The role of gender, age, and income in demand-side management acceptance: A literature review

The world is currently facing a global energy transition from fossil fuels to renewable energy sources. The electrification of various sectors, from transportation to heavy industry, is driving an increasing demand for renewable energy, which adds pressure to the transitional process. At the same time, energy transition has become more urgent in the context of current shortages in energy supply due to global developments such as Russia’s invasion of the Ukraine and associated price escalation. In a system characterized by a high proportion of renewable energy sources, end-users possess the potential to play a significant role in ensuring the stability of the grid (Gelazanskas & Gamage, 2014). Successfully engaged as active participants in the grid, they can become a source of flexibility through close coordination of energy availability and consumption (Ballo, 2015). At the same time, so-called flexible energy consumption does interfere with the everyday life practices of consumers, posing a challenge to the availability of end-user flexibility as a resource (Silvast et al., 2018; Skjølsvold et al., 2017) and grid companies may overestimate public acceptance to shift their energy demand (Winther & Sundet, 2023).

Research in the field of promoting flexible energy consumption among households has primarily concentrated on the provision of economic signals as incentives, often coupled with technological interventions to facilitate the transmission of these signals, with the objective of encouraging individual households to modify their energy consumption patterns. This approach is commonly known as demand-side response (DSR) or demand-side management (DSM). DSM can be done in several ways: manually by using electricity when the price is low, manually by setting up the automation of appliances or by installing a Home Energy Manegment System (HEMS) which allows a third party to manage use (Adams et al., 2021). It is assumed that individuals’ willingness to adapt their energy consumption is driven by economic advantages and rational decision-making (Fell et al., 2014; Fjellså et al., 2021a, 2021b; Throndsen, 2017). This assumption comes from a technical energy utility perspective where the main motivation is to stabilize the grid (Ballo, 2015) and is often based on the vision of the ‘resource man’: a smart energy consumer interested in his own energy data and who acts on price signals to optimize savings (Strengers, 2014). This vision represents the energy industry’s resource bias projected into energy consumers that make the home into a “resource control station” (Strengers, 2014, p. 26). Another similar concept used to describe this set of assumptions made on behalf of users by designers of technology and policy is the concept of ‘imagined lay persons’ (ILP) whereby the expectation of a version of the future—like that of the flexible consumer—creates certain types of innovation projects that yet again facilitate certain action strategies and politics (Maranta et al., 2003).

When it comes to end user flexibility, the imagined lay person becomes a representation affected by unconscious bias about who the user is: “usually white, male, privileged, well-off, and young” (Maranta et al., 2003; Strengers & Kennedy, 2021), leading to the development of solutions that exclude or simply do not appeal to many other energy users. This bias has been addressed by social science researchers who have called for more heterogeneity among participating energy consumers in pilots and studies (Silvast et al., 2018; Skjølsvold et al., 2017; Strengers, 2014), and underlined the need to pay more attention to social relations in energy consumption patterns (Hargreaves & Middlemiss, 2020). These studies are part of an increasing body of research that critically reflects on the complex identities, use patterns, and social contexts of end users (Forlano, 2017). In this way, a houshold’s economic situation and compostion has an impact of its acceptance of DSM.

Through the analysis presented in this paper, we further contribute to this line of research by scrutinizing published studies on automated DSM and the ways in which they engage with end users. Additionally, we propose moving beyond mere critical reflection and advocate for employing intersectionality as a means to expand the scope of the social license to automate concept, making it more inclusive, tangible and usable for the furthering of a socially just energy transition.

The concept of the ‘social license to automate’ refers to the circumstances under which energy users might accept automation within their households. As such, it brings to light the potential gaps between the expectations of energy actors and users’ perspectives and aspirations (Adams et al., 2021). The concept of a social license to automate recognizes that many demand side management (DSM) programs will require intervention at faster speeds than systems of individual decision-making will allow. For example, Frequency Control Ancillary Services at the milli-second level require pre-set permissions, which raises the question as to who provides these permissions as they relate to home appliances, batteries, HVAC systems, electric vehicle charging and more. Granting permissions has important social and political implications insofar as DSM requires users to modify their practices. The social licences to automate concept gives us three key boundaries: legitimacy for acceptance, credibility for approval, and trust for psychological identification with the DSM. In this way, DSM is not merely a question of personal choice but is rather the subject of social conditions required by community stakeholders to address these three barriers of boundaries to gain social license (Boutilier & Thompson, 2011). One of the current limitations of the social license to automate is that it lacks a means to understand the complex, social situatedness of energy users.

To bridge this gap, we combine the social license to automate with an intersectional analytical sensitivity that foregrounds the complex identities of social subjects. When combining SLA with intersectionality, we gain the means to analyse the potentially overlapping ways in which social inequality is produced and reproduced in household contexts and through energy transitions. This is of particular interest to a field like STS where techno-experiences are best understood as shaped by various enactments of gender, class, race, age and disability (Moser, 2006). In this paper we utilize intersectional thinking as a tool to map out the DSM literature in terms of insights and knowledge gaps related to understanding the social situatedness of energy consumers in their household contexts. In other words, it allows us to analyze the review and “become answerable for what we learn how to see” (Haraway, 1988: 583).

Intersectionality is a term coined by Kimberlé Crenshaw in 1989 to describe mutually reinforcing forms of social oppression related to race, class and gender. This novel thinking about identity as being situated at intersections rather than grounded in single categories has since become what some scholars have referred to as the most important insight to emerge from women’s and gender studies in the past thirty years. Since its beginnings, this framework has travelled beyond women’s and gender studies into many other academic disciplines, with little consensus on how it may be used. While intersectionality is commonly understood as an ‘analytical sensitivity’ that spotlights the mutually reinforcing nature of social inequalities grounded in gender, class, race, ability, and ethnicity (to name a few), an intersectional perspective also offers ways to foreground so-called ‘critical inquiry,’ which is what we pursue in this literature review (Collins, 2019). Grzanka et al. (2023) have also noted that although intersectionality is now a widespread framework for “understanding, critiquing, and intervening in complex social inequalities” that has travelled from women’s studies to other disciplines, its potential remains underutilized in STS. Sharing similar social justice priorities, the concept of Social License to Automate allows us to focus on a diverse community of users and the social processes implicated in energy management. Power is critical to most social processes, and energy management is no exception. Intersectional approaches – which seek to combine gender with other axes of difference – have much to offer here. Intersectionality aims to go beyond single category analysis (gender or social class, for example) and instead aims to make visible the unique locations of individuals experiencing compounding marginalisations in what Collins (2019) refers to as the ‘matrix of domination’. Translated into practical terms, we do not aim to suggest that DSM technologies must take into consideration each and every possible intersection of social identity. Rather, it is our aim to highlight the possible risk of DSM to reinforce inequalities or marginalization if intersectional thinking about social difference is not taken into account. In the context of energy consumption for example, an analysis based on the category of ‘gender’ alone may fail to account for the fact that ‘women’ is hardly a homogeneous category and that there are distinct challenges that face different groups of women. Elderly women, for example, may be more often situated by social forces related to age, physical ability and reduced income levels due to loss of a partner. Intersectionality therefore “offers expertise in the study of power, one that is under-utilized yet applicable to many aspects of energy system design, planning, exchange, and use.” (Bell et al., 2020).

Through this literature review, we aim to understand which perspectives are included in the literature on DSM and to what extent these perspectives reflect intersectional thinking. With these insights, different motivations and conditions for crossing the social license boundaries of legitimacy, credibility and trust might be identified and highlighted, allowing for the emergence of a more inclusive and tangible social license to automate. Only by recognizing these social issues can DSM scale up from resource man to a broader range of everyday users. Our review of the literature on DSM has therefore been carried out using intersectionality’s complex thinking about social inequality. This framework allows us to gain insights into three key aspects of home energy consumption:

By adopting this analytical approach that foregrounds gender, power and marginalized identities, we move beyond traditional areas of research, such as those that analyze the experiences of tech-savvy, white, and privileged men as pilot users of DSM and aim to promote energy justice through the lens of recognition justice (Jenkins et al., 2016).

Intersectionality is now one of the most widely used frameworks for understanding social inequality; it therefore plays a crucial role in providing a more nuanced perspective on user experiences, needs, and opportunities within the context of social justice. This approach directs our attention to how program design, when informed by single category thinking, can inadvertently reinforce existing injustices and further marginalize already vulnerable groups during the transition to sustainable energy sources. When we examine the SLA concept through the lenses of intersectionality we expand the concept’s scope to encompass fairness and justice as important dimensions of program development. This expansion enriches its utility as a tool to promote the democratization of energy systems. Furthermore, it allows us to consider how gender and other social inequalities may emerge and persist in relation to Demand-Side Management (DSM) and can inform program design aimed at mitigating such disparities.

Essentially, for an equitable transition to sustainable energy solutions that incorporate automation, we need an SLA in dialogue with intersectionality’s critical inquiry to understand gendered dynamics and the experiences of women, but we also need more insight into the experiences of different kinds of women—those groups whose everyday realities are particularly susceptible to the unpredictable fluctuations in energy prices. It is this intersectional thinking that brings inequality and power to the forefront of the review process, giving us the possibility to critically revisit the existing literature on demand-side management (DSM) with a specific focus on narratives outside of those of the dominant group (e.g., the resource man). Lived experiences can then be considered as part of the pathways leading towards the three social license boundaries (legitimacy, credibility and trust), their possible crossing and translated into concrete recommendations.

In this review, we have used intersectional thinking to pursue complex rather than simplistic thinking about background variables like age, gender, and housing type in the context of DSM. We build from these insights to outline differing motivations and conditions for crossing the social license boundaries of legitimacy, credibility, and trust. In other words, we use intersectionality as a theoretical orientation and framework for examining interconnections and interdependencies between social categories and systems (Atewologun, 2018). The approach of a narrative review was chosen to allow for an exploratory evaluation of the literature and to synthesize in-depth qualitative insights from a variety of perspectives and disciplines (Sovacool et al., 2018).

The review is conducted by an interdisciplinary group of researchers in collaboration within the User-Centred Energy Systems Technology Collaboration Programme under the auspices of the International Energy Agency. The first step was to identify relevant literature, and the researchers initially discussed potential issues in implementing automated DSM in relation to the more complex understanding of identity that intersectionality calls for (Schiebinger et al., (2011–2020)). Through these discussions, the search query was formulated, and the literature search carried out through the scientific databases Web of Science and Scopus between January and March 2023. We searched for articles in English published in peer-reviewed, scholarly journals and conference proceedings. We proceeded with search terms related to social categories that we hypothesized might experience marginalization in the DSM literature. We knew, for example, that gender, age, and income or economic status have been previously addressed in the literature, but we were also curious whether there were any papers on households with disabled users, minority groups and which possibly examined race as a factor in user acceptance. Our keyword search therefore left open the possibility to discover new intersections. Papers were therefore targeted for review if they contained empirical studies (both quantitative and qualitative) explicitly related to all three of the following topics:

Papers were excluded if they i) focused on non-domestic sectors (industry, or individuals as employees), ii) concerned non-empirical results (e.g. modelling or theoretical work), iii) focused solely on technological solutions or demonstrations of technical solutions, and iv) did not include any analysis based on marginalized groups (only simple reporting of such groups in the context of sample description). After removing duplicates, the search produced 255 documents for initial screening outlined in the PRISMA (Page et al., 2021) flow diagram (see Fig. 1).

Full size image

The second step was to screen the search results based on the abstracts. The selection of papers was done independently by the two first authors, who then compared selections and discussed deviations, agreeing on the final list. A close reading of these texts was conducted. The reading was guided by a pre-defined template, where the reader noted for example important aspects of the user/target group, methodology employed, geographical context, and type of flexibility studied. Most importantly for this review, the reader also isolated which marginalised groups and intersections where considered (e.g. female and low income, or age and high income, etc.), and which insights were produced regarding their experiences with DSM: willingness and ability to participate as well as potential consequences of doing so. In this second step of the analysis, we included papers that differed significantly in the extent to which intersectional perspectives or approaches were applied. Three levels could be seen: level one categorization was used to describe the least intersectional studies featuring analysis that simply described socio-demographic factors; level two was used for papers where marginalization and social inequality were theory-based and referenced in the background literature only; and level three was given to those where analysis of marginalized groups is theory-based and serves as the central focus of the paper’s analysis. Table 1 (in appendix) gives an overview of the categorized articles by geographic context, methodology, perspectives, and analytical levels of intersectionality. This analysis yielded clusters of papers based on how the DSM literature had addressed gender, economic status and age. Within these analysis categories, we looked at the extent to which intersectionality came into play (Sect. 4.0). Our literature search confirmed that gender, age, and socio-economic status were indeed social categories covered in the literature – albeit in a way that primarily addresses them as single variables – and in addition, we found that there is a knowledge gap on DSM technology’s use by persons living with disabilities and racialized minority groups.

In a third step, additional articles suggested by reviewers and commentators on the draft were incorporated and further analysis was carried out to uncover the nuances and themes related to each perspective, specifically focusing on the articles that had incorporated an intersectional thinking to the greatest extent. The implications with regards to the social license concept are considered throughout the analysis and recommendations building on these are discussed as part of the conclusions.

We did not define any geographical scope, however most papers were published in a western context including Europe, USA, and Australia. Additional studies carried out outside of a western context are also included (see Table 1 (in appendix)). The methods include both quantitative and qualitative but in general with a mix of qualitative approaches dominating those papers with some form of intersectional analysis. Surveys are most commonly used in papers that have engaged to a lesser degree with marginalised perspectives. This is unsurprising as qualitative research is most often used in intersectionality’s aim to ‘give voice’ to such groups.

The reviewed literature reveals that few articles have engaged with or sought out marginalized perspectives in an in-depth, qualitative manner in their analysis. Less than a third of the articles (17 out of 55) included detailed analysis focusing on marginalized groups and their experiences. The analysis revealed that three main perspectives have been considered in the literature: (1) gender, (2) economic capital, and (3) age. Out of these, most focus has been on gender. The following sections will present findings from the literature related to those three categories.

Studies report that women may be more willing to engage with energy conservation behaviors and flexibility practices (Clancy & Roehr, 2003; Grünewald & Diakonova, 2020; Khalid & Razem, 2022), or at least more willing to follow through on their intention (Tjørring et al., 2018), than men. However, in practice, the realities of family life present create obstacles to such behaviors, including efforts undermined by other members (Johnson, 2020).

The social license to automate concept highlights three key boundaries: legitimacy for acceptance, credibility for approval, and trust for psychological identification with the DSM (Adams et al., 2021). Adding insight from intersectionality brings the context around these three key barriers into better focus, and we see that questions around legitimacy, credibility and trust may be more effectively addressed by research that sees DSM entangled with household dynamics and power (who carries out domestic tasks and who surveils them), and complex social identities as they concern both privilege and disadvantage. In other words, the social license to automate is not just about one individual’s legitimacy but rather about gaining credibility and trust more widely from the entire household and wider society with its diverse groups of users.

Our methodology in this review has been driven by a commitment to broaden the understanding of the domain of DSM and to map out the extent to which marginalized voices have been taken into consideration in developing solutions that do not reinforce existing social inequalities. As a result, we initially employed multiple categories to identify relevant literature on DSM. However, our findings indicate that existing DSM literature predominantly focuses on certain demographic groups, such as men, high-income households, and middle-aged individuals in addition to the more restricted number of counter studies that concern gender, age and economic status that are mentioned above. We will use the next section to discuss how the SLA 2.0 with the help of intersectionality can aid us in lifting other imaginaries and values than the ones that have traditionally dominated DSM technology development. Some examples of this include values that address collaboration, diversity and equality in the everyday domestic energy lives of people (Pink et al., 2023). Such values can be helpful for designers to actively use to open up their own biases concerning how they frame the user and what role automation has in the DSM project (Verkade & Hoffken, 2018).

This review uncovers a scarcity of studies addressing DSM in relation to social factors and the ways in which household social dynamics influence consumer energy behaviour and its potential for adaptability. We also show that, although many studies point out that gender, income, and age have an impact on the flexibility of energy consumption by impacting daily practices and material design such as DSM equipment, only a small number have approached this in a theory-based, in-depth, qualitative manner that goes beyond simply differentiating results according to demographics and “genuinely give voice to lived experiences”.

We enhance this line of research by scrutinizing published studies on automated energy systems and the ways in which end-users interact with them. Additionally, we propose moving beyond mere critical reflection and advocate for employing intersectionality to expand and practically apply the concept of social license to automate. While the social license to automate has been focused on understanding the (mis)alignments between the expectations of actors with the energy system on one hand and the everyday life of households on the other (Adams et al., 2021), our review takes this concept a step further by focusing on who we bring together in pilot testing and research on flexibility, and how we do it. Putting intersectionality into dialogue with social license to automate shows the need for automated DSM program designers to acknowledge power relations and context. This means designing in such a way as to acknowledge the ways in which material and non-material resources (e.g. digital alienation, everyday life routines, gender inequality not as single variables but interconnected) work to form systematic differences in access and knowledge to DSM solutions.

To build legitimacy with all user groups, designers need to take a closer look at consumers’ varying motivations and the lived realities in households: who carries which loads, who can access which type of flexibility under which conditions and which of their needs DSM can meet. Only with such an expansion of perspective can consumers beyond the typical, male, technology-afficionado and well-to do user group be reached and included. To achieve credibility, accessibility needs to be considered much more carefully, in medium, in language, in framing, and with the involvement of middle actors that can support the flow of communication to groups that are harder to reach. Finally, to establish trust, these communication pathways need to be solidified, the true involvement of diverse user groups needs to be made possible and inviting, ensuring that marginalized voices are heard and given weight. The concept of the social license to automate can therefore be applied by developers of DSM by reflecting critically upon question like:

Including social inequality, class, and gender in R&D projects for DSM will not only give project proponents a competitive advantage, but we argue it to be absolutely necessary to scale DSM beyond the scattering of pilots and technology-specific programs to the centerpiece of energy policy. This inclusion cannot be an afterthought. Users, rather than the grid, must be the starting point of DSM solution design. Such a shift and the associated changes of design and framing of solutions have the potential to open up access for less technology-savvy users and enable user-level connection with solutions through personal tasks and experiences. This opening of access would allow for greater transparency regarding personal impact and a conversation on equal footing. The results further show that to enable participation of the energy-poor, their needs and limitations should be integrated from the beginning into program design, finding ways to avoid disadvantaging and enabling benefits for users with little flexibility to offer and little to invest on their own.