality was growing pre-pandemic, while some countries also saw stalling or reversing poverty reduction despite increases in labour productivity or GDP per capita (UNDESA, 2020). In addition, climate change may affect inclusion, for example through increasing energy and food prices, or contributing to poverty persistence (Leichenko and Silva, 2014; Soergel et al, 2021; Diwakar and Lacroix, 2021). Climate change could also affect the pace of growth, for example through damaged infrastructure, lower productivity, and negative effects on health and mortality (Fankhauser and Tol, 2005; Moore and Diaz, 2015). At the same time, growth itself may proceed in ways unsustainable for planetary boundaries.

As such, there is an urgent need to develop a joined-up understanding of and decision-making on these issues (Colenbrander et al, 2022). Already, there is increasing recognition that other measures of wellbeing need to be considered alongside growth in contributing to welfare, and that these measures do not necessarily flow from economic transformation. Instead, what is needed is a more holistic prioritisation of people and planet, such that pillars of transformation, inclusion and sustainability can go hand in hand. This in turn requires an evidence-based understanding of country progress in these domains.

The objective of this analysis is to explore advances made in low- and middle-income countries around economic transformation, poverty and inequality reduction, and environmental sustainability, and how these dimensions may have worked together or against each other especially over the last two decades. We rely on the definition of some of these terms provided by Colenbrander et al (2022), as its joined up understanding is a useful way to examine tripartite outcomes. This offers the following definitions:

• Economic transformation – moving employment to higher-productivity and higher-value activities that enable increases in human and physical capital. This may also encompass improvements in productivity within sectors through innovations and efficiencies.

• Environmental sustainability – economic and social activity happens while conserving biodiversity and ecosystem function, reducing pollution (including GHGs) and using natural resources in ways that take account of the needs of future generations.

• Poverty and inequality reduction – benefits of transformation reach the bottom of the distribution, such that it contributes to poverty eradication and reduction in inequalities including those experienced by people at the bottom of the distribution.

These three pillars together are identified in this analysis as tripartite outcomes. Our focus is not on explaining key drivers of tripartite outcomes, which requires a more in-depth contextually differentiated analysis of pathways (for example, see Pickard and Lemma, 2022). Instead, we use a range of country-level data across these three dimensions to examine synergies and trade-offs between domains and the state of progress since the turn of the century (see Table 1 for data sources). The study presents a typology of low- and middle-income countries that identifies constellations of tripartite outcomes based on data across these three dimensions from 2000 to 2018. We focus on low- and middle-income countries as they may not be able to mobilise domestic revenues as easily as high-income countries to address tripartite outcomes. The results point to the emergence of five clusters that represent different degrees of tripartite outcomes, with a ‘triple-win’ not readily evident but certain countries performing relatively well across dimensions.

The analysis is organised as follows: the second section lays out key frameworks that motivate the choice of dimensions and indicators employed in the analysis, the third section outlines these dimensions and indicators, and the fourth section presents the methods guiding the analysis presented in this article. The fifth section presents the results of the cluster analysis and descriptive complements underpinning the analysis of tripartite outcomes. The sixth section offers emerging policy recommendations and concludes.

Frameworks and literature

In this article, we identify ‘tripartite outcomes’ as the joining up of economic transformation, poverty and inequality reduction, and environmental sustainability. Perhaps the most prominent framework related to tripartite issues today is the doughnut economics model (Raworth, 2012; 2017), which suggests a safe and just space for humanity that recognises an embedded economy and pursues distribution by design while remaining agnostic about growth. Two foundational principals of the model are that basic human rights should be met, and that the ecological stability of the planet needs to be protected (Raworth, 2012). Some scholars have employed doughnut thinking as a means to enhance the transformative potential of the UN Sustainable Development Goals (SDGs) (Hajer et al, 2015). Though the SDGs include several dimensions linked to these tripartite pillars, these lack joining up.

Most frameworks that characterise tripartite issues in inclusion, sustainability and growth, like the doughnut model, typically integrate social wellbeing and aspects of economic transformation as a means to an end within the planetary boundaries concept applied at different scales (Rockström et al, 2009; Leach et al, 2013; Dearing et al, 2014). In these efforts, most place human wellbeing and its core needs as central (Gough, 2015). Even the Sustainable Development Index which includes indicators of environmental sustainability alongside life expectancy, education and income, is focused on ‘the ecological efficiency of nations in delivering human development’ (Hickel, 2020). Frameworks typically recognise that people affected adversely by the ecosystem and climate are often the most vulnerable and marginalised, thus exacerbating poverty and inequality (Corvalan et al, 2005; Leichenko and Silva, 2014; Hallegatte et al, 2016; Diwakar and Lacroix, 2021; Wackernagel et al, 2021). There is also recognition that climate policies could create a financial burden on people in poverty through increased energy and food prices (Soergel et al, 2021). At the same time, there are a myriad of direct and indirect pathways marking the complexity of the relationship between environmental sustainability and poverty and inequality (Leichenko and Silva, 2014; Hallegatte et al, 2020).

The role of economic growth and transformation within ecological frameworks is varied. Ecological economics proponents often argue that there needs to be a shift in attention from growth to redistribution (Vandenhole, 2018). In some cases they also argue for promoting degrowth through rich nations scaling down growth in an effort to limit ecological breakdown (Wiedmann, 2015; Steinmann et al, 2017). There is also recognition in these discussions that reduction of carbon emissions in low- and middle-income countries require reducing carbon intensity as well as changing the historical relationship between energy use and economic growth (Steckel et al, 2013). Others also argue in favour of ‘green growth’ and a decoupling of GDP from raw material consumption (World Bank, 2012; Hammer and Hallegatte, 2020). Hickel and Hallegatte (2021) presents some of this argument of degrowth versus decoupling.

Across these dimensions, various risks such as climate-related disasters and conflict permeate and can structurally affect the degree of poverty and inequality reduction, environmental sustainability and economic transformation. For example, there is a bidirectional relationship between conflict and poverty linked intrinsically to weak institutions, which limits the effectiveness of poverty reduction strategies (Diwakar et al, 2023). As such, if we did not consider the role of risks or of governance alongside the three tripartite dimensions, we would have a limited understanding of the contextual forces that in practice may nurture or constrain tripartite outcomes.

Though there have been important advances in the conceptual understanding of tripartite issues as outlined earler, there is typically not equal consideration of priorities around economic transformation, environmental sustainability, and poverty and inequality reduction in existing frameworks. There is also limited engagement with understanding the degree of change over time between dimensions, and more generally limited empirical engagement around the extent to which countries may experience synergies or trade-offs in tripartite dimensions. Finally, when these topics are brought together such as in the doughnut model or the Sustainable Development Index, there is narrow consideration of contextual risks beyond the environmental domain, such as linked to conflict or the underlying governance conditions that may influence the degree of tripartite outcomes. This article tries to fill a gap in knowledge by attempting to operationalise concepts through an empirical, embedded analysis of tripartite outcomes over the last two decades, using a set of variables commonly identified to represent the dimensions in the wider literature.

Dimensions and indicators

We rely on an analysis of a range of indicators (Table 1) used to operationalise concepts of economic transformation, poverty and inequality reduction, and environmental sustainability over the period from 2000 to 2018. Based on the definitions of the dimensions noted in the introduction, we employ a set of indicators presented in Table 1.² Indicators are selected to proxy different aspects of the dimension they represent, with considerable data points over 2000 to 2018, while limiting correlation between dimensions. We rely on data since the turn of the century, when the Millennium Development Goals were first implemented and the SDGs subsequently introduced. This represents the continuation of a period when issues beyond economic growth, especially around human development and capabilities, began to gain attention in international development discourse (Sen, 1992; Nussbaum, 2000; Robeyns, 2005). Of course, there are a range of other variables that could also be proposed; however, our indicators are selected to balance concerns on conceptual validity, with data reliability and availability.

Methods and descriptive data

In our assessment, we rely on a pooled k-means cluster analysis to identify groups (clusters) of countries that had variable performance on the tripartite outcomes and to understand changes in countries’ alignment with the clusters over time. In this method, k-means clustering groups n observations into k sets S = {S₁, S₂, …, S_k} to minimise the within-cluster variance by identifying:

View Expanded

for S sets represented by a parameter of indicator variables. This process follows a constellations of fragility analysis undertaken on a different topic (to assess state fragility) by Ziaja et al (2019). It also resonates with cluster analysis from the climate literature, where for example Lamb et al (2014) used clustering techniques to group countries according to drivers of life expectancy and carbon emissions. We apply this process to our analysis given our similar interest in understanding constellations of outcomes, albeit on a different set of issues. Given that some of our risk indices also comprise elements that relate to inclusion or sustainability, we exclude it from the calculation of cluster scores.

Like Ziaja et al (2019), we pool all country-years in the sample to increase the number of observations. To limit data gaps, we rely on interpolated poverty data provided by PovcalNet (2021), and linearly interpolate missing inequality data to balance the indicators within the inclusion dimension. The low variance of this indicator pre-transformation, as measured through its standard deviation across countries (0 to 0.06), suggests that these data edits are unlikely to drastically affect the resulting data credibility. The labour productivity indicator is strongly positively skewed, and so we take its logarithm assuming that marginal effects may be lower at higher values. We then standardise the raw data scores, which means that we rescale the data to have a mean of 0 and a standard deviation of 1. We do this to ensure that variables with larger scales do not dominate how the clusters are categorised. Next, we invert relevant indicators such that all follow a consistent gradation from weak to strong outcomes. Table 2 summarises the types of transformation performed for the set of indicators used in the cluster analysis, and summary statistics for the indicators after imputation, while Table 3 presents a correlation matrix of the indicators.

Our pooled country-year database comprises a sample size of 1,908 observations after removing HICs and countries without adequate data from which to derive dimension scores. These observations come from 108 countries from sub-Saharan Africa (40.6% of observations), Europe and central Asia (17.8% of observations), Latin America and Caribbean (15.8% of observations), East Asia & Pacific (10.9% of observations), Middle East and North Africa (9.9% of observations), and South Asia (5.0% of observations).

The next question in our cluster analysis is how to determine dimension scores for our three dimensions and how to select an optimal. The most common approach to indices is to select the dimension score corresponding to the average of the dimension-specific indicators, which is our chosen method.⁴ For example, the dimension score for Poverty and Inequality Reduction is the average of the poverty headcount ratio and inequality in the bottom half score. Figure A1 in the annex presents the matrix chart for dimension scores. Finally, to select the optimal k-means cluster solution for our dimensions, we rely on the curve from the within sum of squares, its logarithm, the proportional reduction of error coefficient, and the η2k which is a coefficient that captures the largest ratio of the within-cluster sum of squares to the total sum of squares (Makles, 2012). These methods together suggest that clustering with k=5 is the optimal local solution (Figure 1).⁵

Selection of clusters

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Selection of clusters

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

Selection of clusters

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

There are some limitations worth discussing. As noted earlier, the data transformations required some imputation due to limited coverage. In addition, grouping indicators into dimensions may hide nuances, where, for example, labour productivity might be high in spite of low diversification. Relying on two indicators per dimension may also be inadequate in capturing the complexity of the phenomenon being investigated. Our reliance on these indicators however is due to the availability of time trend data in our pooled country-year database to enable sufficient observations for our analysis. In addition, other forms of aggregation or local solutions may have arrived at different clusters, but we proceed with the group of five clusters as a solution that is at a local maximum and also parsimonious.

Finally, this cluster method of pooling country-years assumes clusters to be constant; to mitigate this concern, we examine the underlying data to understand the extent to which countries move between groups over time and comment where differences emerge. This helps assess the sensitivity of results and offers additional nuance. In this process, we distinguish different country income groups based on present-day status to recognise the different pace of economic growth that could affect degrees of transformation, inclusion and sustainability. Alongside this, we also examine correlation in the rates of change in key indicators in different dimensions over time, to identify where strong performers emerge within clusters and country income groups. Finally, we explore the risk and governance profiles as potential aspects of the enabling context that can nurture tripartite outcomes.

Our focus on aggregating by country income groups while examining certain tripartite indicators, alongside recognition of structural conditions through drawing attention to risk and governance profiles, is just one way of approaching this analysis; there are other forms of aggregation that would likely yield different results. Even so, the emphasis on change also inherently draws attention to conditions that might enable change, such as the degree of human or physical risk a country faces, as well as its capacity for governance.

Constellations of tripartite outcomes

Clusters of inclusive, sustainable economic transformation

Figure 2 presents the properties of our cluster analysis as boxplots (see Annex Table A1 for country-year clusters). The clusters are not ordinal, insofar as there is no obvious ordering of clusters based on their performance overall. It also does not capture the degree of change over time but only levels, given that the data is pooled across countries and years. Although the risk and governance dimension is not included in the determination of our clusters, we assess its cluster-based properties as potential contextual variables that may influence the degree of inclusivity, environmental sustainability and economic transformation observed over time.⁶

Dimension scores per cluster, 2000–2018 (N=1,908)¹

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Dimension scores per cluster, 2000–2018 (N=1,908)¹

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

Dimension scores per cluster, 2000–2018 (N=1,908)¹

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

From the cluster analysis, we identify cluster A as predominantly UMICs exhibiting strong economic transformation among the set of countries included, though this comes with a trade-off of low sustainability scores. Cluster B, a mix of mainly LMICs and UMICs, fares relatively much better on metrics of sustainability and inclusion, but with a trade-off in terms of lower economic transformation scores relative to cluster A. Cluster C (predominantly UMICs) comprises a smaller number of countries with particularly low sustainability outcomes, typically falling over two standard deviations below the mean, alongside moderate inclusion and transformation scores just below average. Clusters D and E (mainly LICs and LMICs) share weak transformation scores, though cluster D on average performs much better in inclusion and sustainability. Even so, both include a number of outliers performing particularly weakly on sustainability metrics.

A triple win is not readily evident, though cluster B would be the closest with a ‘triple moderate’ outcome across the three dimensions of inclusion, sustainability, and economic transformation. A stronger result could occur if the inclusion performance of cluster B (moderate tripartite gain) and the mean sustainability performance of cluster D (weak transformation) were aligned with the economic transformation potential in cluster A (strong transformation), though in practice these examples are hard to identify. Another way to think about this is the extent to which countries may intentionally or inadvertently balance slightly lower levels of transformation for strong sustainability and inclusion. At the moment, most LICs fall in clusters D (weak transformation) and E (weak inclusive transformation), with relatively better environmental sustainability but moderate or low inclusion and low economic transformation.

Changes over time by income group

Even though most LICs have lower economic transformation scores in terms of absolute values, the change over time that they experience can be strong. Recognising this, we examine the extent to which countries have experienced mobility between clusters and complement this with descriptive analysis of rates of change in individual indicators by country income group. There are cases where results from the two methods may diverge, with a transition country not emerging as such in the descriptive analysis of rates of change. This points to heterogeneity of country experiences and offers insights into potential cases for further research.

Most LICs fell under cluster E (weak inclusive transformation), followed by cluster D (weak transformation). Table 4 highlights certain LICs (for example, movements from cluster E to D) improving their cluster standing towards more inclusive outcomes that are on average at slightly improved levels of economic transformation. The practical significance of these movements is that improvements in tripartite outcomes are possible without necessarily compromising development pillars. Sierra Leone and Ethiopia, especially at the turn of the century, have vacillated between these two clusters. The instability may reflect contexts of insecurity and violence that act to undermine development. Conflict-affected South Sudan and Syria are included in clusters A (strong transformation) and B (moderate tripartite gain), respectively, in some years. However, in both countries, the dimension scores deteriorate, particularly in recent years, though only in South Sudan is this consistently large across dimensions to result in a transition from cluster A to E.

Table 4:

Countries experiencing transitions between clusters, LICs

Region	Country Yr:	00	01	02	03	04	05	06	07	08	09	10	11	12	13	14	15	16	17	18
LCN	Haiti												E	D
MEA	Syria				D	B
SSF	Burkina Faso								E	D
SSF	Burundi										D	E
SSF	Ethiopia	E	D	D	E	D
SSF	Guinea										E	D
SSF	Guinea-Bissau				D	E
SSF	Liberia											E	D
SSF	Malawi	D	E
SSF	Mali							E	D
SSF	Niger								E	D
SSF	Rwanda																		E	D
SSF	Sierra Leone	D	E	D	E	E	D
SSF	South Sudan															A	E
SSF	Togo							D	E
SSF	Uganda						E	D

Note: Blank cells refer to cases where the cluster remained stable (for example, Haiti was in cluster E until 2011, and then transitioned to cluster D). Only countries with transitions are listed in the table.

Figure 3 next goes beyond levels to explore the relationship between changes in poverty rates and labour productivity, weighted by change in material footprints per country income group. We weight the data by material footprints rather than GHG emissions, as the former has been increasing on average across all country income groups since 2000. These effectively reflect common indicators of inclusive economic development and the extent to which this may correlate with sustainability changes. Our analysis points to high labour productivity increases being associated with stronger poverty reduction, especially in LICs, where initial levels of poverty are typically higher. However, countries which are doing particularly well on both poverty reduction and labour productivity (bottom right quadrant of Figure 3) often have higher material footprints. There are nevertheless some countries where poverty is reducing and labour productivity increasing faster than the income-group averages that display comparatively lower environmental impacts: Ethiopia, Guinea, Rwanda and Uganda.⁷ These are some of the countries that have moved from clusters E (weak inclusive transformation) to D (weak transformation) over the period of analysis.

Poverty reduction, material footprint and labour productivity – LICs (bubble size represents annual change in footprint)

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

Note: Lines cutting x- and y-axes represent country-weighted average change. Axes refer to percentage point changes.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Poverty reduction, material footprint and labour productivity – LICs (bubble size represents annual change in footprint)

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

Note: Lines cutting x- and y-axes represent country-weighted average change. Axes refer to percentage point changes.

• Download Figure
• Download figure as PowerPoint slide

Poverty reduction, material footprint and labour productivity – LICs (bubble size represents annual change in footprint)

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

Note: Lines cutting x- and y-axes represent country-weighted average change. Axes refer to percentage point changes.

• Download Figure
• Download figure as PowerPoint slide

Even so, among LICs, it is worth stressing that footprints are low, even if slightly increasing, and may be intuitively observed to constitute a fair share of consumption for a more decent standard of living (particularly if equitably distributed), and not reaching ecological thresholds as in other income contexts. As such, other countries that have moved from cluster E to D and are at the bottom of quadrant IV but with a slightly larger bubble size in Figure 3 also similarly reflect improvements in inclusive growth without exceeding ecological limits. So, there seems to be a transition between LIC and LMICs where countries start to exceed their ‘fair share’ of consumption, and where accordingly sustainability considerations would need to be strongly embedded in policy.

The LICs named here (Ethiopia, Guinea, Rwanda and Uganda) perform better relative to others in their income group, especially in terms of changes in risk (particularly measured by the INFORM index) and perceived government effectiveness over the period (Figure 4). The degree of change in government effectiveness is particularly pronounced in Rwanda, for example, moving from -0.91 in 2002 to 0.21 by 2018, reflecting perceived improvement for example in the country’s quality of public and civil service and the quality of policy formulation and implementation. However, it is worth noting that several indicators in the socio-economic vulnerability dimension of the INFORM risk index constitute factors intrinsic to inclusion, such as the Multidimensional Poverty Index and GINI index of inequality, and so represent a source of endogeneity. Even so, the analysis would still suggest that government effectiveness might play an important role in helping support a road towards inclusive, sustainable economic transformation among LICs. However, as noted elsewhere, there are various other enabling factors, with the role of risks and governance forming only part of the factors that may support tripartite outcomes.

Change in risk and governance indicators

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Change in risk and governance indicators

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

Change in risk and governance indicators

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

Among LMICs, there are also country transitions between constellations of inclusive, sustainable, economic transformation. LMICs fell most commonly within cluster D (weak transformation), followed by clusters B (moderate tripartite gain) and E (weak inclusive transformation). Some countries in this set, especially in East Asia and Pacific, Europe and Central Asia and South Asia have transitioned from cluster D to B on account of improvements in inclusion and economic transformation. Many countries have also churned between clusters D and B, particularly in Latin America and the Caribbean and in South Asia, signalling some volatility in inclusion and growth trajectories, in particular worsening for those moving from B to D (Table 5). In sub-Saharan Africa, movements into or out of group E are also particularly common, improving inclusion outcomes similar to the majority of LICs when transitioning out of cluster E.

Table 5:

Countries experiencing transitions between clusters, LMICs

Region	Country	00	01	02	03	04	05	06	07	08	09	10	11	12	13	14	15	16	17	18
EAS	Lao PDR	E	D												D	B
EAS	Mongolia							D	B	B	C	B	C
EAS	Myanmar			E	D						D	B	B	D
EAS	Timor-Leste		D	B	B	D					D	B	B	D			D	B	D	B
EAS	Viet Nam					D	B
ECS	Kyrgyzstan		D	B
ECS	Moldova			D	B
ECS	Ukraine						B	A		A	B	A			A	B
ECS	Uzbekistan	E	D	D	E	E	D								D	B
LCN	El Salvador						E	B
SAS	India				E	B
SAS	Nepal						D	B		B	D
SAS	Pakistan										D	B	B	D	B	D
SSF	Angola	D	E					E	A	B	A					A	E
SSF	Benin					D	E
SSF	Eswatini						E	B
SSF	Ghana							E	D	E	E	D
SSF	Nigeria																		E	D
SSF	Tanzania											E	D

Examining rates of change alongside this cluster analysis, there are some LMICs that have seen poverty reduction and structural economic change while retaining footprint changes below the group average: Bangladesh, India, Myanmar, Uzbekistan, Eswatini, Nigeria and PNG – though the latter three with much higher poverty rates, and Eswatini also with a much higher footprint in the latest survey year (Figure 5). Many of these countries (Myanmar, Uzbekistan, Nigeria, Eswatini and India) experienced movements out of cluster E and into clusters D or B in the cluster analysis (Table 5), all signalling some improvements in poverty and inequality reduction at certain points over the period, though with variable improvements in growth, and for those moving from E to B also typically sacrificing environmental sustainability in the process. More generally, though, LMICs in Central and South Asia tend to do quite well in terms of poverty reduction accompanied by smaller per capita footprint increases (India, Nepal, Myanmar, Kyrgyzstan and Uzbekistan are in the bottom two quadrants of Figure 5) though with variable increases in productivity.

Poverty reduction, footprint and labour productivity – LMICs (bubble size represents annual change in footprint)

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

Note: Lines cutting x- and y-axes represent country-weighted average change. Axes refer to percentage point changes.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Poverty reduction, footprint and labour productivity – LMICs (bubble size represents annual change in footprint)

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

Note: Lines cutting x- and y-axes represent country-weighted average change. Axes refer to percentage point changes.

• Download Figure
• Download figure as PowerPoint slide

Poverty reduction, footprint and labour productivity – LMICs (bubble size represents annual change in footprint)

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

Note: Lines cutting x- and y-axes represent country-weighted average change. Axes refer to percentage point changes.

• Download Figure
• Download figure as PowerPoint slide

These South Asian countries, however, have varied human and natural hazard risks and varied government effectiveness. For example, while India and Myanmar saw improvements in government effectiveness scores between 2002 and 2018, scores worsened in Bangladesh, Nepal and Kyrgyzstan. The situation in Myanmar has also changed considerably since 2018. In terms of risk, according to our analysis of the datasets, South Asia has the lowest regional ND-GAIN readiness score (together with sub-Saharan Africa – both 0.30), and the highest average INFORM risk score on average over the last two decades, though also the highest rate of improvement in its INFORM score (Figure 6). This improvement in a multi-dimensional risk profile (albeit from a high initial level) could signal a South Asian LMIC context where some of its transformations are less hard on the environment but still poverty reducing.

Change in risk and governance indicators

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Change in risk and governance indicators

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

Change in risk and governance indicators

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

Some of the basic measures to reduce risk – piped water, all-weather roads, stormwater drains, healthcare, emergency services – also contribute to poverty reduction. This once more underscores the need for complementary policies and investments to deliver inclusion and reduce risk. Economic transformation is rarely enough on its own; and even where it is, faster poverty reduction with less environmental degradation could be achieved with additional measures. Indeed, many of these countries struggle with very degraded local environments with severe consequences for human health.

Finally, UMICs were most present in clusters A (strong transformation) and C (weak sustainability), followed by B (moderate tripartite gain). In terms of transitions, most countries moved into or out of group B (Table 6). Some countries in Europe and Central Asia moved from clusters B to A, prioritising economic transformation and trading off inclusion and sustainability in the process. There was also considerable volatility in the Latin American and Caribbean countries across the range of clusters.

Table 6:

Countries experiencing transitions between clusters, UMICs

Region	Country	00	01	02	03	04	05	06	07	08	09	10	11	12	13	14	15	16	17	18
EAS	China				B	A
EAS	Indonesia				B	A	B	A	B
EAS	Malaysia																A	C
EAS	Thailand			B	A
ECS	Armenia	D	B	D	D	B
ECS	Azerbaijan				D	B
ECS	Belarus										B	A
ECS	Bosnia and Herzegovina							B	A
ECS	Bulgaria				B	A
ECS	Georgia						D	B
ECS	Kazakhstan	A	C
ECS	Serbia						B	A
ECS	FYR Macedonia		A	B	A
LCN	Colombia		E	B		B	A
LCN	Ecuador						E	B
LCN	Guatemala	B	D	B								B	D	B			B	D	D	B
LCN	Paraguay			E	C	D	C
LCN	Peru			E	D	E					E	D	B
LCN	Venezuela (Bolivarian Republic of)								E	A	E
MEA	Iraq			B	D	D	B

When examining rates of change, there is a mixed relationship between poverty reduction and labour productivity among UMICs. However, countries faring better on both dimensions tend to have higher material footprints (Figure 7). Indonesia is the only country in this set that is below the group average for changes in footprint and above average for poverty reduction and labour productivity increases. It churned between groups B and A in the early 2000s, with improvements especially in the inclusion and sustainability metrics in the years that followed. Armenia, Georgia and Indonesia (the latter at a relatively lower rate of change in labour productivity) stand out as three UMICs that have managed to have higher rates of change in labour productivity and poverty reduction than the country-weighted income group average while also having average footprints over the period lower than the income group average. All transitioned into group B in the early 2000s, though from different starting points (from group D in Armenia and Georgia, and group A in Indonesia). These may be regarded as more positive examples of tripartite outcomes.

Poverty reduction, footprint, and labour productivity – UMICs (bubble size represents annual change in footprint)

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

Note: Lines cutting x- and y-axes represent country-weighted average change. Axes refer to percentage point changes.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Poverty reduction, footprint, and labour productivity – UMICs (bubble size represents annual change in footprint)

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

Note: Lines cutting x- and y-axes represent country-weighted average change. Axes refer to percentage point changes.

• Download Figure
• Download figure as PowerPoint slide

Poverty reduction, footprint, and labour productivity – UMICs (bubble size represents annual change in footprint)

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

Note: Lines cutting x- and y-axes represent country-weighted average change. Axes refer to percentage point changes.

• Download Figure
• Download figure as PowerPoint slide

Although Georgia, Armenia and Indonesia have varied risk profiles, all three countries see improvements in government effectiveness, while Georgia and Indonesia also see improvements in their risk scores over the period (Figure 8). There are other countries that have such experiences, particularly in terms of improving the risk context, suggesting that these improvements in risk and governance may be necessary but not sufficient conditions to promote tripartite outcomes. Instead, building on the analysis presented earlier, other social and environmental policies are likely to be necessary.

Change (bottom) in risk and governance indicators

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Change (bottom) in risk and governance indicators

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

Change (bottom) in risk and governance indicators

Citation: Journal of Poverty and Social Justice 31, 3; 10.1332/175982721X16845094996449

• Download Figure
• Download figure as PowerPoint slide

A way forward

The analysis suggests that while labour productivity improvements and poverty reduction are closely correlated, these processes have typically not been environmentally sustainable. This reinforces the need to embed mainstream conceptualisations of sustainable development within ecological frameworks such as the doughnut model, but moreover go beyond this to acknowledge a wider array of risks and enabling conditions that affect tripartite outcomes on a more equal footing. COVID-19 has heightened this risk profile by reversing progress in poverty and inequality reduction (World Bank, 2022), constraining growth (IMF, 2022), and in some cases leading to a prioritisation of economic recovery over environmental sustainability (GlobeScan, 2021). Renewed efforts will be needed to address tripartite outcomes in the years ahead.

A key question regarding tripartite outcomes is how to enhance productivity improvement and poverty reduction while moving towards a fairer distribution of carbon and material budgets. Based on our analysis of pre-pandemic data, changes in the development of economic sectors and infrastructure may help rebalance tripartite outcomes – especially in LMICs and UMICs, but also starting these processes in LICs. For example, performance could be improved on sustainability of agriculture, forestry and fisheries, moving countries towards greener industrial processes, and improving the sustainability of the tourism service sector. Indeed, there is a case to be made for more sustainable forms of economic growth and transformation with inclusive social policies across all countries, rather than waiting for poverty, economic growth and environmental degradation having reached certain levels to necessitate them.

By income group, among LICs, the relationship between growth and poverty reduction has been particularly strong, and material footprint and GHGs remain low with these gains (though air pollution is a problem). Some successes are Ethiopia, Rwanda, Guinea and Uganda, with strong labour productivity increases and poverty reduction over the 18-year period, with relatively low environmental impacts. When focusing on tripartite outcomes, government effectiveness and improved risk profiles emerges as a potential contextual enabler to these successes. At the same time, there are likely to also be other factors also constitutive of this enabling environment (for example, related to norm change, diet, education of a middle class, aspects of political change).

In LMICs, the relationship between labour productivity increases and poverty reduction is also strong. Their relationship with environmental degradation is not as strong as in LICs and UMICs, but economic transformation still corresponds to increased consumption and pollution. This suggests a need to work intensively on improving sustainability in LMICs before they become UMICs. There are some relative successes within LMICs over the period analysed, many located in South Asia. However, many of the relative successes perform particularly badly on air quality, suggesting that economic transformation and poverty reduction have been accompanied by degraded local environments (even if global footprints remain small). Governance conditions are also extremely varied and affect tripartite outcomes. Complementary measures to mitigate risk and promote pro-poor governance could contribute to more inclusive and sustainable outcomes.

The relationship between productivity increases and poverty reduction in UMICs is weak. Countries that are doing well in this group tend to see improvements in risk management and government effectiveness, though these appear to be necessary but not sufficient enabling conditions to promote tripartite outcomes. These findings suggest that additional factors such as social policies beyond economic development may play a stronger role in driving poverty reduction in UMICs. It also means that UMICs could theoretically shift away from economic growth-oriented policies (which drive significant increases in material footprint and GHGs for limited human gains) in favour of measures oriented to poverty and inequality reduction with smaller environmental impacts.

What might these social policy measures to disrupt the relationship between poverty reduction and ecological degradation look like? An example would be to strengthen secondary education provision, which is observed to consistently be associated with sustained escapes from poverty (Diwakar and Shepherd, 2022), yet often account for just a small share of carbon footprints of countries (Sun et al, 2021). Social protection and in particular cash transfers also hold promise in increasing recipients’ willingness to pay for environmental services (Nawaz and Gul, 2022). Environmental conditional cash transfers are observed to contribute to increase environmental management activities and reduce deforestation without harming household-level socioeconomic indicators (Alix-Garcia et al, 2019). At the same time, wider efforts are needed that are aimed at growing sustainably in terms of consumption and production patterns. In this context, supporting norm change could involve: 1) choices that involve demand and supply side changes, such as selection of low-carbon technologies for heating or travel, and 2) demand side choices, for example towards lower carbon diets and limiting aviation demand (CCC, 2022). Since poverty reduction appears to rely less on productivity changes in UMICs, however, it may be possible to disrupt its relationship with ecological degradation through these types of careful social and environmental policies alongside norm change.

Going forward, more in-depth analysis to understand reasons behind countries’ improvements in tripartite outcomes is a natural way forward. This process has been initiated in select countries, for example by Pickard and Lemma (2022) in a companion study. Drawing lessons from tripartite processes in Sri Lanka, Dominican Republic, and Thailand, they identify key drivers of tripartite outcomes, namely: windows of opportunity in responding to crises, perceived ‘apolitical’ authorities through increased alignment with international frameworks such as the SDGs, and ‘bottom-up or outside-in routes to the seat of power’ whereby tripartite ‘ideas can either form at lower levels of government or begin in one nexus realm and then move to the centre of policy making’ (Pickard and Lemma, 2022: 48). Finally, examining the extent to which these tripartite processes and outcomes are evidenced in high-income countries offer additional avenues for further research.

Funding

This work was supported by Sida. The findings and conclusions do not necessarily reflect the position of Sida. Any errors remain the author’s own.

Conflict of interest

The author declares that there is no conflict of interest.