Social life-cycle assessment frameworks: a review of criteria and indicators proposed to assess social and socioeconomic impacts

Abstract

Purpose

In addition to the Guidelines for the social life-cycle assessment of products (S-LCA) (Benoît and Mazijn 2009), there are several other methodological frameworks in this field. In recent S-LCA literature reviews, much attention has been paid to how performances or impacts should be or are measured (i.e., life-cycle impact assessment) in existing S-LCA studies. In this review, we focus on what is measured (i.e., assessment criteria and indicators [C&I]) and on the definition and selection of these C&I.

Methods

We conduct a review of existing S-LCA frameworks in order to understand (i) the origin, selection, and applicability of C&I; (ii) the purpose of the assessment and the assessed phenomena as reflected in the indicators; and (iii) the scope of C&I of the topics, life-cycle stages, and stakeholders.

Results and discussion

Based on our review, we identify 14 distinct S-LCA frameworks, for which we propose a classification according to the rationale behind the definition and selection of C&I: value-based, context-oriented, theory-structured, impact-based, and applicability-oriented. While authors of the frameworks agree with the purpose of supporting decision-making, the assessed phenomena are quite diverse among the frameworks. However, given the mixed character of the indicators, we cannot draw a clear line between frameworks assessing practices or performances and frameworks assessing effects or impacts. Lastly, our review highlights the uneven coverage of the stakeholder, life-cycle stages and topics in the S-LCA frameworks and confirms that the use stage and the relations between value chain actors receive less attention than the production stage.

Conclusions

Our comparative review not only confirms the diverse nature of S-LCA frameworks but also highlights their specificities, common features, and potential areas for improvement. We encourage the use of assessment criteria that are legitimate and meaningful for stakeholders. In addition, given the S-LCA promise to provide a holistic assessment, the variables included should be envisaged as elements of a product system and that must be branded according to their position in relation to other elements. In addition, we recommend for LCIA the combination of type I and type II assessment, including further research on impact pathways, that could link these meaningful and legitimate criteria with further impacts and related stressors in order to strengthen the capacity of S-LCA to contribute to sustainability management.

1 Introduction

1.1 Context and definitions

Social life-cycle assessment (S-LCA) is a methodology that aims to assess social and socioeconomic (S&SE) impacts of products. It is meant to complement the environmental LCA (E-LCA) and life-cycle costing (LCC) in the development of life-cycle sustainability assessment (LCSA) (Valdivia et al. 2011). Although in continuous development, E-LCA and LCC can be considered as consolidated methods (ISO 2006; Swarr et al. 2011). However, this is not the case for S-LCA which, despite the significant steps that have been taken, including the availability of Guidelines for S-LCA (Benoît and Mazijn 2009), still faces a number of methodological challenges including matching the method to the life-cycle thinking framework (e.g., link with functional unit), setting of system boundaries, the definition and selection of assessment criteria and indicators, and defining impact assessment methods.

Research into S-LCA started in the mid-1990s and developed significantly from 2005. In 2009, the Guidelines for S-LCA were published under the umbrella of the UNEP/SETAC life-cycle initiative (Benoît and Mazijn 2009). The Guidelines for S-LCA (henceforward referred to as “the Guidelines”) provide recommendations on how to conduct the first two phases of S-LCA (i.e., goal and scope definition and life-cycle inventory). The research on the third phase (life-cycle impact assessment, LCIA) was, at that time, not considered sufficiently mature to be included in the Guidelines. The Guidelines propose subcategories, which can be considered as assessment criteria, i.e., “distinguishing element[s] which a thing is judged by” (Namkoong et al. 2002) and which qualify what is aimed for, e.g., assessment of sustainable development or human well-being. A separate publication (Benoît et al. 2013) proposes indicators, which (Prabhu et al. 1999) define as “any variable […] used to infer the status of a particular [assessment] criterion.” While the Guidelines are an important reference methodological framework¹ for S-LCA, they currently co-exist with a plethora of other S-LCA methodological frameworks and methods (Wu et al. 2014) proposing alternative criteria and indicators (C&I).

1.2 General aim

As part of evolving to an agreement over how S-LCA should be conducted, there is a need to identify and understand characteristics and differences of what is assessed among S-LCA methodological frameworks which have been developed in the 2005–2016 period. Recent reviews (Parent et al. 2010; Chhipi-Shrestha et al. 2014; Wu et al. 2014; Russo Garrido et al. 2016) have focused on differences among S-LCA frameworks in the LCIA phase and on the question of how to measure S&SE impacts. A more recent review (Iofrida et al. 2016) classifies S-LCA approaches according to the research paradigm applied, chiefly on the basis of the LCIA methodology. Whereas the latter was not explicit, various criteria have been used, from the involvement of the researcher and of stakeholders, to the choice of impact categories and of inventory data. This review comes up with three groups of paradigms: postpositivist-oriented (considering impact pathways and corresponding to E-LCA paradigm), interpretivism-oriented, and studies following both paradigms.

In the same vein, our research seeks to make a classification of S-LCA approaches and to provide justifications for differences. However, in this review, our perspective is different since our main parameter is not the LCIA methodology, but the definition and selection of what is being measured in different S-LCA frameworks (i.e., the rationale behind C&I lists). More specifically, we focus on where the lists of C&I put forward by different frameworks are derived from (e.g., CSR criteria, UN conventions, people’s values), and how the selection of C&I is made (i.e., selection method, e.g., by the authors themselves, by experts, by stakeholders). We believe reflecting on these methodological aspects is fundamental as the method to define and select C&I can have as profound implications on a study scope and results as the LCIA method used. Lastly, grappling with the question of what C&I to select for SLCA is an all too common and recurring question for which few, if any, comprehensive and analytical approaches have been proposed to enlighten SLCA practitioners’ decision-making.

In addition to (i) the underlying rationale behind the selection of the C&I, we focus on two other related characteristics: (ii) what observed phenomena are being assessed (e.g., company practices, impacts, and their position on the impact pathway), and (iii) the topics, stakeholders, and life-cycle stages that they cover.

We propose a typology of S-LCA frameworks through which we seek to reduce the complexity of the booming S-LCA research field by looking for convergences and divergences between frameworks. On the basis of this review, we further seek to identify and develop research perspectives and recommendations for S-LCA. In the next section, we specify the materials that have been used to conduct the review. We then detail the rationale for focusing on the three issues listed above, as well as the methods implemented. We explain our results in a third section, before presenting some perspectives and conclusions.

2 Materials and methods

2.1 Materials

We started our literature review on S-LCA frameworks from various existing literature reviews on S-LCA (Arcese et al. 2016; Chhipi-Shrestha et al. 2014; Delcour et al. 2015; Fan et al. 2015; Iofrida et al. 2016; Jørgensen et al. 2007; Macombe et al. 2013; Mattioda et al. 2015; Parent et al. 2010; Petti et al. 2016; Russo Garrido et al. 2016; Wu et al. 2014). We further identified additional publications by bibliographic searches with our internal institutional search engine (CIBLE+, ULB, Belgium) and conventional international engines Scopus and Google Scholar. We also looked into conference proceedings and websites of consultancy companies. These searches resulted in peer-reviewed (conference) articles, book chapters, research reports, and PhD theses.

Of all the papers found through these searches, a number were selected for further review if (1) the outcome was described as guidelines, guide, handbook, tool, instrument, or framework for S-LCA; or (2) they defined general principles for conducting S-LCA; (3) they proposed a list of criteria, indicators, or impact categories for evaluating impacts of products/industries in S-LCA; and (4) they proposed a method for selecting C&I that is different from existing frameworks and can be applied to other case studies.

The following studies were excluded from further analysis: studies applying/adapting an existing framework (e.g., studies applying the Guidelines), studies proposing C&I specifically for one product/industry, studies proposing a list of C&I that did not specify or detail the origin and/or the way of selecting the C&I, studies assessing a single indicator, and studies focusing on another methodological development (e.g., LCIA, setting of the system boundary).

The next section explains the rationale for focusing on the three selected review objectives/issues and the review method implemented.

2.2 Rationale, state of the art, and methods

2.2.1 The rationale behind the selection of assessment criteria and indicators

The set of criteria predefined by a methodological framework influences the social issues covered by a study that applies the framework—it determines the topics that will be looked at in the study. Choosing a certain methodological framework, that includes certain criteria (and not others), thus has implications.

The selection of assessment criteria is a subject of debate in the research field of social impact assessment and sustainability indicators (Bell and Morse 2001; Hák et al. 2012; Mccool and Stankey 2004; Vanclay 2002). While the issue of the origin and selection of assessment C&I seems to have arisen less prominently in S-LCA in recent years, the issue seems to be inherently part of some ongoing discussions (e.g., area of protection [AoP] for S-LCA, assessment of impacts) and has been raised as an element determining the paradigm taken by S-LCA studies (Iofrida et al. 2016).

For S-LCA, there is yet a common list of S&SE impacts agreed, in the sense that proposed lists are still debated. One agreed constraint is that potential S&SE impacts must be assessed (Benoît and Mazijn 2009), with a view to contributing ultimately to human well-being. The latter has indeed been defined by researchers as the AoP for S-LCA (Reitinger et al. 2011), even while it remains undefined and imprecise (Feschet 2014). Most studies mention the link with sustainable development without clarifying or describing the approaches considered for this vague concept (Feschet 2014, p. 159).

There have been efforts to establish a list of C&I to guide the collection of data: the Guidelines propose a list of 31 subcategories, defined as “S&SE issues of concerns and relevant characteristic or attribute to be assessed” (Benoît and Mazijn 2009, p. 71). One aim behind this list is “to prevent using S-LCA results on a few limited topics for social marketing aims while not addressing core issues” (Benoît and Mazijn 2009, p. 71).

Although it has been broadly used since 2009 (Wu et al. 2014; Chhipi-Shrestha et al. 2014), this list of subcategories has been criticized by some researchers, e.g., because there is not always a linear and straightforward causal relationship (or impact pathway) between subcategories and sustainable development (Arvidsson et al. 2014; Jørgensen et al. 2009). This is why some authors investigate impact pathways that link product processes with social impacts that are to be included in S-LCA (Macombe 2013b; Feschet et al. 2012; Bocoum et al. 2015).

There have also been calls to make the adopted conceptual model underlying an S-LCA method more explicit and to use theoretical models as a basis (Feschet 2014). While a conceptual model defines concepts, a theoretical model helps to clarify “the sense given to concepts” and seeks to explain phenomena (Feschet 2014, p. 157), as does Amartya Sen’s capabilities theory. Yet, the Guidelines would be based on CSR, a rather static and descriptive approach which does not offer an understanding of the implications of companies’ practices (Feschet 2014). A theoretical model is argued to help identify what is important to assess and would bring stability to the criteria and indicators used among studies regardless of different geographical and time contexts as well as study sponsors or data availability (Feschet 2014; Macombe 2013a), thus enhancing comparability of studies. In addition, it would help to propose a coherent set of C&I, instead of individually unrelated C&I.

Meanwhile, other methodological S-LCA frameworks have been put forward, such as private-sector-driven frameworks (e.g., Product Social Metrics), with other lists of assessment C&I. Some authors propose contextualized lists of C&I (Mathe 2014), arguing that social indicators are context-dependent.

These discussions, developments, and propositions call for a close examination of how assessment C&I should be selected within S-LCA and on what basis (i.e., the rationale behind their selection), and what the existing frameworks propose in this regard.

Method

For each framework, we identified the different rationales behind the selection of C&I in these frameworks, i.e., the origin of the C&I proposed (e.g., UN conventions) and the ways in which these were selected (e.g., expert consultation). Answering these two questions will help us understand why certain frameworks propose certain criteria. We also looked at whether the list of C&I was meant to be applicable in any context or if it was meant to be adapted according to the specific sector under study or to the local context (as specified by authors of frameworks). On this basis, we propose a classification of frameworks with regard to the origin of C&I and their selection. Following this review, we describe each framework.

The origin of C&I and the ways that C&I are selected result in different lists of C&I, which potentially reflect different phenomena.

2.2.2 The purpose of the assessment and the assessed phenomena

S-LCA should assess impacts. However, given the complexity of assessing S&SE impacts (Slootweg et al. 2001), this objective is nuanced in the Guidelines, which define S-LCA as “a social impact (and potential impact) assessment technique that aims to assess the S&SE aspects of products and their potential positive and negative impacts along their life cycle” (Benoît and Mazijn 2009, p. 37). Assessment C&I proposed can then reflect both aspects such as practices (e.g., level of paid wages) or impacts (e.g., health impacts of working conditions). There are two trends in S-LCA LCIA practice: type I S-LCA, which assesses performances, and type II S-LCA, which assesses impact pathways between two variables or directly impacts (Chhipi-Shrestha et al. 2014; Parent et al. 2010; Wu et al. 2014), with performances upstream of the impact pathway, and impacts downstream. However, in some type I studies, some performances could be located further down the impact pathway (Russo Garrido et al. 2016), blurring the distinction between types I and II.

Method

With this in mind, various frameworks will be reviewed, focusing on the position of the proposed C&I on the impact pathway. For this task, we use the same classification as Russo Garrido et al. (2016), distinguishing three positions on the impact pathway (see Fig. 1): activity on the product system, effects, or impacts. A social effect is “a social phenomenon caused by [a] change that could have impacts,” while an impact is a “consequence of [a] change that is felt by people directly in life” (Macombe et al. 2013, p. 206). Some studies also use context data in S-LCA “as contextual information providing insight on potential effects of corporate social performance, as a proxy for a phenomenon occurring within the product system and/or its resulting causality chains, or as an element that may affect/condition activities at the company level” (Russo Garrido et al. 2016, p. 6). We will thus evaluate if a certain criterion or indicator acts on or describes a context and other stressors, an activity, an effect, or an impact.

Fig. 1

Positioning of indicators on the impact pathways. Source: Russo Garrido et al. (2016)

Relating to the assessed phenomena, the purpose of the study as described by the authors seems of interest and therefore will be assessed and reported for each framework.

In addition to the rationale behind the selection of C&I (Section 2.2.1), the identification of the purpose of the assessment and of the assessed phenomena will enable a more comprehensive picture of S-LCA frameworks. This understanding will be completed with the implications these factors can have for topic, life-cycle, and stakeholder coverage.

2.2.3 The topic, life-cycle, and stakeholder scope of C&I

As a method based on the life-cycle thinking frame and intended to provide a systemic and holistic assessment of S&SE impacts (Benoît and Mazijn 2009), the S-LCA set of C&I should be able to assess performances or impacts of all life-cycle stages (ideally from cradle to grave). Since S-LCA, together with E-LCA and LCC, is intended to provide a full sustainability assessment (Valdivia et al. 2011), the issue of topic coverage is essential. This is necessary for identifying potential impact transfers between life-cycle stages and stakeholders, and between assessment criteria, when comparing different product life cycles.

Method

In order to address this issue, we examine the lists of C&I proposed in each framework, and we compare them with an improved classification by stakeholder based on the one from the Guidelines. We use the Guidelines as our reference since they have been built by a range of researchers and are meant to evolve within the Social LC Alliance (Mazijn n.d.). In this way, we will be able to identify differences among C&I lists of the frameworks.

From this, we examine the life-cycle stages to which the C&I are referring (e.g., working conditions: production phase), and we consider whether the whole life cycle can be covered by the framework.

From these three perspectives, common features and differences among frameworks will be highlighted, as well as gaps regarding the life-cycle thinking approach and S-LCA initial objectives.

3 Results and discussion

Fourteen frameworks were identified for further review and analysis (see Table 1). We describe each of these frameworks with its main features in relation to the three aforementioned issues.

Table 1 Selected frameworks for review

3.1 The rationale behind the selection of C&I—a proposed classification of S-LCA frameworks

The results on the origin and selection of C&I in the different frameworks are shown in Table 2. The first set of columns sets out the starting materials upon which C&I were selected. Seven types of starting materials were identified: international treaties (e.g., UN conventions) and policy documents (e.g., Green Papers); voluntary standards (e.g., CSR); other assessment tools (e.g., social impact assessment), including the Guidelines; literature on well-being; on impact of economic activity/product; theoretical models, and available database(s). The second set of columns sets out how the selection was made: consultation of stakeholders, from the private sector or experts. The final set of columns sets out the corresponding area of applicability (universal, sectoral, contextual).

Table 2 Origin, selection, and applicability of assessment C&I in reviewed S-LCA frameworks

As a first result, most of the frameworks reviewed use a combination of starting materials and ways to select C&I. Most frameworks use between two and three different types of starting materials to elaborate their list of C&I, in addition to a selection process involving experts or stakeholders. This highlights that the selection of C&I replies not to one logic but to several.

Secondly, we did not find two frameworks that used the same starting material and ways of selection (or combination of these). Based on our analysis, we could group the 14 frameworks in five types (Table 2). This typology is based on the features that are brought out most strongly and on the features that distinguish each framework from the others. The first type consists of value-based frameworks. These use international agreements and voluntary standards, but also consultations with stakeholders at an international level (Benoît and Mazijn 2009), with the main company or with the industry (Dreyer et al. 2005; Kruse et al. 2008) to select C&I. A second type of framework defines assessment criteria also on the basis of values, but where these are specific to a context. Within these frameworks (Mathe 2014; Wangel 2014), assessment criteria are selected through the consultation of various stakeholders (including NGOs, consumers, local communities, etc.). A third type uses theoretical models to structure and select criteria, but uses other materials to build the lists of C&I (voluntary standards in (Garrabé and Feschet 2013), other assessment tools in (Gaasbeek and Meijer 2013), Guidelines in (Reitinger et al. 2011)). A fourth type groups the impact-based frameworks, where C&I are defined by backtracking from assumed or observed effects/impacts to social stressors (Macombe 2013b). The applicability of C&I depends on the conditions for use of each impact pathway. A fifth and final type includes frameworks focusing on their applicability; half of these frameworks are based on the work done through the Guidelines. Other starting materials include standards, other assessment tools, existing databases, and easily accessible information. The validation of the sets of C&I is done through consultation with experts as well as the private sector, presumably the main S-LCA users. The resulting list is generally meant to be applicable to any context.

Between applicability-oriented frameworks and the ones defining context-specific C&I, a number of frameworks propose a two-layer C&I list, universal and context-specific (Dreyer et al. 2005; Kruse et al. 2008), or provide the flexibility to be adapted according to the context (Benoît and Mazijn 2009; Garrabé and Feschet 2013).

In the next section, we present the various framework types and corresponding frameworks in more detail.

3.1.1 Value-based frameworks

The Guidelines for S-LCA (Benoît and Mazijn 2009) represent one of the main references on progress regarding the definition of criteria for S-LCA. A list of 31 subcategories (or criteria) has been developed on the basis of a consultation process with a wide range of stakeholders (worker and employer trade unions, consumer and private sector associations, NGOs, UN bodies), mainly from Europe but also from other continents (Mazijn 2010), and has been peer-reviewed by CSR, LCA, and sustainable development experts (Benoît and Mazijn 2009). Subcategories reflect basic UN conventions on human and workers’ rights that are thought “to go beyond personal and cultural subjectivity or political orientation” (Benoît and Mazijn 2009, p. 48) and best practices criteria such as “international instruments, CSR initiatives, model legal framework, social impacts assessment literature.” Subcategories are classified according to five types of stakeholder that can be affected by the practices of companies (workers, consumers, value chain actors, local community, society) and can be classified according to the six impact categories proposed by the Guidelines (human rights, working conditions, governance, cultural heritage, health and safety, and socioeconomic repercussions). However, the Guidelines do not specify links between subcategories and impact categories. As a follow-up, The Methodological Sheets for Subcategories in Social Life Cycle Assessment (Benoît et al. 2013) discusses the link between each subcategory and sustainable development and proposes corresponding indicators and sources.

Meanwhile, Dreyer et al. (2005) propose a two-layer S-LCA: “an obligatory, normative, predetermined set of categories expressing minimum expectations to conducting responsible business (based on UN conventions and on local and national norms), and an optional, self-determined set of categories expressing interests specific to the product manufacturer” (p. 92). These two sets are to be related to a suggested AoP for S-LCA: “Human dignity and well-being.” Similarly, Kruse et al. (2008) propose having two sources of indicators: a first range of indicators that is “representative of broadly recognized societal values” and based “on various international conventions, agreements, and guidelines”; and a second range that is based on industry or stakeholder interests and/or data availability.

3.1.2 Context-oriented frameworks

Within a project focusing on the evaluation of fish farming in three different countries, Mathe (2014) proposes a participatory approach to identify impact categories. It is justified as a way of considering the evolution of two paradigms—the viewpoint of the firm, which has seen the extension of the range of actors to be taken into account, and the evaluation viewpoint, which favors participatory approaches (Mathe 2014). The first step was to undertake “interviews with stakeholders about their representations of the social aspects of the activity concerned” (Mathe 2014, p. 1510). These data are analyzed to reveal lists of main principles and impacts, which are complemented through review of literature and international conventions, consolidated by S-LCA practitioners and adapted again through focus groups with stakeholders. Ultimately, the object assessed is the level of well-being induced by the ecosystem services. The article presents a way of choosing and integrating stakeholders but does not present a list of impacts and indicators.

Wangel (2014) proposes a framework to assess impacts of an alternative oyster value chain. The starting point is the theoretical framework defined by Nussbaum, which is a further development of Sen’s capability approach and which has defined a list of 10 central universal capabilities.² On the basis of these capabilities, which are constitutive of well-being, he proposes to use a participatory approach for stakeholders in the value chain to define for each capability the valuable “functionings” (see earlier footnote) that emerge or are enhanced through the value chain.

3.1.3 Theory-structured frameworks

Like Wangel (2014), other authors have taken the path of using a theoretical model to structure their proposed framework. The starting point of Reitinger et al. (2011) is the methodological gaps identified by the Guidelines, specifically on the LCIA. They propose a theoretical framework to spell out the AoP, “namely the general concept of human well-being and the impact categories,” as a means of disclosing “our own normative assumptions” (Reitinger et al. 2011, p. 380). They use the frameworks of Alkire (2002) and Finnis et al. (1987) who defined dimensions of life, following Sen’s work, to define impact categories. According to Reitinger et al. (2011), subcategories of the Guidelines are easy to convert into capabilities and can be linked to the proposed impact categories.

Gaasbeek and Meijer (2013) propose a framework to assess sustainability impacts linked to the introduction of a new technology in the process, within the broader PROSUITE project, including social impacts. This work originates in the observation that existing methodologies “do not cover all dimensions of sustainability, they do not cover them in a comparable manner, or do not include a rigorous treatment of cause-effect relations towards impacts” (Gaasbeek and Meijer 2013, p. 6). One objective of the approach was to use mainly quantitative indicators, in order to avoid “subjective or ad-hoc judgments as much as possible” (Gaasbeek and Meijer 2013, p. 6), but in the end, five indicators out of 11 are actually qualitative. They use as their basis four impact categories proposed by Weidema (2005)—autonomy; safety, security, and tranquility; equal opportunity; and participation and influence—which are linked with the AoP well-being and with indicators. These indicators are defined through literature review of social indicators, first selection with a protocol combining the normative-functional model of sustainability with the S-LCA, test against specific metacriteria, and final selection by experts in a Delphi group. The framework has been applied in four case studies.

Garrabé and Feschet (2013) propose a methodology based on the multiple-capital model and on Sen’s capabilities approach. With this model, the effects of an economic activity on potential capabilities of actors are assessed. These variations in capabilities can affect the stocks of five capitals: human, technical, financial, social, and institutional. An advantage of the notion of capital is thus to be able to consider flows as well as stocks, with this corresponding to a certain conception of sustainability based on stock of various capitals/resources that are to be passed on to future generations (Feschet 2014; Stiglitz et al. 2009). Each process can be considered as an articulation of various capitals, even if not all relations between these capitals are known (Feschet 2014, pp. 204–205). For human and social capital, authors have identified subclasses on the basis of prescriptions of “GRI, ISO 8000 and 26000, [UN] Global Compact, OECD and EC Green papers” (Feschet 2014, p. 253), and subclasses of institutional capital are based on the work of Rodrik (2000) (Feschet 2014). For other capitals, the authors looked for main factors contributing to “development and thus ceteris paribus to wellbeing variations” (Feschet 2014, p. 253). Within these subclasses, the authors propose categories of effects.

3.1.4 Impact-based frameworks

This category gathers several studies (Feschet et al. 2012; Bocoum et al. 2015; Di Cesare et al. 2016) whose general approach has been described by Macombe (2013a, b). The objective is not to assess a comprehensive range of aspects or impacts but to develop impact pathways between an impact and the origin of this impact. In order to select assessment C&I, the proposition is to proceed in reverse, to backtrack, from effects/impacts to social stressors.³ As explained by Macombe, impacts caused by changes in life cycles of products have been described by scientists in the literature. Through literature review, effects/impacts stemming from a change can be identified, as well as social stressors. The work done by researchers investigating impact pathways is then to verify and model the link between social stressors and impacts. If validated as social stressors, these variables could then constitute the basis for a life-cycle inventory in S-LCA.

3.1.5 Applicability-oriented frameworks

The framework of Labuschagne et al. (2005) aims to assess sustainability of projects and technologies in the processing industry, but can be used for the assessment of products as well. On the basis of a literature review mainly of other assessment tools and voluntary standards, the authors built a set of social criteria classified into four categories (internal human resources, external population, macrosocial performance, and stakeholder participation). Managers and assessment practitioners of a large South African petrochemical company assessed the suitability of the framework and the relevance of criteria though a survey. In a further publication, Brent and Labuschagne (2006) developed a set of indicators and tested them through 10 case studies, which revealed the difficulty of applying the framework given data availability for some criteria.

The private sector has also developed other frameworks for S-LCA. The first one has been elaborated by BASF with the Universities of Karlsruhe and Jena as well as the Öko-Institut e.V., initially for all products: SeeBalance (Schmidt et al. 2004), and later specifically for agricultural products, AgBalance (Schoeneboom et al. 2012). The concept underlying the approach is socio-efficiency, which represents the social benefits throughout the entire life cycle of a product in relation to the costs for the end customer for buying, using, maintaining, and finally disposing of or reselling the product (Schmidt et al. 2004). When SeeBalance was developed, a literature review was conducted to look for social goals and indicators. The selection among this literature was made according to applicability of identified social goals for product and process assessment. Indicators were elaborated for the development of AgBalance only, through a consultation “with international stakeholders, experts and decision makers,” and BASF had the last word on the basis of “relevance, inclusiveness, practicality of quantification and availability of data sources” (Schoeneboom et al. 2012, p. 6).

Similarly, PRé Consultancy started in 2013 to develop a methodology through the Roundtable for Product Social Metrics, gathering multinationals such as Ahold, BASF, or BMW Group, resulting in a handbook: the Products Social Metrics (Fontes 2014). While the tool is presented as one that builds on existing initiatives such as the Guidelines, and CSR corporate level standards (GRI, ISO), it is stated that these latter lack “harmonization across peer-to-peer approaches,” which the Roundtable aims to address. The definition of C&I was also supported by the review of other assessment tools (OECD, UN) and expert consultation. One of the selection criteria was the availability of data in public or private databases.

It has to be noted that some preliminary reports have been published by the World Business Council for Sustainable Development (WBCSD) to provide guidance to measure socioeconomic impacts (WBCSD 2013) and social capital impacts (KPMG 2015), but with no reference to S-LCA or to assessment with the life-cycle approach.

On the basis of the Guidelines, two databases have been built that propose statistical data for indicators related to some of the Guideline’s subcategories. The Social Hotspot Database (SHDB, Benoit-Norris et al. 2012) has been built in order to provide practitioners with generic data to identify social hotspots in value chains, i.e., processes with high social risks, which require site-specific data to be collected. Social hotpots are identified though the determination of most important country-specific sectors (CSS) or processes in the product supply chain, based on the number of worker hours, and identification of social issues in these CSS. Twenty social issues and around 100 indicators are included in the SHDB and were selected from the Guidelines by the advisory board of the consultancy company that developed the database (New Earth). Obviously, this choice was guided by the availability of statistical data at the national and sectoral level.

The other database is the Product Social Impact Life Cycle Assessment (PSILCA) database developed by the consultancy company Greendelta. It provides statistical data for 88 qualitative and quantitative indicators under 25 subcategories from the Guidelines. For the indicators, several sources were used (Gaasbeek and Meijer 2013; Fontes 2014; Benoît et al. 2013; Ciroth and Eisfeldt 2016).

In conclusion, various methods are used to select C&I in proposed frameworks. While looking at these methods, we can better understand the different rationales behind the choice of C&I, some looking at legitimate C&I sets—whether scientifically or ethically—while others are looking for feasible C&I sets. But we can also highlight common bases among them with almost every starting material used by several frameworks, as well as interlinkages, with many of them using the Guidelines or other assessment tools as a basis. To have a deeper understanding of these frameworks, it seems of interest to look at their purpose and at the object which is effectively assessed.

3.2 The purpose of the assessment and assessed phenomena

For each framework, Table 3 describes (1) the specific purpose of the proposed framework as specified by the authors of the framework (column A); (2) what phenomena the authors aim to assess (column B); (3) what is being actually assessed by the framework (column C); and (4) how the articulation with impacts was carried out, in cases when this was done (column D).

Table 3 Purpose and assessed phenomena

For most of the frameworks, S-LCA is meant to support decision-making, with some of them also mentioning the identification of areas for improvement, and communication (Benoît and Mazijn 2009; Fontes 2014; Schoeneboom et al. 2012) (column A). However, the assessed phenomena, as specified by authors (column B), vary greatly, from company practices and social aspects of products, to the level of well-being generated by an industry or the variation in capabilities of stakeholders. Finally, the PSILCA and SHDB frameworks have a completely different focus (and purpose), since their objective is to identify hotspots or processes where social risks may be present, in order to prioritize site-specific data collection.

Going from column B to column C, it should be noted that there are discrepancies in what is described by some authors and what indicators effectively reflect. In most cases, it is because the term “impact” is used instead of “practices,” “performances,” or “effects” (Benoît and Mazijn 2009; Fontes 2014; Kruse et al. 2008; Schoeneboom et al. 2012).

As highlighted in column D, most frameworks consider C&I separately, apart from impact-based framework which are meant to investigate relations between indicators. Some frameworks mention impact categories, but they do not detail how to relate C&I to these (Benoît and Mazijn 2009; Gaasbeek and Meijer 2013). A last group, composed of the frameworks based on the capabilities approach (Garrabé and Feschet 2013; Reitinger et al. 2011; Wangel 2014) as well as the framework of Mathe (2014) describe a path to impacts, in a more and less detailed form.

First, it needs to be noted that for some of these frameworks (Garrabé and Feschet 2013; Mathe 2014; Reitinger et al. 2011; Wangel 2014), no publication of a complete application is available, while the Guidelines (Benoît and Mazijn 2009) are extensively applied by practitioners (Chhipi-Shrestha et al. 2014; Wu et al. 2014). Finally, some have not been applied but have been used by others as a basis to further build on (Benoît and Mazijn 2009; Brent and Labuschagne 2006; Dreyer et al. 2005).

From our review, we can conclude the following: on one hand, value-based frameworks and applicability-oriented frameworks assess mainly company practices and context variables, apart from a few indicators reflecting effects, impacts, and contextual information (for further details, see Russo Garrido et al. 2016). In these frameworks, impacts are thus conceptually not considered, except in the Guidelines, which propose impact categories assessment but do not provide guidance on LCIA.

On the other hand, the other three framework types (context-oriented, theory-structured, and impact-based) generally assess or aim to assess effects or impacts of practices/attributes of different entities (companies, activity, and value chain) or of different developments (introduction of a technology, change in the life cycle).

Four frameworks (Brent and Labuschagne 2006; Gaasbeek and Meijer 2013; Garrabé and Feschet 2013; Macombe 2013b) aim to assess effects deduced from comparing a situation with a reference situation. The results of these ex ante assessments cannot be used on their own but must be used with the results of the reference situation. In these frameworks, an effect is assessed (an effect being the difference between two situations/scenarios), but the assessment does not necessarily go up to the impact at midpoint or endpoint levels and is not an impact pathway type neither.

For example, the PROSUITE framework (Gaasbeek and Meijer 2013) proposes to assess effects for 11 indicators. The effects on these indicators are then not connected with one of the four identified areas of protection but are weighted and aggregated to come up with a final social well-being score. Thus, in this framework, assessing effects does not imply the use of cause-effect relationships.

In addition, in the PROSUITE framework, indicators are not all located at the same position on the impact pathway; some reflect activities on the product system (e.g., child labor occurrence) and some are positioned further down the impact pathway (e.g., effect on people’s trust).

The Multiple Capital Model (MCM) Capacities S-LCA acts differently as it describes the path between an activity related to the product (i.e., indicators of conditions of potential effects of capability) and the final impact (i.e., net marginal real effect of capability), but acknowledges that it is only able to assess up to the social effect (i.e., effective potential marginal effect of capability) (Feschet 2014). This framework does not claim to give tools to assess final impacts, since it depends on the further investigation of impact pathways, and is in line with the suggestion to combine both approaches (i.e., the assessment of practices or performances and the assessment of impacts) (Chhipi-Shrestha et al. 2014).

While trends emerge in terms of the assessed phenomena according to the type of framework, the analysis carried out on indicators reveals that a clear line cannot be drawn between frameworks assessing practices/performances and effects/impacts. Apart from context-oriented and impact-based frameworks which otherwise lack application and development, each type of framework takes or shares parts from the others, resulting in a mix of indicators assessing practices, effect, and impacts in almost every framework.

3.3 Scope of assessment C&I

Based on the life-cycle thinking frame, the S-LCA promise is to provide a systemic and holistic assessment of S&SE impacts. The scope of S-LCA in terms of topics, life-cycle stages, and stakeholders is thus a relevant issue to look at to better understand the proposed S-LCA frameworks.

Table 4 summarizes the C&I included in the identified frameworks, according to the type of stakeholder and the topics considered. This table has been based on the information gathered in the table located in the Electronic Supplementary Material (ESM) that details the topical scope of each framework proposing C&I (see Table S1 in the ESM).

Table 4 Topical and stakeholders coverage

Overall, while value-based and context-oriented frameworks provide some flexibility for including C&I deemed relevant for stakeholders, it is worth noting that theory-structured frameworks help to extend the scope of the set of criteria. For example, the MCM Capacities S-LCA (Garrabé and Feschet 2013) includes C&I relating to social, technical, financial, and institutional capitals which are not all included in the Guidelines. Similarly, through the stakeholder approach, the Guidelines allow the inclusion of a broad set of topics.

3.3.1 The production stage as the main focus of most frameworks, but unevenly covered

More specifically, the production stage is that most covered in the reviewed frameworks, and workers are the stakeholder that receives most attention, as highlighted by Jørgensen et al. (2007). However, workers are not covered similarly according to the frameworks.

While most frameworks include criteria addressing basic standards to be respected, the Product Social Metrics (Fontes 2014) adds positive criteria linked to workers’ well-being that can fit better the context of Northern countries.

Also, three frameworks (Garrabé and Feschet 2013; Kruse et al. 2008; Schoeneboom et al. 2012) include economic C&I relating to companies of the product chain, giving information on its state, costs, and profitability.

3.3.2 The consumption stage insufficiently covered, particularly on the consumer side

On the contrary, the consumption and the use stages are the least covered. An exception to this is the framework of Wangel (2014) which studies a consumer-driven production system and in which C&I have been defined by consumers themselves. But some other frameworks also include criteria relating to product utility and to consumer satisfaction (e.g., experienced well-being of consumers in Fontes 2014) and health impact and functional product characteristics in Schoeneboom et al. (2012).

However, in the Guidelines and in the MCM Capacities S-LCA, even though some criteria regard consumers (e.g., health and safety, transparency, feedback mechanisms), some significant aspects might be overlooked (e.g., product utility, accessibility, affordability). Regarding the Guidelines, one explanation for this is that product utility is to be included in the functional unit. Another reason could be that only criteria relating directly and entirely to company practices toward other stakeholders (in Benoît and Mazijn 2009) or to effects of an economic activity on other stakeholders (in Garrabé and Feschet 2013) are assessed. Thus, C&I that are relevant for consumers and rely on contextual information as well (e.g., purchasing power for affordability) would be overlooked. Lastly, the assessment of the affordability criteria requires economic data to be drawn from life-cycle costing, and these are rarely used in S-LCA.

Based on this observation, it can be questioned whether these frameworks are able to correctly assess products from nonmarket-oriented production and consumption activities: e.g., impacts linked to a product grown, processed, and consumed on a farm. From Table 4, it appears that the perspective taken by context-oriented, impact-based, and other theory-structured frameworks, which aim to assess effects or impacts and with indicators located further down the impact pathway (Mathe 2014; Wangel 2014), makes it possible to assess the whole life cycle, including the consumption stage.

3.3.3 The relations between value chain actors and economic aspects at the product level overlooked

Noticeable differences can also be seen regarding the inclusion of C&I relating to the relations between value chain actors. While the Guidelines propose subcategories regarding competition or the relationship with suppliers, the SHDB and the Product Social Metrics that have been built on the Guidelines assess company practices of a value chain without considering their relationship to one another. Yet, relations with suppliers and clients, as well as market context (competition, concentration), may be important stressors for S&SE impacts. In this regard, approaches of Kruse et al. (2008), the Guidelines, and the MCM Capacities S-LCA of using criteria relating to attributes of the value chain and to value chain governance allow potential additional stressors to be taken into account.

In addition, among the 14 frameworks reviewed, only two allude to fair prices in the criteria to be assessed (Garrabé and Feschet 2013; Schoeneboom et al. 2012). Yet, the Guidelines recognize that socioeconomic processes, such as the pressure for low prices, are causes of social impacts, in addition to company’s behavior (Benoît and Mazijn 2009).

In conclusion, several of these frameworks do not provide a holistic coverage of life-cycle stages and, without considering consumers and relations between actors in the value chain, do not provide a framework capable of taking a systemic approach. Through the analysis of these frameworks, we can identify how the list of subcategories of the Guidelines could be developed to expand its coverage.

4 Conclusions and perspectives

Our review first highlights that since the beginning of research on S-LCA in the mid-1990s, many methodological S-LCA frameworks have been proposed. These present quite different visions on the purpose of S-LCA and of what an S-LCA should assess. This can be seen in the origin of the C&I part of the frameworks and in the diverse ways in which they are selected.

In this paper, we have identified 14 frameworks, which can be classified in five framework types. With this typology, we identify the main common features and divergences among frameworks regarding the rationale behind the selection of C&I but also regarding the assessed phenomena, with respective strengths and weaknesses: Value- and context-oriented can help build C&I sets that are meaningful and relevant for people and that are legitimate, in contrast to other sets that are defined with one type of stakeholder. Applicability-oriented approaches can help feed frameworks with useable indicators, including for the least-covered stages, such as the use stage. Theory-structured frameworks set a frame that can help to broaden the set of C&I and to articulate C&I with each other, including on the impact pathway. Finally, impact-based frameworks help define C&I that are scientifically legitimate and could help validate/invalidate other C&I sets. Thus, the differences in the selection of assessment C&I result in different sets of C&I that are used for assessment and, thus, in different evaluation exercises.

While many of the frameworks mainly focus on practices of companies toward other stakeholders, some others include C&I positioned further down the impact pathway not directly linked to companies’ practices and more context-related. The topics covered by the selected C&I are also further linked to the assessed phenomena. Generally, phases that correspond to an economic activity are better covered than the use phase. This could be due to the company perspective used (as in Benoît and Mazijn (2009) and Garrabé and Feschet (2013)). In addition, most C&I refer to individual stakeholders, which overlook governance and economic aspects of value chains.

Based on this review, we were also able to identify areas for potential improvement and ways for future development.

4.1 Perspectives and recommendations

4.1.1 Selecting C&I at the impact level

The review shows that the ways of selecting C&I are very diverse among the various frameworks, resulting in different C&I sets. The diversity is further enhanced by the practitioners of the frameworks who can adapt the C&I set and select C&I for a particular study. Firstly, we would like to call for the rationale for selecting C&I (on framework and application level) to be made more clear, since “too often, the list of indicators is not justified at all” (Iofrida et al. 2016, p. 12). Secondly, we would like to put forward the participatory approach as a potential method for C&I selection. We would encourage the use of assessment criteria that reflect people’s values and that are legitimate and meaningful to stakeholders. In fact, if recognized as important by stakeholders, it seems that there is no need of another rationale to justify the inclusion of one criterion or another. We assume that stakeholders will select assessment criteria for different reasons, but there are mainly two: first, they want to protect the elements themselves that are reflected by criteria: these elements have an intrinsic value (Jolliet et al. 2004; Weidema 2005) and correspond to S-LCA midpoint or endpoint categories (e.g., decent working conditions, human health); or, second, they believe that certain elements have a positive/negative effect on the elements they want to protect: these elements have an instrumental value and could be considered as stressors of social and socioeconomic impacts (e.g., fair competition).

4.1.2 Positioning C&I on impact pathways

In context, we could propose that C&I be classified according to their positions on impact pathways: as midpoint/endpoint impacts or as stressors. This exercise can be useful in revealing the rationale for the inclusion of one criterion or another in S-LCA: is it included because it is an impact/area to be protected, or because it is thought to have effect on an impact? This exercise could be included in the goal and scope phase, as part of the definition and selection of the C&I to be assessed. This classification would also have implications for the LCIA phase: for midpoint impact variables, it could be argued that evaluation using impact pathway methods is preferable. If impossible (or unavailable), a performance assessment remains possible. The relevance of stressor variables should be checked with the investigation of impact pathways linking them with the midpoint and endpoint impacts. This would help to build knowledge about the levers that need to be activated to improve social impacts in product chains.

In fact, as highlighted by our review, several of the reviewed frameworks assess or consider impact pathways, but most of the reviewed frameworks consider assessment criteria separately. In the latter frameworks, it is not always clear what the practitioner intends to assess with a certain indicator and where the indicators are positioned on the impact pathway. Our proposal to classify C&I according to their positions on impact pathways is in accordance with and complementary to the proposal of Russo Garrido et al. (2016) for practitioners to specify what phenomena they intend to assess with each indicator, especially for variables that are not found on an impact pathway (i.e., context variables). In this way, practitioners could specify whether the indicators are used “as contextual information providing insight on potential effects of corporate social performance, as a proxy for a phenomenon occurring within the product system and/or its resulting causality chains, or as an element that may affect/condition activities at the company level” (Russo Garrido et al. 2016, p. 5).

4.1.3 Integrating and investigating the link between impact variables and stressor variables

C&I classified as stressor variables should not be left apart but should actually be integrated in the assessment. Indeed, most frameworks identified in this review propose to go up to the assessment and reporting of (midpoint) impacts (i.e., company practices toward other stakeholders or the Guidelines subcategories in type I LCIA, e.g., working conditions). However, it has been argued that impacts affected by these midpoint impact categories (e.g., well-being impact) should be assessed as well, e.g., through the investigation of impact pathways (Feschet 2014; Macombe 2013b). We argue here that in addition to assessing affected (downstream) impacts, it would be useful to look into the reasons why (upstream) a certain midpoint impact is at a certain impact level or performance (e.g., socioeconomic processes, such as pressure for low prices, can be the reason for poor working conditions). We believe that looking into these reasons could increase S-LCA’s potential to identify improvement options. This would imply looking for and integrating stressors of midpoint impacts (i.e., instrumental variables) in the assessment. Therefore, we would like to encourage the “impact pathways research” to look into the link between midpoint impact variables and related (upstream) stressors.

Among the reviewed frameworks, we already found C&I that could be considered as social stressors, i.e., elements influencing or constraining company practices and therefore affecting and explaining impacts: e.g., relation between value chain actors (Benoît and Mazijn 2009; Garrabé and Feschet 2013), market context (Kruse et al. 2008), economic aspects throughout the value chain (e.g., fair price), and of separated value chain actors (e.g., profits) (Garrabé and Feschet 2013; Kruse et al. 2008; Schoeneboom et al. 2012). In fact, from a systemic approach, S&SE impacts are obviously drawn not only from social and organizational aspects of chains but also from governance and economic aspects. By integrating such stressors into the S-LCA methodology, we could strengthen the capacity of S-LCA to contribute to sustainability management and further the fulfillment of the S-LCA promise to provide a holistic assessment where the variables are considered as elements of a system that are linked with other elements of the system.