Received: 2 December 2019 Revised: 14 November 2020 Accepted: 13 July 2021 DOI: 10.1002/eet.1960 RESEARCH ARTICLE Pro-environmental behavior: Social norms, intrinsic motivation and external conditions Mariateresa Silvi | Emilio Padilla Department of Applied Economics, Universidad Auto noma de Barcelona, Abstract Bellaterra, Spain Pro-environmental behaviors (PEBs) have been linked in the literature to social Correspondence norms, intrinsic motivation and external conditions. However, no study has jointly Emilio Padilla, Department of Applied analyzed these factors on a cross-country dataset and given a holistic explanation of Economics, Universidad Auto noma de Barcelona, 08193, Bellaterra, Spain. the variance observed in the adoption rates of PEBs across countries. Using a dataset Email: emilio.padilla@uab.es measuring individual adoption of eight PEBs in the European Union's 28 member Funding information states (in 2018), we econometrically test these three groups of factors on a wider ERDF, Grant/Award Numbers: BES- scale. We assess the importance of intrinsic motivation as a dominant factor and 2016-078156, RTI2018-095484-B-I00; Spanish Ministry of Science, Innovation and show how differing levels of intrinsic motivation influence the effectiveness of exter- Universities nal conditions, such as monetary incentives and green infrastructures. The results suggest that two-pronged policies, which take into account intrinsic motivation and external conditions, are needed to reach a high observance rate in the population in the short and in the long term. The wider significance of these results for policy is discussed. KEYWORDS economic incentives, environmental attitudes, green infrastructure, intrinsic motivation, pro- environmental behavior 1 | I N T RO DU CT I O N run a microeconometric analysis using a database counting 28,000 individual observations evenly distributed across EU's 28 members,1 Understanding the determinants of pro-environmental behavior (PEB) and covering 8 PEBs. in different contexts is important to designing policies that can pro- The literature on the determinants of PEBs is vast, yet the near mote more sustainable lifestyles across the European Union's entirety of contributions focuses on local contexts, which limits the (EU) socio-economic and culturally varied nations. Data show that potential applicability of findings to other areas. Additionally, while environmentally friendly behaviors differ widely among the population the literature has already reached robust conclusions regarding the of the EU member states. As an example, in 2014 the actual rate of contribution of either intrinsic motivation or external conditions taken municipal recycled and composted waste ranged between Germany's singularly, studies which account for both are rare. They reached con- 64% and Slovakia's 10% (European Environment Agency, 2017); while trasting findings and mostly focused on interaction effects—that is, 60% of Swedes reported buying green label products as opposed to whether providing extrinsic incentives can crowd out intrinsic motiva- 9% of Portuguese citizens (European Commission, 2015). This article tion to act pro-environmentally—, while disregarding that heterogene- explores such heterogeneity by empirically investigating the contribu- ity across individuals can lead them to react differently to the same tion of different determinants of PEB in the EU. To this purpose we policy. Consequently, the existing literature cannot entirely explain This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2021 The Authors. Environmental Policy and Governance published by ERP Environment and John Wiley & Sons Ltd. Env Pol Gov. 2021;1–14. wileyonlinelibrary.com/journal/eet 1 2 SILVI AND PADILLA the variety of results in the EU, as socio-cultural differences, intrinsic differences, and the reasons behind them. To the best of our knowl- motivation and external conditions—such as the green infrastructures edge there are only two cross-cultural studies in the literature that available—have a role in jointly determining different outcomes in have included both sets of variables. Cecere et al. (2014) concluded the EU. that waste reduction behavior is more strongly driven by intrinsic This article addresses this gap in the literature by: motives as opposed to extrinsic incentives across the EU-27. How- ever, their paper identified external incentives with the social visibility 1. Using a cross-country dataset to offer generalizable conclusions and desirability of a given behavior, a narrow definition that excludes applicable to different European geographical areas. the material factors that facilitate or hinder a behavior. We instead 2. Simultaneously including multiple factors affecting PEB—such as consider external conditions as the availability of green infrastructures social and individual values, and the green infrastructures locally or monetary incentives to facilitate PEB. Ferrara and Missios (2012) available—to identify the most impactful ones. assess the relevance of intrinsic motivation and different waste collec- 3. Defining whether the effectiveness of external conditions and tion policies for ten countries located across the five continents. They green infrastructures are subject to individual heterogeneity— find that intrinsic motivation and the availability of recycling facilities assessing whether individuals with differing intrinsic motivation are highly relevant in determining behavior, while the presence of respond differently to the same green infrastructure. monetary incentives to recycle—such as weight and volume based pricing schemes—are not very effective. The authors acknowledge The remainder of this article is organized as follows: Section 2 that the validity of the latter result may however be undermined by synthesizes the most relevant existing empirical studies; Section 3 the scarcity of data available for each pricing scheme. Furthermore, introduces our analytical outline; Section 4 describes our data sources while recognizing differences among countries, the study cannot iden- and includes a qualitative data analysis. Section 5 reports and dis- tify whether such differences are due to divergent socio-cultural cusses results and Section 6 concludes with key messages and policy values or institutional differences (i.e., regulations and policies). By implications. including indexes of predominant socio-cultural values in our econo- metric model we can capture whether national differences are due to cultural factors. Following Liobikienė et al. (2016), who proved the rel- 2 | LITERATURE REVIEW evance of social values in explaining green purchases in different countries, we use five of the six socio-cultural dimensions identified Like most human behaviors, PEB is jointly determined by cognitive by socio-psychologist Geert Hofstede (Hofstede & Minkov, 2010). processes that are internal to the individual and by the external con- Some of the studies accounting for both intrinsic motivation text that surrounds her/him. However, PEB has traditionally been and external conditions focused on a particular aspect of the rela- studied either as: (i) the outcome of an internal process of moral delib- tionship between these two groups of variables —analyzing eration in which the individual supposedly acts in complete autonomy whether providing external incentives crowd out the moral motiva- from her/his external context (Bamberg & Möser, 2007; Black tion to sort waste (Ferrara & Missios, 2012; Heller & Vatn, 2017; ska-Jurczak, 2003; Heberlein, 1981; Hopper & et al., 1985; Grodzin Thøgersen, 1994; Thøgersen, 1996; Thøgersen, 2003). While they Nielsen, 1991; Liobikienė et al., 2016; Sidique et al., 2010); or (ii) the analyze an important aspect, the relationship between intrinsic and consequence of an external stimulus to which the individual responds external conditions extends to other related and unexplored as an automaton regardless of her/his own convictions (Ferrara & research questions, such as whether the two groups of variables Missios, 2005; Jacobs & Bailey, 1983; Linderhof et al., 2001; Palmer & are effective conditionally to each other (Guagnano et al., 1995). Walls, 1997).2 Consequently, policy recommendations have typically For example, external incentives may be ineffective if not focused on one of these two aspects, favoring either educational supported by a basic level of intrinsic motivation, but strong intrin- interventions or alterations of external conditions through incentives sic motivation may also be ineffective on its own with a complete or taxes (Fehr & Gintis, 2007; Guagnano et al., 1995; Turaga lack of green infrastructures. et al., 2010). While both approaches have demonstrated some validity, This article contributes to the understanding of the conditions later evidence that human behavior is determined by both internal under which internal motivation is conducive of PEB and when pro- and external factors and their interaction has supported the develop- viding external incentives—such as the provision of monetary incen- ment of integrated frameworks (Jackson, 2005; Kirakozian, 2016; tives and green infrastructures—can be effective. The literature has Turaga et al., 2010; van den Bergh, 2008). observed that individual heterogeneity can result in different Previous research accounting for both intrinsic motivation and responses to external conditions (Beretti et al., 2019; Gneezy the external context have focused on case studies and field experi- et al., 2011). By clustering individual responses based on their level of ments (De Young, 1985; Derksen & Gartrell, 1993; Guagnano intrinsic motivation we can record whether the effectiveness of green et al., 1995; Heller & Vatn, 2017; Humphrey et al., 1977; Katzev & infrastructures and economic incentives to recycle vary across groups. Pardini, 1987)—restricting the potential applicability of the findings to This approach has the advantage of enriching the debate by highlight- the particular socio-cultural context involved. Extending the analysis ing why individuals respond differently to the same external to several countries offers two advantages: it highlights national conditions. SILVI AND PADILLA 3 3 | THEORETICAL FRAMEWORK 3.2 | The Attitude-Behavior-Context model The Norm Activation Model (NAM) is a theory designed to The ABC model postulates that the prevalence of a certain behavior in understand how pro-social behavior is affected by internalized a given population will depend from both their attitudes towards that norms (Schwartz, 1977). It is one of the most prominent social behavior (internal factors) and external conditions, intended as “any psychology theories that have been applied to the understanding external source of support or contrast to behavior, whether physical, of PEB (Onwezen et al., 2013). It models altruistic behavior as the financial, legal or social” (Guagnano et al., 1995, p. 702). Intrinsic moti- result of a cognitive process within the individual. Developed by vation (referred to by the authors as attitudes) can vary on a scale from socio-psychologist Shalom H. Schwartz between 1968 and 1977, the extremely negative—the person would perform the behavior only if theory analyzes the gap between an intention to act and actual forced to—, or very positive—the person would autonomously apply behavior. It identifies those key emotions that, if anticipated, help an the behavior. Likewise, external conditions can be very negative— intention to translate into actual behavior. While this approach helps making it difficult to perform the behavior—, or they can facilitate the explaining individual heterogeneity in altruistic behavior, it cannot behavior—making it easy or pleasant. In our article, we define external account for the effect of external conditions; nor for their interaction conditions as the combination of the physical facilities available to facil- with intrinsic motivation. The Attitude–Behavior Context (ABC) model itate PEB (e.g., biking lanes; recycling bins; public transport lines; etc.), completes the NAM, by incorporating it into a broader framework and the economic incentives put in place to promote PEB. that also includes external conditions such as physical structures, According to this theory, the effectiveness of a policy aimed to social institutions and economic incentives (Guagnano et al., 1995). either influence attitudes or external conditions will depend on both Our theoretical framework applies the ABC model, by maintaining intrinsic motivation and external conditions, rather than on the magni- the intrinsic factors specified in the NAM. In the next paragraphs we tude of a single policy intervention. According to these authors, if the summarize the two theories and describe how they complement each external conditions are extremely unfavorable or favorable to PEB, other. intrinsic motivation will not affect behavior, since regardless of how intrinsically motivated they are; people would be unable to follow the behavior in the first case or would do it anyway in the latter. For 3.1 | The norm activation model example, an educational program to improve environmental aware- ness will not have an impact on the population's littering behavior in a The NAM aims to explain the cognitive process and the circum- context in which there are no trash bins or, on the contrary, if there stances under which a personal moral norm is activated into are already bins everywhere and heavy fines for not using them. Simi- behavior. The NAM has proven more fit to explaining PEBs com- larly, Derksen and Gartrell (1993) found that intrinsic motivation can pared to alternative socio-psychological theories of behavior, enhance recycling rates but cannot overcome the barriers represented both conceptually (Thøgersen, 1996) and empirically (Cordano by a lack of infrastructures alone. et al., 2011). Numerous papers have corroborated NAM's main In this model, the opposite also applies: a policy focused on hypothesis in its application to PEB (Black, 1978; Cordano changing the external conditions of a behavior will not be particularly et al., 2011; Stern et al., 1985). The PEBs studied in the context effective on individuals that have a very negative attitude towards the of NAM include: recycling, energy conservation, automobile use behavior, nor it will affect individuals that were already very moti- or travel-mode choice, support for environmental laws and regu- vated and would have applied the behavior nonetheless. Empirical lations, and willingness to pay for environmental protection or for studies have supported the idea that the response to policy interven- greener products (see Guagnano et al., 1995, for a review of the tions differs based on user motivation (Abrahamse et al., 2005) and literature). that communication campaigns can be ineffective if targeting those The theory posits that an individual will behave in accordance who are already intrinsically motivated (Arkesteijn & Oerlemans, with a norm if she/he acknowledges the norm (i.e., she/he has 2005). Previous studies have highlighted that easier, less costly behav- internalized a social norm) and if the following two conditions iors are more likely to be adopted (Attari et al., 2011) and that in these apply simultaneously: (1) the person must have some awareness cases intrinsic motivation can be a strong predictor of PEB (Black that her/his potential acts may have consequences for the wel- et al., 1985). Conversely, higher costs of compliance act as a limiting fare of others—awareness of consequences; (2) the person must factor that impedes intrinsic motivation from translating into actual ascribe some responsibility for these acts and their consequences behavior (Black et al., 1985). to herself/himself—ascription of responsibility (Schwartz, 1968). Thus, the ABC model enables to identify boundary conditions We capture these three indicators within our econometric model. that must be met, for a policy to be effective on behavior. It guides The NAM mentions that the external context may have a role in the policy maker to adopt a more holistic approach, by helping her/- materially facilitating certain behaviors but it does not specify the him to anticipate the potential limitations of a single policy interven- conditions under which it happens nor the existing relationships tion. Are citizens of a country, sufficiently motivated to recycle, such between them. The ABC theory was developed 15 years later to that they would respond to a small monetary incentive to do it? Is the enhance the NAM theory by addressing these points. public transport network sufficiently developed such that an 4 SILVI AND PADILLA educational campaign on its environmental benefits could be enough and ascription of responsibility); and (iii) external conditions facilitating to shift behavior? the behavior. The three factors are interrelated among each other. Social norms affect behavior in two ways: indirectly, through internalized norms that are socially derived, but also directly, through peer-pressure to com- 3.3 | Causal model of relationships ply with observable social norms. Finally, the level of intrinsic motivation affects the effectiveness of external conditions. The figure reflects the Both models refer to social norms as forces that influence behavior. focus of our article although it is clearly not exhaustive, as other factors The NAM proposes that internalized norms are derived by the or relations may also influence PEB. social norms predominant in the society the individual lives in. The ABC model instead, considers social norms as a source of external influence, affecting behavior directly. To account for this compo- 4 | M A T E R I A L S A N D M ET H O D S nent, in our model we include country level social norms, taken from Geert Hofstede's six cultural dimensions theory indicators. 4.1 | Data sources The six dimensions were identified between 1963 and 2010 and they register the main socio-cultural differences among countries We access data from the Eurobarometer's 2014 survey on “attitudes (Hofstede & Minkov, 2010). Figure 1 summarizes our theoretical of Europeans towards environment” (European Commission, 2015), framework. We propose that PEB is determined by three classes of which surveyed 28,000 respondents in the EU's 28 member states. factors: (i) social norms, (ii) internalized individual norms and inter- The respondents who did not reply and replied “do not know” to at nal factors leading to their activation (awareness of consequences least one of the questions relevant to this analysis were excluded F I G U R E 1 Causal model of relationships Source: Own figure partly based on Schwartz (1968), Hopper and Nielsen (1991), and Guagnano et al. (1995) F I G U R E 2 Self-reported average number of PEBs in EU-28 by country (2014) Source: Own computations based on data from the European Commission (2015) [Color figure can be viewed at wileyonlinelibrary.com] SILVI AND PADILLA 5 F I G U R E 3 Ascription to personal individual responsibility for the environment in the EU-28 by country (2014) Source: Own computations based on data from the European Commission (2015) F I G U R E 4 Self-reported average number of PEBs in EU-28 by country (2014) Source: Own computations based on data from the European Commission (2015) [Color figure can be viewed at wileyonlinelibrary.com] from the data set; this reduced the sample size to nearly 23,000– in the previous month: waste separation for recycling; reduction of 25,000 observations, depending on the analysis carried out. waste—by avoiding over-packaged products and buying products with The Eurobarometer data set provided our response variables. It a longer life; reduction of domestic water consumption; reduction of registers whether respondents reported having performed eight PEBs domestic energy consumption; purchase of green-label products; 6 SILVI AND PADILLA The average number of PEBs adopted by individuals (on a 0–8 scale) per country varies between 2 in Bulgaria and 4.3 in Luxemburg (Figure 2). With a few exceptions, we identify patterns, with same color countries being close to each other in the figure. On average, Western and Northern Europe countries tend to have a higher num- ber of PEBs adopted compared to Southern and Eastern Europe coun- tries which are concentrated towards the right hand side of the figure. However, the ranking of countries changes depending on the behav- ior observed; for example, Southern Europe countries are the most dedi- cated to saving water whereas Northern Europeans are more likely to shop green label products. These differences could be due to diverging national priorities and environmental worries (i.e., water scarcity is more likely to plague Southern Europe); but also different economic possibili- ties (green label products are usually pricier and may be more accessible F I G U R E 5 Individual assessment of local green infrastructures in to on average wealthier Northern European nationals). EU-28 by country (2014) The level of ascription to personal individual responsibility for the envi- Source: Own computations based on data from the European Commission (2015) [Color figure can be viewed at ronment also varies widely among countries (Figure 3). The percentage of wileyonlinelibrary.com] individuals ascribing to the highest level of personal responsibility to pre- serve the environment on a 0–4 scale varied between 74% in the Nether- purchase of local products; choice of a greener way to travel; and lands and 22% in Poland. While the proportion of individuals who reject diminished car use. The same dataset also provided the following any responsibility ranged between 1% in Sweden and 8% in Hungary. types of causal variables: indicators of intrinsic motivation; proxies By plotting these last percentages against the average number of measuring the availability of green infrastructures; self-reported eco- PEBs adopted by the population we find a negative relationship nomic problems; demographic controls—age, gender, and years of between these two (Figure 4). This negative correlation may be a sign education. that national level differences in the adoption rates of PEBs is We integrated the database with two additional sources. First, channeled by differing level of individual responsibility. This negative a variable measuring the availability of green infrastructures that correlation could signify that countries with lower adoption rates of considers whether a country has adopted a container deposit col- PEBs may fail to instill a high sense of individual responsibility for car- lection system, that is, BottleBill, which rewards individuals eco- ing about the environment. We further notice regional patterns with nomically for returning used bottles and vessels. The list of Eastern Europe countries clearly gathered in the bottom right of the countries that had adopted this legislation by 2014, the year of the figure; Southern countries mainly clustered in the middle; while West- Eurobarometer survey, was taken from the website BottleBill.org, ern and Northern Europe countries are distributed within the upper maintained by the non-profit organization Container Recycling left corner. Regional proximity likely reflects socio-cultural proximity, Institute (Container Recycling Institute, 2016). The list was in the next section we use indexes of predominant socio-cultural incremented using evidence from a European Parliament report on values to test whether they influence PEB adoption rates. refunding schemes for drink containers (Schneider et al., 2011). It With regard to external conditions variables, individual assessments represents in our model a monetary incentive to recycle at the regarding the availability of green infrastructures vary with only 24% of national level. Second, the Hofstede cultural dimension country Greek respondents considering that their local government is doing level indicators were taken from Geert Hofstede own website enough for the environment as opposed to 65% of Luxembourgers (Fig- GeertHofstede.com (Hofstede, 2015). ure 5). Also in this figure, a clear regional pattern emerges with Western Europeans showing more satisfaction with the green infrastructures avail- able to them and Southern Europeans showing a higher discontent. 4.2 | Qualitative analysis Moreover, at the time of the Eurobarometer survey (2014), only 11 of the then 28 EU's countries had adopted a Container Deposit sys- In this subsection we first qualitatively analyze the variance of tem that rewards economically individuals who returned used vessels PEBs within EU countries. We then observe the distribution across for reuse or recycling, they are nearly evenly distributed across regions. countries of two classes of determinants, intrinsic motivation and external conditions, with the aim of identifying regional patterns that may explain such variation. We classified and color-coded 4.3 | Econometric approach countries based on their geographical area to highlight possible similarities between countries that share similar geographic and In this section we run separate regressions using a logit regression for cultural features: Eastern, Western, Northern, Southern Europe and each of the eight PEBs, to assess how the odds that an individual (i) will the British Islands. adopt a certain PEB (b) are affected by the joint influence of predominant SILVI AND PADILLA 7 TABLE 1 Explanatory variables definition and description Variable Description Intrinsic motivation variables NORMði,bÞ 0–1 dummy variable measuring the recognition of the environmental norm connected to the PEB examined; for example, the response to the question “Do you think it is a priority for people to separate waste for recycling?” is used for recycling behavior.a For b = 2–9 it is set to 0 if the respondent did not recognize the behavior specific norm and 1 otherwise. For b = 1 it is set to 0 if the respondent did not recognize any environmental norm and 1 otherwise.b ENVWORRY ði,bÞ 0–1 dummy variable measuring individual concern with the environmental aspect connected to the PEB examined; for example, the response to the question “Are you worried about the growing amount of waste?” is used for recycling behavior.c For b = 2–9 it is a 0–1 dummy variable, set to 0 if the respondent is not concerned and 1 otherwise. For b = 1 it is a 0–5 scale corresponding to the amount of environmental concerns expressed by the respondent (set to a maximum of 5 in accordance with the survey design). HIGHRESPðiÞ 0–1 dummy variable measuring whether the individual ascribes completely to personal responsibility for caring about the environment. It is set to 0 if the respondent does not and 1 otherwise. MEDIUMRESPðiÞ 0–1 dummy variable measuring whether the individual accepts a medium level of personal responsibility for caring about the environment. It is set to 0 if the respondent does not and 1 otherwise. NORESPðiÞ 0–1 dummy variable measuring whether the individual rejects entirely personal responsibility for caring about the environment. It is set to 0 if the respondent does not and 1 otherwise. External conditions variables ECOINFRAðiÞ 0–1 dummy variable measuring whether the individual believes that her city is fulfilling its duty in preserving the environment. It can be considered as a proxy for the availability of green infrastructures at the local level, enabling citizens to behave pro-environmentally (e.g., the presence of recycling bins, public transport and cycling tracks). It is set to 0 if the respondent does not and 1 otherwise. BOTTLEBILLðcÞ 0–1 dummy variable measuring whether the countryd c where the individual resides has adopted a container deposit law that organizes the collection of cans and bottles and rewards users with a voucher for fuel or groceries. It primarily represents the availability of a green infrastructure tied to an economic incentive to recycle, but it can also be considered as a proxy for the availability of other green infrastructures (e.g., if a country has adopted it, it may be more likely to have adopted other national-level green infrastructures as well). It is set to 0 if the country where the individual resides has not adopted it and 1 otherwise. Hofstede socio-cultural values variablese POWERDIST ðcÞ 0–100 index measuring the extent to which the less powerful members in country c accept an unequal distribution of power. The higher it is, the more important are dependence and subordination. INDIVIDðcÞ 0–100 index measuring how loose ties are among individuals in country c. The higher it is, the most important are independence, competition, personal achievement and self-reliance. MASCULðcÞ 0–100 index measuring how strongly emotional gender roles are set in country c. The higher it is, the most important are assertiveness, success and achievement. LONGTERMðcÞ 0–100 index measuring the extent to which members in country c are oriented towards future rewards as opposite to immediate gains, prioritizing saving, persistence and adaptation to changing circumstances. INDULGðcÞ 0–100 index measuring the extent to which members in country c feel free to pursue subjective happiness and have a sense of control over their own life. Socio-demographic controls FEMALEðiÞ 0–1 dummy variable capturing whether the individual is female. Set to 0 if the respondent is a male and 1 otherwise. AGEðiÞ Continuous variable reporting the respondent's age at the time of the interview. EDUðiÞ 1–10 index reporting the respondent's age when she/he left full-time education. 1 corresponds to “no full-time education” and 10 to “22 years or more” or “still studying.” FINPROBðiÞ 0–1 dummy variable measuring whether the individual reported problems paying her bills most of the times in the last year. It is set to 0 if the respondent did not and 1 otherwise. BOTTLEBILLðcÞ  FINPROBðiÞ Interaction term between the two dummy variables BottleBill and FinProb. It measures whether the fact of having economic problems and living in a BottleBill country has a positive effect on self-reported PEB. If significant, it would imply that economic incentives are effective in increasing PEB for individuals who are more sensitive to small economic rewards. Note: A full list of survey questions extracted from the Eurobarometer database and utilized for this model is available in Annex 1. a A list of the environmental norms corresponding to each behavior is available in Annex 2. b The ideal question to measure recognition of a norm would have been “Do you think people should separate waste for recycling?” The question contained in the survey instead, measures whether the individual considers a given behavior a priority. If the individual responds affirmatively, we consider that she/he implicitly acknowledges the behavior as a norm, as something that ought to be done. If the individual responds negatively, it may be that she/he acknowledges the behavior as a norm but does not identify it as ‘urgent’ or that she does not identify the behavior as something that ought to be done at all. Therefore, there is a possibility, depending on how the respondent interpreted the survey question, that our indicator may underestimate the number of individuals that recognize the norm. On the other hand, the opposite error (inclusion error) would have been worse. We are confident that the individuals who responded affirmatively, are individuals who recognized the norm. At best, the risk is that our estimated parameters for this indicator, that are in all cases, statistically significant, may be slightly underestimated. c A list of the environmental concerns corresponding to each behavior is available in Annex 2. d As of 2014, the year of the Eurobarometer survey, only 11 of the 28 surveyed countries had implemented a Bottle Bill system: Austria, Belgium, Croatia, Cyprus, Denmark, Estonia, Finland, Germany, Hungary, the Netherlands and Sweden. Despite their exclusion from the (Container Recycling Institute, 2016), we decided to include Hungary and Cyprus in the list following evidence reported by (European Parliament, 2011) and (BiPRO/CRI, 2015). e The socio-cultural dimensions identified by Hofstede are six, however, for the sake of this analysis we dropped one, namely “uncertainty avoidance,” because of multicollinearity with other variables. 8 TABLE 2 Impact of internal motivation, external factors and socio-cultural values on eight PEBs OLS LOGIT Sum of all Sum of all Sum of all Waste Reduce energy Buy green Buy local Greener way of PEBS 0–8 PEBS 0–8 PEBS 0–8 Recycling Recycling reduction Reduce water use use products products traveling Use car less 0.a 0.b 0.c 1 2 3 4 5 6 7 8 9 Norm 1.442*** (0.133) 1.442*** (0.133) 1.392*** (0.131) 1.007*** (0.031) 1.076*** (0.032) 0.476*** (0.035) 0.879*** (0.040) 0.811*** (0.029) 0.884*** (0.036) 1.177*** (0.034) 1.033*** (0.030) 0.543*** (0.034) EnvWorry 0.324*** (0.009) 0.320*** (0.009) 0.291*** (0.009) 0.120*** (0.031) 0.198*** (0.032) 0.141*** (0.000) 0.186*** (0.032) 0.157*** (0.030) 0.253*** (0.034) 0.145*** (0.042) 0.091*** (0.031) 0.150*** (0.047) HIGHRESP 0.602*** (0.021) 0.601*** (0.021) 0.518*** (0.021) 0.391*** (0.019) 0.299*** (0.020) 0.320*** (0.020) 0.264*** (0.019) 0.287*** (0.018) 0.369*** (0.024) 0.251*** (0.020) 0.245*** (0.019) 0.211*** (0.023) EcoInfra 0.049** (0.021) 0.012 (0.021) 0.228*** (0.065) 0.182*** (0.032) 0.070** (0.030) 0.004 (0.030) 0.042 (0.029) 0.183*** (0.034) 0.046 (0.030) 0.034 (0.030) 0.123*** (0.035) BOTTLEBILL 0.212*** (0.023) 0.074*** (0.026) 0.065*** (0.032) 0.353*** (0.041) FinProb 0.288*** (0.036) 0.283*** (0.036) 0.608*** (0.044) 0.566*** (0.052) 0.209*** (0.051) 0.165*** (0.046) 0.034 (0.045) 0.188*** (0.060) 0.104** (0.048) 0.110*** (0.047) 0.073 (0.057) FinProB * 0.133* (0.077) 0.281*** (0.077) 0.257** (0.110) BOTTLEBILL POWERDIST 0.005*** (0.001) 0.002** (0.001) 0.006 (0.004) 0.006 (0.004) 0.000 (0.004) 0.006 (0.004) 0.007* (0.004) 0.005 (0.004) 0.001 (0.004) individ 0.003* (0.001) 0.005*** (0.001) 0.020*** (0.005) 0.003 (0.005) 0.001 (0.005) 0.011* (0.006) 0.007 (0.005) 0.001 (0.005) 0.012** (0.006) MASCUL 0.002*** (0.000) 0.004*** (0.001) 0.008*** (0.003) 0.000 (0.003) 0.006** (0.003) 0.001 (0.003) 0.007** (0.003) 0.012*** (0.003) 0.004 (0.003) LONGTERM 0.005*** (0.001) 0.005*** (0.001) 0.010*** (0.003) 0.001 (0.003) 0.003 (0.003) 0.012*** (0.004) 0.007** (0.004) 0.006* (0.004) 0.011*** (0.004) INDULG 0.017*** (0.001) 0.034*** (0.001) 0.018*** (0.005) 0.018*** (0.005) 0.018*** (0.004) 0.025*** (0.005) 0.002 (0.005) 0.015*** (0.005) 0.030*** (0.005) FEMALE 0.242*** (0.021) 0.255*** (0.021) 0.277*** (0.021) 0.128*** (0.031) 0.175*** (0.032) 0.208*** (0.030) 0.267*** (0.029) 0.204*** (0.028) 0.314*** (0.034) 0.234*** (0.029) 0.197*** (0.029) 0.188*** (0.034) AGE 0.009*** (0.001) 0.007*** (0.001) 0.006*** (0.001) 0.011*** (0.001) 0.010*** (0.001) 0.005*** (0.001) 0.009*** (0.001) 0.005*** (0.001) 0.000 (0.001) 0.008*** (0.001) 0.005*** (0.001) 0.002** (0.001) EDU 0.058*** (0.004) 0.046*** (0.004) 0.047*** (0.004) 0.041*** (0.006) 0.045*** (0.006) 0.044*** (0.006) 0.004 (0.006) 0.040*** (0.006) 0.089*** (0.007) 0.063*** (0.006) 0.042*** (0.006) 0.040*** (0.007) RURAL 0.133*** (0.037) 0.148*** (0.045) Country fixed N N N N N Y Y Y Y Y Y Y effects Pseudo R2 0.136 0.143 0.175 0.127 0.206 0.109 0.119 0.115 0.183 0.173 0.143 0.079 F-test/2 log 529.176 358.273 306.231 25709.949 24231.579 27130.916 28319.956 29929.788 22353.527 27725.687 27950.623 22373.026 likelihood No. of obs. 23,516 23,516 23,087 23,107 23,107 23,088 23,088 23,088 23,088 23,088 23,088 23,088 Notes: Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.10. For regressions (1)–(10), the Nagelkerke R2 and McFadden 2 log likelihood are reported. Equation (0) shows an adjusted R2 and F-test of overall significance. SILVI AND PADILLA SILVI AND PADILLA TABLE 3 Impact of external conditions on three PEBs conditional on the degree of internal motivation Recycling 1 Reduce water use 2 Buy green products 3 High Resp (a) Mid Resp (b) No Resp (c) High Resp (a.2) Mid Resp (b.2) No Resp (c.2) High Resp (a) Mid Resp (b) No Resp (c) High Resp (a) Mid Resp (b) No Resp (c) Norm 0.946*** (0.054) 1.113*** (0.044) 1.379*** (0.164) 0.860*** (0.047) 1.033*** (0.038) 1.423*** (0.139) 0.795*** (0.062) 0.910*** (0.055) 1.243*** (0.212) 0.736*** (0.050) 1.034*** (0.053) 1.486*** (0.245) EnvWorry 0.137*** (0.054) 0.181*** (0.044) 0.027 (0.162) 0.090** (0.047) 0.153*** (0.038) 0.069 (0.139) 0.123*** (0.047) 0.252*** (0.045) 0.204 (0.187) 0.270*** (0.047) 0.231*** (0.050) 0.333 (0.232) BOTTLEBILL 0.080* (0.049) 0.024 (0.040) 0.014 (0.141) EcoInfra 0.121** (0.055) 0.045 (0.045) 0.171 (0.166) 0.280*** (0.048) 0.188*** (0.038) 0.304** (0.139) 0.025 (0.044) 0.028 (0.041) 0.046 (0.176) 0.253*** (0.049) 0.117** (0.050) 0.224 (0.242) FinProb 0.377** (0.080) 0.334*** (0.068) 0.253 (0.223) 0.592*** (0.068) 0.611*** (0.058) 0.574*** (0.184) 0.116* (0.069) 0.206*** (0.065) 0.134 (0.236) 0.164** (0.082) 0.245*** (0.092) 0.271 (0.346) Female 0.196*** (0.053) 0.196*** (0.053) 0.067 (0.164) 0.244*** (0.047) 0.086** (0.038) 0.139 (0.139) 0.288*** (0.043) 0.242*** (0.041) 0.461*** (0.174) 0.373*** (0.048) 0.258*** (0.050) 0.463* (0.237) Age 0.011*** (0.002) 0.011*** (0.002) 0.000 (0.005) 0.013*** (0.001) 0.013*** (0.001) 0.003 (0.003) 0.009*** (0.001) 0.009*** (0.001) 0.001 (0.005) 0.002 (0.002) 0.003** (0.002) 0.002 (0.007) Edu 0.055*** (0.011) 0.055*** (0.011) 0.023 (0.032) 0.029*** (0.008) 0.042*** (0.097) 0.023 (0.024) 0.003 (0.009) 0.014* (0.008) 0.013 (0.034) 0.087*** (0.010) 0.093*** (0.010) 0.088*** (0.046) SOCIAL VALUES Y Y Y N N N Y Y Y Y Y Y CONTROLS Country Fixed Eff. Y Y Y N N N Y Y Y Y Y Y Obs 10,816 12,192 881 10,816 13,400 1008 10,027 12,192 881 10,027 12,192 881 R-squared 0.245 0.255 0.313 0.050 0.064 0.103 0.101 0.121 0.194 0.175 0.153 0.187 p > chi2 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Log-likelihood 9092.293 12884.654 985.472 5603.9995 7962.3422 626.099 12828.016 14507.877 905.310 11033.685 10661.343 557.204 2 Notes: Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.10. The Nagelkerke R and McFadden 2 log likelihood are reported at the end of the table. Equation (1.a2)–(1.c2) excludes social values indicators and country fixed effects due to collinearity with BottleBill. 9 10 SILVI AND PADILLA socio-cultural values in her/his country (c), by her/his personal level of of ascription of responsibility—react differently to the same external intrinsic motivation to undertake that specific behavior, by her/his own conditions. This analysis contributes to the understanding of whether assessment of the green infrastructures locally available and by the pres- intrinsic motivation is a precondition for PEB and whether its absence ence of a container deposit scheme in her/his country. We also run an has an impact on the effectiveness of monetary incentives and green OLS regression to assess how the total number of PEBs adopted by each infrastructures. individual is affected by the same set of variables.  PEBði,bÞ ¼ F socio  cultural valuesðcÞ , intrinsic motivationði,bÞ , external conditionsði,cÞ 5 | RESULTS AND DISCUSSION ð1Þ The response variables PEBði,bÞ capture whether an individual (i) 5.1 | Relative importance of intrinsic motivation, reported having engaged in one of the eight behaviors (b) in the previ- external conditions and socio-cultural values ous month. Results of the regression performed on the entire dataset are reported • For b = 1 it is a 0–8 scale measuring how many of the 8 consid- in Table 2. In regressions (1)–(9), we analyzed the eight PEBs sepa- ered PEBs the individual reported having performed in the last rately, capturing the intuition—supported by Oskamp et al. (1991)— month. that environmental efforts and attitudes are fractioned into specific • For b = 2–9 it is a 0–1 dummy variable measuring whether the components that are peculiar to each behavior; that is, the anteced- individual has performed PEB b in the last month. It is set to 0 if ents of a PEB and their relative importance vary with each behavior. the respondent has not and 1 otherwise. Nevertheless, there may also be findings that are generalizable to most PEBs. The latter are captured in regression (0.a–c), in which the The full list of explanatory variables and their description is response variable represents the sum of the PEBs adopted by included in Table 1. the individual. We first run the model against the entire dataset, with the sole As shown by the R2 values reported in the table, the parameters in exclusion of Cyprus, for which we lacked data on socio-cultural the model were successful in explaining 7.9%–20.6% of the variation in values. We then re-run the model using logit regression after cluster- the response variable, a level that, given the complexity of human ing groups depending on their level of ascription of responsibility for behavior, is considered significant for studies with individual persons as caring for the environment. Three levels of responsibility are identi- units of analysis and a heterogeneous sample (Langbein, 2015). The fied, HighRespðiÞ , MediumRespðiÞ , and NoRespðiÞ . Comparing the results checks for collinearity did not reveal near dependencies among the across the three groups for a given PEB should give an indication of regressors used. Pearson pairwise correlation showed low correlation whether different individuals—as identified by varying degree in all cases with two exceptions. However, all the VIF and Condition F I G U R E 6 Policy makers roadmap to incentivize pro- environmental behavior Source: Produced by the authors SILVI AND PADILLA 11 Index values were well below the threshold values for multicollinearity source used, and they generally have an important fixed compo- of 10 and 30, respectively. nent that leads to decreasing average prices for the consumer. The The most notable finding is that, across all the PEBs observed, the user is more likely to control the final bill by changing the usage estimated parameters for intrinsic motivation—that is, Norm, HighResp, times rather than by reducing the consumption per se. For example, and EnvWorry—dominate the external factors; they have the highest Filippini (2011) found that households are highly responsive to the absolute value among the estimated regressors.3 All the intrinsic moti- changes in off-peak and high-peak energy prices and adapt their vation indicators are statistically significant and have the expected energy use accordingly. On the other hand, water tariffs in OECD sign: recognition of the norm (Norm), environmental worry (EnvWorry) countries tend to follow constant volumetric pricing, with a grow- and personal responsibility (HighResp) correlate positively with the ing trend to apply increasing block tariff systems, while the rele- dependent variable. The biggest impact is represented by the recogni- vance of fixed charges has declined significantly (OECD, 2009). tion of the norm, followed by the personal responsibility coefficient. Volumetric pricing schemes, and particularly increasing block tar- The relative predominance of intrinsic motivation over external iffs, encourage a reduction in water consumption. Although, as conditions and socio-cultural factors is also demonstrated by pseudo- income rises, a volumetric tariff scheme might be an ineffective R2 values in regression (0.a–c). In (0.a), intrinsic motivation alone mean to induce water-saving behavior among richer individuals. 4 accounts for 13.6% of the variation in the response variable. Incre- In the second regression for recycling (2), we introduce the mentally adding external conditions (regression 0.b) and socio-cultural interaction term BottleBill*FinProb. The coefficient is statistically sig- values (regression 0.c) leads to marginal increases to the pseudo-R2 nificant with a positive sign. This suggests that economic incentives values, 14.3% and 17.5% respectively. are effective in limiting the negative impact of economic difficulties External conditions are represented by the EcoInfra and BottleBill, on PEB. This last point is corroborated by the descriptive statistics their coefficients are statistically significant and have positive sign in from the database (European Commission, 2015) showing that: regression (0.b), showing that the presence of green infrastructures (i) Overall, individuals with economic problems are much less likely and monetary incentives increase the numbers of PEBs adopted. to recycle per se: only 59% of them recycle compared with 73% in These coefficients change to a non-statistically significant and nega- the rest of the sample. (ii) However, if an individual with economic tive coefficient respectively once socio-cultural values are included problems resides in a Bottle Bill country, she/he is 2 percentage (regression 0.c), possibly due to the high correlation between these points more likely to recycle her/his trash (60%) than if she/he two classes of variables. In the recycling regression (1), both coeffi- resides elsewhere (58%). cients are statistically significant with positive sign, suggesting that The five 0–100 indexes of socio-cultural values we included both green infrastructures and the presence in the country of a con- showed that predominant values in societies can explain a meaningful tainer deposit legislation positively affect the probability that individ- part of PEB adoption rates. Particularly, indulgent societies are more uals will recycle. Additionally, the coefficient for green infrastructures likely to engage in almost any PEB. Individuals living in more long- is not statistically significant or has negative sign when the other term oriented societies are also more likely to adopt a higher number behaviors are considered in isolation, thus leading to inconclusive of PEBs. They are more likely to adopt any behavior except for reduc- results. This shows that the presence of an external apparatus, which ing water and energy use. Both findings are expected since more facilitates certain PEBs, does not necessarily induce the adoption of indulgent societies are usually linked to a greater sense of control other unrelated PEBs, and reflect the fact that some PEBs, such as (Hofstede & Minkov, 2010), which can instill a greater sense of reducing water and energy use do not depend on green responsibility; and caring for the environment implies the ability to infrastructures. take into consideration future consequences. In individualist societies, Financial constraints affect most PEBs negatively with the excep- people are less likely to reduce car use, purchase green label products tion of green traveling, diminished car use and energy and water sav- and reduce waste. ing. While it is unsurprising that financially distressed individuals are Additionally, socio-demographic controls show the following: less likely to purchase green products or energy-saving appliances, Being a female correlates positively with all the PEBs with the which are often more expensive, the negative correlation with exception of “use car less,” suggesting that women are more likely to income-neutral behaviors, such as recycling, was unexpected. How- adopt any PEBs except renouncing to drive their car. This finding is ever, it is consistent with the hypothesis that, if an individual is dis- supported in the literature on altruistic behavior, such as PEBs, which tracted by more pressing personal circumstances, she/he is less likely finds that women are more likely to engage in other-regarding behav- to act in a norm-consistent way (Cialdini et al., 1990, p. 204). The ior (Hunter et al., 2004; Kollmuss & Agyeman, 2002) and hold more FinProb coefficient is statistically significant and positive for “reduce pro-environmental attitudes (Dietz et al., 1998; Vaske et al., 2001). water consumption,” and “green traveling” which is also expected Age correlates positively with all the PEBs observed except for green considering that these behaviors reflect positively in savings. Unex- traveling and reduced car use, suggesting that older people are in gen- pectedly, the FinProb coefficient is not statistically significant in the eral more observant of environmental norms, except when it comes “reduce energy consumption.” Several European countries have to considering alternatives to their own car. Predictably, the level of adopted pricing schemes whereby energy tariffs are determined by education also increases engagement in nearly all the PEBs, except for market prices, which vary depending on the time of the day and the water saving. 12 SILVI AND PADILLA 5.2 | Clustering individuals by level of attributed to differences in the level of intrinsic motivation in its national responsibility ascription populations and to heterogenous responses to the same policy. In particular, ascription of responsibility appears to be an essential Table 3 summarizes the estimated results for the same model with the precondition for an individual to respond positively to external incen- exception that survey respondents were divided into clusters. Logit tives, for example, being less prone to negative influences (such as split regressions are performed on three groups clustered by their economic constraints) and more receptive to enabling conditions level of personal ascription of responsibility for preserving the envi- (such as the availability of economic rewards or green infrastructures). ronment (a equals to high responsibility, b to medium responsibility This finding highlights a potential limit for the effectiveness of exter- and c to no responsibility at all). The three regressions are performed nal conditions: they can only be effective in promoting PEB among on three PEBs that differ in the way in which they affect income: individuals who have a minimum level of intrinsic motivation. (1) recycling (income neutral); (2) water saving (income positive); and Based on our results we advise that policy makers should start by (3) purchase of green-label products (income negative). assessing the level of intrinsic motivation—specifically ascription of The rationale is to assess whether people with differing levels of personal responsibility and recognition of the environmental norm— ascription of responsibility react differently to external conditions and for a given behavior in their target population. These data are regu- whether changes apply to different kinds of PEB, as defined by their larly surveyed within EU (Eurobarometer) and elsewhere, and they effect on income. The regressions include country fixed effects and include reports with summary statistics. If these average values are social values indicators. The regressions for recycling, however, are also low or if a consistent proportion of individuals rejects any environ- computed including BottleBill but excluding country and social values mental responsibility; the policy should focus primarily on increasing controls due to collinearity with this variable (Equation 1.a2–1.c2). both responsibility ascription and norm recognition. Policy makers Recognition of the norm (Norm) is the most important factor should, however, be conscious that changing intrinsic motivation is a across groups and behaviors (statistically significant coefficient and long-term goal (Thøgersen & Ölander, 2002). greatest in magnitude). In all three behaviors, the coefficient for There are, however, other policy avenues that can increase PEB EnvWorry is not statistically significant for the NoResp group. This adoption rates in the shorter term and that can run in parallel with the would suggest that, without a minimum level of ascription of responsi- long-term objectives mentioned above. If the target population is bility, environmental concern alone is not conducive to adopting PEB. highly motivated but norm observance is low, it may be an indication The availability of “green infrastructures”—captured by EcoInfra— that interventions should focus on facilitating the behavior financially, is relevant only to high/medium levels of ascription of responsibility, physically or socially. the coefficient is not statistically significant in the regressions for the Influencing social norms: National level socio-cultural values are NoResp group. This implies that, without a minimum level of motiva- the second factor affecting PEB. Social values and norms influence tion, green infrastructures alone are not effective. Green infrastruc- behavior indirectly, by influencing the norms that individuals internal- tures appear to negatively affect green label purchases for medium/ ize, but also directly through compliance with what individuals per- high levels of responsibility, a result of difficult interpretation. Con- ceive as social expectations. While cultural values are hard to change, tainer deposit schemes for used bottles positively impact recycling ample evidence supports that social norms can be influenced. Norm rates—the BottleBill coefficient is statistically significant with positive nudges are behavioral interventions aimed at changing social expecta- sign—but only for the HighResp group (Equation 1.a2–1.c2). tions. They can induce shifts in behavior by changing what people think others will do or approve of (Bicchieri and Mercier, 2014). Using economic incentives: Monetary rewards to recycle proved to be 6 | C O N CL U S I O N S effective to increase recycling rates, particularly for highly motivated indi- viduals and for financially strained ones. We observed that financially dis- This article used a cross-country dataset to empirically investigate the tressed individuals are less likely to adopt PEBs that imply higher costs reason behind different PEB adoption rates across the EU. To do so, (such as buying green label products) or that are cost neutral (such as our empirical model, based on survey data of 28,000 individuals recycling); but that they are more likely to engage in PEBs when they imply across the EU's 28 member states (in 2018), assessed the relative clear cost savings (such as reducing water usage) or rewards (such as strength of predominant socio-cultural values, intrinsic motivation and returning empty vessels to container deposit collection schemes). green infrastructures on eight self-reported PEBs. Reframing the choice context into economic terms—by attaching an eco- PEB depends on all the three classes of factors analyzed. However, nomic value to the behavior—appeared to be successful in raising the we find that intrinsic motivation—mostly internalized environmental observance of the environmental norm in this group. norms, but also awareness of environmental consequences and ascrip- Beyond providing economic incentives, policy makers should ana- tion to personal responsibility towards the environment—is the leading lyze the pre-existing ones. Pricing schemes of water and energy utilities force behind the eight PEBs considered. This finding is consistent across can discourage excessive consumption through volumetric pricing the eight PEBs examined and it is consistent with conclusions in Cecere schemes or increasing block tariffs. Conversely, with contracts that et al. (2014). Consequently, since intrinsic motivation varies widely across offer market-prices, the user is more likely to try to reduce the final the countries analyzed, diverging outcomes in the EU can be mainly bill by using electricity at off-peak times rather than by diminishing SILVI AND PADILLA 13 total usage.5 As another example, individuals may refrain from ENDNOTES energy-efficiency investments, even when is it economically conve- 1 We include in the count United Kingdom, which was still a EU member nient to do so in the long-run, for lack of financial means. Financing state at the time our dataset was created and at the time this paper was being written. schemes that highlight the economic benefits of energy-efficiency 2 investments would extend this opportunity to more people. A dicothomy first defined by Guagnano et al. (1995). 3 Providing or improving infrastructure: Some behaviors—such as reduc- Since all the intrinsic motivation and external conditions variables are dummies, the analysis concerning the magnitude of the estimated coeffi- ing water and energy consumption—do not need green infrastructures to cients among these variables is straightforward. be carried out, while others are highly dependent on the availability of 4 The pseudo-R2 result is 11.9% when excluding demographic control safe biking lanes, recycling bins, and so forth. Green infrastructures variables. increase the number of PEBs adopted for this latter group and it can 5 A related future research topic could be aimed at clarifying whether it is increase adoption rates among individuals that are already motivated. more environmentally beneficial to smooth energy consumption through They are, however, less likely to be effective on those who do not off peak-times or to just diminish total usage, given different composi- tions of the energy source mix. ascribe to at least a minimal level of responsibility for the environment. Strengthening intrinsic motivation: Individuals that recognize the moral norm related to the behavior—for example, "One should reduce RE FE RE NCE S car use"—and that ascribe to responsibility for the state of the environ- Abrahamse, W., Steg, L., Vlek, C., & Rothengatter, T. (2005). A review of ment become more likely to perform the behavior if they are also aware intervention studies aimed at household energy conservation. Journal of Environmental Psychology, 25, 273–291. of the consequences of not performing it—for example, bad air quality, Arkesteijn, K., & Oerlemans, L. (2005). The early adoption of green power by congestion and GHGs emissions. Information campaigns targeting this Dutch households: An empirical exploration of factors influencing the early aspect will not, however, be effective on individuals that do not ascribe adoption of green electricity for domestic purposes. Energy Policy, 33, to any personal responsibility for the environment. 183–196. Attari, B. S. Z., Dekay, M. L., Davidson, C. I., & de Bruin, W. B. (2011). The flowchart below summarizes these steps (Figure 6). Changing household behaviors to curb climate change: How hard can The considerations above lead us to conclude that raising intrinsic it be ? Sustainability, 4(1), 9–11. motivation should in the long term decrease the number of non- Bamberg, S., & Möser, G. (2007). Twenty years after Hines, Hungerford, and intrinsically motivated individuals—that is, defectors—, while progres- Tomera: A new meta-analysis of psycho-social determinants of pro- environmental behaviour. Journal of Environmental Psychology, 27(1), sively extending the effectiveness of the other policy tools to a 14–25. broader portion of the target population. Consequently, the ideal Beretti, A., Figuières, C., & Grolleau, G. (2019). How to turn crowding- approach should combine long-term educational efforts with short/ out into crowding-in? An innovative instrument and some law- medium-term policies aimed at facilitating PEB materially and socially. related examples. European Journal of Law and Economics volume. 48, 417–438. https://doi.org/10.1007/s10657-019-09630-9 The main limitation in this study is linked to data availability. The Bicchieri, C., & Mercier, H. (2014). Norms and beliefs: How change occurs. lack of municipal level data regarding the different green infrastruc- In M. Xenitidou & B. Edmonds (Eds.), The complexity of social norms tures in Europe obliged us to rely on proxies. The development of a (pp. 37–54). Springer International Publishing. coherent database of green infrastructures in Europe in the future BiPRO/CRI. (2015). Assessment of separate collection schemes in the 28 capitals of the EU—National factsheet Hungary. https://www.municipalwasteeurope. would offer a chance to refine the analysis. eu/sites/default/files/HU%20National%20factsheet.pdf To date, no comprehensive research has been undertaken on Black, J. S. (1978). Attitudinal, normative, and economic factors in early responses the impact of economic problems on PEB. Future research avenues to an energy-use field experiment (Vol. 1). University of Wisconsin. may include an overview of how PEBs' observance rates have Black, S., Stern, P., & Elworth, J. (1985). Personal and contextual influences on evolved since the beginning of the Great Recession and a study on household energy adaptations. Journal of Applied Psychology, 70(1), 3–21. Cecere, G., Mancinelli, S., & Mazzanti, M. (2014). Waste prevention and the conditions and extent to which individuals trade environmental social preferences: The role of intrinsic and extrinsic motivations. Eco- and economic priorities. logical Economics, 107, 163–176. Cialdini, R. B., Reno, R. R., & Kallgren, C. A. (1990). A focus theory of norma- tive conduct: Recycling the concept of norms to reduce littering in pub- ACKNOWLEDGMENTS lic places. Journal of Personality and Social Psychology, 58(6), 1015–1026. We thank Lidia Andrés, Michael Brody, Daniel J. Fiorino, Joe Container Recycling Institute. (2016). BottleBill.org—Worldwide. Retrieved Greenblott, Tommaso Luzzati, Josep Lluis Raymond, and Claudia May 5, 2016, from http://www.bottlebill.org/legislation/world.htm Schneider for helpful comments and inspiring conversations. We also Cordano, M., Welcomer, S., Scherer, R. F., Pradenas, L., & Parada, V. thank two anonymous reviewers for their valuable comments and (2011). A cross-cultural assessment of three theories of pro- environmental behavior: A comparison between business students of suggestions. This research was supported by the Spanish Ministry of Chile and the United States. Environment and Behavior, 43, 634–657. Science, Innovation and Universities, and ERDF under grants De Young, R. (1985). Encouraging environmentally appropriate behavior: The RTI2018-095484-B-I00 and BES-2016-078156. role of intrinsic motivation. Journal of Environmental Systems, 15(4), 281–292. Derksen, L., & Gartrell, J. (1993). The social context of recycling. American Sociological Review, 58(3), 434–442. ORCID Dietz, T., Stern, P. C., & Guagnano, G. A. (1998). Social structural and social Mariateresa Silvi https://orcid.org/0000-0003-2316-0210 psychological bases of environmental concern. Environment and Behav- Emilio Padilla https://orcid.org/0000-0002-9510-8262 ior, 30(4), 450–471. 14 SILVI AND PADILLA European Commission. (2015). Eurobarometer 81.3 (2014). TNS Opinion, OECD. (2009). Managing water for all: An OECD perspective on pricing and Brussels [Producer]. GESIS Data Archive. financing. OECD. European Environment Agency. (2017). Waste recycling. Retrieved Onwezen, M. C., Antonides, G., & Bartels, J. (2013). The norm activation February 23, 2018, from https://www.eea.europa.eu/data-and-maps/ model: An exploration of the functions of anticipated pride and guilt in indicators/waste-recycling-1/assessment pro-environmental behaviour. Journal of Economic Psychology, 39, European Parliament. (2011). A European refunding scheme for drinks con- 141–153. https://doi.org/10.1016/j.joep.2013.07.005 tainers. Directorate general for external policies. http://www.europarl. Oskamp, S., Harrington, M. J., Edwards, T. C., Sherwood, D. L., europa.eu/RegData/etudes/note/join/2011/457065/IPOL-AFET_NT Okuda, S. M., & Swanson, D. C. (1991). Factors influencing household (2011)457065_EN.pdf recycling behavior. Environment and Behavior, 23(4), 494–519. Fehr, E., & Gintis, H. (2007). Human motivation and social cooperation: Palmer, K. L., & Walls, M. A. (1997). Optimal policies for solid waste dis- Experimental and analytical foundations. Annual Review of Sociology, posal: Taxes, subsidies, and standards. Journal of Public Economics, 33(1), 43–64. 65(2), 193–205. Ferrara, I., & Missios, P. (2005). Recycling and waste diversion effectiveness: Evi- Schneider, J., Karigl, B., Reisinger, H., Oliva, J., Sübenbacher, E. & Read, B. dence from Canada. Environmental & Resource Economics, 30(2), 221–238. (2011). A European refunding scheme for drinks containers. https:// Ferrara, I., & Missios, P. (2012). A cross-country study of household waste www.europarl.europa.eu/RegData/etudes/note/join/2011/457065/ prevention and recycling: Assessing the effectiveness of policy instru- IPOL-AFET_NT(2011)457065_EN.pdf ments. Land Economics, 88(4), 710–744. Schwartz, S. (1968). Awareness of consequences and the influence of Filippini, M. (2011). Short and long-run time-of-use price elasticities in Swiss moral norms on interpersonal behavior. Sociometry, 31(4), 355–369. residential electricity demand (Working Paper No. 76). Schwartz, S. (1977). Normative influences on altruism. In L. Berkowitz Gneezy, U., Meier, S., & Rey-Biel, P. (2011). When and why incentives (don't) (Ed.), Advances in experimental social psychology (Vol. 10, pp. 221–279). work to modify behavior. Journal of Economic Perspectives, 25(4), 191–210. Academic Press. Grodzin ska-Jurczak, M. (2003). The relation between education, knowl- Sidique, S. F., Joshi, S. V., & Lupi, F. (2010). Factors influencing the rate of edge and action for better waste management in Poland. Waste Man- recycling: An analysis of Minnesota counties. Resources, Conservation agement & Research, 21(1), 2–18. and Recycling, 54(4), 242–249. Guagnano, G. A., Stern, P. C. P., & Dietz, T. (1995). Influences on attitude- Stern, P. C., Dietz, T., & Black, J. S. (1985). Support for environmental pro- behavior relationships: A natural experiment with curbside recycling. tection: The role of moral norms. Population and Environment, 8(3-4), Environment and Behavior, 27(5), 699–718. 204–222. https://doi.org/10.1007/bf01263074 Heberlein, T. A. (1981). Environmental attitudes. Zeitschrift fur Umweltpolitik, Thøgersen, J. (1994). Monetary incentives and environmental concern. Journal of Environmental Policy, 81(2), 241–270. Effects of a differentiated garbage fee. J Consum Policy, 17, 407–442. Heller, M. H., & Vatn, A. (2017). The divisive and disruptive effect of a https://doi.org/10.1007/BF01022912 weight-based waste fee. Ecological Economics, 131, 275–285. Thøgersen, J. (1996). Recycling and morality: A critical review of the litera- Hofstede, G. J. (2015). Six dimension of culture by country—Data matrix. ture. Environment and Behavior, 28, 536–558. https://doi.org/10. Retrieved February 27, 2018, from http://geerthofstede.com/ 1177/0013916596284006 research-and-vsm/dimension-data-matrix/ Thøgersen, J. (2003). Monetary incentives and recycling: Behavioural and Hofstede, G. J., & Minkov, M. (2010). Cultures and Organizations: Software psychological reactions to a performance-dependent garbage fee. of the mind (Rev. 3rd ed.). McGraw-Hill. Journal of Consumer Policy, 26, 197–228. https://doi.org/10.1023/A: Hopper, J., & Nielsen, J. (1991). Recycling as altruistic behavior: Normative 1023633320485 and behavioral strategies to expand participation in a community Thøgersen, J., & Ölander, F. (2002). Human values and the emergence of a recycling program. Environment and Behavior, 23, 195–220. sustainable consumption pattern: A panel study. Journal of Economic Humphrey, C. R., Bord, R. J., Hammond, M. M., & Mann, S. H. (1977). Atti- Psychology, 23, 605–630. tudes and conditions for cooperation in a paper recycling program. Turaga, R. M. R., Howarth, R. B., & Borsuk, M. E. (2010). Pro- Environment and Behavior, 9(1), 107–124. environmental behavior: Rational choice meets moral motivation. Hunter, L. M., Hatch, A., & Johnson, A. (2004). Cross-national gender variation Annals of the New York Academy of Sciences, 1185, 211–224. in environmental behaviors. Social Science Quarterly, 85(3), 677–694. van den Bergh, J. C. J. M. (2008). Environmental regulation of households: Jackson, T. (2005). Motivating sustainable consumption. Sustainable Devel- An empirical review of economic and psychological factors. Ecological opment Research Network. Economics, 66(4), 559–574. Jacobs, H. E., & Bailey, J. S. (1983). Evaluating participation in a residential Vaske, J. J., Donnelly, M. P., Williams, D. R., & Jonker, S. (2001). Demo- recycling program. Journal of Environmental Systems, 12(2), 141–152. graphic influences on environmental value orientations and normative Katzev, R., & Pardini, A. (1987). The comparative effectiveness of reward beliefs about national forest management. Society and Natural and commitment approaches in motivating community recycling. Jour- Resources, 14, 761–776. nal of Environmental Systems, 17, 93–113. Kirakozian, A. (2016). One without the other? Behavioural and incentive policies for household waste management. Journal of Economic Sur- SUPPORTING INF ORMATION veys, 30(3), 526–551. Additional supporting information may be found online in the Kollmuss, A., & Agyeman, J. (2002). Mind the gap: Why do people act envi- Supporting Information section at the end of this article. ronmentally and what are the barriers to pro-environmental behavior? Environmental Education Research, 8(3), 239–260. Langbein, L. (2015). Public program evaluation: A statistical guide (first). Routledge. Linderhof, V., Kooreman, P., Allers, M., & Wiersma, D. (2001). Weight- How to cite this article: Silvi, M., & Padilla, E. (2021). based pricing in the collection of household waste. Resource and Pro-environmental behavior: Social norms, intrinsic motivation Energy Economics, 24(4), 359–371. and external conditions. Environmental Policy and Governance, Liobikienė, G., Mandravickaitė, J., & Bernatonienė, J. (2016). Theory of 1–14. https://doi.org/10.1002/eet.1960 planned behavior approach to understand the green purchasing behav- ior in the EU: A cross-cultural study. Ecological Economics, 125, 38–46.