Do individuals perceptions of international climate policy affect their climate-friendly activities?

Transcription

1 Do individuals perceptions of international climate policy affect their climate-friendly activities? Joachim Schleich, Fraunhofer ISI Claudia Schwirplies, University of Kassel Andreas Ziegler, University of Kassel Preliminary version: Please do not cite or circulate without the author s permission. April 2014 Joachim Schleich Fraunhofer Institute for Systems and Innovation Research Breslauer Str. 48, Karlsruhe, Germany joachim.schleich@isi.fraunhofer.de Phone: , Fax: Grenoble Ecole de Management 12 rue Pierre Sémard, BP 127, Grenoble Cedex 01, France Claudia Schwirplies (corresponding author) University of Kassel, Department of Economics Nora-Platiel-Str. 5, Kassel, Germany claudia.schwirplies@uni-kassel.de Phone: , Fax: Andreas Ziegler University of Kassel, Department of Economics Nora-Platiel-Str. 5, Kassel, Germany andreas.ziegler@uni-kassel.de Phone: , Fax:

2 Do individuals perceptions of international climate policy affect their climate-friendly activities? Abstract A substantial body of economic literature provides potential explanations for individuals voluntary contributions to public goods, focusing on the roles of social norms and feelings of warm glow or responsibility. This paper provides additional theoretical insights and empirical evidence on the effect of people s perceptions of international climate policy on the provision of the public good climate protection. We distinguish between perceptions on aspects of perceived justification, procedural justice, and trust in international climate negotiations and agreements. Our empirical results on individuals stated future and past adoption of a wide range of climate-friendly activities in China, Germany, and the US suggest that the perceived justification of international climate policy motivates future climate protection efforts. In a broader interpretation, the propensity to voluntarily contribute to the public good is higher the more efficient the public good providing process. We further find that citizens in Germany (but not in China or the US) tend to compensate a perceived lack of procedural trust by additional climate-friendly activities. In comparison, citizens in the US (but not in China or Germany) tend to compensate a perceived lack of procedural justice. Our findings are robust to various model specifications. Keywords: climate policy; climate change; individual climate protection efforts; climate-friendly activities; justification; procedural justice; procedural trust; JEL: H41, Q54, Q58 1

3 1. Introduction Since the Copenhagen Accord in 2009, world leaders acknowledge the 2 C target to limit the dangerous impacts of anthropogenic climate change. While the topic of global warming has received considerable public and political attention, international climate policy has only made limited progress so far. Pledges to the Copenhagen Accord (COP 15) and decisions reached during COP 16 in Cancun are unlikely to be consistent with a path towards reaching the 2 C target (e.g. Höhne et al., 2012). Similarly, despite more than two decades of climate negotiations, international climate policy has often been criticized for failing to achieve an effective global agreement. The second commitment period of the Kyoto Protocol involves binding emission targets for a few countries only (notably the EU and Australia). Since these countries account for no more than about 14% of global carbon emissions, the immediate impact on climate change will be negligible. In addition, progress on a new global climate agreement, which is scheduled to enter into force after 2020, has been slow. Countries disagree, in particular, on the distribution of mitigation cost. Industrialized countries (in particular the US) fear that emission targets effect the competitiveness of their economy negatively (e.g. Pauwelyn, 2007), while emerging and developing countries (e.g. China) fear that emission targets inhibit future economic growth ( cap on development ) (e.g. Banerjee, 2012). The process of international climate negotiations has further been criticized for its inefficiency due to the large number of involved parties, as well as for developing countries underrepresentation in core UN- FCCC committees, and for developed countries dominance in guiding the decision-making processes (e.g. Okereke, 2010). Some critics even term the process of international climate negotiations as nontransparent and undemocratic (e.g. IISD, 2009). In this regard, justification, legitimacy, and effectiveness of international climate policy are called into question. At the same time, economic theory suggests that international climate policy may be beneficial to all parties. Since global emissions of all sources determine the concentration of greenhouse gases in the atmosphere, combating climate change is inherently a public goods problem (e.g. IPCC, 2001). This characteristic provides the essential justification of international climate policy, because collective climate protection efforts are more cost-efficient compared to activities of individual countries aimed at reducing greenhouse gas emissions. These issues are also strongly related to procedural justice and trust. In the context of international climate policy, the process faces a climate of mistrust about other countries intentions and motives with regard to their engagement in international climate negotiations. In addition, it may be questioned whether commitments made at international climate conferences will be kept since every nation has an incentive to free-ride and profit from the mitigation efforts of other countries. For example, the US never ratified the Kyoto Protocol and, in 2011, Canada formally withdrew from it. Similarly, at the 2012 climate summit in Warsaw, Japan reneged on its pledged reduction commitment under the Copenhagen Accord and Cancun Agreement. Likewise, China long time refused to have its greenhouse gas emissions monitored by others, and current dispute evolves about procedures to verify the pledges to the decisions reached in Copenhagen and Cancun by third parties. These examples emphasize that trust is essential in the context of climate policy as no authority with enforcement power exists. Most of the literature concerning international climate policy has focused on distributive justice (e.g. Carlsson et al., 2013; Dannenberg et al., 2010; Lange et al., 2007), while procedural justice has attained remarkably less attention (e.g. Okereke, 2010). Procedural justice traces back to the work of Lind and Tyler (1988), who showed that individuals are often as concerned about the justice of the process as with the outcome itself. If the process is not considered to be fair and transparent, the individual will not grant its legitimacy and accept the outcomes. In the context of climate policy, procedural justice incorporates how allocation decisions are made. This requires adequate representation of all countries 2

4 in the climate negotiations, in particular of those countries that are most vulnerable to climate change (e.g. Okereke, 2010; Klinsky and Dowlatabadi, 2009) as well as the inclusion of the sub-national level, i.e. the interests of those being represented who are actually affected by climate change. Procedural trust is defined as the confidence in a structure or process (e.g. Furlong, 2005). Thus, the belief in the process of international climate policy may lead to binding agreements even if parties distrust each other. Pittel and Rübbelke (2013), for example, point out that increased trust in the process of international climate policy and the perceived intentions of others may raise the probability of cooperative behavior concerning climate protection, while Levi (1997) argues that citizens are, in general, more likely to support policies if they believe that government actors and other citizens are behaving fairly toward them. Several empirical studies show that people s attitudes towards policy instruments are influenced by their perception of fairness and trust in politicians or governments (e.g. Jagers et al., 2010; Jagers and Hammar, 2009; Torgler and García-Valiñas, 2007). Additionally, the experimental literature provides evidence that people are more willing to cooperate only if others are willing to do the same (conditional cooperation, e.g. Gächter, 2007). Since decisions at the international level get executed at the domestic level, citizens contribution to climate protection may be affected by their perceptions of justification as well as procedural justice and trust. According to Oberholzer-Gee et al. (1997), a higher acceptance of international climate policy may mean that citizens are more willing to take financial burdens associated with national implementations of climate policies. Otherwise, citizens contributions to climate protection may also be a substitute for lacking action at the international level and, thus, the effect of a higher acceptance of international is generally not clear. Since pioneer works of Samuelson (1954) or Olson (1965), for instance, there have been numerous attempts to theoretically and empirically explain voluntary provisions of public goods. According to the standard theory, individuals have no incentive to contribute to a public good like climate protection (e.g. Holländer, 1990). Contrary to these expectations, numerous free-rider experiments and related studies show that individuals do not only act egoistically and make positive contributions to the public good (e.g. Blanco et al., 2012; Kerr et al., 2012; Andreoni, 1988). One strand of literature focuses on the role of social objectives like prestige, respect, reputation, or the contempt of others to explain voluntary contributions to public goods (e.g. Banerjee and Shogren, 2012; Becker, 1974; Olson, 1965), while another strand of literature carves out the impacts of environmental concern as well as a feeling of responsibility and warm glow (e.g. Ballet et al., 2007; Kotchen, 2005; Andreoni, 1990). In this respect, public policy may have the potential to strengthen or weaken norms of the private contributions to a public good (e.g. Interis and Haab, 2011). To the best of our knowledge, citizens perspectives on justification, procedural justice, and trust in the context of international climate policies and their impact on voluntary efforts in climate protection have not been analyzed yet. Similarly, these aspects have not been explored empirically in a cross-country comparison. This paper attempts to take a first step towards closing this gap by investigating the effect of people s perceptions of international climate policy on the voluntary contribution to climate protection in three countries, which in terms of emissions and political clout are considered to play a key role for the success of future climate policy: China, Germany, and the US. We distinguish between perceptions on aspects of perceived justification, procedural justice, and trust in international climate negotiations and agreements. Firstly, we extend the conventional model of allocating resources among public good and private consumption by individuals. While our paper is related to the literature of private provision of public goods (e.g. Andreoni, 1990; Bergstrom et al., 1986), we explicitly consider the effect of perceived justification, procedural trust and justice on individuals decisions to engage in climate-friendly activities. 3

5 Based on this theoretical modeling, we econometrically examine the interrelation between individuals perception of international climate policy (i.e. justification of climate policy in general, procedural justice, and trust) and their stated willingness to engage in eight domestic individual climate-friendly activities. Our explorative studies are based on a survey, which was conducted simultaneously in these countries in May and June Our findings provide new insights into the individual decision-making process concerning voluntary efforts for climate protection and the role of the perception of international climate policy. The remainder of this paper is organized as follows. Section 2 sets up the theoretical model and derives implications on the interrelation between the perception of international climate policy and the domestic engagement in climate-friendly activities. Section 3 presents the methodology including a description of the survey and variables. Section 4 provides the results of our econometric analysis and the concluding Section 5 finally discusses the main findings and points to future research needs. 2. Model In order to investigate the effects of justification, procedural justice, and trust on the individual willingness to voluntarily engage in climate protection, we adapt the standard model provided by (Bergstrom et al., 1986) for extending the conventional problem of allocating resources among public good and private consumption. Consumer i maximizes her utility from private consumption and her contribution to climate protection by choosing between three market goods: one conventional good, which generates only, one green good, which generates and a contribution to jointly, and a direct donation to (e.g. Chan and Kotchen, 2012). In our empirical analysis, carbon offsetting can be interpreted as a direct donation and the green good is represented by seven further climate-friendly activities, such as using energy from renewable sources, reducing the number of flights, and buying energy-efficient appliances. The maximization problem of each consumer i is then: The effective provision of the public good climate protection is defined as, where is i s contribution and is the contribution of all other individuals. Here, is an increasing and quasi-concave utility function and all prices are normalized to one. Note that i s contribution to the public good depends on the perceived justification of international climate policy (with ) as well as i s perceived procedural justice of and trust in the process of international climate negotiations (with ). The parameter reflects i s uncertainty about the public good itself and, thus, i s belief in the justification of international climate policy. This particularly concerns the presumption that climate protection can be characterized as a public good and that future international agreements are important for combating climate change. Hence, means that international climate policy is supposed to be more efficient compared to individual climate protection activities of each country. The parameter reflects i s uncertainty concerning the amount of and, thus, her confidence in other countries intentions, since the outcomes of climate negotiations and the intentions of other countries (e.g. whether promises are kept or the process is abused by others) are unknown. High trust in the process of international climate policy and the intentions of others, however, may motivate cooperative behavior with regard to climate protection (e.g. Pittel and Rübbelke, 2013). Hence, the first- and second-order conditions for i s optimal contribution to the public good climate protection are 4

6 Note that Totally differentiating the first-order condition yields In both cases, the denominator is negative because of the second order condition. The numerator, however, cannot be signed without further assumptions about the functional form of. For illustration and also for deriving testable hypotheses, we use a common specification from the experimental economics literature where consumer i has preferences according to the following utility function (e.g. Chan et al., 2002):. Substituting and the budget constraint yields. Solving the first-order condition with respect to results in. Note that if international climate policy is considered to be justified without restriction and enjoys absolute trust, i.e., i s optimal contribution to climate protection is the same as without consideration of perceived justification, procedural justice, and trust. Taking the partial derivative with respect to and indicates at an increase of the individual contribution to the public good climate protection as the perceived justification of international climate policy increases and the procedural justice and trust decrease. From these findings two hypotheses can be derived: (i) The propensity of individuals to engage in climate-friendly activities increases with perceived justification of international climate policy. That is, individuals are more willing to engage in climate protection if the process of international climate policy is perceived to be inefficient. 5

7 (ii) The propensity of individuals to engage in climate-friendly activities decreases with perceived confidence/trust in other countries contribution to climate protection. That is, individuals compensate for other countries slack. In our econometric analysis, we test these hypotheses by examining the relationship between the perception of international climate policy and the stated willingness to engage in eight domestic climate-friendly activities of citizens in China, Germany, and the US. 3. Econometric methodology The data for our analyses were collected from three representative online surveys among a total of 3445 citizens aged 18 and older in Germany, the US, and China. All surveys were carried out in May and June 2013 by the market research company GfK SE (Gesellschaft für Konsumforschung). In total, 1005 respondents in Germany, 1010 respondents in the US, and 1430 respondents in China completed the questionnaire. In Germany and the US, the sample was drawn from the GfK Online Panel of the particular country. Respondents were invited via to attend a self-administered web-based interview. In China, respondents were recruited by employees of GfK China in eleven core regions, invited to centrally located test studios and interviewed in a face-to-face situation. About one half of the respondents in China come from rural areas and the other half lives in city areas. The questions of the survey referred to general personal assessments of climate change, specific attitudes towards international climate policy and negotiations, and individual engagement in climate-friendly activities and carbon offsetting. For completing the survey, respondents in Germany took 31.8 minutes on average, in China 28.5 minutes and in the US 30.4 minutes. With respect to our dependent variables in our micro-econometric analysis, the respondents were asked, which of the following climate-friendly activities they are planning to take on in the future: buying energy-efficient appliances, saving energy at home, buying a car with lower fuel consumption, reducing the consumption of meat or dairy products, reducing car use, reducing the number of flights, using or purchasing energy from renewable sources, and compensating carbon emissions (offsetting). Based on the binary structure of the response options, we construct eight dummy variables (i.e. energy-efficient appliances, saving energy, car with lower fuel consumption, reducing meat or dairy products, reducing car use, reducing number of flights, energy from renewable sources, and carbon offsetting) that take the value one if the respondent stated to undertake the certain activity in the future. In order to measure individual s perceptions of international climate policy, we asked participants to specify their level of agreement with particular statements, their subjective assessment of the importance of international climate policy, and their level of information on a symmetric scale with five ordered response levels. Potential problems associated with this kind of scale are central tendency bias especially in China, acquiescence bias, and social desirability bias. These issues were addressed by don t know/no answer options to distinguish true neutral from unsure responses, a scale design involving balanced keying, and closed ended and neutral wording of the items. For a more detailed discussion see also Schleich et al. (2013). Our main explanatory variables for perceived justification (i.e. justification) and procedural trust (i.e. procedural trust) are constructed by calculating the median value of the statements presented in Table 1. Please note that negatively keyed items indicating procedural trust were reversed so that higher values of the parameter estimates of this variable indicate a higher level of trust. In the same way, we compose an indicator for the self-reported informational level of the respondent (i.e. level of information) consisting of two items also reported in Table 1. Results of standard factor analysis (using 6

8 varimax rotation) confirm the construction of these indices. Furthermore, we included two dummy variables to control for perceived procedural justice and success of international climate policy. The variable concerning perceived procedural justice takes the value one if the respondent agreed rather strongly or very strongly to the statement that all countries have the same opportunities to represent their interests at international climate conferences. The indicator for the perceived success takes the value one if the respondent considered climate negotiations to be rather successful or very successful. In addition, we included two dummy variables, high contribution and financial advantage, which indicate that the respondent believes that the climate-friendly activity contributes rather a lot or a lot to climate protection and provides rather financial advantages for her personally. Since climate-friendly behavior and consumption is supposed to be mainly driven by environmental attitudes, we construct four dummy variables which indicate a raised awareness for environmental and climate change issues. Firstly, we used six items from the New Ecological Paradigm (NEP) Scale for measuring environmental concern (e.g. Dunlap et al., 2000). The NEP Scale is a standard instrument in the social and behavioral sciences and is also more and more common in the economic literature (e.g. Kotchen and Moore, 2007). The corresponding variable environmental concern takes values from zero (lowest level of environmental awareness) to five (highest level of environmental awareness) and is constructed by calculating the median value of the statements after reversing negatively keyed items. Our second dummy variable feeling responsible for climate protection takes the value one if the respondent agrees rather strongly or very strongly to the statement I feel responsible for making a contribution to climate protection on a symmetric five-stage ordinal scale from very weakly to very strongly. Another dummy variable member of environmental organization takes the value one if the respondent is a member of a group or organization that engages in the preservation and protection of the environment and nature. Finally, we consider the dummy variable identifying with green politics, which takes the value one if the respondent agrees rather strongly or very strongly to the statement I identify myself closest with green politics on a symmetric five-stage ordinal scale from very weakly to very strongly. Please note that this statement had not been included in the Chinese questionnaire. In addition, we control for several socio-demographic characteristics of the respondents, namely the variable age of the respondent (in years), the gender dummy variable female that takes the value one if the respondent is a woman, the variable number of own children, the variable household size, which indicates the number of people living in the household of the respondent, the dummy variable living with a partner for the marital status of the respondent, and the dummy variable highly educated if the respondent received the general qualification for university entrance. Additionally, the dummy variable high household income takes the value one if the monthly net income of the household (the underlying question was based on several income intervals) amounts to at least Yuan (i.e. about 1647 US-Dollar at the exchange rate applicable in December 2013), 3000 Euros (i.e. about 4120 US-Dollar at the exchange rate applicable in December 2013), and 4000 US-Dollars, respectively, and, thus, exceeds the corresponding median income interval of the country sample. For the dependent variables car with lower fuel consumption, reducing car use, and reducing number of flights we control for the level of activity, and include the number of miles or number of kilometers driven on average each year and the number of flights in the last three years, respectively. Finally, to control for regional heterogeneity, the set of explanatory variables includes regional dummies which differ by country. Tables 2 to 4 report several descriptive statistics (i.e. mean, standard deviation, minimum, maximum) for the three countries. Considering the high number of missing income data, especially in Germany and the US, we use single imputation methods for the variable high household income to reduce the number of missing observations and check the robustness of our results. For this purpose, missing income data are substituted by the median income interval before constructing the 7

9 dummy variable for household income above the sample median income interval. The missing values were captured by an additional dummy variable, which takes the value one if income is not reported, and zero otherwise. Since this method may result in biased estimators for high household income, we compare the estimated parameters with those of the model without imputation and assert that relationships at the one and five percent significance level as well as the majority of the effects significant at the ten percent level are robust in both models. Therefore, in section 4 we only report the estimation results for the model without imputation 1. To corroborate our findings and to limit the effect of a potential social desirability bias, we also used the self-reported past engagement in the eight climate-friendly activities as the dependent variable, rather than planned activities. Due to the binary structure of the dependent variables, we apply standard probit models to estimate the determinants of the stated willingness to take the eight climate-friendly activities. The corresponding parameters are estimated via maximum likelihood method, correcting for heteroskedasticity (White, 1982). In addition, we discuss the estimates of the discrete probability effects. To allow for the possibility that decisions to engage in different climate-friendly activities are not made independently, we also apply multivariate probit models, accounting for possible correlations between the dependent variables in the corresponding error terms. If these correlations are not captured, the estimation of univariate probit models might lead to biased and inconsistent parameter estimates (e.g. Greene, 2012). Since the questions on reducing the number of flights and reducing car use were not posed to respondents who did not fly or drive, we excluded the binary model for reducing the number of flights, car with lower fuel consumption, and reducing car use from the estimation of the multivariate probit models and estimated a bivariate probit model for the two dependent variable concerning car use. While univariate binary probit models can easily be estimated by the maximum likelihood method, the simulated counterpart of this method, which incorporates the Geweke-Hajivassiliou-Keane (GHK) simulator (Geweke et al., 1994; Keane, 1994; Börsch-Supan and Hajivassiliou, 1993), has to be applied for the estimation in the multivariate probit models. This simulation method is implemented in STATA. For our calculations and estimations we employ weights, which were constructed by the survey institute to ensure offline representativeness of our results for the whole population of the country. 4. Results of the econometric analyses Tables 5 to 7 report the ML estimates of the parameters (including robust z-statistics) in the binary probit models for the determinants of the stated willingness to engage in the eight climate-friendly activities in the future. Please note that the determinants of carbon offsetting are not reported for China since citizens in China did not have the possibility to compensate for their carbon emissions in a way comparable to Germany and the US at the time they were interviewed. With respect to our first hypothesis, justification is statistically significant at least at the 5% level and also has the expected positive effect on four activities in Germany and five activities in the US, respectively. In Germany, we also find one statistically significant effect of perceived justification on the willingness to reduce the number of flights with a negative sign. In China, we only find one statistically significantly positive impact of justification on the stated willingness to reduce car use. Overall, the perceived justification of international climate policy seems to rather motivate citizens in Germany and the US to engage in climate-friendly activities, while it seems to play a minor role in China. With regard to our second hypothesis, the role of procedural trust for the willingness to engage in climate-friendly activities is less clear in the three countries. We find three negative effects in Germany, two positive effects in the US and 1 Findings for these alternative specifications are available upon request. 8

10 three positive effects in China, all statistically significant at least at the 5% level. Interestingly, the estimated parameters in Germany, on the one hand, and in the US and China, on the other hand, exhibit opposite signs. A higher level of perceived procedural trust decreases the willingness to engage in some climate-friendly activities in Germany, but increases the willingness to engage in some climate-friendly activities in China and the US. For procedural justice, we find two effects with opposite signs on the planned activities in the future in Germany, only one positive effect in China, and four positive effects in the US, all statistically significant at least at the 5% level. Although higher perceived justification of international climate policy stimulates the willingness of the citizens in all three countries to behave in a climate-friendly way, lower perceived procedural trust rather seems to be compensated by a higher domestic engagement in climate protection only in Germany and higher perceived procedural justice motivates US citizens even more to carry out climatefriendly activities. In the US, there is also a statistically significant positive effect of a higher informational level on three climate-friendly activities planned for the future. In Germany and China, our estimation results do not allow drawing a clear conclusion about the role of respondents informational level. Respondents perception of the success of international climate policy plays practically no noteworthy role, since in all three countries there is at most one significant effect. The cross-country comparison reveals that in Germany and the US reducing meat or dairy products are not significantly affected by perceived justification as well as procedural trust and procedural justice. Possible explanations for this finding are that, in both these countries, the majority of the population does not associate their consumption of meat and dairy products with climate change. Also, German and US citizens appear to account for justification, procedural trust and procedural justice of climate policy when purchasing rather than using their car. The perceived effectiveness and financial advantages associated with the climate-friendly activities are of paramount importance in Germany. The parameter estimates for high contribution suggest significantly (p<0.1) positive impacts on five and those of financial advantage on four of the climate-friendly activities. Similarly, financial advantage positively affects the stated willingness of US respondents to engage in six climate-friendly activities, statistically significant (p<0.05). While two parameter estimates for high contribution are positive and statistically significant, this variable exhibits an unexpected negative effect on the planned purchase of energy efficient appliance in the US. In China, we only find two estimated parameters that are significantly different from zero at common significant levels. In Germany, feeling responsible for climate protection and being a member of an environmental organization are also highly relevant drivers of climate-friendly activities. Interestingly, parameter estimates for feeling responsible are statistically significantly for all activities but for energy-efficient appliances and saving energy. This result is in line with former experimental studies (e.g. Menges et al., 2005) and might be rationalized by EEG apportionment, i.e. mandatory charges on household electricity bills, which are used by the German Government to invest in renewable energies and, thus, may compensate the feeling of responsibility. Additionally, environmental concern measured by the NEP Scale is slightly more important in China and the US than in Germany, since we find three significantly positive effects in China and the US but only one in Germany. With respect to the sociodemographic characteristics, only some occasional significant effects can be detected except for four significantly positive effects of being female in Germany, three contradictory significant effects of being female and three significantly positive effects of household size in the US as well as three contradictory significant effects of a higher educational level and four contradictory significant effects of living together with a partner in China. Since the indicators for environmental attitudes as well as the socio-economic variables might be highly correlated, we additionally conducted several Wald tests to carve out joint impacts 2. The results of these tests show that in the majority 2 Findings for these tests are available upon request. 9

11 of activities the groups of variables concerning environmental attitudes have a common statistically significant effect on the probability to conduct these activities in China and Germany, but not in the US. For the socio-economic characteristics we also detect joint influences for several measures in all three countries. Interestingly, the indicators of citizens perceptions of international climate policy have a common statistically significant effect on the stated willingness to conduct the majority of activities in Germany and the US, while in China this common impact can only be detected on two climate-friendly measures. In order to evaluate the magnitude of the discussed effects, Tables 8 to 10 report the estimates of average probabilities at minimum and maximum values of explanatory variables with a statistically significant effect. In Germany, the estimated average probability of buying energy-efficient appliances, for example, increases by more than 40 percentage points from 49.72% at the minimum level of perceived justification to 90.16% at the maximum level and decreases by nearly 40 percentage points from 93.84% at the minimum level of procedural trust to 54.52% at the maximum level. Similarly, the estimated average probability of carbon offsetting increases by more than 40 percentage points from 24.50% at the minimum value of perceived justification of international climate policy to 67.42% at the maximum level. In the US, perceived justification increases the estimated average probability of carbon offsetting by more than 30 percentage points from 42.02% at the minimum value to 74.55% at the maximum level, procedural trust raises the probability of buying a car with lower fuel consumption by more than 30 percentage points from 60.47% at the minimum level to 92.54% at the maximum level and a higher informational level increases the propensity of reducing the number of flights by more than 40 percentage points from 32.85% at the minimum level of information to 75.88% at the maximum level of information. In China, the probability of reducing car use increases by more than 30 percentage points from 51.21% at the minimum level of perceived justification to 84.74% at the maximum level and the propensity to use energy from renewable sources increases by more than 30 percentage points from 31.56% at the minimum level of procedural trust to 62.73% at the maximum level. These results imply that the perception of international climate policy is of considerable importance for citizens willingness to engage in domestic climate-protecting measures. As mentioned in section 3, we conducted several robustness checks to assess the reliability of our results. With regard to our main indicators for citizens perceptions of international climate policy, the simulation results of the multivariate and bivariate binary probit models hardly differ from the estimation results of the univariate binary probit models, although the majority of the correlations between the dependent variables in the corresponding error terms are significantly different from zero. Furthermore, we examine the determinants of the past implementation of the eight climate-friendly activities and report the estimation results in Tables 11 to 13. Unlike for future activities, statistically significant parameters may not reflect a causal relation between the associated variable and past activities. Instead, we refer to correlation rather than causality. In China, we find one more statistically significant positive correlation of perceived justification and the positive signs of two statistically significant parameter estimates of procedural trust get negative, statistically significant at the 1% level. We also detect two statistically significant positive correlations of procedural justice, while perceived success of international climate policy is negatively interrelated to two of the climate-friendly activities. For activities carried out in the past, the estimation results also place a greater emphasis on a perceived high contribution to climate protection with five significantly positive correlations. Regarding the determinants of the past implementation of the eight climate-friendly activities in Germany, the roles of perceived justification and procedural trust get slightly less important, since we find only three and two statistically significant correlations, respectively, but the signs of these interrelations are the same as 10

12 for stated future activities. Most notable is the negligible role of feeling responsible for climate protection in contrast to the estimation results of the stated willingness to engage in future climate-friendly activities. Finally in the US, there is no interrelation of procedural trust and the past engagement in climate-friendly measures and, similarly, perceived justification is of less importance as we detect only two statistically significant positive correlations. In contrast, a higher informational level about international climate policy and identifying with green politics play a more important role with four and five significantly positive interrelations, respectively. Surprisingly, environmental concern and the implementation of four climate-friendly activities in the past are significantly negatively correlated in the US. Overall, our estimation results appear to be fairly robust with regard to our main indicators for individual perceptions of international climate policy. 5. Conclusions A substantial body of economic literature has carved out factors explaining individual voluntary contributions to public goods, focusing on the roles of social norms and feelings of warm glow and responsibility. To augment this literature, this paper provides theoretical insights and empirical evidence that citizens perceptions of international climate policy may also affect individuals contribution to the public good climate protection. Our empirical results on individuals stated future and also past adoption of a wide range of climate-friendly activities in China, Germany and the US suggest that especially the perceived justification of international climate policy motivates future climate protection efforts. In a broader interpretation, individuals propensity to contribute to a public good is the higher the more efficient the publicgood providing process. In addition, a perceived lack of procedural trust is compensated by a higher stated willingness to engage in climate-friendly activities in Germany, while in the US citizens appear to offset a perceived lack of procedural justice by a higher stated willingness to engage in climate-protecting measures. In other words, citizens across countries differ in their stated willingness to compensate for others lack of performance in providing a public good. Our findings imply that the process of providing public goods should receive as much attention as the outcome of this process when studying individuals voluntary contributions to public goods. Future research could explore this relation for other applications than climate policy. Future research could also analyze the link between individuals perception of climate policy and climate-friendly activities more deeply with a richer set of items. Acknowledgements This research has been carried out within the project titled The Relevance of Voluntary Efforts and Fairness Preferences for the Success of International Climate Policy: A Theoretical and Empirical Analysis at the Individual Level (VolFair). VolFair is supported by the German Federal Ministry of Education and Research (BMBF) under the funding priority Economics of climate change. 11

13 References Andreoni, J. (1988): Why free ride?: Strategies and learning in public goods experiments, Journal of Public Economics 37 (3), Andreoni, J. (1990): Impure altruism and donations to public goods: A theory of warm-glow giving, The Economic Journal 100 (401), Ballet, J.; Bazin, D.; Lioui, A.; Touahri, D. (2007): Green taxation and individual responsibility, Ecological Economics 63 (4), Banerjee, P.; Shogren, J. F. (2012): Material interests, moral reputation, and crowding out species protection on private land, Journal of Environmental Economics and Management 63 (1), Banerjee, S. B. (2012): A climate for change? Critical reflections on the Durban United Nations Climate Change Conference, Organizational Studies 33 (12), Becker, G. S. (1974): A Theory of Social Interactions, Journal of Political Economy 82 (6), Bergstrom, T.; Blume, L.; Varian, H. (1986): On the private provision of public goods, Journal of Public Economics 29 (1), Blanco, E.; Lopez, M. C.; Coleman, E. A. (2012): Voting for environmental donations: Experimental evidence from Majorca, Spain, Ecological Economics 75 (0), Börsch-Supan, A.; Hajivassiliou, V. A. (1993): Smooth unbiased multivariate probability simulators for maximum likelihood estimation of limited dependent variable models, Journal of Econometrics 58 (3), Carlsson, F.; Kataria, M.; Krupnick, A.; Lampi, E.; Löfgren, Å.; Qin, P.; Sterner, T. (2013): A fair share: Burden-sharing preferences in the United States and China, Resource and Energy Economics 35 (1), Chan, K. S.; Godby, R.; Mestelman, S.; Andrew Muller, R. (2002): Crowding-out voluntary contributions to public goods, Journal of Economic Behavior & Organization 48 (3), Chan, N. W.; Kotchen, M. J. (2012): A generalized impure public good and linear characteristics model of green consumption, Working Paper. Dannenberg, A.; Sturm, B.; Vogt, C. (2010): Do equity preferences matter for climate negotiators? An experimental investigation, Environmental and Resource Economics 47 (1), Dunlap, R. E.; Van Liere, K. D.; Mertig, A. G.; Jones, R. E. (2000): New trends in measuring environmental attitudes: Measuring endorsement of the new ecological paradigm: A revised NEP scale, Journal of Social Issues 56 (3), Furlong, G. T. (2005): The conflict resolution toolbox. Models & maps for analyzing, diagnosing, and resolving conflict, Mississauga, Ontario: John Wiley & Sons Canada, Ltd. Gächter, Simon (2007): Conditional cooperation: Behavioral regularities from the lab and the field and their policy implications. Published in B. S. Frey and A. Stutzer (eds): Economics and Psychology. A Promising New Cross- Disciplinary Field: The MIT Press. Geweke, J.; Keane M.; Runkle D. (1994): Alternative computational approaches to inference in the multinomial probit model, The Review of Economics and Statistics 76 (4),

14 Greene, W. H. (2012): Econometric analysis, Boston, London: Pearson. Höhne, N.; Taylor, C.; Elias, R.; Den Elzen, M.; Riahi, K.; Chen, C.; Rogelj, J.; Grassi, G.; Wagner, F.; Levin, K.; Massetti, E.; Xiusheng, Z. (2012): National GHG emissions reduction pledges and 2 C: comparison of studies, Climate Policy 12 (3), Holländer, H. (1990): A social exchange approach to voluntary cooperation, The American Economic Review 80 (5), IISD (2009): A brief analysis of the Copenhagen Climate Change Conference. An IISD Commentary, International Institute for Sustainable Development. Interis, M. G.; Haab, T. C. (2011): Woodsy the optimal owl: Environmental campaigns, norms, and implications for public goods policy, Ecological Economics 70 (12), IPCC (2001): Climate change: Mitigation. Contribution of working group III to the Third Assessment Report (TAR) of the Intergovernmental Panel on Climate Change, Cambridge, UK: Cambridge University Press. Jagers, S.; Hammar, H. (2009): Environmental taxation for good and for bad: on individuals reluctance to mitigate climate change via CO2-tax vis-à-vis alternative policy instruments., Environmental Politics 18 (2), Jagers, S. C.; Löfgren, Å.; Stripple, J. (2010): Attitudes to personal carbon allowances: political trust, fairness and ideology, Climate Policy 10 (4), Keane, M. (1994): A Computationally Practical Simulation Estimator for Panel Data, Econometrica 62, Kerr, J.; Vardhan, M.; Jindal, R. (2012): Prosocial behavior and incentives: Evidence from field experiments in rural Mexico and Tanzania, Ecological Economics 73 (0), Klinsky, S.; Dowlatabadi, H. (2009): Conceptualizations of justice in climate policy, Climate Policy 9 (1), Kotchen, M. J. (2005): Impure public goods and the comparative statics of environmentally friendly consumption, Journal of Environmental Economics and Management 49 (2), Kotchen, M. J.; Moore, M. R. (2007): Private provision of environmental public goods: Household participation in greenelectricity programs, Journal of Environmental Economics and Management 53 (1), Lange, A.; Vogt, C.; Ziegler, A. (2007): On the importance of equity in international climate policy: An empirical analysis, Energy Economics 29 (3), Levi, Margaret (1997): Consent, dissent, and patriotism, Cambridge, New York: Cambridge University Press (Political economy of institutions and decisions). Lind, E. Allan; Tyler, Tom R. (1988): The social psychology of procedural justice, New York: Plenum Press (Critical issues in social justice). Menges, R.; Schroeder, C.; Traub, S. (2005): Altruism, warm glow and the willingness-to-donate for green electricity: An artefactual field experiment, Environ Resource Econ 31 (4), Oberholzer-Gee, F.; Bohnet, I.; Frey, B. (1997): Fairness and competence in democratic decisions, Public Choice 91 (1),

15 Okereke, C. (2010): Climate justice and the international regime, Wiley Interdisciplinary Reviews: Climate Change 1 (3), Olson, Mancur (1965): The logic of collective action. Public goods and the theory of groups, Cambridge: Harvard University Press. Pauwelyn, J. (2007): U.S. Federal Climate Policy and Competitiveness Concerns: The Limits and Options of International Trade Law, Working Paper, Nicolas Institute for Environmental Policy Solutions. Duke University. Pittel, K.; Rübbelke, D. (2013): International climate finance and its influence on fairness and policy, The World Economy 36 (4), Samuelson, P. A. (1954): The pure theory of public expenditure, The Review of Economics and Statistics 4 (36), Schleich, J.; Duetschke, E.; Schwirplies, C.; Ziegler Andreas (2014): Citizens perceptions of justice in international climate policy - empirical insights from China, Germany and the US, Working Paper. Torgler, B.; García-Valiñas, M. A. (2007): The determinants of individuals' attitudes towards preventing environmental damage, Energy Economics 63 (2 3), White, H. (1982): Maximum likelihood estimation of misspecified models, Econometrica 50 (1),

16 Appendix Table 1: Indicators of respondents perceptions of international climate policy, indicators were constructed by calculating the median values of all corresponding items. Indicator Justification Items How important do you consider future international agreements are for combating climate change? Responses range from "very unimportant" to "very important" on a symmetric five-point scale. How important do you consider the following issues to be for international climate negotiations? - Comprehensive quantitative targets to reduce global greenhouse gas emissions - Measures to reduce global greenhouse gas emissions - Adaptation measures to the consequences of climate change (e.g. dams for flood protection) Responses range from "very unimportant" to "very important" on a symmetric five-point scale. This question concerns your personal perception of international climate negotiations and agreements. How strongly do you agree with the following statements? - All countries can benefit from international climate agreements" Responses range from very strongly to very weakly on a symmetric five-point scale. Trust This question concerns your personal perception of international climate negotiations and agreements. How strongly do you agree with the following statements? - "The richer (industrialized) countries use international climate negotiations to push through their own economic interests yis-a-vis other countries" - "The poorer (developing) countries use international climate negotiations to push through their own economic interests vis-a-vis other countries" - "Governments use international climate negotiations to pacify their citizens instead of reducing global greenhouse gas emissions" - "Commitments made at international climate negotiations will not be kept anyhow" Responses range from very strongly to very weakly" on a symmetric five-point scale. Level of information How well informed do you feel about these climate conferences? Responses range from "very badly" to "very well" on a symmetric five-point scale. How well do you think your personal position is represented at international climate negotiations? Responses range from "very badly" to "very well" on a symmetric five-point scale.