Identifying Sustainability and Knowledge Gaps in Socio-Economic Pathways Vis-à-Vis the Sustainable Development Goals

Abstract

With the Sustainable Development Goals (SDGs), the global community has set itself an ambitious development agenda. Current analytical and quantitative modeling capabilities fall short of being able to capture all 17 SDGs and their targets. Even highly ambitious and optimistic pathways currently used in research, such as SSP1/SSP1-2.6, do not meet all SDGs (sustainability gaps) and fail to provide information on some of them (knowledge gaps). We show that for research and modeling purposes, the SDG targets can serve as a basis but need to be operationalized to reduce complexity and also to account for long-term sustainability concerns beyond 2030. We have explored here the requirements for assessing more comprehensively the sustainability of development pathways, guided by holistic interpretation of the SDGs to enable an assessment of the potential embedded synergies and trade-offs between the economic, social and environmental objectives. We see this as call for action for science to work on filling these gaps. At the same time, this is also a call for policy makers and the global community to close the sustainability gaps that emerge from such analysis. We anticipate that such analysis will provide useful information for policy advice and investment decisions during implementation of the UN 2030 Agenda.

1. Introduction

In 2015, the United Nations General Assembly adopted the Sustainable Development Goals (SDGs) (UN 2015), which define an ambitious development agenda to be achieved by 2030. Comprised of 17 goals and 169 associated targets, the SDGs provide quantitative and qualitative specifications of the economic, social and environmental dimensions of sustainable development. While the SDGs should be pursued in a comprehensive and integrative manner, there exist currently no integrated development pathways that achieve the 17 SDGs. Proving the feasibility of sustainable pathways that meet all SDGs can provide helpful guidance to the global community during the SDG implementation process. The key question is: how can all 17 SDGs be achieved together?

The potential for diverse trade-offs and synergies across SDGs has been recognized (Griggs et al. 2014; ICSU and ISSC 2015; McCollum et al. 2017; Nilsson et al. 2016). It has been shown, for example, that through tackling multiple goals, such as climate mitigation, air pollution control and energy access, at the same time, overall costs can be lowered substantially (McCollum et al. 2011). Technological innovation, efficiency gains and transformation of behavior can help resolve some of the trade-offs, which are currently observed or expected across social, economic and environmental policy objectives. Hence, trade-offs should not necessarily be viewed as permanent, but rather as an indication of transformation needs.

With the Shared Socio-economic Pathways (SSPs) (Riahi et al. 2017) the global change research communities have developed scenarios for integrated assessments, which are used for example by the Intergovernmental Panel on Climate Change (IPCC) and which already cover several of the SDG targets (e.g., SDG13). However, the SSPs do not comprehensively cover all SDGs, which illustrates the need to develop novel integrated pathways.

In this paper, we provide a starting point for a more comprehensive assessment of sustainable development pathways by describing the sustainability and knowledge gaps within the SSPs. We show that even the most optimistic pathway currently used by scientific community falls short of several dimensions of sustainable development. To support the UN 2030 agenda scientifically and prove the SDG’s feasibility, we call on the research community to expand existing pathways and methodological approaches to systematically assess what is required to achieve the SDGs and sustainable development for all thereafter. This analysis is intended as a contribution to The World in 2050 (TWI2050) initiative1, which seeks to scientifically assess the feasibility of achieving the SDGs in unison and which will build parts of its analysis on our work.

Section 2 provides background on the methodology. Section 3 introduces the results. Section 4 provides a discussion and concluding remarks.

2. Methodology

2.1. Applying the Sustainable Development Goals (SDGs) in the Long-Run

The SDGs constitute the foundation of our analysis, as they define the politically agreed social, economic and environmental aspirations of the international community (UN 2015). The majority of the SDGs have 2030 as their running period, but some of the targets should be realized already by 2025 or even 2020. While the SDGs provide a first reference point, assessments of the sustainability of development pathways require longer time horizon. This applies especially when looking at earth-system processes where changes might only materialize over decades (or even centuries). Climate change is such an example, where a long-term challenge of stabilizing temperature increase has implications for near term decisions concerning for example energy and transportation systems and the fulfilment of several other SDGs and vice versa (Sperling et al. 2016).

To reflect this,

Table 1. Sustaining and strengthening Sustainable Development Goal (SDG) targets beyond 2030.

SDG Target Classification beyond 2030	Rationale	SDG Targets
A. Sustain 2030 SDG target levels	A. Future efforts have to focus on sustaining the SDG target levels. Most of these targets relate to universal access to services (e.g., health) or eradicating adverse conditions (e.g., illnesses).	1.1, 2.1, 2.2, 3.3, 3.8, 4.1, 5.1, 6.1, 6.2, 7.1, 8.7, 9.1, 10.3, 11.1, 11.2, 12.2, 14.4, 15.2, 16.2, 17.1 *
B. Increase ambition to close sustainability gap	B1. Further effort required to close remaining gap (e.g., protect cultural heritage).	1.2, 4.4 , 8.6, 9.3, 10.2, 10.3, 11.3, 11.4, 15.6, 16.1, 16.4
	B2. Dynamic update required to ensure progress along target dimensions (e.g., maternal mortality, efficiencies).	1.5, 2.3, 3.1, 3.2, 3.4, 3.5, 3.6, 3.9, 7.3, 10.6, 11.5, 13.1, 13.2, 13.3, 16.5, 16.8
	B3. Further effort required to stay within Planetary Boundaries (environmental & earth systems) (e.g., pollution).	6.3, 6.4, 7.2 , 8.4, 9.4, 11.4, 11.6, 12.3, 12.4, 12.5, 12.6, 12.7, 14.1, 14.2, 14.3, 14.5
C. 2030 SDG targets might not apply	C. Reevaluation required to decide on ambition level beyond 2030.	8.1, 8.2, 9.2, 10.1,

* All targets that are not listed in rows B and C fall within A.

is categorizing the possible evolution of the numeric SDG targets (Appendix A) beyond 2030.2 This is a preliminary assessment of how the targets relate to long-term sustainability. Several SDG targets do not come with absolute target values but provide only an indication whether the related indicator value should increase or decrease, thereby requiring further interpretation. Many SDG targets cover multiple issues and indicators, which poses a challenge when classifying them. We identify three categories of SDG targets:

• (A) SDG targets where the level of progress reached by 2030 (or before) should be sustained indefinitely thereafter (e.g., universal access to services meeting basic human needs such as food, water, energy, education, and health services).
• (B) SDG targets that constitute not a complete achievement by 2030, requiring further increases in ambition thereafter, as they are (B1) not ambitious enough considering current best practices (e.g., protecting cultural heritage); (B2) do not consider possible future improvements (e.g., efficiency); or (B3) do not consider planetary boundaries (e.g., environmental targets).
• (C) SDG targets, which will have to be revisited post 2030 as they may not apply in the long run.

Several targets and suggested indicators are not directly suitable for operational research and hence require further interpretation. Furthermore, the SDGs are very broad in scope. This poses a challenge for research groups, which need to arrive at a concise set of indicators representing sustainable development in 2030 and 2050 and beyond in their work. It will require reducing the complexity of the full set of SDG targets, while minimizing information loss. Targets selected need to be suited for scientific analysis, enable inter-comparison of results, while also being recognizable to decision-makers as representative of the SDGs.

2.2. Identifying Sustainability and Knowledge Gaps in Shared Socio-economic Pathways (SSPs)

Scenarios and pathways provide alternative futures and can be used to evaluate different policy options (Gea 2012; Nakicenovic et al. 2000). The SSPs (Riahi et al. 2017) are one example of such a framework. The research community developed the SSPs originally to inform climate change research for the assessments of the Intergovernmental Panel on Climate Change (IPCC). The SSP framework represents a suitable starting point for building more comprehensive analyses of sustainable development pathways, as it has undergone rigorous review, has been developed by a diverse research community, is known to policy makers and has been applied by modeling teams globally. The SSP framework provides five broad development trajectories that describe alternative futures. These differ in population development, economic growth, degree of cooperation, environmental awareness and several other characteristics.

We take SSP1: Sustainability—taking the green road (O’Neill et al. 2017) as a starting point. By doing so, we chose a transition to a low-emissions future with a narrative that comes closest to the intention of the SDGs: “The world shifts gradually, but pervasively, toward a more sustainable path, emphasizing more inclusive development that respects perceived environmental boundaries. Increasing evidence of and accounting for the social, cultural, and economic costs of environmental degradation and inequality drive this shift. […]” (full narrative in Appendix B). SSP1 describes a world that is quite different from the past development trajectory and political trends the world is currently experiencing. In SSP1 the global community works together to fulfill development targets. In more detail, scenario runs of SSP1 that meet the Paris Climate Agreement goal of holding warming to approximately 2 °C (above pre-industrial levels) through limiting anthropogenic forcing of the climate system to 2.6 W/m² by 2100 are followed (van Vuuren et al. 2011, 2017). We will refer to these runs as “SSP1-2.6”. SSP1-2.6 runs were chosen as they already meet the challenge of mitigating climate change and as it is the most ambitious and sustainable possible future within the SSP framework.

We will match SSP1/1-2.6 with the 126 numeric of the 169 SDG targets to see which SDG targets are met by SSP1/1-2.6 and where sustainability and knowledge gaps remain. We compare SDG targets with quantifications and model outputs based on SSP1/1-2.6. The SSP Public Database (Version 1.1 (IIASA 2016)), the Wittgenstein Centre Data Explorer (Lutz et al. 2014) and literature using the SSP framework provide the data for this assessment. Model runs available in the database have been contributed by the following model frameworks: AIMES/CGE, GCAM4, Image, MESSAGE-GLOBIOM, REMIND-MAGPIE and WITCH-GLOBIOM3. Depending on the model, some of the SSP data we assess are exogenous (e.g., population and education projections), others are model outputs (e.g., share of renewable energies).

We initially scanned the numeric SDG targets and SSP database and literature to identify where matching information was available. Many SDG targets cover several topics that need more than one indicator or dimension to be represented appropriately. SDG targets can also be indicators themselves. In some cases these are the same or similar to the indicators suggested by Inter-Agency and Expert Group (IAEG) on Sustainable Development Indicators (United Nations Economic and Social Council 2017). We chose to focus on the SDG targets as closely as possible as they were also adopted together with the UN 2030 agenda. The closest matching indicator available from the SSP database or literature was used. Where target values are missing, it is a judgement call if a target has been achieved (e.g., SDG target 7.2 “By 2030, increase substantially the share of renewable energy in the global energy mix,”). In general, the SDG targets leave ample room for interpretation. Appendix C gives an overview of the selected indicators and for which SDG target we used them as a proxy.

A sustainability gap shows that the select indicator within SSP1/1-2.6 does not meet a specific SDG target or indicator by 2030. A knowledge gap indicates that no quantitative information is available within the SSPs to provide information on an SDG target. As noted above, classifying and assessing SDG targets in this context poses several challenges. The classification provided in Section 3 has to be viewed as an approximation. The following sections provide detail on the individual indicators assessed.

3. Results

Figure 1 shows how SSP1/1-2.6 aligns with the SDGs and their targets. It shows which SDG targets are reached by SSP1/1-2.6 (category I) in time. It shows the sustainability gaps (e.g., addressing hunger or education) (category II) that remain in 2030 and in some cases even by 2050 or 2100 within SSP1/1-2.6. Figure 1 also introduces the knowledge gaps (category III) in the SSP framework, which does not quantitatively cover the majority of SDG targets (e.g., access to water or gender equality) or which only implies them within its narratives. This provides an indication of where future work is needed for research groups developing pathways and tools, going beyond a climate-only focus, to capture all SDG dimensions. In many cases, this will not be possible via quantitative indicators only. Narratives can be used to cover qualitative elements such as governance, cooperation, or behaviors.

We now provide more detail on this assessment of how SSP1/1-2.6 pathways perform vis-à-vis the SDGs and their targets. We will indicate which representative indicator we are referring to for each target and which assumptions we base our assessment on. We will present background on each target that is met (category I) and where sustainability gaps persist (category II).

3.1. Category I Where SSP1/1-2.6 Runs Are Fully or Partly in Line with SDG Targets

3.1.1. SDG Targets 3.1–3.7, 3.9 on Decreasing Mortality and Access to Reproductive Services

SDG 3 and its 13 targets focus on tackling child and maternal mortality, the major infectious diseases, non-communicable diseases, sexual and reproductive health, as well as providing universal health coverage. Possible future trajectories of these individual causes of death can be translated to aggregate mortality trends where achievement of the SDG targets can be well represented by a general low mortality trajectory (Abel et al. 2016). SSP1 represents an overall low-mortality pathway, which is seen through increases in life expectancy globally, covering close enough SDG targets related to decreasing mortality (3.1–3.6, 3.9). Fertility is low in SSP1 (Kc and Lutz 2017), which is connected to the targets on sexual and reproductive health-care service (3.7) and (5.6). SDG 3 is also closely linked to SDG 5 on gender equality and SDG4 on education. Lutz et al. (2014) show that higher educational attainment of women leads to lower fertility rates.

3.1.2. SDG Targets 7.2 and 7.3 on Renewable Energy and Energy Efficiency

SSP1-2.6 runs mostly meet the energy efficiency (7.3) and the renewable energy targets (7.2). The latter does not provide a concrete target value quantification but calls for a “substantial increase” (UN 2015). The share of renewable energy in secondary energy increases twofold in most SSP1-2.6 runs, to 17.6% to 28.4% in 2030 (IIASA 2016), where the majority of runs is above the 2015 level of around 18% (UN 2016). Annual energy intensity falls from 5.7 MJ/US$2005 in 2012 (UN 2016) to 3 to 4 MJ/US$2005 in 2030 in SSP1-2.6 runs (IIASA 2016). In the long run, improvements in energy intensity outpace historic rates (van Vuuren et al. 2016).

3.1.3. SDG Targets 8.1, 8.2, 8.5 on Economic Growth and Unemployment

SSP1 follows a contraction and convergence approach: income levels of less developed countries will increase faster than those of more developed economies (Dellink et al. 2017). The growth rate of per capita GDP in least developed countries in SSP1 is 7.3% in 2030 (IIASA 2016), which meets the 7% limit of SDG target 8.1. If one however considers a longer-term average, i.e., 2025 to 2035, the per capita growth rate is 6.5% in LDCs. The global growth rate of 3.9% per capita (IIASA 2016) indicates higher levels of economic productivity (8.2). Unemployment is close to 2% in the long run in SSP1 (Dellink et al. 2017), which partly covers 8.5 on work for all, yet fails to cover the quality of work.

3.2. SDG Target 9.4 on CO₂ Intensity

SSP1-2.6 achieves SDG target 9.4 on more sustainable industries in terms of a decrease in CO₂ intensities. From around 0.45 MtCO₂/year/billion US$2005 in 2015 (UN 2016) SSP1-2.6 runs reach 0.12–0.22 MtCO₂/year/billion US$2005 in 2030 (IIASA 2016). However, no indicator beyond CO₂ intensities, such as recycling rates or circular economy that could characterize sustainable industries is available in the SSPs.

3.2.1. SDG Target 10.1 on Income Inequality

SSP1 provides information on income inequality across countries. The income Gini across countries falls from 0.66 in 2015 to 0.43 in 2030 (Dellink et al. 2017; IIASA 2016), as the income per capita of poorer economies grows stronger than that of more industrialized countries. SSP1 projections on income inequality within countries also see declining Gini coefficients in all countries (Gidden et al. n.d.).

3.2.2. SDG Target 11.6 on Air Quality in Cities

SSP1-2.6. includes data on population living in areas affected by dangerous air pollution, partly covering target 3.9 on reducing the number of deaths and illnesses from pollution and contamination and 11.6 on air quality in cities. SSP1 accounts for significant improvements of air quality based on “faster rate of pollution control technology development, with greater effectiveness as compared to current technologies” (Rao et al. 2017). In addition, medium- and low income countries will quickly adapt best available technologies and experience strong enforcement of environmental laws. Despite large improvements in climate mitigation policy and energy access in SSP1-2.6, parts of the population are still exposed to levels of pollution above WHO recommended levels of PM 2.5 in 2030, especially in Asia (13% of population exposed to PM2.5 levels of above 35 μg/m³), the Middle East and Africa (8% above 35 μg/m³) (Rao et al. 2017). Details on urban pollution levels alone are not available. We conclude that the reduction in people affected by air pollution in general will result in improved air quality in cities and reduced deaths from air pollution, meeting part of the targets 11.6 and 3.9.

3.2.3. SDG Target 15.2 on Halting Deforestation

SSP1-2.6 runs perform well in terms of land-use change and meeting global aspirations under the SDGs. In most runs deforestation comes to a halt, in some forest cover increases slightly (start value in 2015: ~4000 million ha. SSP1-2.6 range for 2030: 3881–4135 million ha) (IIASA 2016). However, while halting deforestation is an explicit target, this leaves considerable room for interpretation. Given that the SDGs are global goals, should this target only be applied at the global level or also at regional or national levels? How should the ambition level with regards to conserving and managing natural resources change between 2030 and 2050 to ensure long-term environmental sustainability? In this context, further consideration may be given to assessing changes by forest types or biomes and where such changes occur. This will also be important with regards to assessing progress towards other environmentally relevant SDG targets, such as halting biodiversity loss (target 15.5). There is the need to combine the quantitative analysis with additional qualitative considerations. The recently launched Forests, Agriculture, Biodiversity, Land, and Energy Project: Pathways for Sustainable Land Use (FABLE) (IIASA 2017) will further help in assessing interlinkages between development and environmental objectives in the land-use context across scales.

3.2.4. SDG 13 on Climate Change

The SDGs need to be achieved while limiting global warming and ensuring resilience to climatic change, which poses an additional constraint when assessing potential synergies and trade-offs between the various goals (Sperling et al. 2016). SDG 13 specifically refers to the United Nations Framework Convention on Climate Change (UNFCCC) as primary decision-making body on international climate change policy (UN 2015). The Paris Agreement reached in 2015 set the objective of limiting global warming to well below 2 °C above pre-industrial levels, while also encouraging efforts to further confine the warming to only 1.5 °C (UNFCCC 2015). SSP1-2.6 runs generally show a high probability of staying below 2 °C, but a scenario with a more stringent carbon budget would be needed to meet the additional ambition level. The targets under SDG 13 on education, resilience and policies fall under category III of our assessment.

3.3. Category II Where SSP1/1-2.6 Runs Do Not Meet SDG Targets in 2030 (or Later) (Sustainability Gaps)

3.3.1. SDG Targets 1.1 and 1.2 on Absolute and Relative Poverty

Assuming a log-normal distribution of income and its relation to poverty (Lopez and Servén 2006), national GDP (IIASA 2016) and population projections of SSP1 (Lutz et al. 2014), household consumptions shares of GDP (which were held constant) (World Bank 2017a) together with national income inequality projections (Gidden et al. n.d.) were used to calculate the share of population living in absolute and relative poverty, respectively. The absolute poverty threshold of 1.90 $2011/person/day (World Bank 2015) was also assumed for 2030. The number of people living in absolute poverty then falls from around 700 million (9.6% of global population) in 2015 (PovcalNet 2015 ) to around 350 million (4%) in 2030 in SSP1, missing the SDG target of ending absolute poverty.

Relative national poverty is based on nationally defined threshold values. For illustrative purposes we apply the widely used poverty measure of proportion of people living below 60% of their country’s median income (Eurostat 2017). In countries where this would correspond to below 1.9 USD per day, the absolute poverty line represents the minimum. The population living in relative poverty according to this measure increases in SSP1 from around 1.8 billion in 2015 (26% of global population) to 2 billion in 2030 (25% of global population). Thus applying such a threshold value would result in missing the SDG target of halving the proportion of population living in relative poverty in 2015 by 2030, which would correspond to 13% or around 1 billion people. SSP1 achieves both poverty targets only in the second half of the 21st century.

3.3.2. SDG Target 2.1 on Undernourishment

SSP1-2.6 runs do reduce the people living at the risk of hunger by 2030, yet around 180 million people remain undernourished (Hasegawa et al. 2015), failing the SDG target 2.1. By 2050 though, hunger is eradicated in SSP1.2-6. With increasing economic prosperity within SSP1, malnourishment should also be regarded in terms of the risk of obesity. This needs to be further assessed within the context of the SSPs, including the quality of the food intake, beyond caloric value, such as micronutrients.

3.3.3. SDG Target 4.1 on Universal Secondary Education

SDG 4 focuses on “ensuring inclusive and equitable quality education and promote lifelong learning opportunities for all” and is split into ten targets. Target 4.1 on primary and secondary education for all by 2030 is the most prominent and specific one. It also covers content of other targets (e.g., 4.5 on equal access, 4.6 on literacy). Higher educational attainment is linked with several benefits of human development, from higher productivity, income and health to stronger political agency, environmental awareness, risk resilience and gender equality (Lutz et al. 2014).

The population projections of SSP1 rely on high levels of investments in education across all countries which accelerates the demographic transition (Kc and Lutz 2017). In the rapid development case of SSP1, these investments result in high advances in education rates which are based on a combination of the medium future trajectory for progression rates rooted in the experiences of the past four decades of all countries and the assumption that educational advances are accelerating with South Korea, which experienced the most rapid expansion recently, severing as a benchmark.

In 2010, more than 1.1 billion adults (17% of the population) had no education at all or had not completed primary school (Lutz et al. 2014). With 57% of global population with primary, 44% with lower and 27% with upper secondary education in 2010 (Lutz et al. 2014), target 4.1 cannot be fulfilled anymore in time as too many cohorts have fallen out of the education system already to catch up in 15 years, independent of interpreting “secondary education” as completing lower or upper secondary education. Also, enrollment and graduation rates do not provide any insights on quality of education. The World Bank (2017b) estimates that about 250 million children in the world cannot read or write despite having attended primary school for three years or more.

According to SSP1 (Kc and Lutz 2017; Lutz et al. 2014) in 2030, 71% of global population will finish primary education, 60% lower secondary and 44% upper secondary education, leaving 10% with no or incomplete primary education and another 11% with primary only. By 2100, the SSP1 comes very close to achieving target 4.1 with 91% finishing primary education, 87% lower secondary and 81% upper secondary education and 52% post-secondary, leaving only 1% without any education or incomplete primary (the remainder 8% are children and youth below 15 years). SSP1 is more comprehensive in terms of education than SDG 4 as it presents also high shares of tertiary education, including for women, which is not covered quantitatively by SDG 4.

3.3.4. SDG Targets 4.3 and 4.5 on Education and Gender

Education is the only area of the SSPs where implications on gender equality are included. It has been shown that higher educational attainment of women has a direct effect on fertility rates (Lutz et al. 2014). This is reflected in SSP1. The trickle-down effect via fertility and mortality on population growth will impact several other SDGs, especially concerning basic human needs (access to food, water and energy) and human impacts on the earth systems. Within SSP1, the gender gap in the number of mean years of schooling among the whole adult population is decreasing from 0.78 in 2015 to 0.5 years in 2030 (Kc and Lutz 2017; Lutz et al. 2014).

3.4. Category III Wherenot Enough Quantitative Information is Available within the SSPs (Knowledge Gaps)

The SSP framework or related analyses do not cover all of the SDG targets quantitatively. Instead several SDG targets (Category III) are qualitatively described in the SSP-narrative (e.g., technological innovation, cooperation, environmental awareness, access to services, see Appendix B). Some targets can be assumed implicitly included in the framework as they are closely related to other indicators. One example would be 3.8 on universal health coverage. The low mortality rates and increasing life expectancy within SSP1 provides an indication that health coverage and health services will improve. In the narrative increasing health investments are also mentioned. Yet no clear indicator is available within the SSPs to assess the degree of coverage of health services. In other cases, there is no information on an entire SDG in the SSPs. This is for example the case with regard to global goals dealing with gender (SDG 5) or the marine environment (SDG 14).

4. Discussion

The UN 2030 Agenda frames an ambitious global development agenda with a holistic view of sustainability and calls for policy coherence. Scientific assessments can provide guidance to the SDG implementation process, helping to identify options for mitigating potential trade-offs and maximizing synergies between the multiple social, economic and environmental objectives. However, current analytical and quantitative modeling capabilities fall short of being able to capture all 17 SDGs and their targets. In addition, long-term sustainability concerns need to be considered beyond the 2030 time-horizon of the SDGs.

IAMs implementing the SSPs are especially strong in certain environmental and resource use areas and are weak in governance and human systems as can be seen in our assessment. But the SSPs are well positioned to serve as the basis for a comprehensive framework that allows different research fields to work together and increase interactions and linkages across disciplines where they are being adapted to the respective needs and capabilities of the models (see for example Hegre et al. (2016) who use the SSPs to forecast conflicts). Yet, we believe it is timely that the following aspects are also being considered in any further enhancement of IAMs and their scenario framework of the SSPs, so that the work remains relevant for policy makers and provides scientific findings supporting policy coherence across the SDG domains.

We have explored here the requirements for more comprehensively evaluating the sustainability of development pathways, guided by holistic interpretation of the SDGs to enable an assessment of the potential embedded synergies and trade-offs between the economic, social and environmental objectives. We have shown that sustainability and knowledge gaps persist in the pathways currently used by the global change community. Examining the overlap of SSPs and SDGs benefits the research community currently using the SSPs. The SSPs consist of narratives that guide quantitative scenarios of future socio-economic development, foremost scenarios on population and economic development that are then used by models, especially IAMs, to provide estimates of future development in areas such as the energy or land-use sectors. Despite SSP1/SSP1-2.6 being a very ambitious and optimistic scenario for future global socio-economic development, the pathway does not meet several SDG targets.

The long list of SDG targets, which cannot be assessed by the current SSPs provides us with a diverse research agenda. For research and modeling purposes, the SDG targets can serve as a basis but need to be operationalized to reduce complexity. We see this as call for action for science to work on filling the knowledge gaps. Research is needed to develop integrated sustainable development pathways to include as many SDG dimensions as possible and to test policy options that achieve them. Without adequate consideration of potential trade-offs, mitigation policies, for example, may lead to undesired outcomes, such as threatening lower income groups by constraining energy access (Cameron et al. 2016). The UN 2030 agenda requires going well beyond the climate policy focus of the SSPs. Both the models used for quantitative assessments applying the SSPs as well as the scope of the SSP narratives would benefit from broadening.

First, we propose broadening the SSP narratives to cover more domains relevant to the SDGs. This already means a big step forward, going beyond a climate change focus. Advances can be reached by more systematically developing the interfaces, where quantitative and qualitative analyses can talk to each other. Yet, broadening the narratives alone will not be sufficient in analyzing them quantitatively, if these domains are not reflected in models using the SSPs.

Second, we are also suggesting that the capabilities of the models used for the quantitative assessment, foremost IAMs, need to be enhanced. The current generation of IAMs can be regarded as ‘best compromise’ between different desirable properties in the model based analysis of system(s) transformation. On the one hand, they describe technological options and other sectoral characteristics in sufficient detail to be able to assess substantially different technological futures over the long run. On the other hand, their system boundaries are broad enough to capture endogenously most to all relevant effects of interest and interactions at the system level. For instance, IAMs typically cover all sources of emissions at global level. Partial coverage only (ignoring regions or emission sources) would not be sufficient to assess climate change outcomes. However, new challenges require these models to be further enhanced in the future—regarding both the level of detail and the system boundaries. Capturing more SDG dimensions in a modeling framework requires to expand the system boundaries of the models significantly compared to the status quo, since these sectors—e.g., health—typically are not yet covered by IAMs. Some SDG sectors require the development of new concepts that can be translated into mathematical equations to become accessible for modeling, such as concepts for theories for societal change, analogous to concepts of technological change that are already incorporated in IAMs.

Moreover, IAMs need to improve on the representation of sectorial structures that affect decision making in reality. Traditionally rooted in energy system planning IAMs classically have adopted a central planner logic, which may not accurately represent current market structures and decision making in the energy and other sectors. In order to enhance the degree of market realism, models may need to operate for example with a higher time resolution or better reflect the economic incentive structure by different actors. These often also depend on policy frameworks, which may also imply a higher regional resolution.

Trade-offs also arise from computational constraints, i.e., having too large and/or too detailed models with too many variables at some point is not anymore computationally feasible, but also causes too high complexities for the modeling analyst, i.e., managing the modeling workflow gets overly complex and interpreting the model behavior difficult. Including a plethora of indicators presents the risk of adding noise to the analysis. Endogenizing everything will not be possible and is also not our objective. Ideally, IAMs can provide some information for possible future developments for each of the SDGs based on a set of representative indicators.

So what could be a suitable way forward? A framework that can integrate both sectorial details with cross-sectorial linkages and macroeconomic completeness is known as top-down-bottom-up hybrid modeling (Holz et al. 2016; Hourcade et al. 2006). However, even if advances in computational power and solver algorithms would allow to build a single hybrid model that captures everything this may not be desirable due to the other mentioned drawbacks. Instead one could implement all the needed details in the models, but keep them at a manageable size and link them to each other as separate modules within a top-down-bottom-up hybrid framework, where the top level model(s)—e.g., a computable general equilibrium model—acts both as an accounting framework and reference model for the identification of equilibrium outcomes. The linkage would not have to be done manually, but could be controlled by some learning algorithm. Recent development in IAM modeling with the development of modeling frameworks rather than single models, are already a step in this direction.

The objective is neither to arrive at one single model that covers all SDG targets nor that all models should be expanded to cover all SDG areas. Different IAMs provide valuable details in different sectors and at different levels of granularity (e.g., global versus national models, sectoral versus macroeconomic models), crucial information that should not be lost. While the SDGs will have to be implemented by nation states, there is benefit in linking and comparing regional, national and global level assessments in order to capture potential knock-on effects across scales and review implications for sustainability. We do not imply that all 126 (numeric) SDG targets (or even all suggested indicators) should be covered by either models or policy makers in their work.

At the same time, policy makers are urged to implement the 2030 Agenda to close the sustainability gaps that emerge from our analysis, as even within an optimistic case several dimensions are not achieved in time. Our analysis underlines the importance of the SDG notion of “leaving no SDG behind”.

This analysis is a contribution to the TWI2050’s objectives of framing the solution space for sustainable development pathways. We anticipate that such integrative analyses put forth by initiatives such as TWI2050 and the IAM community will inform the implementation of the UN 2030 Agenda. Together with regional and national level assessments, this can strengthen the foundation for strategic policy decisions, which integrate sustainability concerns across spatial and temporal scales.