Quantifying lifestyle based social equity implications for national sustainable development policy

The aim of this research is to address the challenge of achieving more equitable social outcomes through a reduction and fairer allocation of environmental burdens, and in doing so, contributing to national sustainable development policy. This novel study demonstrates the nature of societal outcomes through the lens of inequity with respect to lifestyle related environmental footprints and stakeholder preferences. Footprints are derived using input-output analysis, while environmental issue preferences and potential remedial actions are identified using a national survey. To highlight the value of the broadly applicable framework, here we demonstrate a case study of Japan, which is interesting due to shifting demographics engendering an aging, shrinking population. Key findings include that the mitigation of environmental footprints in line with household preferences can positively influence both societal equity outcomes and contribute to closing the gap between rich and poor. Importantly, broad participation, i.e. participation irrespective of income level, is shown to be more effective than participation from a single sector. These findings can assist policymakers to develop policies which are responsive to societal preferences and demographic trends while also furthering the debate toward clarifying norms for acceptable levels of social equity.


Introduction
Demand-side or consumption-based approaches focusing on people's lifestyles and the factors that influence them are essential, both for climate change mitigation and also for sustainable development which integrates environmental, economic and social issues (Bertram et al 2018, Creutzig et al 2018. Consumption-based accounting, a widely accepted, recognized methodology may offer a way to quantify lifecycle environmental pressures via international supply chains (i.e. environmental footprints) and their allocation to final consumers (Peters 2008, Wiedmann 2009, Nansai et al 2012, Hertwich and Wood 2018. The application of this methodology has also shed light on the socio-environmental inequalities engendered by uneven distribution of social and environmental burdens (Boyce et al 2016, De Schutter et al 2018, across nations and regions (Peters et al 2011, Simas et al 2015, Moran et al 2018, Wiedmann and Lenzen 2018, Zhang et al 2018, Nansai et al 2020b. For example, consumption by only 5% and 10% of the population in richer countries contributed to global energy consumption and greenhouse gas (GHG) emissions at a higher rate than for the entire bottom 50% (Hubacek et al 2017a, Oswald et al 2020). These inequalities were identified using environmentally extended input-output analysis (EEIOA).
Our lifestyles, expressed via household consumption, impart a substantial contribution toward our environmental footprints (Tukker et al 2010, Ivanova et al 2016, leading to focused investigations over the last two decades to detail these contributions (Hertwich 2011, Wiedenhofer et al 2018. A key finding over this period highlights that household environmental footprints are not distributed evenly across income levels at both the national and global scale. For example, the difference in per-capita carbon footprints between the highest and lowest income groups was about 1.4 times in Japan (Shigetomi et al 2016), 4.5 times in the UK (Chitnis et al 2014), 13 times in China (Wiedenhofer et al 2017), and 14 times at the global level (Hubacek et al 2017b).
Alleviation of both environmental pressures and resultant inequalities is a prerequisite for achieving the 2015 United Nations Sustainable Development Goals (SDGs; United Nations 2015). For policymakers, it is essential to consider how to mitigate intricate environmental and social issues resultant from economic activity considering synergies and trade-offs (Liu et al 2018, Parkinson et al 2019, Vanham et al 2019. It is also crucial to identify how to encourage members of society to engage with policy so as to reduce environmental impacts while also yielding a personal benefit, consistent with their ideals (Vallet et al 2019). The footprinting approach contributes to illustrating the structure of pressures and inequalities linked with final consumption as presented in the above precedential studies. However, these studies have not assessed inequality under multiple environmental pressures. They also overlook stakeholder environmental preferences (in this case, those of household consumers) that will likely have an impact upon preferred abatement measures, in line with environmental justice ideals (Jenkins 2018). By addressing these challenges simultaneously, an improvement in environmental quality and amenity through a reduction in environmental burdens and a fairer allocation among stakeholders can be engendered, resulting in positive social equity outcomes (Finley-Brook and Holloman 2016).
Against this backdrop, this study develops a novel, holistic, inclusive approach which allows for the identification of emerging societal inequities due to disparity in the allocation of multiple environmental pressures and differing preferences and priorities for this allocation among stakeholders. Specifically, we aim to quantify social equity outcomes resulting from household consumption considering future trends of various environmental footprints, and by incorporating stakeholder values and priorities related to each footprint. By considering household consumption reflective of stakeholder behavior and preferences for the environment and society in the future, we can begin to extract the factors which shape stakeholder's environmental preferences, and how these may impact upon future consumption behavior. In this research, consistent with our evaluated stakeholder group's values, we consider a future society which reduces its environmental footprints and avoids their allocation to lower income groups to be a 'more equitable' outcome (i.e. vertical equity; Mooney and Jan 1997).

Framework for quantifications of environmental burden and social equity
This study proposes the Input-Output Analysis Sustainability Evaluation Framework (IOSEF), consists of three steps: (1) estimation of households' environmental impacts, (2) measurement of stakeholder preferences to mitigate those impacts, and, (3) quantification of the distributional equity impacts of environmental burdens across households, based on potential future remedial actions. Note that in this framework, environmental impacts, stakeholders, and distributional equity impacts are defined as the environmental footprints, householders who create the footprints by income bracket, and the size and bias of those footprints, respectively.
The procedures for the first step of the IOSEF are based on EEIOA in combination with a consumer expenditure survey, in line with precedential studies where both i and j denote goods and services (i.e. commodities) in the input-output table (IOT). Superscript t denotes the analyzed year. q (t) ik represents the direct environmental pressure k (k = 1…K) per unit of output for commodity i. L ik represents the direct environmental pressure from unit expenditure (e.g. CO 2 emissions generated by driving).
In the second step, household preferences for year t, p (t) bk , are incorporated as a proxy for stakeholder 'malleability' , representing potential future behaviors which reduce environmental footprints based on perceived personal or community merit (e.g. health information promoting footprint mitigation; Herrmann et al 2017). Hence, p (t) bk are used as weightings for each of the household environmental footprints. Each range of p (t) bk is defined by a large-sample, representative survey which quantified householder importance weightings for issues around sustainability. The questions in this survey focused on the analyzed footprints, but also addressed aspects of lifestyle convenience and desirable income levels, enabling intelligence gathering on issues outside of the environment. A detailed explanation of the methods for determining household preferences using a national survey is elaborated in the supporting information (SI (https://stacks.iop.org/ERL/15/084044/mmedia)).
In the final step, an environmental burden score for each household bracket is calculated by dividing each expected footprint ) ) during the analyzed period and the number of footprints (K). Next, we prepare household consumption share by bracket ( S . Following this, area weighted centroid coordinates ( of the resulting polygon are calculated using E , and the geometric operation for area weighted centroid calculation, as shown in equations (2) and 3.
Finally, the burden distribution (BD (t) = X (t) − 50), and social inequity score (SI (t) = 2Y (t) ) are determined based on these centroid coordinates. A BD (t) identifies the household brackets which are causing the greatest level of environmental burden in society (due to their consumption), relative to average consumption (i.e. the 50% mark on the x-axis). Hence, a positive BD (t) indicates that the majority of environmental burdens are generated by higher income households. SI (t) denotes the amount of environmental burdens in each household bracket and the bias between brackets. SI (t) has a maximum value of 100 and incorporates stakeholder importance weightings for each environmental footprint. A low SI (t) is desirable, as this indicates improving social equity outcomes due to reduced environmental burdens and discrepancy among households. The method for quantifying these scores builds on and adapts previous scholarship which outlined social equity outcomes of energy system and policy issues including fossil-fuel retirement (Chapman et al 2018), mega-solar siting (Fraser and Chapman 2018), bottom-up energy transitions (Chapman and Pambudi 2018), and social inequities considering consumption and shifting demographics (Chapman and Shigetomi 2018b).

Case study of japan using multiple environmental footprints and stakeholder behavior
To demonstrate the operation of the IOSEF, this study details trends in social inequity of household consumption in Japan during 1990-2040 and how they may be influenced by changing household behaviors over time. Japan is facing the social issues of an aging, shrinking population. It is therefore anticipated that shifting demographics, along with changing social and economic conditions have implications on both the values and priorities of stakeholders in different age and income groups (Chapman and Shigetomi 2018a). In this study, prediction of environmental footprints by income bracket was conducted based on a previous study (Chapman and Shigetomi 2018b). In addition, we defined 'active' and 'proactive' desirable future scenarios to demonstrate how burden distribution and social inequity may be affected by household preferences and responsive behaviors. The dataset utilized, desirable future scenario settings and assumptions are detailed in the SI.

Demographic trends and household footprint variation
An estimation of household environmental footprints from 1990-2040 is shown in figure 2. These estimates incorporate trends of household income group composition and consumption expenditure but do not account for weighting or stakeholder preferences. This estimation is utilized in the business as usual (BaU) scenario. Details of household income group composition and consumption expenditure trends from 1990-2040 can be found in the SI.
Between 1990 (Inagaki 2013, IPSS 2018, MHLW 2016. Notably, the mining risk household footprint (Nansai et al 2015) rapidly increased from 1990 to 2005 in response to an increase in the demand for information and communication devices such as personal computers and cellular phones. All footprints except for blue and green water peak around 2020, with the water footprint peaking in 1995.
Post-2020 household footprints are also projected including stakeholder preference weightings. Note that scenarios based on stakeholder preferences are applied from 2020 onwards, reflective of the timing of the national survey. Figure 3 demonstrates the potential difference in footprints between the BaU and active and proactive desirable future scenarios.
In comparing the BaU scenario with the other scenarios, footprints are estimated to be up to 9.7% lower under the active desirable future scenario and up to 29% lower under the proactive desirable future scenario by 2040. The largest difference is observed for the water footprint, followed by industrial waste, PM 2.5 and carbon. For neodymium mining risk footprints, no difference is observed between the BaU  and other scenarios as no preference to reduce them was expressed among households (i.e. all weighting factors were 0 during the studied period). As seen in figure 3, only the very low, low, and very high-income households are likely to achieve a reduction in carbon footprint due to their high level of environmental consciousness. These income brackets include the majority of elderly households who are likely to have higher consciousness about environmental protection and addressing climate change (Chapman and Shigetomi 2018a). It was also identified that very high-income households tend to have a higher education level (these two factors are heavily inter-related), which is linked to higher environmental knowledge and therefore consciousness. Importantly, if the very low-income bracket reduce their carbon footprint in line with their reported level of responsiveness to conducive public policy (i.e. by 5%), the total footprint would be decreased by some 8.5 Mt-CO 2 eq/y (0.91% overall). This anticipated reduction is 2.7 times larger than that of very high-income households due in part to a larger number of lower income households. Although all income groups show some propensity toward reducing waste, water, and PM 2.5 footprints, the low-income group had the largest contribution to overall potential future reductions.

Social inequity impacts considering stakeholder environmental burden preferences
Based on the results for environmental burdens as estimated in the previous section, burden distribution and relative social inequity scores were calculated as shown in figures 4 and 5 respectively. Burden distribution scores peak above 0, taking a positive value in 1995, however, after 1995 scores rapidly decline out to 2010, with a small increase occurring between 2010 and 2025. This trend as observed in figure 4 is partially due to marked increases in contributions from lower income households toward overall environmental footprints.  . Trends for burden distribution scores between 1990-2040. Household preference weightings are applied to desirable future scenarios from 2020 onward. The burden distribution score represents the origin of the majority of environmental burdens in terms of household income group relative to average consumption, and the shift in burden allocation over time. A negative score (as shown post-2000) indicates that the majority of environmental burdens are generated by lower income households, and an improving score indicates progress toward societally preferred levels of burden allocation.
According to national statistics, Japanese household's average income increased in the late 90's, compared to 1990 levels. However, after 1999, incomes reduced until 2010. After 2010, average income was slightly increased to 2015 (MHLW 2016). Without exogenous shocks to economic growth (e.g. policy intervention or natural disasters), average incomes are likely to decrease due mainly to the aging, shrinking population (Inagaki 2013). It is notable that although burden distribution scores are comparatively higher under active and proactive desirable future scenarios, they worsen even when incorporating householder preference weightings during 2025-2040. The peak burden distribution score occurs in 2025 under the . Trends for social inequity scores between 1990-2040. Household preference weightings are applied to desirable future scenarios from 2020 onward. The social inequity score represents the level of overall environmental burdens imparted on households, with a lower score as shown here indicating improving social outcomes in line with reducing environmental burdens overall and reduced variance among household brackets. active desirable future scenario, at approximately the same level as the BaU scenario in 2005. The proactive desirable future scenario is approximately 0.6 points higher at the same time. These results indicate that reducing environmental burdens in line with householder preferences and merit-seeking behavior engenders the co-benefit of alleviating lower income households' level of burden.
Environmental burdens arising from household consumption influence both burden distribution and social inequity scores. Considering the results shown in figure 5, social inequity peaks in 2010 even though the majority of household footprints are projected to peak in 2020. This confirms that the social inequity score reflects not only the level of environmental burdens but also the balance of consumption across society. For example, contradictory to the result seen for burden distribution, social inequity is expected to reduce after 2010, indicating that social equity is improved by a fairer distribution of environmental burdens (i.e. burden distribution scores increase) and a reduction in footprints in certain income levels. Demographic changes influence scores, along with changing levels of environmental burden. Results suggest that a continuous decline in social inequity will occur, mainly due to the expected decrease in environmental burdens, which outweigh the worsening of their distribution post 2025. In comparing the active desirable future scenario to the BaU scenario, social inequity scores are expected to decrease by between 3.7-4.3 points during 2020-2040. Under the proactive desirable future scenario, social inequity scores drop markedly in 2020, and continue decreasing until 2040 to a level lower than that observed in 1990 for the BaU scenario. The active and proactive desirable future scenarios achieve superior social equity outcomes 15 and 20 years earlier than the BaU scenario, respectively, due to stakeholder preferences, malleability toward policy, and footprint reducing behaviors.

Usefulness of social inequity indicators proposed in this study
The IOSEF, as proposed in this study, demonstrates how disparity in household consumption causes societal distortion via the supply chain, in terms of consumption distribution, environmental burdens and household preferences. The IOSEF has the potential to be a useful tool to aid in measuring social inequity and burden distribution allocation across time and demographics. Based on the results shown in the Japanese case study, the framework can also provide a strong evidence base for policy intervention. This will engender the achievement of policy goals, including the reduction of undesirable environmental burdens (e.g. environmental footprints-the root cause of social inequity), in line with household preferences.
Combining household preferences and responsiveness with demographic and household consumption trends alongside EEIOA, it is possible to investigate broad footprint categories and identify specific goods and services which generate the most environmental burdens, as well as those who are consuming them (Chapman and Shigetomi 2018b). Owing to the high resolution of commodity sectors in the inputoutput model utilized in this study (~400 sectors), it is possible to distinguish details within environmental burdens and to detail the impact of changes in household consumption on burden distribution and social inequity scores. The usefulness of the framework in this respect is demonstrated in figure 6, detailing a sensitivity analysis of burden distribution and social inequity scores with respect to potential changes of 10% for consumption of commodities based on the BaU Scenario in 2040.
A salient example from figure 6 shows that if all households reduced 10% of gasoline consumption for private transport, burden distribution and social inequity scores would decrease by 0.05 and 0.40 respectively. This indicates that although the impact of environmental burdens 'felt' by society are diminished (i.e. the social inequity score drops), those borne by lower income groups increase (i.e. the burden distribution score is increasingly negative). However, when electricity consumption is reduced by 10%, the social inequity score decreases by 0.28 while the burden distribution score increases by 0.01, implying positive outcomes for both social inequity and burden distribution. Here we can deduct that a reduction in electricity consumption would be   Figure 6. Sensitivity analysis of (a) burden distribution score and (b) social inequity score. Colored bars above the axis denote a positive effect on scores when consumption levels increase (red) or decrease (blue) by 10% compared to the BaU scenario in the year 2040. The ten most influential commodities are labeled.
preferable to a reduction in gasoline in terms of the improvement of social equity and the alleviation of burden allocation simultaneously. The framework, therefore, highlights the importance of stakeholder participation in improving social equity outcomes, i.e. through bottom-up policy implementation and the devolution of detailed footprint mitigation policy responsibility to stakeholders.

Policy implications considering social equity associated with lifestyle consumption
The approach posited in this study offers the following benefits for sustainable policymaking. First, it enables the evaluation of social equity incorporating multiple footprint quantification associated with lifestyle consumption. Policymakers have identified the need for national sustainable development progress monitoring, which can be complex, particularly utilizing quantitative models (Allen et al 2016).
Under the posited approach, mitigation of social inequity related to SDGs 10.2 and 10.3 can contribute toward progress in achieving SDG 3.9 (PM 2.5 ), 4.3 and 7.2 (mining risk for Nd), 6.4 (water), 12.5 (industrial waste), and 13.2 (GHG). The combination of IOSEF outputs with environmental burdens and their respective breakdown provides policymakers with a snapshot of progress toward goals, pathways toward improvement, as well as key sectors which require attention for redress from the demand-side. In addition, the IOSEF can identify underlying demographic and income trends which impact upon both SDGs (Dasgupta et al 2015) and stakeholder responsiveness. Second, the consideration of social inequity and burden distribution scores allows for comprehensive policymaking which recognizes the benefits of reduced environmental burdens, and of a societally desirable level of burden sharing. One way to close the income-footprint gap in the context of climate change mitigation is through income transfer from higher to lower income levels utilizing taxation revenue (Abrell et al 2018, Oishi et al 2018). Another way to achieve this goal is by using targeted policy. An example relative to the findings of this research is the encouragement of stakeholders who are concerned about water resources to reduce their consumption of high water footprint products such as meat for example, through conducive policies or subsidies. It is also essential to motivate stakeholders to be aware of the impacts of consumption by increasing education surrounding simple actions to reduce footprints. These may include such things as reducing excessive calorie intake and food loss, increasing walking and bicycle use, participating in community activities (e.g. volunteering), and sharing and repairing goods (Vita et al 2019).
Through the development and case study implementation of the IOSEF utilizing stakeholder engagement, we identified that there is no 'one-sizefits-all' policy approach. Additionally, the IOSEF identifies quantitatively that the participation of all stakeholders at varying levels, i.e. some effort from each income level, engenders better outcomes than the same reduction made by a single income level group. Further, a modal shift from automobiles to walking/cycling could be beneficial not only for various footprints that worsen social inequity but also for people's health, which can positively affect their propensity to take appropriate action (Herrmann et al 2017, Malik et al 2018, Nansai et al 2020a. It would also be relatively easy for low-income households to undertake these simple activities which have the dual benefit of reducing transport related costs and taxes as well as shrinking environmental burdens. Considering the difference in footprints between demographics with similar preferences gives an additional policymaking option which may be fit for both demographic and preference (i.e. carbon footprint preferences between low and very high-income households; see figure 3). The IOSEF allows policymakers to test various approaches prior to their implementation, in order to improve societal equity and burden distribution allocation outcomes considering lifestyles, consumption and stakeholder propensity for participation in environmental footprint reducing activities.

Conclusion
The IOSEF, a new sustainability assessment framework which considers fairness and lifecycle perspectives, allows for the consideration of policy implications to improve social equity with respect to multiple lifecycle environmental pressures and household consumption. In addition, the tool is also useful in its ability to help highlight priority SDGs requiring redress. Through a case study application to Japan, results demonstrate that there is no onesize-fits-all policy which will improve social equity and burden allocation outcomes for all stakeholders (e.g. households). For this reason, multi-faceted, fit for purpose policy is required. This research builds on this finding to enable policy which is both appropriate and responsive to stakeholder preferences and demographics. Developing policies which consider preference and demographics (i.e. income level, age etc) will lead to desirable outcomes whereby each sector of society is able to contribute in their own way, seeking their own perceived benefits. The overall goal of this approach is improved social equity, which the proposed framework can measure quantitatively.
This study is unique, as it demonstrates for the first time the nature of societal outcomes through the lens of inequity underpinned by lifestyle related environmental burdens. In achieving this goal, there are several limitations to be aware of which may affect overall results arising from use of the framework. To improve accuracy, it is essential to consider the factors underpinning footprint estimates. In the case study, factors are fixed based on 2005 footprint intensities, consumption patterns and policy interventions, in order to clarify the impact of demographic changes. Incorporating scenario approaches which consider technology innovation (Barrett and Also, the methodology employed to determine proactiveness levels using a national survey is in the early stages of development. Environmental awareness cannot be directly translated from the degree of pro-environmental behavior, although there is some evidence that increasing awareness engenders a reduction in environmental burdens (Li et al 2019). Additionally, the survey method employed and the influence of different ways of asking questions also has an impact on responses (Pew Research Center 2015).
To overcome these methodological limitations, further investigations are required to validate the applicability of the proposed framework in other jurisdictions, particularly with regard to the availability of IO tables within and across regions (e.g. Faturay et al 2017, Stadler et al 2018, Wakiyama et al 2020.