Energy Prosumers’ Spillovers and the Policy Effect: Comparing Two Alpine Valleys in Styria and South Tyrol

Almost all human activities are directly (electricity, heat, mobility, etc.) or indirectly (food, clothing, etc.) linked to energy consumption (Pothitou et al. 2016). Thus, to overcome the energy and climate crises, we need a new energy lifestyle, not just technical solutions to reduce the dependence on fossil fuels. Therefore, technical solutions (e.g., reducing CO₂ emissions per unit of energy produced, increasing energy efficiency) are as important as changing human energy behavior (Rosenbloom 2017; Steg et al. 2021). However, unlike technical solutions, it is more difficult to change energy behavior as it comprises a broad variety of patterns related to the consumption or generation of energy. These are ranging from everyday routines (e.g., turning on the lights) that require little cognitive effort as they are mainly shaped by habits and social practices, to one-shot behaviors (e.g., installing a new energy system) that require more cognitive effort because they involve a conscious decision-making process (Dütschke et al. 2021). In this respect, people’s everyday energy behaviors occur mostly unconsciously, in the context of repetitive, habitual behaviors, or indirectly through other behaviors (Axon et al. 2018; Verplanken and Roy 2016). Consequently, people are usually not even aware of how much energy they consume (Pothitou et al. 2016). As the interaction between technologies, practices, and norms locks people in “old” behavior patterns (Axon et al. 2018; Maréchal 2010), sustainability initiatives face the challenge of having to change habitual behavior (Axon et al. 2018). And that is the crux: Changing habits is not easy (Moberg et al. 2019). Researchers recommend focusing on the context in which habitual behaviors occur – if the context changes, new behaviors are more likely to be established (Baum and Gross 2017; Maréchal 2010; Miller et al. 2015; Verplanken and Roy 2016).

In this regard, photovoltaic (PV) systems are promising as they turn electricity consumers into so-called prosumers (producers and consumers of electricity combined in one person), and thus, completely change people’s energy context (Kotilainen and Saari 2018; Palm et al. 2018). For prosumers, electricity becomes more tangible and touchable as a commodity, in contrast to consumers, who are usually unaware of where and how the electricity they consume is produced (Braito et al. 2017). In addition, prosumers typically receive feedback from the electricity-generating entity (Dütschke et al. 2021). This allows prosumers to observe, reflect, and reconsider their energy lifestyles (Stedmon et al. 2013).

For understanding changes in energy lifestyle, the “material participation” theory (Marres 2012) has its merits. It is an object-oriented perspective and assumes that using a novel energy technology can lead to new energy practices. For example, when PV panels become objects of participation and engagement, they can promote material participation as energy citizens (Ryghaug et al. 2018). This shift from a passive, habitual energy consumer to an active energy citizen is a significant change in people’s energy reality (Palm et al. 2018) that can trigger additional behavioral effects, so-called spillovers¹, which can lead to more or less compliance in the transition to a more sustainable (energy) lifestyle. Positive spillover effects on the environment can support the energy transition (Carrico 2021).

From a scientific perspective, spillovers are interesting because they make lifestyle change observable by directly focusing on holistic relationships between behavioral change within and across contexts (Galizzi and Whitmarsh 2019). Prosumers may experience energy-related and non-energy-related spillover effects.

In the literature, spillovers are discussed ambiguously – some studies find positive, others adverse, and even others no such effects at all (Dütschke et al. 2021; Geiger et al. 2021; Nash et al. 2017; Sintov et al. 2019; Stikvoort et al. 2020; Truelove et al. 2014). For instance, Palm et al. (2018) failed to demonstrate that prosumers in Sweden changed their energy consumption to a reasonable extent. Sorrell et al. (2020) found rebound effects, which erode energy and emission savings. Steinhorst et al. (2015) analyzed the energy-saving behavior of German consumers and found no spillover effects when the behavior change was economically framed. Similarly, Dütschke et al. (2021) studied Swedish prosumers and concluded that spillovers depend on the design of financial incentives. According to Maki et al. (2019), the ambiguous empirical findings on spillovers are due to the fact that different studies measure miscellaneous things: actual behavior, behavioral intentions, self-reported behavior, or the support for policies. Instead, Truelove et al. (2014) assert that studies are disjointed and come from different disciplines, and Palm et al. (2018) consider the numbers of studies that measure prosumers’ electricity consumption to be too small.

In this article, I present and test an alternative explanation. The underlying idea stems from a previous study (Braito et al. 2017), in which we found that energy policies have not only primary effects on encouraging PV investments but also secondary ones, such as crowding-in/-out segments of a society. We found that generous monetary incentives attract (crowding-in) households that pursue PV investments for economic considerations and an anthropocentric relationship with nature but tend to discourage (crowding-out) those groups that embrace non-monetary considerations as well as the motivations for collective energy projects.

The research question I now face is whether the ambiguous picture of spillovers may also be related to how households were induced to become prosumers in the first place. In other words: How do prosumers’ spillovers differ across policy regimes? The hypothesis is that higher financial incentives reduce spillovers. The idea that policies have not only the primary (intended) effect but also secondary effects is not entirely new. Baum and Gross (2017) classified secondary behavioral effects of policies and developed a behavior change framework that explicitly incorporates secondary behavioral effects. However, prosumers from different policy contexts have not yet been compared to test the effects of policy measures. In this regard, the article aims to examine descriptively and comparatively whether the hypothesis that prosumers in varying incentive schemes exhibit different spillover effects is valid.

Contributing to the sparse literature on spillover effects of prosumers, this article offers two novel contributions. First, the analysis assesses not only electricity-related behavior of prosumers, but also behaviors from other areas. So far, behavioral studies of prosumers focused mainly on electricity consumption patterns (Maki et al. 2019). Broadening the research perspective to include spillovers provides a better understanding of how successful behavior change intervention might be to disrupt habitual behaviors and stimulate more sustainable lifestyles. Second, the analysis explores the role of policies in disentangling the mixed and confusing results that research on prosumer spillover effects produced so far. A better understanding of what strengthens or weakens spillovers is valuable for the transition literature and can be helpful to practitioners considering behavior change interventions.

The analysis presented here is based on a data set from a previous study (Braito et al. 2017). It consists of a large-scale questionnaire survey of prosumers and non-prosumers conducted in two Alpine regions in 2015: i) Puster valley (Pustertal) in the province of South Tyrol, (Italy, IT), and ii) Mur valley (Murtal) in the province of Styria (Austria, AT). The two neighboring countries pursued different strategies to stimulate private investment in PV. Italy supported private investments with generous financial subsidies until 2015 (Orioli and Di Gangi 2016). Austria offered much lower financial support, but emphasized the societal and ecological added value of energy production through PV (Braito et al. 2017). Besides private investment in PV, this also encouraged the establishment of citizen solar power plants (CSPP), which allow investment costs to be shared and are seen “… as a form of empowerment, giving bottom-up initiatives control over energy infrastructures and channeling revenues to local communities” (Schreuer 2016, p. 126). Apart from the differences in PV policies, the two study regions are very similar geographically, culturally, historically, and socioeconomically. Typically, it is difficult for behavioral science disciplines to conduct comparative cross-national studies because too many contextual factors interact with the research objectives. However, the cultural-historical identity of South Tyrol² and the Austrian Alpine region is comparable. So, it is possible to conduct a “quasi” ex-post field experiment³, where the two different PV policies are the experimental intervention.

The Mur valley was selected for comparison with the Puster valley in an iterative exclusion procedure. The questionnaire survey was conducted between October 2014 and March 2015. The language of the questionnaire was German, and only German-speaking respondents in South Tyrol were interviewed to ensure comparability between the two subsamples and to control for cultural heterogeneity (for a detailed description of the selection of case studies and recruitment of participants, see Braito et al. 2017).

The sampling design distinguishes five respondent groups (see Fig. 1). In both regions, individual prosumers (PVAT, PVIT) and non-prosumers (nPVAT, nPVIT) were surveyed. Collective prosumers involved in citizen solar power plants (CSPP) could only be studied in Austria, as such collective PV initiatives had only just become established in Austria at the time of data collection (Braito 2017). This group was surveyed to understand whether being an individual or collective prosumer triggers comparable spillovers. The two groups of non-prosumers represent the control groups of the quasi-experimental research design. Spillover effects can be identified by comparing prosumers and non-prosumers within the same country. In addition, by comparing the two control groups across countries allows to examine which behavioral elements are admissible. Although, the two regions are very similar, it must be assumed that citizens from different countries do not have quite the same behavioral options, whether for infrastructural reasons (e.g., no public transportation network, no renewable energy supply) or simply because of different social practices. Thus, if the cross-country comparison of non-prosumers showed significant differences, it is assumed that the behavior of society in general is not comparable, and therefore the comparison of prosumers is also not valid.

When comparing the behavior of prosumers and non-prosumers, some spillover effects stand out. Table 2 provides the test statistics where the group comparison between prosumers and the control group of non-prosumers shows statistically significant differences. The effect size r of the differences is less than .3 (i.e., relatively small), but this does not detract from significance. As mentioned in sec. 3.2, a value of .2, for example, means that 20 % of the difference is due to the independent variable (belonging to a group). The results reflect previous studies: The spillover effects of prosumers are likely to be relatively small (Palm et al. 2018), negligible (Geiger et al. 2021), or even negative (Dütschke et al. 2021).

This research sought to enhance our understanding of prosumer spillover effects and thus contribute to the ambiguous literature. The results indicate that overly generous financial incentives not only carry the hidden costs of crowding-out intrinsic motivation (Braito et al. 2017), but also increase the likelihood of adverse spillover effects. Although market-based strategies can effectively promote environmentally friendly behavior (Maki et al. 2016) because they increase (economic) self-interest in, for example, energy-saving behavior (Pothitou et al. 2016), questions are also repeatedly raised about whether they support stable and sustainable behavioral change, because they may reward selfish behavior (Thøgersen 2003). Indeed, according to Bolderdijk et al. (2013), appealing only to economic self-interest is not always effective in securing behavior change. While PV investments increased tremendously during the government subsidy in South Tyrol, they stopped after the government subsidy ended in 2015 (Braito et al. 2017). Thus generous financial incentives do “… rarely lead to sustainable behavior change” (Axon et al. 2018, p. 583). Moreover, the PV policy in South Tyrol missed the opportunity to promote positive spillover effects by accompanying financial subsidies with a communication strategy.

Lanzini and Thøgersen (2014) attest that spillover effects of monetary support are at least as strong as those of verbal encouragement and praise. According to Lacasse (2016), highlighting environmentally friendly behavior helps to strengthen environmental self-identity, which in turn supports positive spillovers. Thus, labeling PV investments and environmental behaviors could prevent negative and reinforce positive spillovers (Geng et al. 2019). This is evident in the study case in Styria. It shows that more substantial positive spillovers arise when monetary incentives are complemented by additional instruments that appeal to intrinsic values. In particular, the case of CSPP shows that knowledge and communication, along with engagement, are potential factors that can change habitual energy behavior (Pothitou et al. 2016). For sustainability initiatives, the idea of behavioral spillover effects is attractive because such effects can lead to holistic behavioral changes in a cost-effective manner. Furthermore, it is tantamount to a missed opportunity if policies promote negative spillover effects, such as rebound effects.