To Trust or Not to Trust in National Climate Policy A Quantitative Macro-Analysis on The Role of Social- and Political Trust in Determining National Policy Outcomes of Climate Taxation

(1)

To Trust or Not to Trust in National Climate Policy

A Quantitative Macro-Analysis on The Role of Social- and Political Trust in Determining National Policy Outcomes of Climate Taxation

Petter Bjersér

Political Science C (Bachelor Thesis) Department of Government

Uppsala University, Spring 2019

Supervisor: Maria Nordbrandt Bergström Words: 13925

Pages: 46

(2)

Abstract

This thesis examines the role of aggregated mean levels of Social Trust and Political Trust in national climate policy outcomes of Climate Taxation. The aim of the study is to contribute with the first quantitative macro-level comparative study of how Social Trust and three different forms of Political Trust¹, the independent variables, affect national climate policy outcomes of Climate Tax Revenue, the dependent variable. In pursuit of this aim the study utilise fixed effects regression models based on bi-annual Time-Series-Cross-Sectional data of 24 OECD countries over 12 years.

To operationalise the independent variables the study draws on individual-level survey-data from the European Social Survey Round 2-7 (2004-2014) and to operationalise the dependent variable the study uses tax revenue from Energy and Transports from the OECD-database Policy Instruments for the Environment.

The theoretical framework draws on behavioural approaches to social dilemmas in theorising that Social- and Political Trust facilitates cooperation and lowers transaction costs in order to overcome Social Dilemmas based on separate mechanisms of Networks and Institutional Safeguards. The study finds significant evidence for two of its hypotheses, as higher levels of Social Trust is correlated positively with higher Climate Taxes, and higher levels of Trust in International Organisations correlate negatively with Climate Taxes. Moreover, the study finds significant but contradictory evidence concerning two of its hypotheses, as higher levels of Trust in Political Institutions and Trust in Implementing Authorities correlate negatively with Climate Taxes.

However, there are some indications that the assumption of no perfect multicollinearity does not hold for the fixed effects regression model and as such the results could be compromised.

To the knowledge of the author, this thesis outlines the first macro-level comparative study on the role of Social- and Political Trust in determining national climate policy outcomes of Climate Taxation that are generalisable in an OECD-country context. Future research is encouraged to scrutinise these preliminary findings in comparative analysis of both OECD and non-OECD countries.

Keywords

Trust, Social Trust, Political Trust, Social Dilemmas, Environmental Taxes, Climate Taxes, Environmental Policy, Climate Policy, Policy Outcomes

1 Trust in Political Institutions, Trust in Implementing Authorities and Trust in International Organisations

(3)

Acknowledgements

Maria Nordbrandt Bergström, my supervisor. Your guidance and inspiration have been truly invaluable.

Oscar Theblin, for giving uncompromising notes in times of dire need.

Filip Wästberg, for introducing me to the magical wonders of data analysis in R.

(4)

1. Introduction

1.1 Background

Solving the climate crisis is arguably one of the most paramount social dilemmas² that has ever faced mankind (Hardin, 1968; Olson, 1965; Ostrom, 1990). Both social- and political trust have been argued to be essential in overcoming social dilemmas throughout the social sciences (Easton, 1965; Fukuyama, 1995; Ostrom, 1998; Lubell, 2002; Hardin, 2002; Putnam, 2000). The Paris Agreement (2015) is the most comprehensive international response to climate change and achieving its goal of limiting global warming is dependent on enough governments committing to voluntary emissions reductions. The success of the Paris Agreement has been argued to depend on the ability of domestic actors to mobilise public opinion to pressure national governments into implementing ambitious national climate policies (Keohane & Oppenheimer, 2016).

However, governments trying to implement national climate policies risk being punished at the polls or confronted in the streets by violent protestors that perceive for instance flat rate fuel taxes as economically unjust (Klenert et. al, 2018; theGuardian.com, 2018). Simultaneously, across the world millions of young people are referring to the Paris Agreement as motivation to participate in the school-strikes-for-climate. The movement’s call for intergenerational- and climate justice convey solidarity with the peoples and communities on the frontlines of the climate crises. Further, it demands action from those with the power to change the course of escalating greenhouse gas emissions that could risk the prosperity of young and future generations (Time.com, 2019; IPCC, 2018).

1.2 Previous Research & The Research Gap

Previous research has found that the public hold strong aversions and general distrust toward environmental taxation whilst economists consider climate taxes as the first-best policy alternative alongside subsidies and trading schemes. Climate taxes are understood to be a highly efficient and effective policy as they set out to internalising the cost of carbon pollution in the market price, this is known as a Pigouvian tax (Baranzini et. al, 2014; Baumol & Oates, 1971; Brannlund & Persson, 2012; Nordhaus 2007; IPCC, 2014; Kallbekken et. al, 2011; Kallbekken & Aasen, 2010; Jagers and Hammar 2009; Harring, 2016; Harring, 2018).

2 “Social dilemmas” are used interchangeably with “collective action problems” throughout the article.

(7)

Furthermore, research has found that low public support for carbon- or energy taxes is based on perceptions that price-mechanisms such as climate taxes are ineffective (Nordhaus, 2007; Baranzini et. al, 2014: 11; Kallbekken et. al, 2011; Kallbekken & Aasen, 2010; Brannlund & Persson, 2012).

Some researchers argue that this is partly caused by general distrust in politicians and the political institutions among the general public (Rafaty, 2018; Baranzini et. al, 2014).

There are several studies that compare national climate policies on a macro-level, but very few studies have considered the role of trust as one determinant of environmentally related policy outcomes despite its theoretical importance (Bättig & Bernauer, 2009; Benauer & Böhmelt, 2013;

Jahn, 2016; Neumayer, 2003; Scruggs, 2003; Steves & Teytelboym, 2013). The study most closely related to this study is the macro-comparative study by Rafaty (2018), as he finds that corruption, which is closely interconnected with public distrust, affects the stringency of environmental policies negatively.

1.3 The Aim, Research Question and Results

This thesis purposes to examine the role of social- and political trust in national government policy outcomes of climate taxation. To the knowledge of the author, this thesis provides the first macro- level analysis of the effects of different forms of aggregated trust on national climate policy. The theoretical framework posits how social- and political trust could affect policy outcomes of national climate taxes in a cross-national comparative context. In pursuit of the aim of this thesis, the research question reads.

How does citizens’ aggregated levels of social- and political trust respectively affect national policy outcomes concerning climate taxation?

The research question will be answered through quantitative analysis using an Ordinary Least Squares fixed effects regression model. The analysis is based on Time-Series-Cross-Sectional (TSCS) panel data from the OECD Policy Instruments for the Environment database (2017), the European Social Survey Round 3-8 (2006-2016) and the Quality of Government Basic Dataset (2019).

This thesis finds significant evidence of an effect of trust on policy outcomes of climate taxation that are both supportive and contradictory of its different hypotheses. In the discussion, I relate the results to previous research, outline potential for future research and argue that these come

(8)

with relevant insights to national policy makers in relation to the prospects of engaging in collaborative behaviour to solve the climate crises both on the national and the international level.

1.4 Contribution

The contribution to previous research is both theoretical and empirical. The thesis makes a theoretical contribution by elaborating more extensively compared to previous studies on the theoretical mechanisms that link social and political trust to solutions of social dilemmas in general and climate taxation in particular. The empirical results contribute to the research field on how different forms of trust affect policy outcomes on climate taxation empirically. In terms of external relevance, the results are of interest to policy makers, professionals, activists and citizens interested in addressing climate change on all levels of society. To exemplify, social trust is the only trust variable that is positively correlated to climate taxation. However, social trust is essentially limited to an exclusive group of countries, such as the Nordics, the Netherlands and Switzerland (Holmberg & Rothstein, 2017). Consequently, policy makers interested in implementing climate taxes in other societies are encouraged to compromise on the effectiveness of economic policy on behalf of public acceptability (Nordhaus, 2018; Klenert et. al, 2018).

(9)

2. Theoretical Framework

This section outlines how trust affect policy outcomes based on the literature on trust, rational choice and behavioural approaches to social dilemmas. Firstly, the thesis defines its understanding of social and political trust. Secondly, it draws on the empirical research on trust and climate policy.

Thirdly, it makes the theoretical link between trust and social dilemmas by contrasting rational choice perspectives with behavioural approaches on trust. This elaborates more thoroughly on the theoretical mechanisms that link trust to environmental policy as compared to previous empirical research. Lastly, the hypotheses on the relationship between social- and political trust and climate tax revenue are presented.

2.1 Definitions of Trust In Previous Research

Contemporary research is concerned with two main categories of trust, interpersonal- or vertical trust and institutional- or horizontal trust. This thesis will understand interpersonal- or horizontal trust as social trust and institutional- or vertical trust as political trust (Bretzer, 2005; Rothstein &

Stolle, 2008; van Deth 2000; Zmerli & Newton, 2017).

2.1.1 Social Trust

The concept of social trust emerged out of Tocqueville’s (1835) Democracy in America, as he believed that it would be the shared experience of the American people rather than the constitution that would uphold civil- and political freedoms in the newly established democracy (Tocqueville, 2000).

Social trust is defined as “…the expectation of one person about the actions of others that affects the first person’s choice, when an action must be taken before the actions of others are known”

(Dasgupta, 1997: 5). Social trust may take many forms, depending on how far an individual is capable of extending trust beyond themselves. This thesis will concern itself with the most extended form of social trust found in the trust research, generalized trust. This form of social trust is derived from consistent positive encounters by a large diversity of others and it makes it possible for people to relate and connect to people who are not necessarily similar to themselves (Brewer 1981; Delhey et. al, 2011; Stolle 2002; Uslaner, 2017).

2.1.2 Political Trust

Political trust has been central to the art of government since Confucius (551-471BC) stated that government rule relied on weapons, food, and trust. Also, in western philosophy this has been a prevalent thought. Hobbes (1651) argued in Leviathan that a state of nature necessitated individuals to entrust political power to a sovereign in order to safeguard mutual benefits in a context where

(10)

distrust was otherwise flourishing as the result of competition over scare resource (Listhaug &

Jakobsen, 2017). Whilst disagreeing with Hobbes about the nature of this trust, contract theorists of the Renaissance elaborated on political trust, as Rousseau (1762) argued in The Social Contract that individuals are only bound by laws supported by the general will of the people (Freeman, 2012).

Political trust denotes the confidence of citizens that their government will act to do what is right and perceived fair (Easton, 1965). Political trust is divided into sub-categories based on the belief that individuals relate differently to different state institutions, this has been outlined in theory and confirmed empirically (Harring, 2018; Harring & Jagers, 2013; Rothstein & Stolle, 2008). The scope of this thesis will be limited to trust in political institutions, trust in implementing authorities and trust in the international organisations. Trust in political institutions relates to the institutions determining the political agenda. Trust in implementing authorities regard trust in the state institutions that are considered as neutral and impartial (Bretzer, 2005; Rothstein & Stolle, 2008;

Zmerli & Newton, 2017). Trust in international organisations relate individuals’ trust in international, and supra-national institutions responsible for the implementing or determining a political agenda (van Deth, 2000).

2.2 Trust and Climate Policy in Empirical Research

Remarkably few empirical studies address the relationship between trust and environmental public policy. However, one debate has concerned the flow of causality between trust and the creation of environmental policy. Some argue that implementing policies to address environmental degradation and climate change would have an increasing effect on political trust as these problems become more visible and thus more central to public opinion (Huseby, 1997; Kvaløy et. al, 2012).

This originates in the understanding of political performance within standard political behaviour models, where political trust is considered a consequence of “the government delivering results and outcomes that are in demand by citizens” (Listhaug & Jakobsen, 2017: 5).

To the contrary, other studies have provided evidence of the opposite causal relationship based on statistical analysis of how perceptions of perceived corruption, understood as distrust in politicians, result in weaker climate policies (Harring, 2014; Rafaty, 2018). Based on time-series-cross-national analysis Rafaty (2018) argues that corruption, or distrust, causes a reduced stringency of environmental policies. This as perceptions of corruption, the equivalent to distrust in politicians, is argued to weaken the position of politicians to regulate climate change which strengthens the relative leverage of business opposed to the climate agenda. Ultimately, this provides strong

(11)

incentives for policy makers to adopt a business-as-usual approach (Rafaty, 2018). Concurring with this understanding, distrust has been argued to cause low public support for environmental taxes and to result in weak energy and carbon tax policies in qualitative studies (Baranzini et. al, 2014).

2.3 The Theoretical Link Between Trust and Social Dilemmas 2.3.1 Social Dilemmas

This thesis draws on the insight that political activity on behalf of a group is a public good, as the actions of an individual are consequential to the collective (Olson, 1965). Further it elaborates on the understanding outlined by Elinor Ostrom in Governing the Commons (1990), as she defines social dilemmas as conflicts between individual utility and collective utility in relation to the provision of public goods and management of common pool resources.

“Social dilemmas occur whenever individuals in interdependent situations face choices in which the maximisation of short-term self-interested yields outcomes leaving all participants worse off

than feasible alternatives” (Ostrom, 1998: 1).

As such social dilemmas emerge as there is a necessity of mutual cooperation to achieve optimal utility for the collective. Consequently, this incentivises individuals, states or organisations to renege on cooperative behaviour to maximise its own utility in the short-term. This is known as free riding which is detrimental to the collective since it leads to long-term sub-optimal outcomes (Keohane, 1984; Lubell 2002; Ostrom, 1990; Ostrom, 1998).

2.3.2 First-Order and Second-Order Social Dilemmas of Climate Change

The first order-social dilemma is the original collective action problem. The second-order social dilemma is the collective action problem that emerges when there is a need to enforce the commitments to the original social dilemma. As such, mitigating climate change is a global social dilemma of the first-order. The conventional collective action perspective posits that cooperation to address climate change cannot be achieved if there is no external authority to determine the course of action, monitor its implementation and sanction non-collaborative behaviour (Hardin 2002, Ostrom, 2010).

This perspective comes with significant challenges to the prospect of successfully addressing climate change under the Paris Agreement. This is because the agreement lacks a legally-binding enforcement mechanism and mitigating efforts are depending on countries voluntary contributions in the form of national climate policy. The only institutional safeguard of the Paris Agreement

(12)

depends on norm-formation among states that would facilitate collaboration in terms of more ambitions national climate policies (Kemp, 2018; Keohane and Oppenheimer, 2016; Victor, 2016).

Consequently, international collaboration to mitigate climate change under the Paris Agreement makes up a second-order social dilemma.

2.3.3 Behavioural Perspectives on Rational Choice Theory of Trust

Perspectives of strong rationality in the economic sciences hold that “trust is irrelevant to commercial exchange” as potential risk or the presence of safeguards will always be reflected in the price (Williamson, 1993: 469). This strong rational choice perspective considers the state as the essential authority in overcoming non-cooperative, self-serving and short-term behaviour. It assumes that different agents act according with strong rationality based on cognitive evaluations derived on narrow self-interest and complete information (Cook & Santana, 2017). This perspective, ultimately making the perspective of trust redundant, is context-dependent and others have argued that the role of trust is made more relevant in contexts of increased conflict between individual and social utility (Balliet & Van Lange, 2013; Samuelson, 1954).

The behavioural approach has grown out of a critique of the rational choice theory. In behavioural research, reciprocity, reputation and trust are considered pre-requisites for long-term, cooperative behaviour in situations of social dilemmas across different contexts (Ostrom, 1998). The behavioural model consider trust as central to cooperation. This as trust is understood to be interconnected with both reputation and reciprocity “ … at the core of a behavioural explanation are the links between the trust that individuals have in others, the investment others make in trustworthy reputations, and the probability that participants will use reciprocity norm.” (Ostrom, 1998: 12).

The model ascribes a “bounded” or “thin” rationality to actors and assumes that they are constrained by cognitive limitations, imperfect information and social norms (Ostrom, 1998;

Simon 1955). The presence of trust is considered necessary to overcome the strong temptations of acting according with the short-run self-interest held by an individual, an organisation or the state (Keohane 1984; Ostrom, 1998). Both social- and political trust have been found to be fundamental to address collective action problems as individuals are hesitant to act collectively if they believe others will not (Lubell, 2002; Malka et. al, 2009; White & Gatersleben, 2010). According to the behavioural approach, higher trust is considered to lower transaction costs in any social, economic and political relationship and to facilitate collaborative behaviour in both first-order and second

(13)

order social dilemmas (Fukuyama, 1995; Heckathorn, 1989; Ostrom, 1990; Ostrom 1998;

Yamagishi, 1986; Yamagishi, 1992).

2.4 Hypotheses About Social- and Political Trust and Climate Related Tax Revenue Democracy theorists and social capital theorists have long believed that social trust allows states to develop functional institutions. As a result, they see social trust as being decisive to the emergence of political trust (Fukuyama, 1995, Putnam et. al, 1993, Putnam, 2000; Tocqueville, 2000). On the contrary, recent empirical studies utilising time-series-cross-national analysis have defined a reverse flow of causality from political trust to social trust, in several European countries (Sønderskov &

Dinesen, 2016; Seifert, 2017). There are diverging views in the literature about the causal relationship between social and political trust. In this thesis, I will not attempt to disentangle the direction of causality. Instead different forms of trust are theorised to have independent effects on climate tax revenue through different theoretical mechanisms and this will furthermore be tested empirically.

2.4.1 Networks and Institutional Safeguards

Concerning the role of different forms of trust in affecting policy outcomes, social trust is a pre- requisite for the formation of networks and political trust is fundamental in establishing and upholding the credibility of institutional safeguards. Networks are dependent on social trust as they facilitate social cohesion between groups of individuals across geographical scales. Furthermore, networks decrease uncertainty and facilitate cooperation based on reciprocity as well as reputation and as result people engage in deepened social-, political- and economic transactions with members of their network (Cook, 2005; Granovetter, 1985; Hardin, 2002; Molm et. al, 2009; Ostrom, 1998;

Walker & Ostrom, 2009).

Institutional safeguards are understood to be created and upheld by organisations and institutions such as the state or an international organisation, e.g. in the form a national tax system or international agreements (Hardin, 2002; Cook et. al, 2005). An institutional safeguard is understood to “align individual and collective interests” (Cook et. al, 2005: 17) by serving as “trustworthy third parties that certify exchange and enforce expectations” (Cook & Santana, 2017: 17-18).

In order to overcome social dilemmas at increasing scale of the national or international level, some have argued that institutional safeguards are a necessity and that this makes trust in networks obsolete (Cook et. al, 2005; Hardin 2002). These perspectives need not to be considered as mutually exclusive, as research has found that individuals are capable of extending social trust beyond the

(14)

nation state and as such networks can be considered to be effective at increasing scale (Delhey et.

al, 2011; Newton et. al, 2017). Therefore, both networks and institutional safeguards are theorised to be mechanisms that lower transaction costs and facilitate cooperation in relation to how different forms of trust affect the implementation of national climate policy. This is further theorised to be true across scales spanning the local to the global (Cook et. al, 2005; Fukuyama 1995, Newton et. al, 2017, Tilly 2004; Yamagishi, 1992).

2.4.2 Institutional Safeguards: Trust in Political Institutions and Trust in Implementing Authorities

Trust in political institutions is understood as relevant in relation to individual’s trust in the legitimacy of national legislatures to implement effective and fair institutional safeguards to address the first-order social dilemma. In turn, trust in implementing authorities regard individuals’

perceptions concerning the legitimacy of impartial government agencies to enforce compliance with institutional safeguards and as such avoid the emergence of a second-order social dilemma (Hardin, 2002; Harring, 2018; Cook, 2005, Ostrom 1998). Consequently, trust in political institutions is relevant as it relates people’s trust in the national legislative to implement climate taxes that are fair and effective. Trust in implementing authorities facilitate the credibility of the institutional safeguard as it relates to citizens’ perceptions that the taxes are managed fairly and effectively (Scholz & Lubell, 1998; Svallfors, 2013).

2.4.3 Networks and Social Trust

Social trust is applicable in relation to the perception among individuals’ that others will comply with the first-order social dilemma. As such higher levels of social trust decrease the risk of free riding and the emergence of a second-order dilemma. Furthermore, if the second order dilemma emerges social trust have been attributed to facilitate collaborative action vis-à-vis the original social dilemma, rather than dedicating precious resources to sanctioning behaviour (Cook & Santana, 2017; Ostrom, 1998; Walker and Ostrom, 2009; Yamagishi, 1986; Yamagishi, 1992). Therefore, social trust is relevant to determine individuals’ willingness to comply with climate taxes determined by their perceived trustworthiness of their fellow citizens to pay their fair share of taxes (Scholz &

Lubell, 1998; Svallfors, 2013). As such the capacity of governments to collect climate taxes to address climate change will depend on social trust, trust in political institutions and trust in implementing authorities. Consequently, I hypothesise that;

Hypothesis 1

Governments in societies with higher level of social trust collect more environmentally related tax revenue than governments in societies where individuals perceive low levels of such trust.

(15)

Hypothesis 2

Governments in societies with higher level of trust in political institutions collect more climate related tax revenue than governments in societies where individuals perceive low levels of such

trust.

Hypothesis 3

Governments in societies with higher level of trust in implementing authorities collect more climate related tax revenue than governments in societies where individuals perceive low levels of

such trust.

2.4.4 Trust in International Organisations and National Climate Policy

The Paris Agreement’s voluntary approach is fundamentally different to the previous international climate regime of Kyoto, which relied on a top-down legally binding framework. However, both lack the capacity to sanction non-collaborative behaviour (Keohane and Oppenheimer, 2016). In formulating its hypothesis on how trust in international organisations could affect national climate policies the thesis will consider the conditions valid to the top-down approach of the former climate regime, as the available data is from the pre-Paris analysis. The old climate regime can be considered as an institutional safeguard that both implements and enforces emissions reductions that are dictated to national governments (van Deth, 2000; Keohane and Oppenheimer, 2016).

Based on the understanding that trust facilitates cooperation and lowers transaction costs, high aggregated trust in international organisations could mean that countries rely more on international global warming mitigation mechanisms than national policy (Fukuyama, 1995; Yamagishi, 1992).

As such, I hypothesise that:

Hypothesis 4

Governments in societies with higher levels of trust in international organisations collect less climate related tax revenue than governments in societies where individuals perceive

low levels of such trust.

(16)

3. Research design

This section presents the analytical framework, choice of method, case selection as well as data and measures including the dependent variable, independent variables and control variables.

3.1 Analytical Framework

The analytical framework, see Figure 1, is derived from the theoretical framework and is considered a visual aid in understanding the conceptualisation of the effects of social and political trust on policy outcomes respectively. The internal relationship between social and political trust is not the subject of empirical analysis. Instead, the thesis aims to test whether both forms of trust have independent effects on environmental tax revenues from a comparative perspective. To conclude, it deserves to be emphasised that the statistical analysis does not set out to determine a causal relationship between trust and policy outcomes.

3.2 Choice of Method

Ordinary Least-Squares (OLS) linear regression is used to study the effects of trust on policy outcomes of climate taxation. Multivariate regression enables to draw on theory to define control variables to isolate the correlation between the independent variables and the dependent variable for any spurious influencers, factors that could explain the relationship. This method is chosen as the independent and dependent variables are measured on a continuous ordinal scale. Furthermore, it allows to estimate the effect on the dependent variable of changing one independent variable and keeping all other independent variables constant (Teorell & Svenson, 2007: Stock & Watson, 2015).

However, the explanatory potential of the OLS-method is limited, as it is unable to define causal mechanisms³ and time-order concerning the relationship between the independent and dependent

3 Also known as a counter-factual relationship (Teorell & Svensson, 2007)

Policy Outcomes Social Trust

Political Trust*

?

Figure 1 – Analytical framework

*Trust in Implementing Authorities, Trust in Political Institutions and Trust in International Organisations.

(17)

variables. Consequently, the statistical analysis cannot provide further evidence as to whether trust precedes climate taxes or if there is a causal relationship between the two (Teorell & Svensson, 2007).

This thesis utilises a form of multivariate OLS-regression analysis called fixed effects regression model. This method is used because the available data is structured in the form of Time-Series- Cross-Sectional panel data. Fixed effects are included to the fixed effects regression model in the form of dummy-variables for both year and country. The year dummies isolate the correlation for inter-temporal variations (year specific conditions) and the country dummies isolate the correlation concerning contextual factors (country-specific conditions) (Stock & Watson, 2015).

In the fixed effects regression model, time fixed effects control for changes in unobserved variables that that are constant across countries but change over time and country fixed effects control for changes in unobserved variables that are constant over time but differ across countries.

Unobserved variables are conditions that are not included in the model and other factors that could not be modelled. Such factors constitute what is usually referred to as unobservable time-invariant individual heterogeneity. To conclude, fixed effects are included in the multivariate regression as a way to mitigate bias in pooled models where many observations belong to the same country (Arellano, 2003; Stock & Watson, 2015).

The regression model relies on a set of basic assumptions, one of those is that there can be no perfect multicollinearity, the variance influence factor (VIF) is used to study if this assumption hold. A second assumption is that the residuals, the sum of the error terms, need to be allocated according with the normal distribution in the data (Teorell & Svenson, 2007; Stock & Watson, 2015). To remove concerns that results of the fixed effects regression models are skewed due to heteroscedasticity⁴ and autocorrelation⁵, cluster-robust standard errors by country are used (Stock

& Watson, 2015). I set out to describe two additional assumptions of the regression-model in the section on data and measures. However, as the assumptions of the fixed regression model and the OLS-multivariate regression are slightly different, the thesis will limit its scope to the most relevant assumptions (Stock & Watson, 2015).

4 When the residuals are not distributed according to the normal distribution (Stock & Watson, 2015).

5 Also understood as serial correlation with values of lagged values (Stock & Watson, 2015).

(18)

3.3 Case Selection

The study bases its case selection on available data on OECD-countries.⁶ Also, most countries are either members of the European Union or the Single Market except for Israel, Switzerland and Turkey. Membership in the EU or the Single market comes with the delegation political authority between supra-national and national level as decided by the constitutional treaties of the Union, the Treaty of Rome (1958) and Treaty of Maastricht (1992) as amended by the Treaty of Lisbon (TFEU) (2007) that entered into the force on 1 December 2009 (Panizza, 2018). This allows for cross-national and cross-temporal analysis of a set of countries that share legally binding frameworks dictating the functioning of a single market and a commonly agreed policy on energy and climate. Furthermore, the TFEU defines that taxation is the authority of national governments (European Union, 2007).

This case selection allows for analysis of 24 countries based on bi-annual observations over 12 years (2004-2014) making up 115 observations in the fixed effects regression models with the complete set of control variables. This includes complete Time-Series-Cross-Sectional data on 16 countries⁷ and incomplete time-series data on 8 countries⁸. The fixed effects regression models without controls contains 150 observations of 28 countries with complete data on 14 countries⁹ and incomplete data on 14 countries¹⁰ over 14 years (2004-2016).

3.4 Data and Measures

3.4.1 Dependent Variable: Climate Tax Revenue

The dependent variable is operationalised by using data from the OECD database Policy Instruments for the Environment (2017) on “environmentally related tax revenues” covering over 1400 different taxes of both OECD and non-OECD countries spanning 1991-2016. To operationalise climate related taxes, the thesis only use revenue from “energy products (including

6 Except for Bulgaria and Cyprus that are not members of the OECD but are members of the EU.

7 Belgium, Denmark, Estonia, Germany, Spain, Finland, UK, Hungary, Ireland, Netherlands, Norway, Poland, Portugal, Slovenia, Sweden, Switzerland.

8 Austria (2004-2006, 2014), Bulgaria (2006-2012), Croatia (2008-2010), Greece (2004, 2008-2010), Iceland (2004, 2012), Lithuania (2010-2014), Luxembourg (2004), Italy (2012).

9 Belgium, Estonia, Finland, Germany, Hungary, Ireland, Norway, Poland, Portugal, Slovenia, Spain, Sweden, Switzerland, United Kingdom.

10 Austria (2004 – 2006, 2014-2016), Bulgaria, (2006-2012), Denmark (2004-2014), Croatia (2008-2010), Cyprus (2006-2012), Croatia (2008-2010), France (2004-2014), Greece (2004, 2008-2010), Iceland (2004, 2012, 2016), Israel (2008-2014), Lithuania (2010-2016), Luxemburg (2014), Italy (2012, 2016), Turkey (2004, 2008).

(19)

vehicle fuels)”¹¹ and “motor vehicles and transport services”¹². This excludes streams of revenue that are less relevant to national climate policy, such as water and wastewater, waste management and mining and quarrying. This limitation is used as the environmentally related tax revenue from energy and transports are most closely related to the concept of climate taxes, as outlined below, and as such they are considered to make up climate related tax revenue¹³.

Previous research has operationalised the price of carbon emissions to compare national climate policies (Klenert et. al, 2018). Climate policies that regulate the price on carbon set out to increase the cost of emitting greenhouse gases (e.g. carbon dioxide or methane). The main climate policies that regulate this price are trading-schemes and carbon taxes, understood as climate taxes (Goyal et. al, 2018). Furthermore, an accurate description of climate taxes enables cross-country analysis of national climate policy outcomes. This is because climate taxation is a policy area that are within the authority of each respective governments. This as compared with international trading schemes such as the European Trading Scheme that have been established on the supra-national level in Brussels (Goyal et. al, 2018). National policy outcomes are understood in relation to climate related tax revenue and this study is limited in scope as to making the connection between how these taxes reduce the emissions of greenhouse gases, however this is included as a control variable (Bättig and Bernauer, 2019).

The operationalisation of climate tax revenue is measured in tax revenue per capita expressed in purchasing power parity (PPP) of the USD in the year 2010, as this adjusts for the size of the population and price-levels between countries (OECD, 2017). The measurement is expressed in relative terms as the logarithm of the real GDP per capita to allow for OLS-linear regression analysis (Stock & Watson, 2015). There are both benefits and draw-backs to this operationalisation.

This as energy and transport related tax revenue include taxes on different forms of greenhouse gases. However, this measurement includes taxes that are indirectly leveraged on the emission of greenhouse gases (OECD, 2017). The former could contribute to higher validity of the measurement as it not only includes taxes on carbon emissions, whereas the latter could contribute to lower validity as it includes a broad set of tax-instruments that are not equally efficient in limiting greenhouse gas emissions (Goyal et. al, 2018).

11 Referred to as tax revenue from energy.

12 Referred to as tax revenue from transports.

13 ”Climate Related Tax Revenue” is used interchangeably with Climate Tax Revenue.

(20)

The thesis argues that this operationalisation of climate taxes is in accordance with the theoretical definition and allows for cross-national comparison of national climate tax policies. Therefore, it holds that the measurement meets the ideals of high validity and high reliability (Teorell &

Svensson, 2007).

3.4.2 Independent Variables

The independent variables on trust come from the European Social Surveys (ESS) from Round 2 (2004) to Round 8 (2016). The ESS is a bi-annual international survey that collects individual-level data using face-to-face-interviews based on common sampling principles with a minimum of 70%

response rate and coordinated country specific sampling designs (Lynn et. al, 2007). The ESS-data enables to safely assume that the survey-data can be considered to make up independently and identically distributed random variables, the third assumptions of the regression model. The final assumption of the model that is considered is that the OLS-analysis is sensitive to outliers (Stock

& Watson, 2015).

Furthermore, the individual-level data from the ESS have been treated with design weights and population size weights. The former corrects for the differences in the probability of selection of individuals aged 15+ in each country whereas the latter allows for cross-national comparison as it adjusts for the use of similar sample size across countries with vastly different populations (europeansocialsurvey.org, 2014), see Appendix A for a full description of the constructs of the complete set of the independent variables.

3.4.2.1 Social Trust

The independent variable on social trust leans on one of the most commonly used operationalisations of social trust in survey research namely, “most-people can be trusted questions”, see Appendix A (Almond & Verba, 1963; Verba & Nie, 1972). This thesis uses a mean- based index to operationalise generalised trust inspired by the article Davidovic et. al, (2019) and combine three different survey-questions on an ordinal scale (See Appendix A). This as the indexed measurement could mitigate some of the cross-cultural differences of what social trust means in different countries, that otherwise could be a source of variation (Freitag & Bauer, 2013).

Consequently, the social trust index is believed to be of higher validity and reliability compared to applying a single measure or any other combinations of the three survey-questions that capture generalized trust. The internal consistency of the index is controlled with Cronbach’s alpha (α = 0.94). The internal consistency is 0.94. Thus, it passes the rule of thumb of a minimum value of 0.7.

(21)

3.4.2.2 Political Trust

Political trust is operationalised based on three different theoretical definitions of trust. One mean- based index is constructed, and two independently applied variables are utilised, inspired by previous trust research and the “trust-in-government questions” (Levi & Stoker, 2000: 477). The operationalisations are based on the theoretical understanding and empirical application concerning how individuals relate to the political institutions, implementing authorities and international organisations, and are measured on an ordinal scale (Rothstein & Stolle, 2008; Harring

& Jagers, 2013; Harring, 2018; van Deth, 2000). This distinction is inspired partly by the approach used by Davidovic et. al, (2019). However, the definitions used in this thesis draws on different terms to describe similar concepts. The terms used in this thesis are trust in political institutions, trust in implementing authorities and trust in international organisations as this allows for an understanding of institutional trust as an umbrella term, as outlined in theory.¹⁴

Trust in implementing authorities is operationalised by a mean-based index of trust in the police and the legal system (α = 0.82). This measurement can be understood to partly reflect citizens perceptions of the government authorities’ capacity to enforce the second-order social dilemma concerning compliance with climate taxation. There could be problems with validity as the specific government agencies are concerned with either the enforcement of taxation, the protection of the environment or climate change. This is because there is potential discrepancy between the theoretical definition and the empirical operationalisation (Teorell and Svensson, 2007). Since Cronbach’s Alpha does not support indexing the items for trust in political institutional trust and trust in international organisations. Instead, the operationalisations of these forms of trust are based on single survey-items.

To operationalise trust in political institutions, a survey-item on trust in the country’s parliament is used. This as it is the national legislature that is the political institution ultimately responsible for national legislation of national climate policies. To operationalise trust in international organisations, the item on trust in the United Nations is used as it is the international organisations in charge of the international climate change regime. Researchers concerned with the validity of trust-batteries have argued that that trust is situation-specific (Bauer & Freitag, 2017).

14 The term institutional trust is used in the study by Davidovic et. al, (2019) to describe what the thesis understands as trust in implementing authorities and the term political trust is used to describe what the thesis understands as trust in political institutions.

(22)

Therefore, the increased specificity of the operationalisations of trust in political institutions and trust in international organisations respectively, as compared to the indexes, could even come with increased validity. This as researchers argue that trust should be measured as it relates to a behaviour (trust of an individual), referring to an object (the Parliament / the UN), at a specific a time (each time-series observation) (Baier, 1986; Hardin, 1992; Sztompka, 1999).

This depends however, on whether each specific survey-item captures the theoretical definitions.

Is trust in national parliaments representative of individuals' trust in political institutions that determine the political agenda in general? Is trust in the United Nations representative of individuals’ general perceptions of international organisations capable of determining and enforcing political agendas? To conclude, the reliability of these measurements can be considered to be high as they are based on several rounds of ESS-data.

3.4.3 Control Variables

The control variables are derived from The Quality of Government Basic Dataset (2019) as it compiles Time-Series-Cross-Sectional-data (on the country level from a number of internationally

recognised sources (Teorell et. al, 2019). Furthermore, the ESS Round 2-8 makes available an individual-level control variable on environmental values, see Appendix B for full description of the constructs of the complete set of controls.

3.4.3.1 Economic Output per Capita and Tax Revenue

Economic output per Capita, “Log of Real GDP per Capita”, is measured in the log of GDP per capita adjusted by PPP in 2011 USD. “Tax Revenue” is measured as total tax revenue percentage of GDP. These are applied to ensure that the general difference in economic output and propensity to collect tax revenue between countries and at different times, is not reflected in the dependent variable.

3.4.3.2 National Carbon Emissions and Fossil Fuel Energy

“National Carbon Emissions” is measured in metric tonnes per capita and “Fossil Fuel Energy” is operationalised as the share of fossil fuel energy consumption as a percentage of the total energy consumption. These are relevant as climate taxes are directly or indirectly related to energy consumption and emissions of greenhouse gases. As such they isolate for cross-national and inter- temporal variation relating to fossil fuel consumption as percentage of the total energy consumption and carbon dioxide emissions (Goyal et. al, 2018).

(23)

3.4.3.3 Quality of Government

Operationalising “Quality of Government” a mean-value index is used to isolate any spurious variation due to its components “Corruption”, “Law” and “Bureaucracy Quality”. These have been applied in previous research to isolate the relationship between trust and public support for climate taxes on the individual level (Davidovic 2018; Davidovic et. al, 2019). Furthermore, isolating for corruption aligns itself with the understanding of Rafaty (2018) that perceived corruption have a causal effect on the stringency of environmental policy. Further relevance of this control variable is derived based on research on public support for taxation and government spending, which have found that the propensity to pay taxes is conditioned upon the effectiveness and fairness of government institutions (Scholz & Lubell, 1998; Rothstein & Stolle, 2008, Svallfors, 2013).

3.4.3.4 Democracy

The level of “Democracy” is operationalised based on political freedoms and civil rights and draws on a mean-value index from Freedom House on political- and civil rights. This is inspired by the macro-level study by Bättig & Bernauer (2009) as they draw on political freedoms to operationalise the level of democracy and find a positive correlation between the level of democracy and national climate change policy output. Policy output is understood as countries political commitments (Bättig & Bernauer, 2009)

3.4.3.5 National Politics

“National Politics” is operationalised as “share of seats in parliament: green” to control for the influence of green parties on the implementation of national climate policies. This draws on the role of ideology in the formation of public attitudes (Shworm, et. al. 2015). Electoral success of green parties has been found to lead to mainstream parties de-emphasising environmental politics depending on the age of the green party. Mainstream parties tend to be more accommodative to established green parties and this results in an increased vote-share of these parties as the salience of environmental politics increase. Consequently, an increased share of green parties in national legislatures could be translated into climate policies (Abou-Chadi, 2016; Grant & Tilley, 2019).

3.4.3.6 Environmental Values

The control “Environmental Values” is based on individual-level data from the ESS survey-item

“important to care for nature and environment”. This variable is operationalised based on the aggregated mean values of each country. Environmental values have been argued by Inglehart (1971) to be one indicator of post-materialist values. Empirical studies have found that individuals with high levels of environmental values are more supportive of environmental protection and that

(24)

societies with high levels of post-material values are more supportive of environmental protection on the national level (Fairbrother; 2016; Gelissen, 2007).

(25)

4. Results

This section presents descriptive statistics, the fixed effects regression models and discusses the results. Thereafter, the support for the different hypotheses are discussed and interpreted. Lastly, obstacles to drawing generalisable conclusions based on the fixed effects regression model are presented.

4.1 Descriptive Statistics

Figure 1 Social Trust and Climate Taxes Figure 3 Trust in International Organisations and Climate Taxes

Figure 4 Trust in Implementing Authorities and Climate Taxes Figure 5 Trust in International Organisations and Climate Taxes

Figures¹⁵ 2, 3, 4 and 5 visualise the bi-variate relationships between social trust, trust in implementing authorities, trust in international organisations and trust in political institutions respectively, and climate related tax revenue. In all Figures Italy and the Netherlands are outliers as their environmentally related tax revenue is substantially higher than predicted by their respective levels of trust. In Figure 3 and 4, Italy and the Netherlands are joined by Slovenia and in Figure 5, Austria replaces Slovenia to make up the outliers with higher values than predicted by the model.

15 The Y-axis is measured as Climate Tax revenue per Capita (Log of USD 2010 PPP)

(26)

Across all Figures, Lithuania and Poland have lower tax revenues than predicted by the model, in Figure 5, they are joined by Spain.

4.2 The Fixed Effects Regression Models

Table 1 Fixed Effects Regression Models of Trust and Climate Tax Revenue

(27)

4.3 The Results of The Fixed Effects Regression Model

Table 1 describe the complete fixed effects regression models with dummy variables for year and country. In Model 1, 2, 3 and 4 social trust (0.024***) is significant and positively correlated whereas trust in political institutions (-0.017***), trust in implementing authorities (-0.015***) and trust in international organisations (-0.021***) are all significantly and negatively correlated, with the dependent variable climate related tax revenue. Model 5 applies the complete set of controls including the independent variables and the results are significant at the 1% level across all variables, social trust (0.046***) is positively correlated with environmental tax revenue whereas trust in political institutions (-0.015***), trust in implementing authorities (-0.010***) and trust in international organisations (-0.026***) are all negatively correlated.

Furthermore, there are some interesting results concerning trust in political institutions and trust in international organisations when considering the changes between different regression models.

This as the correlation of trust in political institutions and trust in implementing authorities change from being positively correlated to being negatively correlated when applying the complete set of controls (see Appendix C, Table 2 & Table 3). This indicate that there are some spurious influencers that are isolated for when applying the complete model.

4.4 Support for Hypotheses

Hypothesis 1, that higher aggregated levels of social trust correspond to governments collecting higher climate taxes is supported. This as there is a significant positive correlation across all models including the complete model (Table 1, model 5). The significant results can be interpreted as the increase of aggregated social trust of one degree, the equivalent of Sweden increasing its level of social trust to levels of Norway (Figure 2), correspond to an increase of climate related tax revenue of 4.6%.

Hypothesis 2, that higher aggregated levels of trust in political institutions correspond to governments collecting higher climate taxes is not supported. This as there is a significant negative correlation which contrasts the hypothesis (Table 1, model 5). This contradictory evidence can be interpreted as an increase of aggregated trust in political institutions of one degree, the approximate equivalent of United Kingdom increasing trust in international organisations to levels of Germany in the year 2016, corresponds to a decrease in climate related tax revenue of 1.5% (Figure 3).

Hypothesis 3, that higher levels of aggregated trust in implementing authorities correspond to governments collecting more climate related taxes is not supported. However, there is significant

(28)

evidence of a negative relationship (Table 1, model 5). The results can be interpreted as an increase of aggregated levels of trust in implementing authorities of one degree, the equivalent of Spain increasing trust in implementing authorities to levels of Hungary in 2016 (Figure 4), corresponds to a 1% decrease in climate related tax revenue. Concerning both hypothesis 2 and 3, there are interesting questions to why the significant positive correlation disappears when applying the control variables.

Hypothesis 4, that higher levels of aggregated trust in international organisations correspond to governments collecting more climate related taxes is supported. This as there is as significant negative correlation across all models (Table 1, model 5). This provides strong support for the theoretical framework and the thesis argues that this could have implications to the prospects of implementing the Paris Agreement. The results are interpreted as an increase of aggregated trust in international organisations of one degree, the approximate equivalent of Switzerland increasing trust in international organisations to levels of Lithuania in the year 2016 (Figure 5), corresponds to a decrease in climate related tax revenue of 2.6%.

4.5 Obstacles to Drawing Generalisable Conclusions

There are conflicting results as to the effect of trust on policy outcomes of climate taxation, that both confirm and run in contrast to the outlined hypotheses. The most pressing issue in terms of generalising the results is the risk that the results are incompatible with the Fixed Effect Regression- assumption of no perfect multicollinearity. This is indicated both by high VIF-numbers and the negative adjusted R² values of the fixed effects regression model. However, assuming that the results are valid, this enables the thesis to draw generalisable conclusions. Furthermore, it provides support for two of the hypotheses, concerning how changes in the aggregated levels of social trust and trust in international organisations among citizens correspond with changes in climate tax revenue across specific country contexts and for different time periods.

(29)

5. Discussion

This section discusses the support and non-support in relation to the assumptions that the results of the fixed effects regressions models enables to draw generalisable conclusions. Firstly, the prospects for generalising the results to a greater population are discussed. Secondly, the support for the hypothesis on social trust and trust in international organisations are discussed. Thirdly, the contradictory evidence to the hypothesis on trust in political institutions and trust in implementing authorities is evaluated based on the validity of the measurements and the theoretical framework.

Throughout the discussion, the results are related to previous research as a basis for insights into potential future research. Lastly, the conclusion summarise the discussion based on the results, the thesis’ contribution to research and the main prospects for future research.

5.1 Generalisable Results

The significant evidence enables for preliminary generalisation of the results in relation to the OECD-country context. This as the fixed effects regression model has controlled for country- specific and time-specific conditions in addition to applying a comprehensive set of control variables. However, these results are derived based data from a relatively homogenous group of OECD-countries (Stock & Watson, 2015). To the knowledge of the author, this contribution constitutes the first macro-level comparative study providing significant evidence on how different forms of trust affect national climate policy outcomes of climate taxation.

Moreover, this study contributes cumulatively to the field of trust-research, where the effects of social trust are well elaborated on and the role of trust in international organisations remains under- explored. Furthermore, the results are to be considered relevant in the context of the model and as such provide support for the theoretical framework. However, the theoretical framework cannot be confirmed based on fixed effects regression models as they are not capable of confirming specific mechanisms or define that the time order run from trust to climate taxes, this can only be argued based on theory (Teorell & Svensson, 2007; Stock & Watson, 2015). Further research is needed on both non-OECD and OECD country data before any definitive generalisations are made as to the effect of social trust and institutional trust on climate tax revenue in particular and national climate policy in general.

5.2 Implications of the Support for Hypothesis 1 & 4

The support for the hypothesis on social trust provides evidence of a positive correlation between social trust and climate tax in the fixed effects regression model. The results strengthen the argument that the presence of social trust facilitates cooperative behaviour and lower transaction

(30)

costs through networks (the mechanism) and that this have concrete effects on the policy outcomes to address collective action problems in regard to both first-order and second-order social dilemmas (Cook et. al, 2005; Fukuyama, 1995; Ostrom, 1998; Lubell, 2002; Tilly, 2004; Yamagishi, 1992). Furthermore, these results are aligned with the understanding that individuals are capable of extending trust in networks far beyond themselves (Delhey et. al, 2013; Newton et. al, 2017).

The support for the hypothesis on trust in international organisations strengthen the theoretical argument that higher aggregated levels of trust in institutional safeguards on the international level defer climate action on the national level. This as institutional safeguards, as they relate to trust in international organisations, are theorised to be the mechanism capable of determining and enforcing political agendas to address both first-order and second order social dilemmas on the international level. Consequently, collective action on the national level is postponed based on a perception that it is ineffective and insufficient to act nationally to address large scale social dilemmas (Cook et. al, 2005; Hardin, 2002; van Deth, 2000).

5.2.1 Social Trust and The Implications to Research and Policy

The significant evidence of the positive effect of social trust on climate tax revenue aligns itself with previous research that argue that social trust facilitates positive outcomes on a broad range of issues such as democracy, welfare states, political tolerance, and the rule of law to name a few (Putnam 1993; Putnam, 2000; Rothstein & Stolle, 2008; Rothstein & Uslaner, 2005) This thesis provides some preliminary evidence that social trust qualifies as a macro-level explanation of positive national climate policy outcomes as it increases the capacity to implement and collect climate taxes (e.g. Bättig and Bernauer 2009; Rafaty, 2018).

Concerning implications to policy makers, research has confirmed that high levels of social trust is a phenomenon that is limited to very few countries, the Nordic Countries, Netherlands, Switzerland, Australia and New Zealand and that it is inherently difficult to affect levels of social trust (Holmberg & Rothstein, 2017). Therefore, implementing carbon taxes can be more easily achieved through policy-designs that compromise efficiency on behalf of public acceptability instead of relying on high levels of trust (Klenert et. al, 2018).

5.2.2 International Organisations and The Implications to Research and Policy

There is reason to be concerned as to the policy consequences of the significant evidence that higher-levels of trust in international organisations defers climate action on the national level in relation to the prospects to successfully increase mitigating ambitions within the Paris Agreement.

This as national governments reneging on their Paris commitments could in-turn incentivise other

(31)

countries to free ride and as such undermine both the effectiveness and the legitimacy of the Paris Agreement (Kemp, 2018; Keohane and Oppenheimer 2016). As such, non-cooperative behaviour in relation to the Paris Agreement could be detrimental as most of the world’s largest economies are required to decarbonise by 2050 to limit climate change (Ostrom 1998; IPCC, 2018; Kemp, 2018; Victor, 2016). However, as the public is responsive to policy makers there is reason to remain hopeful (Stimson, 1995; Dahl, 1971; Pitkin 1967). Consequently, the negative relationship could turn into a positive one as the understanding that the new international climate regime requires national climate action permeates the view of the public (Wlezien & Soroka, 2016). This opens up for interesting future research on trust in international organisations and perceptions of the Paris Agreement related to national climate policies, both as the Agreement is implemented and as more data become available.

Furthermore, there is interesting potential for future research to elaborate on the relationship between trust in international organisations and sub-international climate policies, such as comparison of regional carbon trading schemes or metropolitan congestion fees (Klenert et. al, 2018, Goyal et. al, 2018). Specifically, as it relates to Ostrom’s (2010) concept of polycentricism, as she argues that in the absence of an institutional safeguard on the international level, multiple actors on the local, regional and national level are capable of cooperating to effectively address large-scale social dilemmas such as climate change (Ostrom, 2010).

5.3 Implications of Non-Support for the Hypothesis 2 & 3

There are several potential explanations to the contradictory results to the hypotheses on trust in political institutions and trust in implementing authorities respectively. This section discusses validity of the measurements and the implications to the theoretical framework as a result of the non-supported hypotheses.

5.3.1 The Validity of the Measurement of Trust in Political Institutions

The significant evidence contradicting the theoretical framework serve as a basis to question the validity of the measurement, trust in country’s parliament. If the measurement is found to be of low validity it could mean that there is a positive relationship between individuals’ trust in political institutions and national policy of climate taxation as an institutional safeguard to address first- order social dilemmas. However, there is little room to criticise the operationalisation of trust in international organisations based on the survey-item on trust in country’s parliament. This as previous research has found that survey items relating to specific branches of government are of higher validity. As the object of this study concern national legislation on climate taxes,

To Trust or Not to Trust in National Climate Policy A Quantitative Macro-Analysis on The Role of Social- and Political Trust in Determining National Policy Outcomes of Climate Taxation