Getting the Public on Side

1. Defining a conceptual framework for analysing public acceptability: What factors matter for inequality-reducing policies and how can they be measured?

Abstract

This chapter explores the role that perceptions, attitudes and other subjective factors play in shaping the demand for and public acceptability of policies designed to reduce inequalities. Taking stock of current OECD research and data collection efforts, it highlights the need for a broader and more realistic approach to better account for the range of factors involved, the interactions between them as well as cross-national differences and the influence of context. Based on the guidance from the interdisciplinary OECD Expert Group on New Measures of the Public Acceptability of Reforms, a conceptual framework is proposed for analysing the formation of preferences for inequality-reducing policies. This framework is designed to help extend OECD research into areas including equality of opportunities, discrimination and horizontal inequalities and could be supported through the setting up of an OECD Perceptions Database.

1.1. What does the OECD’s work on inequality tell us about public acceptability?

The 2021 report Does Inequality Matter? has made an important contribution to the OECD’s study of inequality. The OECD has a long and well-established tradition of work on the trends, causes and consequences of inequality.1 The report Does Inequality Matter? How People Perceive Economic Disparities and Social Mobility extended the analysis into an important new area by looking at the “demand-side” of the issue (OECD, 2021[1]). In recent decades, the wider use of survey-based and experimental methods has allowed for deeper insight into the key perceptions, attitudes and policy preferences relating to inequality. Does Inequality Matter? reviewed the growing body of evidence provided by these perceptual and behavioural data. From there, it highlighted the relevant insights that can be drawn regarding how people perceive inequality and respond to policies designed to reduce it, focusing mainly on income inequality and redistributive policies.

In doing so, it has also underlined the relevance of the OECD’s work on inequality for the broader question of public acceptability. By shedding light on the “demand-side” of inequality, the report has also helped advance the OECD’s analysis of public acceptability. It has done so in two ways: (i) by tracing the impact that the observed trends and evidence on inequality have had on public perceptions and debates; and (ii) by providing an empirical basis for assessing how well the policies designed to tackle inequalities are aligned with the public’s views on the issue and how likely they are to receive sufficient public support. The fact that demand for redistribution is relatively well studied and understood makes it a good starting-point for a broader reflection on the factors that drive public acceptability and the extent to which they can be measured. The OECD has contributed to collect perceptual and behavioural data on inequality through tools such as the Risks that Matter cross-national survey, Compare Your Income web-tool and Trustlab.2

Does Inequality Matter? confirmed many of the key findings from the literature regarding the importance of perceptions and their implications for policy. Most notably, it showed that:

• Perceptions capture real changes and are not pure artifacts: Perceptions and the level of concern about inequality tend to reflect the picture provided by conventional estimates. The link between actual and perceived inequality can be seen in cross-country data: where conventional estimates indicate higher income disparities and intergenerational persistence, on average people also perceive greater income inequality and lower social mobility. This link can also be seen over time, with people’s concerns about income disparities tending to move in line with conventional estimates. As a result, where the latter have risen the most, the former have also tended to increase the most. Furthermore, differences in the level of concern about inequality are not strongly related to other macro-economic conditions: higher employment rates and median incomes reduce concern, but the effect is small and insignificant (OECD, 2021, pp. 40-43[1]).

• Perceptions matter and they contribute to shape policy preferences: The level of concern people express about inequality is an important driver of demand for redistribution. Across the OECD in 2017, 80% of those who found that income disparities were too large also agreed that it was the responsibility of government to reduce them. Furthermore, perceived inequality may be more relevant for understanding policy preferences than actual levels of inequality. In this respect, higher levels of income disparity only tend to translate into greater demand for redistribution insofar as they are also accompanied by increased concern about inequality (OECD, 2021, p. 93[1]).

Does Inequality Matter? also drew attention to two problems that need to be taken into account when designing inequality-reducing policies. First, policy preferences are not easily explained by traditional socio-economic characteristics. In most OECD countries, perceptions of inequality have tended to become more dispersed in recent decades, with a growing gap between those who believe that income inequality is too high and those who believe it is not (OECD, 2021, pp. 129-149[1]). As shown in Figure 1.1, disagreement over the extent of earning disparities is significantly larger than it was 30 years ago. Furthermore, very little of the growing dispersion in perceptions can be explained by traditional socio-economic characteristics. According to OECD calculations, no more than 10% of the total dispersion in perceptions of income inequality and social mobility relates to differences between socio-economic groups in terms of income, education, employment status, gender, age and household type. The remaining 90% reflects differences in perceptions of inequality within socio-economic groups – i.e. increased disagreement between people with similar socio-economic characteristics. A similar conclusion can be observed in other important areas, as highlighted for example by OECD evidence on attitudes towards climate policies.3 At an analytical level, this calls for the exploration of a wider range of potentially relevant socio-economic characteristics, notably the effects of spatial location on people’s perceptions and policy preferences. It also calls for greater focus on “demand-side” factors and the role they play in explaining differences in individual views and preferences.

This problem presents challenges for policymakers and for the international organisations tasked with advising them. The fact that policy preferences cannot easily be explained by traditional socio-economic characteristics has significant implications for the political economy of reform. Identifying and building coalitions in support of policies becomes more challenging if policymakers are not able to target coherent and well-defined socio-economic groups.4 Similarly, policy advice must take account of the key perceptions and attitudes that constitute the “demand-side” of reform and provide a clearer understanding of what views people hold and why in order to be effective in this context. In turn, appropriate data and analytical tools are needed to support policy advice.

Figure 1.1. People disagree more than they did 30 years ago about earning disparities

10^th and 90^th percentiles of the perceived top-bottom earnings ratio, averaged over 8 countries

Note: The lines represent the difference between the 10th and 90th percentile of respondents ranked by their view on the top-bottom earnings ratio. The values are the average of the values for Australia, Germany, Italy, New Zealand, Norway, Slovenia, Switzerland and the United Kingdom. Austria, Hungary, Poland, Sweden and the United States were not part of the last available ISSP wave and have not been included in this analysis. Including these countries does not change the overall pattern: the perception differential is much larger in the late 2000s than in the late 1980s.

Source: OECD (2021[1]), Does Inequality Matter? : How People Perceive Economic Disparities and Social Mobility, OECD Publishing, Paris, https://doi.org/10.1787/3023ed40-en; OECD calculations from ISSP 1987, 1992, 2009, 2019; Australian Survey of Social Attitudes 2019; Norwegian part of ISSP 2019; British Social Attitudes 2019 (Section 4.2).

The second problem consists in explaining persistent cross-country variation in the extent to which concern about inequality translates into actual support for inequality-reducing policies. The fact that concern about inequality does not uniformly and unambiguously give rise to increased demand for redistribution is a well-identified “paradox” in the literature.5 In some cases, a wide gap exists between levels of concern about inequality and support for government intervention. For example, in one-fifth of OECD countries only 60% or less of those who think that income disparities are too large believe that it is the responsibility of government to reduce them (see Figure 1.2). Here again, these observed differences between countries in terms of perceptions, attitudes and policy preferences need to be understood in order to provide policy advice that is effective, adapted to specific national contexts and likely to receive sufficient public support.

Figure 1.2. Concerns about inequality do not necessarily translate into support for redistribution

Fraction of people who believe it is the responsibility of the government to reduce income differences, among those who think that such disparities are too large, 2017

Note: Respondents are asked their opinion about the statements “Differences in income in [your country] are too large” and “It is the responsibility of the government to reduce the differences in income between people with high incomes and those with low incomes”. In Eurobarometer, the statements provided are slightly different: “Nowadays in [your country] differences in people’s incomes are too great” and “The government in [your country] should take measures to reduce differences in income levels”, but the response scale is identical. For consistency, this figure uses data from ISSP where available.

Source: OECD (2021[1]), Does Inequality Matter? : How People Perceive Economic Disparities and Social Mobility, OECD Publishing, Paris, https://doi.org/10.1787/3023ed40-en; OECD calculations from ISSP 2017, apart from Belgium, Estonia, Greece, Ireland, Italy, Luxembourg, Latvia, Netherlands, Poland, Portugal, Slovenia whose data are from Eurobarometer 471/2017.

Finally, Does Inequality Matter? outlined a programme for further research. Overall, Does Inequality Matter? demonstrated the valuable additional insights that can be drawn by combining and comparing the “objective” measures of actual inequality provided by conventional statistical indicators with “subjective” measures reflecting how inequality is perceived, experienced and interpreted. It also pointed to the need for further research to fully exploit the insights from perceptual and behavioural data and support the design of more effective policies for reducing inequality. Following OECD (2021[1]), several priorities can be identified to guide the OECD’s research in this area:

• In order to turn the available perceptual and behavioural data into actionable insights, the ways in which the relevant perceptions, attitudes and preferences interact need to be better understood.

• Moving beyond the narrower focus on income inequality and demand for redistribution, further questions emerge regarding (i) the types of inequality that matter to people; (ii) the broader range of actors and policies that contribute to reducing inequalities; and (iii) the role played by “horizontal” forms of inequality between groups.

• From a more practical perspective, it is also important to consider what types of data and indicators, both objective and subjective, are needed to shed light on these questions and how they can best complement existing OECD indicators on inequality.

Progress is needed on two fronts: (i) richer data on people’s perceptions and attitudes towards inequality; and (ii) better analytical frameworks to explain how their policy preferences are formed. The OECD is contributing to advance the research agenda on both of these fronts. On the data collection side, the OECD Observatory on Social Mobility and Equal Opportunity has extended the scope of analysis beyond the starting-point provided by OECD (2021[1]). In order to do so, it has looked beyond the latter’s focus on income inequality and demand for redistribution and has collected data on perceptions and attitudes towards other types of inequalities and on preferences for a broader range of inequality-reducing policies. These data offer a richer and more complex picture of what drives the demand for and acceptability of inequality-reducing policies. More detailed background on the work conducted under the OECD Observatory on Social Mobility and Equal Opportunity can be found in Annex 1.A. In parallel, the OECD Expert Group on New Measures of the Public Acceptability of Reforms has sought to strengthen the conceptual basis of the OECD’s work by improving its understanding of the process through which policy preferences relating to inequality are formed (see Box 1.1).

Based on the input of the Expert Group, this chapter develops a more realistic and adapted model of the process through which preferences for inequality-reducing policies are formed. This has implied updating the standard economic model used in Does Inequality Matter? to measure and interpret perceptions and preferences (OECD, 2021, pp. 33-35[1]). While this model is broadly in line with the economic literature on perceptions of inequality and preference for redistribution and has demonstrated its utility in the context of Does Inequality Matter?, a multidisciplinary perspective is needed to address several of the remaining analytical and methodological challenges highlighted in the previous section and move the research agenda forward.6 The Expert Group was set up to provide this multidisciplinary perspective. Its objective in developing a conceptual framework for analysing the formation of policy preferences is (i) to help better identify the key subjective factors that contribute to shape public acceptability in the area of inequality; and (ii) to advise on an effective strategy for measuring these factors.

The details of the proposed conceptual framework are presented in Section 1.2. Section 1.2.1 identifies the key factors to include in the framework. Section 1.2.2 discusses additional issues relating to the structure of the framework. Finally, Section 1.2.3 describes the framework.

Box 1.1. The OECD Expert Group on New Measures of the Public Acceptability of Reforms

The Expert Group on New Measures of the Public Acceptability of Reforms was set up in 2021 to help the OECD take stock of recent advances in the measurement and analysis of subjective factors, including perceptions, attitudes, policy preferences and social norms. The aim in doing so was (i) to consolidate the OECD’s efforts to collect relevant perceptual and behavioural data; (ii) to harness the insights from this growing body of evidence on the role that subjective factors play as determinants of public acceptability; and (iii) to improve the OECD’s policy advice and recommendations through a better understanding of the key elements that constitute the “demand side” of reform.

Many of the recent advances in the measurement and analysis of subjective factors have been driven by the use of experimental methods and behavioural approaches that are not always well-represented in standard policy toolboxes. The research agenda on this topic has also taken place across (and often at the intersection of) many different disciplines. Reflecting this, the Expert Group was designed to be interdisciplinary and brought together leading academic experts from relevant fields including economics, game theory, behavioural science, political science and sociology. Given its focus on policy applications, it also included experts from national governments and statistical offices, as well as experts from other international organisations (European Commission, International Monetary Fund and World Bank) and from foundations working on this topic (Ford Foundation). The members of the Expert Group are listed in the Foreword to this document.

The mandate of the Expert Group consisted in helping the OECD develop a theoretical and empirical framework to support policymakers’ efforts to design reforms that are acceptable to the public, as called for at the 2019 OECD Conference on The Changing Political Economy of Reform.

In order to do so, the activities of the Expert Group were organised into two complementary stages:

• The Expert Group started by looking at the question of the demand for and acceptability of inequality-reducing policies. The reasons for choosing this particular question and area as a starting point are two-fold. First, the role played by subjective factors is well-studied and understood in the case of inequality and the OECD has collected extensive evidence on this topic that the Expert Group can contribute to leverage and extend (see section above). Second, inequality is both a highly salient topic and one that is likely to yield broader lessons on public acceptability through the specific focus it puts on fairness considerations and on the distributional impact of policies. The perceived fairness and effectiveness of policies are two essential determinants of public acceptability and key conditions for the successful implementation of reforms. In the case of inequality, both are intrinsically connected. This stage of the Expert Group’s activities is covered in Chapter 1 of this report.

• In a second stage, the Expert Group considered the lessons that can be drawn from the study of inequality-reducing policies for the broader question of the public acceptability of reforms across policy areas. In particular, it sought to explore (i) to what extent the key lessons from the first stage can be generalised and applied to assess public acceptability in other policy areas (How do people perceive distributional impacts?; What type of fairness principles do they use when assessing policies?...); and (ii) how the insights from perceptual and behavioural data can be organised and used more effectively to inform the design and communication of reforms. This stage of the Expert Group’s activities is covered in Chapter 2.

1.2. How are policy preferences formed? A conceptual framework for understanding the public acceptability of inequality-reducing policies

1.2.1. The content of the framework: What are the key factors to consider in the analysis?

Policy preferences are shaped by a range of other perceptions and attitudes and cannot be properly understood in isolation. These perceptions and attitudes may relate to the policies themselves, but also to the issues they are meant to address or to the broader context in which they take place. Based on the input from the Expert Group, policy preferences relating to inequality are analysed here as the complex outcome of interactions between a set of subjective determinants. The key factors identified by the Expert Group include (i) beliefs about inequality; (ii) attitudes towards policies and government; (iii) attitudes towards others and society in general; and (iv) the type of information available to individuals and how it is received.7 This section provides a short overview of these key factors, why they matter for the analysis of policy preferences and what challenges remain to be addressed.

Beliefs about inequality

How people view inequality is essential for understanding whether or not they support policies designed to reduce it. These views themselves are shaped by different types of beliefs. Based on the literature, three main types of belief about inequality can be distinguished (Mijs, 2018[2])8:

i. Perceptions of inequality, which consist in descriptive beliefs about the levels and trends in inequalities (both in terms of outcomes and opportunities, at the bottom and at the top of the distribution…), as well as individuals’ relative position within the distribution.

ii. Explanations of inequality, which consist in causal beliefs about the determinants of inequality. As highlighted by the broader literature and confirmed by OECD data (see previous section), there is an additional important distinction to be made here between two types of causal beliefs that explain inequalities by attributing them to factors either within or beyond an individual’s control. On this basis, beliefs about inequality tend to be organised around (i) a “meritocratic” worldview in which outcomes primarily reflect the individual’s own agency (talent, hard work, effort…); and (ii) a “structuralist” worldview in which external factors play a significant role in shaping individuals’ outcomes despite their agency (luck, socio-economic background…).

iii. Attitudes towards inequality, which consist in a wide range of judgements relating to inequalities (beliefs about the acceptability and fairness of inequalities, what types of inequality matter, whether they should be reduced or not…).

Important questions remain to be addressed regarding beliefs about inequality, their determinants and the role they play in the formation of policy preferences. Further research is notably needed to better understand how these different beliefs interact with one another and with contextual variables. Perceptions, explanations and attitudes towards inequality tend to be closely related and can influence one another in multiple and sometimes contradictory ways. One well-studied example of a complex channel through which perceptions, explanations and attitudes interact can be seen in the dampening effect that meritocratic beliefs have on people’s concerns about inequality and on their support for redistribution.9 “Unpacking” people’s views about inequality is therefore not a straightforward or linear task and a degree of psychological realism is needed in order to properly account for the way in which different types of beliefs interact. Understanding how these different types of beliefs relate to individual characteristics is equally complex and can help explain the limited relevance of traditional socio-economic determinants in predicting people’s views about inequality (OECD, 2021[1]) and about climate policies (Dechezleprêtre et al., 2022[3]). Figure 1.3 provides an overview of what is known about the individual determinants of meritocratic beliefs.

Figure 1.3. Belief in meritocracy and redistribution: What do we know about individual-level determinants?

Note: “+/-“ indicate an established correlation, positive and negative respectively, between individual characteristics and key beliefs relating to inequality. “?” indicates that individual characteristics are not predictors of key beliefs relating to inequality or that the results from the literature are ambiguous.

Source: Jonathan Mijs – Presentation to the OECD Expert Group on New Measures of the Public Acceptability of Reforms.

Further research is also needed to better understand under what conditions people form and change their beliefs about inequality. Evidence from the experimental literature confirms that people do adjust their perceptions and concerns about inequality when provided with new information, but these effects do not always translate through to their policy preferences, notably for redistribution.10 Furthermore, people do not act like ideal “Bayesian learners” actively and continuously updating their beliefs in line with the available information. Evidence from developmental psychology suggests that certain core beliefs about inequality are formed early and remain relatively stable throughout life (Almås et al., 2010[4]; Almås et al., 2024[5]; Dickinson, Leman and Easterbrook, 2023[6]). Similarly, evidence from the political science and sociological literatures suggests that attitudinal change is more likely to happen at a generational level than within cohorts (Janmaat, 2013[7]). As a result, the analytical framework should also seek to take account of the processes through which people update their beliefs about inequality and the conditions under which they do so. Furthermore, the relative importance of belief formation and belief adjustment has significant implications for the value of information, for communication strategies and for the dynamics of attitudinal change.

Finally, the social context plays an important role in shaping the formation of beliefs about inequality and in enabling or hindering the adjustment of beliefs. An individual’s social context influences how they learn about inequality (Mijs, 2018[2]; Mijs, 2023[8]). It can also contribute to reinforce particular beliefs about inequality. For example, high levels of socio-economic segregation may facilitate the formation of meritocratic beliefs which may in turn deepen inequality and segregation (Mijs and Roe, 2021[9]). The analytical framework should therefore seek to take account of the role played by social context as a “filter” through which beliefs about inequality are formed and adjust. Understanding its role will require in turn a greater use of geo-spatial data and further efforts to analyse the impact of household location on perceptions and attitudes. Doing so can help better identify the conditions under which people update their beliefs, what type of information they are likely to integrate into their views and in what form.

Attitudes towards policies and government

The perceived effectiveness and fairness of policies are important determinants of public support across different areas. Empirical evidence confirms that this is notably the case for inequality-reducing policies (Durante, Putterman and van der Weele, 2014[10]; OECD, 2021[1]) and for climate policies (Dechezleprêtre et al., 2022[3]; Colantone et al., 2024[11]; Voeten, 2024[12]). More broadly, trust in government and institutions is an outcome of the perceived effectiveness and fairness of policies and public services (Murtin et al., 2018[13]). It is also a factor that contributes to perceived effectiveness and fairness, as trust in government and institutions tends to increase compliance with a range of public policies (OECD, 2017[14]; 2024[15]), with some evidence suggesting this was the case for compliance with public health measures during the COVID-19 pandemic (Wright, Steptoe and Fancourt, 2021[16]). As a result, all three of these factors (perceived effectiveness, perceived fairness, trust in government) should feature in the analysis of policy preferences on inequality.

Important gaps may nonetheless exist between people’s support for policy goals and for the means through which these goals are to be implemented. Figure 1.4 provides an illustration of this by highlighting the gap in support between climate policies that are publicly acceptable (such as banning the sale of inefficient household appliances – Panel A) and those that economic theory suggests are most effective in reducing emissions (such as a carbon tax – Panel B). The provision of information on the effectiveness and distributional impact of policies can in some cases help shift attitudes and strengthen public support, as shown by (Dechezleprêtre et al., 2022[3]) for a range of climate policies including a carbon tax with cash transfers. However, improving people’s understanding of how policies work and who they affect through the provision of information will not always lead public views to align with the most effective or efficient policies.11 Policymakers will therefore continue to face real trade-offs between efficiency and public acceptability when selecting instruments in a number of areas (Kallbekken, 2013[17]; Cohn, 2016[18]).12 This is more likely to be the case in situations where (i) there is a wide gap between the measures that are efficient from a cost/benefit point of view and those that are acceptable to the public; and (ii) people’s attitudes towards policies or issues are based on strongly-held beliefs that may not be swayed by new information.

Figure 1.4. Across the EU, there are wide gaps in public support between the most acceptable policies and the most effective policies in the context of the Green Transition

Source: European Commission: Directorate-General for Employment, Social Affairs and Inclusion (2019[19]), Employment and social developments in Europe 2019 – Sustainable growth for all – Choices for the future of Social Europe, Publications Office, 2019, https://data.europa.eu/doi/10.2767/305832, based on European Social Survey 2016.

Evidence confirms that the principles used to evaluate fairness tend to be broadly shared, with some variation across countries (see Figure 1.5). However, people may disagree when applying them to specific cases and different principles may be used to evaluate different types of policies.13 As a result, further research is needed to understand (i) which fairness principles are relevant for evaluating policies and their outcomes and in what contexts; and at a more granular level (ii) how different groups converge on a shared understanding of what counts as fair, how they apply these principles when assessing the status quo and what policy responses they consider appropriate when the status quo deviates significantly from what is perceived to be “fair”. The available evidence suggests that specific fairness beliefs are mobilised when evaluating different types of policies (Cavaillé, 2023[20]). In the case of inequality, this implies that particular attention be given to two types of fairness beliefs:

• Meritocratic vs. Structuralist beliefs, which are relevant for understanding attitudes towards redistributive policies (for example, willingness to support progressive taxation) as these causal beliefs about inequality play a central role in shaping people’s assessment of the fairness of market outcomes and the degree to which effort is rewarded;14 and

• Reciprocity beliefs, which are relevant for understanding attitudes towards mutual insurance (for example, willingness to support the expansion of social welfare). Here, evaluations of both the fairness and effectiveness of policies are tied to beliefs about the motivational aspects of social benefits (including potential dis-incentivising effects) as well as the personal characteristics and behaviour of their recipients (risk of “moral hazard” and absence or prevalence of free-riding).15

Figure 1.5. Across the EU, there is broad support for fairness principles combining merit-based elements and basic access

Note: Respondents are asked their opinion on the importance of the following statements as reflections of “what a society should provide”: “Guaranteeing that basic needs are met for all in terms of food, housing, clothing, education, health”; “Recognising people on their merits”; and “Eliminating large inequalities in income between citizens”. % shown in the chart combines those considering these principles to be “very important” or “quite important” as opposed to “not important” or “not at all important”.

Source: European Commission: Directorate-General for Employment, Social Affairs and Inclusion (2020[21]), Employment and social developments in Europe – Leaving no one behind and striving for more – Fairness and solidarity in the European social market economy, Publications Office, https://data.europa.eu/doi/10.2767/478772, based on European Values Study 2017.

Attitudes towards others and society in general

Attitudes towards policies and government are not the only element to take into account when considering the relation between inequality and support for policies designed to reduce it. This idea is well captured in classical and modern political theories of the social contract. These theories include a “vertical” dimension through which government derives legitimacy and the loyalty of citizens from its capacity to deliver the fundamental individual rights and public goods that are part of the social contract. However, theories of the social contract also include a “horizontal” dimension through which individuals commit themselves to follow shared rules that are necessary for the functioning of society and of a common government.16 This horizontal dimension has been formalised in economic and game-theoretic models of the social contract, where it has been used to explain the emergence of social order and highlight the role that mutual expectations play in sustaining shared standards and stable patterns of collective behaviour.17 In the context of welfare and redistribution, it reflects the idea that, while government constitutes one of the main levers through which society can reallocate resources in a way that is designed to promote the common good, in the end of the day redistribution represents a collectively agreed transfer of resources within society itself in addition to a moral obligation towards individuals.

Attitudes towards others are highly relevant for understanding people’s beliefs about inequality and their support for inequality-reducing policies. In the context of welfare and redistribution, support for policies partly reflects self-interest but is often also mediated by the expectations and beliefs that people have of others (Cavaillé and Trump, 2015[22]; Dimick, Rueda and Stegmueller, 2016[23]). In turn, these “other-regarding” expectations and beliefs tend to be “target-specific” and differentiate between groups in terms of deservingness (van Oorschot, 2000[24]; Fong, 2001[25]; Fong and Poutvaara, 2019[26]). The analysis needs to be able to take account of these “target-specific” beliefs, their influence on policy preferences and the key reference groups they are applied to. Significant gaps remain to be filled here. For instance, while there is a well-developed literature on how “the poor” are perceived, what is expected of them and what constitutes pro-social behaviour on their part, more work is needed on how “the rich” are perceived and what people expect of them. Focus should also be put on better understanding the effects that social welfare institutions have on perceptions of recipients and support for policies to help explain cross-country differences in key beliefs and attitudes towards inequality (Larsen and Dejgaard, 2013[27]; Hedegaard and Larsen, 2019[28]).

The impact of discrimination on support for policies, notably through the racialisation of attitudes towards government interventions, is another important area where further research is needed. Alesina, Miano and Stantcheva (2023[29]) provides an interesting illustration of a perception gap of this type (see Figure 1.6). The left panel shows for example that nearly a quarter of French respondents believe immigrants receive at least twice as much in government transfers as native citizens do. In fact, when pension benefits are included, immigrants actually receive less in transfers than native citizens do. The analysis suggests furthermore that the perception gap is largely driven by the belief that immigrants are favoured by the welfare system or take advantage of it, rather than by the more benign belief that immigrants are poorer than native citizens and receive transfers they are entitled to. Evidence from Alesina, Miano and Stantcheva (2023[29]) also points towards widespread and often large misperceptions regarding the size, composition and socio-economic characteristics of immigrant populations.

Figure 1.6. There are large gaps in the perceived reliance of immigrants on government transfers

Share of respondents who think immigrants receive at least twice as many government transfers as non-immigrants

Note: The left panel shows the share of respondents in 6 OECD countries who think that an average immigrant (defined as “somebody legally living in the country of the respondent but born abroad”) receives at least twice as many government transfers as an average non-immigrant. The shaded areas represent the 95% confidence intervals around the average share. The right panel represents the average share across all countries for the following population groups: those who do not work in a high immigration sector; those who work in high immigration sectors and (i) have no college degree or (ii) have a college degree; the college-educated and non-college-educated; high-income and low-income respondents; those who have an immigrant parent and those who do not have an immigrant parent; respondents 18-45 vs. 46-69 years old; male vs. female; and left-wing vs. right-wing. Respondents are classified as being in high immigration sectors depending on whether their current sector of employment (or last sector of employment, if currently unemployed) has an immigrant share that is higher than the national average.

Source: Alesina, Miano and Stantcheva (2023[29]), Immigration and Redistribution, The Review of Economic Studies, https://doi.org/10.1093/restud/rdac011.

Furthermore, attitudes towards others play an important role in determining whether societies can solve social dilemmas and how. Perceptions of others tend to focus on particular salient traits that can give rise to stereotypes and stigma when they are generalised and taken as representative of a whole group (Kahneman and Tversky, 1972[30]; Bordalo et al., 2016[31]; Alesina, Miano and Stantcheva, 2023[29]). The literature on collective action highlights the strong effect that the existence of stereotypes and “labels” can have on the level of coordination and extent of cooperation within societies, alongside other structural conditions such as the size and homogeneity of societies (Sugden, 1986[32]; Ostrom, 2000[33]; Ellingsen et al., 2012[34]). Similarly, shared norms and conventions also help determine how specific societies solve “social dilemmas” where coordinated action is needed and influence the long-term “paths” they take as well as the institutions they develop (Young, 2001[35]; Algan and Cahuc, 2009[36]). Social dilemmas are useful tools for understanding and analysing the “public goods” aspects of welfare and redistribution (Rothstein, 2001[37]). Here again, attention should be given to the role played by horizontal forms of inequality, including group identity and stereotypes, in shaping support for inequality-reducing policies, as well as to the role that institutions (including informal ones such as social norms) play as devices for solving social dilemmas.

The role of information

Information plays an important role in explaining how individuals experience and learn about inequality, as well as the beliefs they form. Information does not only influence an individual’s perceptions, it also has a broader effect on their beliefs about inequality (Trump, 2018[38]) and attitudes towards others (Cetre et al., 2020[39]). This reflects the broader fact that both types of belief are mobilised and partly generated during the process of learning about inequality. The analytical framework should therefore seek to take account of the key characteristics of the information environment, including in terms of access to information, the salience of inequality as an issue and the type of information consumed. The Inequality and Politics survey, for example, sheds light on the salience of inequality across 14 OECD countries (Pontusson et al., 2020[40]). Results suggest that, at least in the countries surveyed, inequality is a highly salient topic (see Figure 1.7). However, while over 50% of respondents in all countries and income deciles consider economic inequality to be important, the issue only features among the 3 most important topics for respondents in one of the countries surveyed (Germany). The survey also finds that the salience of inequality depends on income and social status, but often in a complex and non-linear way.

Figure 1.7. Inequality is a highly salient issue in many OECD countries

% of respondents for whom economic inequality is a very important or extremely important topic

Note: Respondents were asked to rate the importance for them of the following 7 topics: “Economic inequality”, “Crime and terrorism”, “Unemployment”, “Public debt”, “Immigration”, “Public services”, and “Environment”. Answers were on a 5-point scale going from “Not at all important” to “Extremely important”. The figure represents the proportion of respondents who answered “Very important” (4) or “Extremely important” (5).

Source: Pontusson et al. (2020[40]), “Introducing the Inequality and Politics Survey: Preliminary Findings”, Unequal Democracies Working Paper No. 16, University of Geneva, https://doi.org/10.48573/jfa4-cd92.

Further research is also needed to better understand the mechanisms through which individuals collect and process information when forming their views on inequality. There is significant heterogeneity in how individuals “receive” information, notably in terms of search behaviour, reference groups and trusted sources.18 Here again, the social context may play a significant role in “filtering” people’s experience of inequality, the type of information they have access to and how they process it. This can contribute to increase and entrench the dispersion of views on inequality, notably in societies with high levels of segregation (Mijs, 2017[41]; 2018[2]).

The “supply side” also needs to be considered when analysing the informational environment. In particular, this involves looking at the ways in which media and political actors shape the reception of information about inequality through their agenda-setting power, their influence on the salience of inequality and the specific narratives they develop to frame public debates (Cavaillé, 2023, pp. 164-219[20]).These actors may play a significant role in explaining cross-national differences in attitudes towards inequality and relevant policy preferences. Taking account of the supply of information is particularly important in light of the growing challenges raised by misinformation and disinformation spread online.19

Finally, a measure of caution is warranted as more information does not always lead to better outcomes. As highlighted by the literature in game theory and behavioural economics, this is notably due to the impact that information may have at the collective level in terms of facilitating or hindering coordination and cooperation,20 as well as its impact on individual behaviour and the limits of individuals’ capacity to process more information.21 In this respect, policymakers and international organisations may need to consider what types of information to provide, when and with what explanatory context, particularly in situations where there is significant heterogeneity in views.

1.2.2. The structure of the framework: How should the factors be organised?

In addition to identifying the key factors for analysis, a number of structural issues need to be addressed in order to build a model of the formation of policy preferences relating to inequality. The recommendations from the Expert Group imply the need for deeper analysis of several key factors already present in (OECD, 2021[1]) (beliefs about inequality; attitudes towards policies and government; the role of information) and the introduction of a new type of factor into the model (beliefs about others and society). Furthermore, they underline the role that other background elements, including notably the social context and group identity, play in shaping these factors. The Expert Group provided advice on how to structure the framework in a way that can integrate the different factors identified and the main elements that influence them. The present section explains the approach taken in organising the key factors.

The proposed framework is organised around 3 different blocks. These blocks cover the main specific “components” that are necessary to understand policy preferences relating to inequality and what determines them:

• A first block on Inequality would cover the processes through which people learn about and understand the framework’s core issue, including the beliefs they form (the What? question, which serves to define the issue and related policy preferences to be analysed).

• A second block on Policy would cover the processes through which people assess the means that governments use to address the issue and the attitudes that result from these processes (the How? question, which serves to define the policy measures and responses to be considered).

• A third block on Society would cover the processes through which people form expectations of each other in the context of welfare and redistribution (the Who? question), as well as the general vision of society which they rely upon to frame issues relating to inequality (the Why? question). This third block is needed to account for the specific “social contractual” dimension associated with inequality and the policies designed to address it. It also brings into the analysis other structural elements or trends that may be influencing people’s attitudes towards others and through them their beliefs about inequality, their attitudes towards policies and government as well as their policy preferences. These structural elements and trends may include, for instance, changes in group dynamics or in levels of interpersonal trust and the effects of increased resentment along ethnic-racial lines.

The framework must also consider how to account for the social and psychological processes involved in the formation of policy preferences. In order to do so, a number of difficulties must be addressed. First, introducing greater complexity into the model raises the issue of how and to what extent the interactions between the different factors can be adequately described, with a significant risk of endogeneity. Secondly, thought must be given to how the framework can account for the role played by important background or contextual variables (such as group identity, social context or access to information) that contribute to shape individual perceptions, attitudes and preferences relating to inequality.

The strategy adopted in building the framework reflects these structural considerations. It follows two guiding principles:

i. Greater psychological realism in the analysis of the processes through which people form their perceptions, attitudes and preferences relating to inequality: This involves taking perceptions of inequality as a starting point, distinguishing between different types of beliefs and accounting for the main relevant forms of behaviour (including information search, belief formation and belief change). Doing so is necessary in order to advance the research agenda and better address the challenges highlighted in Section 1.1; and

ii. A “thicker” description of the informational and social environment in which people form these perceptions, attitudes and preferences: This is necessary in order to capture important contextual elements that shape the way in which people learn and form their views about inequality and the policies designed to reduce them. Doing so can also shed light on the conditions under which people may change their views.

1.2.3. Description and visualisation of the framework

Based on the guidance of the Expert Group, a conceptual framework was developed for analysing the formation of policy preferences relating to inequality. As described in the previous section, the framework is organised around 3 blocks. These blocks include (i) a key output variable; and (ii) the main explanatory factors that have been identified under each block.

The proposed key output variables are:

• Block 1: Concerns about inequality (measured through surveys covering the different domains of inequality that matter for people: income, wealth, opportunities, between groups…). In addition to surveys and experiments, other methods designed to elicit social preferences relating to inequality include calculating revealed preferences from actual behaviour, notably through labour supply responses to changes in the tax and benefits system using the “inverted optimal tax approach” (Bourguignon and Spadaro, 2012[42]; Bargain et al., 2013[43]; 2014[44]).

• Block 2: Support for policies / policy preferences (measured by levels or ranking, as a proxy for public acceptability). The inequality-reducing policies considered can be distinguished using a policy matrix as done for example in OECD (2023[45]) (see Table 1.1 below). The first axis of the matrix considers the targets of intervention, dividing them into three groups: low-income households, middle-income households and high-income households. The second axis classifies policies in terms of the stage of the economic process at which the interventions take place. On this basis, the relevant types of policies can be distinguished as follows:

  • Pre-production or “pre-distribution”: Policies that influence the initial endowments individuals bring to the marketplace, covering aspects such as education, skills, financial resources, social networks and social capital.

  • Production-stage or “in-market distribution”: Policies that directly influence firms' decisions relating to employment, investment and innovation, and thereby shape the economic landscape from within.

  • Post-production or "redistribution": Policies that are implemented ex-post, involving the transfer of income and wealth after they have already been generated.

• Block 3: Societal preferences regarding inequality. Several key factors relating to the perceptions and attitudes people have towards others, society as a whole and the social contract should be covered here. This may notably include collective attitudes towards risk sharing, people’s conceptions of the social contract (measured qualitatively) and country-specific responses to social dilemmas (captured through behavioural measures or proxies).

Table 1.1. A matrix to classify inequality-reducing policies

	Pre-production	Production	Post-production (redistribution)
Low-income	Improve equal access to education	Increase (or introduce) minimum wages Support low-income areas through investments and subsidies for firms	Expand (or introduce) social benefits and/or services for low-income households
Middle-income	Re-train the unemployed and workers affected by structural change	Introduce or strengthen measures to fight discrimination Strengthen the role of trade unions/collective bargaining and other forms of employee representation	Introduce tax exemptions targeted at mortgage interest, childcare and health expenses Increase the progressivity of the tax system
High-income	Increase (or introduce) taxes on large inheritance, gifts and estates	Protect national businesses from international competition Strengthen anti-trust and competition policies to level the playing field	Increase income taxes on top earners Increase (or introduce) corporate taxes

Note: This matrix is proposed as an analytical tool and example of a structured framework for classifying inequality-reducing policies and understanding how they may be perceived. The policies included in the matrix are based on the options provided to respondents in the Opportunities module of the 2022 Risks that Matter survey and should not be taken as recommended policies or as a reflection on the economic efficiency and effectiveness of these policies in reducing inequalities (“protecting national businesses from international competition” may for example result in higher prices for consumers and lower productivity growth if competition is reduced, with potential regressive effects on the distribution of income). The classification of policies also presents challenges given the diversity of measures, their scope and the possible interlinkages between them. As such, the classification shown here is for illustrative purposes only and is not meant to be exhaustive or definitive.

Source: OECD (2023[45]), “Working hand in hand? Exploring people’s views of the role of different actors in fighting inequality”, OECD Policy Insights on Well-being, Inclusion and Equal Opportunity, No. 13, OECD Publishing, Paris, https://doi.org/10.1787/dbd54315-en, adapted from Rodrik and Stantcheva (2021[46]), “A Policy Matrix for Inclusive Prosperity”, NBER Working Paper Series No. 28736, National Bureau of Economic Research, https://www.nber.org/papers/w28736.

In parallel, the main contextual variables that contribute to shape the explanatory factors and influence the key output variable in each block can be integrated into the analysis. Several “filters” of this kind can be identified:

• Block 1: The social context may influence the way in which people learn about and experience inequality, thereby shaping their perceptions, reference points and attitudes relating to inequality independently of the question of their access to information. Reference groups notably play a role in determining how individuals perceive their position within the income distribution and what forms of inequality they consider to be “fair” or not (Cruces, Pérez Truglia and Tetaz, 2013[47]; Karadja, Möllerström and Seim, 2017[48]; Hvidberg, Kreiner and Stantcheva, 2023[49]). Similarly, when societies are highly segregated, experience and understanding of inequality may differ widely based on social context, contributing to a greater dispersion of views (Mijs, 2018[2]; Mijs and Roe, 2021[9]). By shedding light on the contextual conditions of belief formation and belief change, analysis of the social context can help improve communication on inequality by selecting tools and methods that are more likely to ensure people integrate reliable and generalisable information into their views.

• Block 2: Social identity may contribute to shape policy preferences relating to inequality, due to the fact that (i) people care about “horizontal” forms of inequality between groups as well as “vertical” forms between individuals; and (ii) groups and social identity have a mediating effect on the way in which individuals perceive their own interest and goals (Klor and Shayo, 2010[50]; Costa-Font and Cowell, 2015[51]). Mapping clusters of attitudes and beliefs and exploring the degree to which they overlap or not with socio-demographic characteristics can help shed light on relevant forms of social identity. Connecting social identity, measured in this way, with support for policies may help better identify key interest groups, particularly in a context where socio-economic characteristics do not fully explain the variance in individuals’ beliefs and preferences. Understanding to what extent support for policies is tied to different groups and what are the key characteristics of these groups is likely to provide useful insights for designing reform strategies that elicit public support.

• Block 3: The institutional environment plays an important role in explaining societal preferences regarding inequality and the expectations people have of policies. Institutions both reflect and shape beliefs about society and collective attitudes towards risk sharing (Esping-Andersen, 1990[52]; Alesina, Glaeser and Sacerdote, 2001[53]). They also frame public debates and available policy options. “Informal” institutions, such as conventions and social norms, also need to be included here as they contribute to anchor the mutual expectations people have of each other. Analysing the relation between changes in the relevant policy settings, the evolution of social norms and collective attitudes towards risk sharing can help better understand (i) whether policies designed to reduce inequality are well aligned with societal preferences, both in terms of the scope and type of policies used; and (ii) how these views and expectations evolve over time, including in response to changes in policy settings.

A visual representation of the conceptual framework is presented in Figure 1.8.

Figure 1.8. A conceptual framework for analysing the formation of policy preferences in the area of inequality – Blocks, key factors and contextual variables

Source: OECD Secretariat.

1.3. How can the conceptual framework be operationalised? An agenda for measuring the public acceptability of inequality-reducing policies

Several steps are needed to operationalise the conceptual framework described above. This includes (i) identifying a list of appropriate indicators to “populate” the different blocks; (ii) establishing the key factors for measurement; and (iii) defining a “cost-effective” strategy for doing so. To this end, Section 1.3.1 below sets out priority areas for measurement, as suggested by the Expert Group. Section 1.3.2 presents some of the recommendations made by the Expert Group on how to measure key factors effectively. Taken together, these two sections outline a possible agenda for operationalising the framework and measuring the public acceptability of policies designed to reduce inequalities.

In addition, a short review of the existing data landscape was conducted. The review details the types of indicators available and generally used to measure the different factors included in the framework. It also maps these indicators with relevant OECD data collection instruments (Risks that Matter survey, Compare Your Income webtool, Drivers of Trust in Public Institutions survey, pilot Survey on International Attitudes towards Climate Policies…). The results from this review can be found in Annex 1.B. A preliminary assessment of the quality and reliability of these indicators is provided in Annex 1.C.

1.3.1. Identifying the main factors to measure

The conceptual framework provides a model of the process through which policy preferences relating to inequality are formed and a way to organise the increasingly rich and detailed picture that can be drawn from perceptual and behavioural data. While comparable data may not be easily available for all of the subjective determinants of public acceptability covered in the framework, it is possible to identify key factors for the analysis that should be measured in priority. Based on the input received from the Expert Group, priority variables to measure include, but are not limited to:

• Fairness principles and beliefs: This covers the general principles through which people assess whether situations and policies are fair or not, as well as the associated beliefs on the extent to which “what is” differs from “what ought to be”. Fairness principles apply both to existing policies and to new or proposed policies. An important point to note is that the relevant principles differ across policy areas. As described in the previous section, particular attention should be given to the distinction between “meritocratic” and “structuralist” beliefs, as well as to reciprocity beliefs (see p.1 above).

• Intensity of policy preferences: This variable is important to measure as it may not be linked in straightforward fashion to the expected or actual material effects of a policy on an individual’s own welfare. In this respect, the politically relevant constituencies for any specific reform (e.g. the groups most likely to mobilise in support of or in opposition to it) may not overlap with a simple definition of “winners” and “losers” in economic terms.

• Perceived stakeholders vs. actual stakeholders: Assessment of the expected distributional impact of policies can help identify key stakeholders by providing information on questions such as: “Who is likely to be affected by a policy?”, “How?” and “Where will the costs, benefits, risks and opportunities be concentrated?”. This should be complemented by analysis of public perceptions on these questions in order to identify possible mismatches between the actual and imagined groups that will be impacted by reform.

• The ethnic-racial dimensions of policy issues relating to inequality: Broader structural elements that contribute to shape attitudes towards inequality also need to be brought into the analysis. In particular, group dynamics and ethnic-racial resentment may influence policy preferences and lead to a “racialisation” of issues linked to redistribution and social protection. Evidence of this can notably be seen in the gaps between perceived and actual reliance on government transfers by native and foreign-born populations, which vary significantly across countries and across demographic groups (see Figure 1.6 above).

• People’s understanding and evaluation of “the state of society”: While properly delivered information can lead people to adjust their existing beliefs, evidence suggests that public engagement with quantitative detail is limited. Conversely, people’s understanding and evaluation of the status quo constitute important “reference points” through which they assess the need for and desirability of policies that will affect the state of society. With regard to levels of inequality, OECD evidence indicates (i) that people tend to disagree less on their “vision for society” (i.e. their goals and what the ideal state of society should look like) than on the “actual state of society” and how far removed it is from that ideal (see Figure 1.9); and (ii) that disagreement over the extent of inequality increasingly reflects differences in perceptions within socio-economic groups (see Section 1.1.1).

• A focus on people’s risk preferences: Evidence suggests (i) that people may often care as much about socio-economic risks as they care about inequality per se; (ii) and that they consider the management of these risks to be a key function of the welfare state. This implies that attention be given to economic insecurity and social protection as policy objectives and levers that are likely to strongly influence the public acceptability of policies, as well as people’s evaluation of and trust in government. Furthermore, a large amount of inequality reduction at a particular time comes as a by-product of the reduction of risk over the life-course.

• Policymakers’ perceptions of public attitudes: Evidence suggests that mismatches between actual public attitudes and policymakers’ perceptions of public attitudes may be large in some countries and should be studied. Beyond the issue of how “representative” policymakers are of average citizens, several structural factors may help explain these mismatches. The nature of political debate may lead policymakers to focus on differences between “visions for society”, while the public may disagree more in terms of how they perceive the “state of society” (see point above). Similarly, policymakers may underestimate the importance people give to procedural fairness and equality of opportunities and overestimate the importance of equality of outcomes (see point above on risk preference). A further relevant question to consider is to what extent growing demand for citizen participation may change the traditional and asymmetric relation between “policymakers” and the public seen as simple “policy-takers”.

Figure 1.9. People disagree more on "what is" than on "what ought to be" in terms of inequality

Note: Users were asked to describe (i) the actual distribution of income in their country; and (ii) the distribution of income they would prefer, by choosing between the 4 diagrams on the bottom right. These diagrams were selected to represent typical forms of income stratification that reflect high inequality (darker colours) and low inequality (lighter colours). This information (High inequality / Low inequality labels) was not shared with users. The figure presents the results for Mexico, Norway and the OECD average for the perceived actual distribution of income (Panel A) and for the preferred distribution (Panel B).

Source: Data from OECD (2020), Compare Your Income webtool, https://www.oecd.org/en/data/tools/compare-your-income.html.

1.3.2. Measuring these factors efficiently

The operational value of the framework depends not only on the feasibility but also on the efficiency of measurement. Attention must therefore be given to the way in which data are collected and measurement is administered. This implies assessing the cost-effectiveness of different available methods for measuring the key factors identified, as well as the quality and relevance of the data generated. Here, the Expert Group provided a number of recommendations designed to help guide data collection and measurement efforts. As noted above, a review of available indicators (Annex 1.B) and a preliminary assessment of their quality and reliability (Annex 1.C) are included at the end of this chapter.

In advancing the agenda for measuring key subjective factors, the following recommendations can help ensure greater efficiency and relevance:

• Where relevant, appropriate survey methods should be used to obtain precise and robust information about individuals’ preferences and their decision-making process. For example, discrete choice experiments can be used to measure the relative importance of a set of options, the valuation of multiple dimensions and the trade-offs between them (OECD, 2018[54]) and quadratic voting methods can be used to measure the intensity of attitudes and preferences (Cavaillé, Chen and Van der Straeten, 2024[55]).

• Bringing administrative data and broader economic data into the analysis provides a better sense of the lived experience of the populations whose perceptions and beliefs are studied.22

• In this perspective, better use can be made of existing survey data by integrating it with:

  • Geo-spatial political-economic data to know more about people’s context and the places where they live;

  • Administrative data, which is easy to access in some countries and constitutes an underexploited resource for research, but raises issues in terms of confidentiality and identifying what information to use;

  • Adding survey questions to economic long-term panel data, as for example with added modules in the European Union Statistics of Income and Living Conditions (EU SILC) and the German Socio-Economic Panel (SOEP). These survey questions can be used to assess the relation between people’s life circumstances (including social mobility and economic insecurity) and their views; and

    • Data from survey experiments, which can be useful in evaluating policies and understanding their causal effects and consequences.

• Consolidating and merging existing datasets should take priority over the creation of new surveys. Doing so will maximise continuity and provide a longitudinal perspective that can help better understand trends and patterns over time.

• Lessons can be learnt from good practices, such as for example efforts conducted under the UK Office of National Statistics’ Inclusive Data Taskforce to improve the collection of data on the public’s lived experience of inequality and inclusion and increase the capacity for intersectional analysis.

• The OECD has a role to play in advancing the agenda on measuring policy-relevant subjective factors, notably by facilitating the harmonisation of data on perceptions and attitudes and their integration with national data.

1.4. Possible next steps for this work

As a possible follow up, the application of the conceptual framework could be supported by the setting up of an OECD Perceptions Database to compile key perceptual and behavioural data from different OECD tools and other main sources. The aim of this Perceptions Database would consist in collecting and harmonising data on key factors identified under the conceptual framework as a resource for (i) the comparative analysis of relevant perceptions, attitudes and policy preferences across OECD countries; and (ii) improving the OECD’s advice on policies designed to reduce inequalities through a more detailed and tailored focus on important factors that contribute to influence the demand for and response to these policies in specific national contexts. A Perceptions Database would complement existing “objective” indicators on inequality collected notably through the OECD’s Income Distribution Database, Wealth Distribution Database and How’s Life? Database. In doing so, it would contribute to increase the value and policy relevance of these indicators by providing a more systematic basis for the type of analysis conducted in (OECD, 2021[1]). A Perceptions Database would also contribute to support the OECD’s broader work on the political economy of reform by helping harness the relevant insights on public acceptability from perceptual and behavioural data (see Chapter 2).

References

[149] Alesina, A. and G. Angeletos (2005), “Fairness and Redistribution”, American Economic Review, Vol. 95/4, pp. 960-980, https://doi.org/10.1257/0002828054825655.

[87] Alesina, A. et al. (2003), “Fractionalization”, Journal of Economic Growth, Vol. 8/2, pp. 155–194, https://doi.org/10.1023/A:1024471506938.

[53] Alesina, A., E. Glaeser and B. Sacerdote (2001), “Why Doesn’t the US Have a European-Style Welfare System?”, NBER Working Paper Series, Vol. No. 8524, https://doi.org/10.3386/w8524.

[114] Alesina, A. and E. La Ferrara (2005), “Preferences for Redistribution in the Land of Opportunities”, Journal of Public Economics, Vol. 89/5-6, pp. 897-931, https://doi.org/10.1016/j.jpubeco.2004.05.009.

[29] Alesina, A., A. Miano and S. Stantcheva (2023), “Immigration and Redistribution”, The Review of Economic Studies, Vol. 90/1, pp. 1-39, https://doi.org/10.1093/restud/rdac011.

[124] Alesina, A., A. Miano and S. Stantcheva (2020), “The Polarization of Reality”, AEA Papers and Proceedings, Vol. 110, pp. 324-328, https://doi.org/10.1257/pandp.20201072.

[106] Alesina, A., E. Murard and H. Rapoport (2021), “Immigration and Preferences for Redistribution in Europe”, Journal of Economic Geography, Vol. 21/6, pp. 925-954, https://doi.org/10.1093/jeg/lbab002.

[76] Alesina, A., S. Stantcheva and E. Teso (2018), “Intergenerational Mobility and Preferences for Redistribution”, American Economic Review, Vol. 108/2, pp. 521-554, https://doi.org/10.1257/aer.20162015.

[36] Algan, Y. and P. Cahuc (2009), “Civic Virtue and Labor Market Institutions”, American Economic Journal: Macroeconomics, Vol. 1/1, pp. 111-145, https://doi.org/10.1257/mac.1.1.111.

[86] Algan, Y., C. Hémet and D. Laitin (2016), “The Social Effects of Ethnic Diversity at the Local Level: A Natural Experiment with Exogenous Residential Allocation”, Journal of Political Economy, Vol. 124/3, pp. 696-733, https://doi.org/10.1086/686010.

[145] Almås, I. et al. (2017), “Fairness and Family Background”, Politics, Philosophy & Economics, Vol. 16/2, pp. 117-131, https://doi.org/10.1177/1470594X15618966.

[5] Almås, I. et al. (2024), “Attitudes to Inequality: Preferences and beliefs”, Oxford Open Economics, Volume 3, Issue , 2024, Pages, Vol. 3/Supplement_1, pp. i64–i79, https://doi.org/10.1093/ooec/odae001.

[4] Almås, I. et al. (2010), “Fairness and the Development of Inequality Acceptance”, Science, Vol. 328/5982, pp. 1176-1178, https://doi.org/10.1126/science.1187300.

[109] Arikan, G. and P. Ben-Nun Bloom (2015), “Social Values and Cross‐National Differences in Attitudes towards Welfare”, Political Studies, Vol. 63/2, pp. 431-448, https://doi.org/10.1111/1467-9248.12100.

[111] Balafoutas, L. et al. (2013), “Equality, Equity and Incentives: An experiment”, European Economic Review, Vol. 60/C, pp. 32-51, https://doi.org/10.1016/j.euroecorev.2013.01.005.

[44] Bargain, O. et al. (2014), “Tax-Benefit Revealed Social Preferences in Europe and the US”, Annals of Economics and Statistics, Vol. 113/114/Special Issue on the Economics of Taxation, pp. 257-289, https://doi.org/10.15609/annaeconstat2009.113-114.257.

[43] Bargain, O. et al. (2013), “Comparing Inequality Aversion across Countries when Labor Supply Responses Differ”, International Tax and Public Finance, Vol. 21, pp. 845-873, https://doi.org/10.1007/s10797-013-9277-9.

[98] Becker, B. (2021), “Temporal Change in Inequality Perceptions and Effects on Political Attitudes”, Political Research Exchange, Vol. 3/1, https://doi.org/10.1080/2474736X.2020.1860652.

[112] Benabou, R. and E. Ok (2001), “Social Mobility and the Demand for Redistribution: The Poum hypothesis”, The Quarterly Journal of Economics, Vol. 116/2, pp. 447-487, https://doi.org/10.1162/00335530151144078.

[119] Benson, R. et al. (2021), Attitudes to Inequalities, Institute for Fiscal Studies, https://ifs.org.uk/inequality/attitudes-to-inequalities/.

[78] Bolderdijk, J. et al. (2017), “Understanding Effectiveness Skepticism”, Journal of Public Policy & Marketing, Vol. 36/2, pp. 348–361, https://doi.org/10.1509/JPPM.16.118.

[31] Bordalo, P. et al. (2016), “Stereotypes”, The Quarterly Journal of Economics, Vol. 131/4, pp. 1753-1794, https://doi.org/10.1093/qje/qjw029.

[42] Bourguignon, F. and A. Spadaro (2012), “Tax-Benefit Revealed Social Preferences”, Journal of Economic Inequality, Vol. 10/1, pp. 75-108, https://doi.org/10.1007/s10888-010-9153-0.

[79] Brezzi, M. et al. (2021), “An updated OECD framework on drivers of trust in public institutions to meet current and future challenges”, OECD Working Papers on Public Governance, No. 48, OECD Publishing, Paris, https://doi.org/10.1787/b6c5478c-en.

[74] Bucca, M. (2016), “Merit and Blame in Unequal Societies: Explaining Latin Americans’ beliefs about wealth and poverty”, Research in Social Stratification and Mobility, Vol. 44, pp. 98–112, https://doi.org/10.1016/J.RSSM.2016.02.005.

[85] Buckler, K. and G. Higgins (2016), “Support for Aggressive Stop-and-Frisk Policy in NYC: Does perceived policy effectiveness and perceived disparate treatment explain observed racial and ethnic divides?”, Journal of Crime and Justice, Vol. 39/2, pp. 303–323, https://doi.org/10.1080/0735648X.2014.952320.

[71] Budria, S., A. Ferrer-i-Carbonell and X. Ramos (2013), “Personality and Dislike for Inequality: Typifying inequality aversion with respect to locus of control”, SSRN Working Papers, https://www.semanticscholar.org/paper/Personality-and-dislike-for-inequality-%3A-Typifying-Budr%C3%ADa/fef925848758726c1945a92c82ca4a059150ef39.

[69] Bullock, H. (2008), Justifying Inequality: A social psychological analysis of beliefs about poverty and the poor, Russell Sage Foundation, https://www.russellsage.org/publications/colors-poverty-1.

[146] Bussolo, M. et al. (2019), Toward a New Social Contract: Taking on distributional tensions in Europe and Central Asia, World Bank, https://doi.org/10.1596/978-1-4648-1353-5.

[138] Capozza, F. et al. (2021), “Studying Information Acquisition in the Field: A practical guide and review”, Center for Economic Behavior and Inequality Working Paper Series, Vol. No. 15/21, https://doi.org/10.2139/ssrn.3952561.

[150] Cassio, L. (2024), Delivering on Equality and Inclusion, Publications Office of the European Union, https://doi.org/10.2760/150711.

[20] Cavaillé, C. (2023), Fair Enough? Support for redistribution in the age of inequality, Cambridge University Press, https://doi.org/10.1017/9781009366038.

[55] Cavaillé, C., D. Chen and K. Van der Straeten (2024), “Who Cares? Measuring differences in preference intensity”, Political Science Research and Methods, pp. 1-17, https://doi.org/10.1017/psrm.2024.27.

[22] Cavaillé, C. and K. Trump (2015), “The Two Facets of Social Policy Preferences”, The Journal of Politics, Vol. 77/1, pp. 146-160, https://doi.org/10.1086/678312.

[39] Cetre, S. et al. (2020), “Ethnic bias, economic success and trust: Findings from large sample experiments in Germany and the United States through the Trustlab platform”, OECD Statistics Working Papers, No. 2020/04, OECD Publishing, Paris, https://doi.org/10.1787/f6d0b7dd-en.

[95] Choi, G. (2019), “Revisiting the Redistribution Hypothesis with Perceived Inequality and Redistributive Preferences”, European Journal of Political Economy, Vol. 58, pp. 220-244, https://doi.org/10.1016/j.ejpoleco.2018.12.004.

[152] Ciani, ., L. Fréget and T. Manfredi (2021), “Learning about inequality and demand for redistribution: A meta-analysis of in-survey informational experiments”, OECD Papers on Well-being and Inequalities, No. 02, OECD Publishing, Paris, https://doi.org/10.1787/8876ec48-en.

[18] Cohn, S. (2016), “Reconceptualising Public Acceptability: A study of the ways people respond to policies aimed to reduce alcohol consumption”, Health (London), Vol. 20/3, pp. 203-219, https://doi.org/10.1177/1363459315574117.

[11] Colantone, I. et al. (2024), “The Political Consequences of Green Policies: Evidence from Italy”, American Political Science Review, Vol. 118/1, pp. 108-126, https://doi.org/10.1017/S0003055423000308.

[51] Costa-Font, J. and F. Cowell (2015), “Social Identity and Redistributive Preferences: A survey”, Journal of Economic Surveys, Vol. 29/2, pp. 357-374, https://doi.org/10.1111/joes.12061.

[47] Cruces, G., R. Pérez Truglia and M. Tetaz (2013), “Biased Perceptions of Income Distribution: Evidence from a survey experiment”, Journal of Public Economics, Vol. 98, pp. 100-112, https://doi.org/10.1016/j.jpubeco.2012.10.009.

[128] D’Arcangelo, F. et al. (2022), “A framework to decarbonise the economy”, OECD Economic Policy Papers, No. 31, OECD Publishing, Paris, https://doi.org/10.1787/4e4d973d-en.

[110] Daniele, G. and B. Geys (2015), “Interpersonal Trust and Welfare State Support”, European Journal of Political Economy, Vol. 39, pp. 1-12, https://doi.org/10.1016/j.ejpoleco.2015.03.005.

[121] Davidai, S. (2022), “How Do People Make Sense of Wealth and Poverty?”, Current Opinion in Psychology, Vol. 43, pp. 42-47, https://doi.org/10.1016/j.copsyc.2021.06.010.

[77] de Groot, J. and G. Schuitema (2012), “How to make the Unpopular Popular? Policy characteristics, social norms and the acceptability of environmental policies”, Environmental Science & Policy, Vol. 19/20, pp. 100-107, https://doi.org/10.1016/j.envsci.2012.03.004.

[3] Dechezleprêtre, A. et al. (2022), “Fighting Climate Change: International attitudes towards climate policies”, OECD Economics Department Working Papers, Vol. No. 1714, https://doi.org/10.1787/3406f29a-en.

[113] Desmet, K., I. Ortuño-Ortín and R. Wacziarg (2017), “Culture, Ethnicity, and Diversity”, American Economic Review, Vol. 107/9, pp. 2479-2513, https://doi.org/10.1257/aer.20150243.

[6] Dickinson, J., P. Leman and M. Easterbrook (2023), “Children’s Developing Understanding of Economic Inequality and their Place within it”, British Journal of Developmental Psychology, Vol. 41/2, pp. 81-98, https://doi.org/10.1111/bjdp.12446.

[72] Dietze, P. and M. Craig (2021), “Framing Economic Inequality and Policy as Group Disadvantages (versus Group Advantages) Spurs Support for Action”, Nature Human Behaviour, Vol. 5, pp. 349–360, https://doi.org/10.1038/s41562-020-00988-4.

[23] Dimick, M., D. Rueda and D. Stegmueller (2016), “The Altruistic Rich? Inequality and other-regarding preferences for redistribution”, Quarterly Journal of Political Science, Vol. 11/4, pp. 385-439, https://doi.org/10.2139/ssrn.2453703.

[126] Douenne, T. and A. Fabre (2022), “Yellow Vests, Pessimistic Beliefs, and Carbon Tax Aversion”, American Economic Journal: Economic Policy, Vol. 14/1, pp. 81-110, https://doi.org/10.1257/pol.20200092.

[10] Durante, R., L. Putterman and J. van der Weele (2014), “Preferences for Redistribution and Perception of Fairness: An experimental study”, Journal of the European Economic Association, Vol. 12/4, pp. 1059-1086, https://doi.org/10.1111/jeea.12082.

[120] Easterbrook, M. (2021), “The Social Psychology of Economic Inequality”, UNU-WIDER Working Papers, Vol. 2021/43, https://www.wider.unu.edu/sites/default/files/Publications/Working-paper/PDF/wp2021-43-social-psychology-economic-inequality.pdf.

[34] Ellingsen, T. et al. (2012), “Social Framing Effects: Preferences or beliefs?”, Games and Economic Behavior, Vol. 76/1, pp. 117-130, https://doi.org/10.1016/j.geb.2012.05.007.

[52] Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Princeton University Press, https://press.princeton.edu/books/paperback/9780691028576/the-three-worlds-of-welfare-capitalism.

[19] European Commission: Directorate-General for Employment (2019), Employment and Social Developments in Europe 2019: Sustainable growth for all – Choices for the future of Social Europe, Publications Office, https://doi.org/10.2767/305832.

[21] European Commission: Directorate-General for Employment, Social Affairs and Inclusion (2020), Employment and Social Developments in Europe 2020: Leaving no-one behind and striving for more – Fairness and solidarity in the European social market economy, Publications Office, https://data.europa.eu/doi/10.2767/478772.

[83] Fitzgerald, J. and J. Wolak (2016), “The Roots of Trust in Local Government in Western Europe”, International Political Science Review, Vol. 37/1, pp. 130–146, https://doi.org/10.1177/0192512114545119.

[25] Fong, C. (2001), “Social preferences, Self-Interest, and the Demand for Redistribution”, Journal of Public Economics, Vol. 82/2, pp. 225-246, https://doi.org/10.1016/S0047-2727(00)00141-9.

[26] Fong, C. and P. Poutvaara (2019), “Redistributive Politics with Target-Specific Beliefs”, ifo Working Paper Series, Vol. No. 297, https://www.ifo.de/DocDL/wp-2019-297-fong-poutvaara-redistributive-politcs.pdf.

[67] García-Sánchez, E. et al. (2020), “Attitudes towards Redistribution and the Interplay between Perceptions and Beliefs about Inequality”, British Journal of Social Psychology, Vol. 59/1, pp. 111-136, https://doi.org/10.1111/bjso.12326.

[136] Gauthier, D. (1986), Morals by Agreement, Clarendon Press.

[144] Gigerenzer, G. and R. Selten (eds.) (2001), Fast and Frugal Heuristics for Environmentally Bounded Minds, MIT Press, https://doi.org/10.7551/mitpress/1654.003.0006.

[100] Hacker, J. et al. (2014), “The Economic Security Index: A new measure for research and policy analysis”, The Review of Income and Wealth, Vol. 60/S1, pp. S5-S32, https://doi.org/10.1111/roiw.12053.

[28] Hedegaard, T. and C. Larsen (2019), “Coming to Europe: American exceptionalism and American migrants’ adaption to comprehensive welfare states”, International Journal of Sociology, Vol. 49/2, pp. 130-147, https://doi.org/10.1080/00207659.2019.1582965.

[73] Héricourt, J. and G. Spielvogel (2013), “Beliefs, Media Exposure and Policy Preferences on Immigration: Evidence from Europe”, Applied Economics, Vol. 46/2, pp. 225–239, https://doi.org/10.1080/00036846.2013.844330.

[151] Hobbes, T. (1974 [1650]), Human Nature and De Corpore Politico, Oxford University Press.

[104] Hooghe, M. and S. Marien (2013), “A Comparative Analysis of the Relation between Political Trust and Forms of Political Participation in Europe”, European Societies, Vol. 15/1, pp. 131-152, https://doi.org/10.1080/14616696.2012.692807.

[116] Howarth, D. et al. (2019), “What Do the British Public Think of Inequality in Health, Wealth and Power?”, Social Science & Medicine, Vol. 222, pp. 198-206, https://doi.org/10.1016/j.socscimed.2019.01.006.

[125] Hoy, C. and F. Mager (2021), “Why Are Relatively Poor People Not More Supportive of Redistribution? Evidence from a Randomized Survey Experiment across Ten Countries”, American Economic Journal: Economic Policy, Vol. 13/4, pp. 299-328, https://doi.org/10.1257/pol.20190276.

[130] Hume, D. (1985 [1740]), A Treatise of Human Nature: Book 3 On Morals, Penguin Classics.

[49] Hvidberg, K., C. Kreiner and S. Stantcheva (2023), “Social Positions and Fairness Views on Inequality”, The Review of Economic Studies, Vol. 90/6, pp. 3083-3118, https://doi.org/10.1093/restud/rdad019.

[147] IDDRI / Hot or Cool Institute (2024), Inside the Minds of Citizens: Interpretations of today’s social contract in France and the UK, https://www.iddri.org/sites/default/files/PDF/Publications/Catalogue%20Iddri/Rapport/202412-social%20contract%20perceptions.pdf.

[7] Janmaat, J. (2013), “Subjective Inequality: A review of international comparative studies on people’s views about inequality”, European Journal of Sociology, Vol. 54/3, pp. 357-389, https://doi.org/10.1017/S0003975613000209.

[30] Kahneman, D. and A. Tversky (1972), “Subjective Probability: A judgment of representativeness”, Cognitive Psychology, Vol. 3/3, pp. 430-454, https://doi.org/10.1016/0010-0285(72)90016-3.

[17] Kallbekken, S. (2013), “Public Acceptability of Incentive-Based Mechanisms”, Encyclopedia of Energy, Natural Resource, and Environmental Economics, Vol. 3, pp. 306-312, https://doi.org/10.1016/B978-0-12-375067-9.00013-9.

[48] Karadja, M., J. Möllerström and D. Seim (2017), “Richer (and Holier) Than Thou? The effect of relative income improvements on demand for redistribution”, Review of Economics and Statistics, Vol. 99/2, pp. 201-212, https://doi.org/10.1162/REST_a_00623.

[137] Kimbrough, E., V. Smith and B. Wilson (2008), “Historical Property Rights, Sociality, and the Emergence of Impersonal Exchange in Long-Distance Trade”, American Economic Review, Vol. 98/3, pp. 1009-1039, https://doi.org/10.1257/aer.98.3.1009.

[50] Klor, E. and M. Shayo (2010), “Social Identity and Preferences over Redistribution”, Journal of Public Economics, Vol. 94/3/4, pp. 269-278, https://doi.org/10.1016/j.jpubeco.2009.12.003.

[70] Kluegel, J. and E. Smith (1986), Beliefs About Inequality: Americans’ views of what is and what ought to be, Transaction Publishers.

[91] Konow, J. (2001), “Fair and Square: The four sides of distributive justice”, Journal of Economic Behavior & Organization, Vol. 46/2, pp. 137-164, https://doi.org/10.1016/S0167-2681(01)00194-9.

[101] Kopasker, D., C. Montagna and K. Bender (2018), “Economic Insecurity: A socioeconomic determinant of mental health”, SSM Population Health, pp. 184-194, https://doi.org/10.1016/j.ssmph.2018.09.006.

[117] Kraus, M. et al. (2019), “The Misperception of Racial Economic Inequality”, Perspectives on Psychological Science, Vol. 14/6, pp. 899-921, https://doi.org/10.1177/1745691619863049.

[115] Kraus, M., J. Rucker and J. Richeson (2017), “Americans Misperceive Racial Economic Equality”, Proceedings of the National Academy of Sciences, Vol. 114/39, pp. 10324-10331, https://doi.org/10.1073/pnas.1707719114.

[123] Kuziemko, I. et al. (2015), “How Elastic Are Preferences for Redistribution? Evidence from randomized survey experiments”, American Economic Review, Vol. 105/4, pp. 1478-1508, https://doi.org/10.1257/aer.20130360.

[97] Larsen, C. (2016), “How Three Narratives of Modernity Justify Economic Inequality”, Acta Sociologica, Vol. 59/2, pp. 93-111, https://doi.org/10.1177/0001699315622801.

[27] Larsen, C. and T. Dejgaard (2013), “The Institutional Logic of Images of the Poor and Welfare Recipients: A comparative study of British, Swedish and Danish newspapers”, Journal of European Social Policy, Vol. 23/3, pp. 287-299, https://doi.org/10.1177/0958928713480068.

[92] León, F. (2012), “Reciprocity and Public Support for the Redistributive Role of the State”, Journal of European Social Policy, Vol. 22/2, pp. 198–215, https://doi.org/10.1177/0958928711433657.

[135] Lewis, D. (1969), Convention: A philosophical study, Harvard University Press.

[133] MacPherson, C. (1964), The Political Theory of Possessive Individualism: Hobbes to Locke, Oxford University Press.

[80] Maestre-Andrés, S., S. Drews and J. Van Den Bergh (2019), “Perceived Fairness and Public Acceptability of Carbon Pricing: A review of the literature”, Climate Policy, Vol. 19/9, pp. 1186-1204, https://doi.org/10.1080/14693062.2019.1639490.

[105] Marien, S. and M. Hooghe (2011), “Does Political Trust Matter? An empirical investigation into the relation between political trust and support for law compliance”, European Journal of Political Research, Vol. 50/2, pp. 267-291, https://doi.org/10.1111/j.1475-6765.2010.01930.x.

[143] Marino, M., R. Iacono and J. Möllerström (2024), “(Mis-)perceptions, Information, and Political Polarization: A survey and a systematic literature review”, European Journal of Political Economy, Vol. 85, https://doi.org/10.1016/j.ejpoleco.2024.102578.

[142] Mellers, B. and J. Baron (eds.) (1993), Information, Fairness, and Efficiency in Bargaining, Cambridge University Press, https://doi.org/10.1017/CBO9780511552069.009.

[8] Mijs, J. (2023), “Learning about Inequality in Unequal America: How heterogeneity in college shapes students’ beliefs about meritocracy and racial discrimination”, Research in Social Stratification and Mobility, Vol. 85, https://doi.org/10.1016/j.rssm.2023.100814.

[66] Mijs, J. (2021), “The Paradox of Inequality: Income inequality and belief in meritocracy go hand in hand”, The Socio-Economic Review, Vol. 19/1, pp. 7-35, https://doi.org/10.1093/ser/mwy051.

[2] Mijs, J. (2018), “Inequality Is a Problem of Inference: How people solve the social puzzle of unequal outcomes”, Societies, Vol. 8/3, https://doi.org/10.3390/soc8030064.

[122] Mijs, J. (2018), “Visualizing Belief in Meritocracy, 1930-2010”, Socius, Vol. 4/1, https://doi.org/10.1177/2378023118811805.

[41] Mijs, J. (2017), Institutions as Inferential Spaces: How people learn about inequality, Doctoral Dissertation, Harvard University, Graduate School of Arts & Sciences, https://dash.harvard.edu/handle/1/41140253.

[75] Mijs, J. et al. (2022), “Belief in Meritocracy Reexamined: Scrutinizing the role of subjective social mobility”, Social Psychology Quarterly, Vol. 85/2, pp. 131-141, https://doi.org/10.1177/01902725211063818.

[81] Mijs, J., W. de Koster and J. van der Waal (2022), “Belief change in times of crisis: Providing facts about COVID-19-induced inequalities closes the partisan divide but fuels intra-partisan polarization about inequality”, Social Science Research, Vol. 104, https://doi.org/10.1016/j.ssresearch.2021.102692.

[82] Mijs, J., A. Herrera Huang and W. Regan (2023), “Confronting Racism of Omission: Experimental Evidence of the Impact of Information about Ethnic and Racial Inequality in the United States and the Netherlands”, Du Bois Review: Social Science Research on Race, pp. 1-23, https://doi.org/10.1017/S1742058X23000140.

[68] Mijs, J. and C. Hoy (2022), “How Information about Inequality Impacts Belief in Meritocracy: Evidence from a randomized survey experiment in Australia, Indonesia and Mexico”, Social Problems, Vol. 69, pp. 91–122, https://doi.org/10.1093/socpro/spaa059.

[9] Mijs, J. and E. Roe (2021), “Is America Coming Apart? Socioeconomic segregation in neighborhoods, schools, workplaces, and social networks, 1970–2020”, Sociology Compass, Vol. 15/6, https://doi.org/10.1111/soc4.12884.

[13] Murtin, F. et al. (2018), “Trust and its determinants: Evidence from the Trustlab experiment”, OECD Statistics Working Papers, No. 2018/2, OECD Publishing, Paris, https://doi.org/10.1787/869ef2ec-en.

[94] Nguyen, D. et al. (2022), “Survey design and technical documentation supporting the 2021 OECD Survey on Drivers of Trust in Government Institutions”, OECD Working Papers on Public Governance, Vol. No. 53, https://doi.org/10.1787/6f6093c5-en.

[132] Nozick, R. (1974), Anarchy, State and Utopia, Basic Books.

[134] Oakeshott, M. (1975), Hobbes on Civil Association, Blackwell.

[129] OECD (2024), “Accelerating climate adaptation: A framework for assessing and addressing adaptation needs and priorities”, OECD Economic Policy Papers, No. 35, OECD Publishing, Paris, https://doi.org/10.1787/8afaaeb8-en.

[140] OECD (2024), Facts not Fakes: Tackling Disinformation, Strengthening Information Integrity, OECD Publishing, Paris, https://doi.org/10.1787/d909ff7a-en.

[15] OECD (2024), OECD Survey on Drivers of Trust in Public Institutions – 2024 Results: Building Trust in a Complex Policy Environment, OECD Publishing, https://doi.org/10.1787/9a20554b-en.

[61] OECD (2023), “Enhancing Opportunities by Design: Exploring people’s views of what should be done to fight inequality”, OECD Policy Insights on Well-being, Inclusion and Equal Opportunity, Vol. No.14, https://doi.org/10.1787/dbd54315-en.

[60] OECD (2023), “Hard work, privilege or luck? Exploring people’s views of what matters most to get ahead in life”, OECD Policy Insights on Well-being, Inclusion and Equal Opportunity, Vol. No. 12, https://doi.org/10.1787/e1903ab2-en.

[64] OECD (2023), Main Findings from the 2022 OECD Risks that Matter Survey, OECD Publishing, https://doi.org/10.1787/70aea928-en.

[45] OECD (2023), “Working hand in hand? Exploring people’s views of the role of different actors in fighting inequality”, OECD Policy Insights on Well-being, Inclusion and Equal Opportunity, No. 13, OECD Publishing, Paris, https://doi.org/10.1787/dbd54315-en.

[65] OECD (2023), Worries about affording essentials in a high-inflation environment, OECD Publishing, https://www.oecd.org/en/publications/worries-about-affording-essentials-in-a-high-inflation-environment_1ca84132-en.html.

[139] OECD (2022), Building Trust and Reinforcing Democracy: Preparing the Ground for Government Action, OECD Public Governance Reviews, OECD Publishing, Paris, https://doi.org/10.1787/76972a4a-en.

[1] OECD (2021), Does Inequality Matter? How People Perceive Economic Disparities and Social Mobility, OECD Publishing, https://doi.org/10.1787/3023ed40-en.

[63] OECD (2021), Main Findings from the 2020 Risks that Matter Survey, OECD Publishing, Paris, https://doi.org/10.1787/b9e85cf5-en.

[62] OECD (2019), Main Findings from the 2018 Risks that Matter Survey, OECD Publishing, https://doi.org/10.1787/9266e48a-en.

[156] OECD (2019), Under Pressure: The Squeezed Middle Class, OECD Publishing, https://doi.org/10.1787/689afed1-en.

[59] OECD (2018), A Broken Social Elevator? How to Promote Social Mobility, OECD Publishing, https://doi.org/10.1787/9789264301085-en.

[54] OECD (2018), Cost-Benefit Analysis and the Environment: Further Developments and Policy Use, OECD Publishing, Paris, https://doi.org/10.1787/9789264085169-en.

[56] OECD (2018), Opportunities for All: A Framework for Policy Action on Inclusive Growth, OECD Publishing, https://doi.org/10.1787/9789264301665-en.

[57] OECD (2018), The Productivity-Inclusiveness Nexus, OECD Publishing, https://doi.org/10.1787/9789264292932-en.

[14] OECD (2017), Trust and Public Policy: How Better Governance Can Help Rebuild Public Trust, OECD Public Governance Reviews, OECD Publishing, Paris, https://doi.org/10.1787/9789264268920-en.

[58] OECD (2015), In It Together: Why Less Inequality Benefits All, OECD Publishing, https://doi.org/10.1787/9789264235120-en.

[155] OECD (2011), Divided We Stand: Why Inequality Keeps Rising, OECD Publishing, https://doi.org/10.1787/9789264119536-en.

[154] OECD (2008), Growing Unequal?: Income Distribution and Poverty in OECD Countries, OECD Publishing, https://doi.org/10.1787/9789264044197-en.

[153] OECD et al. (2021), Latin American Economic Outlook 2021: Working Together for a Better Recovery, OECD Publishing, https://doi.org/10.1787/5fedabe5-en.

[33] Ostrom, E. (2000), “Collective Action and the Evolution of Social Norms”, Journal of Economic Perspectives, Vol. 14/3, pp. 137-158, https://doi.org/10.1257/jep.14.3.137.

[148] Piketty, T. (1995), “Social Mobility and Redistributive Politics”, The Quarterly Journal of Economics, Vol. 110/3, pp. 551-584, https://doi.org/10.2307/2946692.

[89] Pitlik, H. and M. Rode (2021), “Radical Distrust: Are economic policy attitudes tempered by social trust?”, Social Indicators Research, Vol. 158/2, pp. 485–506, https://doi.org/10.1007/s11205-020-02317-8.

[40] Pontusson, J. et al. (2020), “Introducing the Inequality and Politics Survey: Preliminary findings”, Unequal Democracies, No. 16, University of Geneva, https://archive-ouverte.unige.ch/unige:135683.

[131] Rawls, J. (1971), A Theory of Justice, Harvard University Press, https://doi.org/10.2307/j.ctvjf9z6v.

[88] Reardon, S. and G. Firebaugh (2002), “Measures of Multigroup Segregation”, Sociological Methodology, Vol. 32/1, pp. 33–67, https://doi.org/10.1111/1467-9531.00110.

[90] Reeskens, T. and M. Hooghe (2008), “Cross-Cultural Measurement Equivalence of Generalized Trust: Evidence from the European Social Survey (2002 and 2004).”, Social Indicators Research, Vol. 85/3, pp. 515-532, https://doi.org/10.1007/s11205-007-9100-z.

[46] Rodrik, D. and S. Stantcheva (2021), “A Policy Matrix for Inclusive Prosperity”, NBER Working Paper Series, Vol. No. 28736, https://doi.org/10.3386/w28736.

[37] Rothstein, B. (2001), “The Universal Welfare State as a Social Dilemma”, Rationality and Society, Vol. 13/2, pp. 213-233, https://doi.org/10.1177/104346301013002004.

[103] Rudolph, T. and J. Evans (2005), “Political Trust, Ideology, and Public Support for Government Spending”, American Journal of Political Science, Vol. 49/3, pp. 660-671, https://doi.org/10.1111/j.1540-5907.2005.00148.x.

[84] Schäfer, K. et al. (2019), “Survey-based Personas for a Target-group-specific Consideration of Elderly End-users of Information and Communication Systems in the German Health-care Sector”, International Journal of Medical Informatics, Vol. 132, https://doi.org/10.1016/j.ijmedinf.2019.07.003.

[141] Schelling, T. (1960), The Strategy of Conflict, Harvard University Press.

[99] Schnaudt, C., C. Hahn and E. Heppner (2021), “Distributive and Procedural Justice and Political Trust in Europe”, Frontiers in Political Science, Vol. 3, pp. 1-18, https://doi.org/10.3389/fpos.2021.642232.

[127] Stantcheva, S. (2021), “Understanding Tax Policy: How do people reason?”, The Quarterly Journal of Economics, Vol. 136/4, pp. 2309-2369, https://doi.org/10.1093/qje/qjab033.

[107] Stolle, D., S. Soroka and R. Johnston (2008), “When Does Diversity Erode Trust? Neighborhood diversity, interpersonal trust and the mediating effect of social interactions”, Political Studies, Vol. 56/1, pp. 57-75, https://doi.org/10.1111/j.1467-9248.2007.00717.x.

[32] Sugden, R. (1986), The Economics of Rights, Co-operation, and Welfare, Blackwell, https://doi.org/10.1057/9780230536791.

[108] Sumino, T. (2014), “Does Immigration Erode the Multicultural Welfare State? A cross-national multilevel analysis in 19 OECD member states”, Journal of Ethnic and Migration Studies, Vol. 40/3, pp. 436-455, https://doi.org/10.1080/1369183X.2013.831548.

[118] Trump, K. (2020), “When and Why Is Economic Inequality Seen as Fair?”, Current Opinion in Behavioral Sciences, Vol. 34, pp. 46-51, https://doi.org/10.1016/j.cobeha.2019.12.001.

[38] Trump, K. (2018), “Income Inequality Influences Perceptions of Legitimate Income Differences”, British Journal of Political Science, Vol. 48/4, pp. 929-952, https://doi.org/10.1017/S0007123416000326.

[24] van Oorschot, W. (2000), “Who Should Get What, and Why? On deservingness criteria and the conditionality of solidarity among the public”, Policy and Politics, Vol. 28/1, pp. 33-48, https://doi.org/10.1332/0305573002500811.

[12] Voeten, E. (2024), “The Energy Transition and Support for the Radical Right: Evidence from the Netherlands”, Comparative Political Studies, Vol. 58/2, pp. 394-428, https://doi.org/10.1177/00104140241237468.

[96] Wiwad, D. et al. (2019), “The Support for Economic Inequality Scale: Development and adjudication”, PLOS One, Vol. 14/6, https://doi.org/10.1371/journal.pone.0218685.

[16] Wright, L., A. Steptoe and D. Fancourt (2021), “Predictors of Self-Reported Adherence to COVID-19 Guidelines: A longitudinal observational study of 51,600 UK adults”, Lancet Regional Health - Europe, Vol. 4, https://doi.org/10.1016/j.lanepe.2021.100061.

[102] Wroe, A. (2016), “Economic Insecurity and Political Trust in the United States”, American Politics Research, Vol. 44/1, pp. 131-163, https://doi.org/10.1177/1532673X15597745.

[93] Young, H. (2015), “The Evolution of Social Norms”, Annual Review of Economics, Vol. 7, pp. 259-387, https://doi.org/10.1146/annurev-economics-080614-115322.

[35] Young, H. (2001), Individual Strategy and Social Structure: An evolutionary theory of institutions, Princeton University Press, https://doi.org/10.2307/j.ctv10h9d35.

Annex 1.A. The OECD Observatory on Social Mobility and Equal Opportunity – Advancing the agenda on measurement

Building on the conclusions from Does Inequality Matter?, the OECD is contributing to advance the research agenda on two complementary fronts. On the data collection and analytical side, the OECD Observatory on Social Mobility and Equal Opportunity has extended the work done in (OECD, 2021[1]) by exploring people’s perceptions, attitudes and preferences relating to equality of opportunity and their implications for public policy. In parallel, the OECD Expert Group on New Measures of the Public Acceptability of Reforms has sought to strengthen the conceptual basis of the OECD’s work by improving its understanding of the process through which policy preferences relating to inequality are formed (see Chapter 1).

The OECD Observatory on Social Mobility and Equal Opportunity aims to fill important knowledge gaps and provide a broader picture of how people perceive inequalities and the policies designed to reduce them. Among the main gaps to address, international surveys do not collect data on people’s preferences for policies relating to “opportunities” and pre-distribution to the same extent as they do for policies aimed at “levelling outcomes” and redistribution. Other important areas to develop include the collection of geospatial data and better data on urban/rural differences. As part of its activities, the Observatory seeks to develop new data and indicators on social mobility, equality of opportunity and the key determinants that contribute to them throughout the life-cycle. These determinants include changes in early childhood investment and inequalities in child well-being, the role of inherited wealth and the impact of discrimination. In addition, the Observatory collects relevant data on public perceptions of social mobility and equality of opportunity, as well as on public preferences regarding the measures needed to promote them. Where possible, it will connect this evidence on public perceptions and preferences with conventional statistical indicators.

Broadening the analysis beyond income inequality is crucial for understanding why inequality matters to people and how it can be addressed. As the OECD has highlighted on many occasions, equality of outcomes and opportunities are two sides of the same coin and they need to be thought of and fostered jointly (OECD, 2018[56]; 2018[57]). The connection between the two is particularly clear in the case of social mobility, as present inequality in outcomes contributes to frame future opportunities (OECD, 2015[58]; 2018[59]). Does Inequality Matter? confirmed that the link between outcomes and opportunities is also reflected in public perceptions and contributes to shape attitudes towards policies for reducing inequalities. In this respect, people care about inequality of both outcomes and opportunities, but their policy preferences tend to depend on which type of inequality is perceived as most salient (OECD, 2021[1]). The work being conducted under the auspices of the Observatory constitutes an important step towards developing a richer and more realistic picture of how the public understands inequality and what shapes support for actual policies designed to reduce inequality.

To collect evidence on public perceptions of social mobility and equality of opportunity, a specific module focusing on opportunities was included in the 2022 wave of the OECD’s Risks that Matter cross-national survey (see Box 1.A1 at the end of this Annex). A series of policy briefs on Measuring opportunities: The role of public perceptions highlights the main insights from the data collected through this module. These data shed light on some of the key questions already raised by Does Inequality Matter? regarding preferences for redistribution:

• The first brief provides evidence on public attitudes towards the role played by effort and other causal factors in explaining social mobility (OECD, 2023[60]);

• The second brief looks at public perceptions of the role of different actors in reducing inequalities (OECD, 2023[45]);

• The final brief in the series explores public preferences for policies and actions (OECD, 2023[61]).

Key takeaways from these three briefs include confirmation that people’s concerns about inequality extend beyond economic resources. According to data from the Risks that Matter Opportunities module, inequalities in income and wealth are of primary concern, with 60% of respondents on average declaring that they are too high or far too high. However, inequalities of opportunity and access also matter, with around half of the respondents considering for example that inequalities in children’s educational outcomes or in political power and representation are too high or far too high (see Annex Figure 1.A1).

Annex Figure 1.A1.  Concerns about inequality are not limited to economic resources

Percentage of respondents thinking that inequality is far too high or too high in their country, by domain of inequality, 2022

Source: OECD (2023[45]), Opportunities module of the OECD Risks that Matter Survey 2022, http://oe.cd/rtm.

Similarly, people’s views about the causes of inequality are complex. While hard work is often viewed as very important, it is rarely considered to be the only condition for success. On average, around 60% of respondents believe that hard work is essential or very important in determining one’s chances to get ahead in life. However, among this group, only a small proportion – one-fifth on average – consider that it is the sole factor of success (see Annex Figure 1.A2). Socio-economic background and individual characteristics relating to identity are also perceived as important determinants of success by a large share of respondents. Furthermore, significant divides can be observed between different groups in terms of their beliefs about the causes of inequality. For example, younger respondents and minorities are much more likely to view traits linked to identity as determinants of success.

Annex Figure 1.A2.  Hard work is often seen as an important factor of success, but rarely as the only one

In your country, nowadays, how important to you think each of the following factors is for an individual to get ahead in life? Distribution of responses, by country, 2022

Note: Data from the Opportunities module relate to respondents aged 18-64. Respondents were able to choose from among the following options: “A person’s ethnicity, skin colour or language”; “A person’s religion”; “A person’s sex”; “A person’s sexual orientation or gender identity”; “Coming from a wealthy family”; “Having well-educated parents”; “Being born in the country where you live”; “The neighbourhood, town or territory where a person grew up”; “A person’s health status or disability”; “Working hard”. Countries are ranked in ascending order of the cumulative share of respondents reporting that hard work is very important or essential. The OECD average refers to the weighted average of the 27 OECD countries for which data are available.

Source: OECD (2023[60]), Opportunities module of the OECD Risks that Matter Survey 2022, http://oe.cd/rtm.

Annex Box 1.A1.  The Opportunities module of the 2022 OECD Risks that Matter (RtM) survey

Launched at the 2018 Social Policy Ministerial, the multi-country OECD Risks that Matter survey asks people about their perceptions of economic risks, their satisfaction with social programmes, and their preferences for social protection going forward. Since 2018, RtM has been conducted every two years and now covers over 27 000 respondents in 27 OECD countries. The latest wave was conducted in November 2024. The results and main findings from previous waves are published in (OECD, 2019[62]; 2021[63]; 2023[64]).

The 2022 wave covered the following 27 OECD countries: Austria, Belgium, Canada, Chile, Denmark, Estonia, Finland, France, Germany, Greece, Ireland, Israel, Italy, Korea, Latvia, Lithuania, Mexico, the Netherlands, Norway, Poland, Portugal, Slovenia, Spain, Switzerland, Türkiye, the United Kingdom and the United States.

Most of the questions in RtM have been asked in prior waves to allow reporting on changes in the overall views of the general public. Other questions are part of ad hoc modules that can be replaced to prioritise topical issues or emerging concerns. For instance, the 2020 wave of RtM included questions on individuals’ experiences during the pandemic. In 2022, three ad hoc modules were introduced, focusing on (a) the cost-of-living crisis (OECD, 2023[65]); (b) climate change; and (c) social mobility and equal opportunities (e.g. the Opportunities module). The 2024 wave continues to ask core questions about perceptions of economic risk and social protection. It also includes new modules on the gender wage gap, technological advances in social programme delivery, and social policy preferences related to the climate and digital transitions.

The Opportunities module focuses on four main thematic areas identified as lacking comparable information: (i) the relative importance of different aspects of inequality, (ii) people’s views on the factors shaping equality of opportunity, (iii) people’s views on the role different actors play in reducing inequality and on the type of policies and actions to prioritise, and (iv) experienced and perceived discrimination and support for pro-inclusion policies. Evidence from the Opportunities module will also contribute to the data collection efforts of the OECD Observatory on Social Mobility and Equal Opportunity and to additional projects of the OECD Centre on Well-being, Inclusion, Sustainability and Equal Opportunity (WISE) – e.g. on measuring equity and discrimination.

The 2022 RtM questionnaire was developed by the OECD Secretariat in collaboration with OECD member-country Delegates and stakeholders who participated in a technical workshop in April 2022. It was subsequently translated into national languages. Consistent with similar surveys, RtM is implemented online using non-probability samples recruited via internet and over the phone. The survey contractor is Bilendi Ltd (formerly Respondi Ltd). Respondents are paid a nominal sum of two euros per survey. Sampling is conducted through quotas with sex, age group, education level, income level and employment status (in the last quarter of 2019) used as the sampling criteria. Survey weights are used to correct for any under- or over-representation based on these five criteria. The target and weighted sample is around 1 000 respondents per country. Financial support for the 2022 survey was provided through voluntary contributions by participating OECD member-countries, the OECD Secretariat, Amundi (which sponsored the development of the Opportunities module), Kings College London and the University of Stavanger.

Annex 1.B. Measuring the factors listed in the conceptual framework: The data landscape

This Annex provides a short review of the types of indicators available and generally used to measure the different factors included in the framework and identifies existing gaps. It also maps these indicators with relevant OECD data collection instruments (Risks that Matter survey, Compare Your Income webtool, Drivers of Trust in Public Institutions survey, pilot Survey on International Attitudes towards Climate Policies).

Block 1: Inequality

Beliefs about inequality

There are two main types of measures used in the literature to assess the perceived magnitude, salience and degree of concern about inequality: simple Likert-scale-based measures and elicited estimates from respondents. The following examples can be found in practice:

• Mijs (2021[66]) asks respondents the extent of their agreement with the statement “Income differences [in my country] are too high”, with answers ranging from strongly agree to strongly disagree. This item measures concerns about inequality.

• García-Sánchez et al. (2020[67]) and the International Social Survey Programme (ISSP) in general ask respondent to provide an estimate of the salary of a low-status worker (e.g. a factory worker) and a high-status one (e.g. a CEO). This allows researchers to compute a measure of the magnitude of perceived inequality.

• In similar fashion, Pontusson et al. (2020[40]) and the Inequality & Politics Survey ask respondents to provide their estimates of the income ratio between the 10th and 50th percentiles, and between the 90th and 50th percentiles.

• Other items like the ones featured in the Inequality & Politics Survey ask respondents to assess whether inequalities in educational opportunities have got worse or better over the last 20 years using a Likert-scale.

• Finally, the Inequality & Politics Survey assesses the salience of inequality in the following way: respondents are presented with a series of politically relevant topics, including economic inequality, and asked to rate them from 1 (not important at all) to 5 (extremely important). See Figure 1.7.

Looking at the measures used in OECD surveys and data collection tools:

• The Compare Your Income webtool asks respondents about their perceptions of and preferences regarding the shape of the income distribution, as well as their perceptions of the level of income of poor households.

• The Risks that Matter survey explores people’s perceived degree of job and income mobility, opportunities and economic insecurity through a range of questions designed to assess the socio-economic risks they face and their evolution over time, the degree of financial support they can count on, and the financial situation and prospects of their household as well as the overall economic situation of their country. The Opportunities module included in the 2022 wave of the survey (see Box 1.A1 for more detail) asks respondents to assess the level of inequality in a wider range of areas covering outcomes and opportunities (income; wealth; political power and representation; educational outcomes for children; digital skills and access to digital technology; exposure to extreme environmental events) using a 5-point Likert scale.

• The Survey on International Attitudes toward Climate Policies features the question “How big of an issue do you think income inequality is in [your country]?”

Questions about causal explanations and justifications of inequality either focus on the belief in the importance of specific structural factors, or on a rating or ranking of multiple factors covering individual, structural and cultural processes.

• For instance, the European Social Survey (ESS) and the Inequality & Politics Survey (Pontusson et al., 2020[40]) both ask respondents whether they agree that rising inequalities are a direct by-product of technological change and whether they agree that inequalities promote economic growth. These are both typical justifications used to support tolerance for inequalities.

• The General Social Survey (GSS) and the ISSP include questions where respondents are asked to rate the importance of several factors explaining why people are either poor or (in other questions) “getting ahead in life”. Factors include individually-controlled meritocratic ones (like “talent” and “hard work”) and structural ones (like ethnic and/or gender discrimination) (Mijs and Hoy, 2022[68]). Risks That Matter features questions about people’s perceived degree of economic mobility and economic opportunities. This matters as these perceptions contribute to the perceived fairness of inequalities, as well as to the causal narratives individuals form about inequality.

Looking at OECD surveys and tools:

• The Opportunities module in the 2022 Risks that Matter survey introduces direct questions designed to elicit people’s beliefs regarding the causes of inequality (“How important do you think each of the following factors is for an individual to get ahead in life?”). Interestingly, more specific questions are also targeted at the “rich” and the “poor” (“In your opinion, if a working-age person’s income is low [high], which is most often the reason?”) focusing on “lack of effort” vs. “circumstances beyond one’s control”.

A third important set of questions cover beliefs and attitudes towards poverty, inequality and equality of opportunity:

• Bullock (2008[69]) gives participants a large questionnaire combining multiple scales in order to get a deeper understanding of the factors individuals attribute poverty to, covering individual, structural, “fatalistic” and cultural factors. Moreover, respondents’ attitudes towards inequality are also evaluated through a battery of questions eliciting beliefs on and attitudes towards inequality from Kluegel and Smith (1986[70]). The latter ask respondents for the extent of their agreement with statements such as “Incomes cannot be made more equal since people’s abilities and talents are unequal”. This work is partly related to psychological studies measuring Locus of Control, Perceived Behavioural Control and Self-Efficacy with validated inventories and scales, as in Budria, Ferrer-i-Carbonell and Ramos (2013[71]).

Information

One aspect to consider and measure, at a macro-level rather than at the individual level, is the general media framing of inequality as an issue:

• An important difference noted by Dietze and Craig (2021[72]) is that individuals engage more with media and express greater support for inequality-reducing policies when inequalities are framed in terms of disadvantages for lower socioeconomic status (SES) individuals rather than advantages for higher SES individuals. However, common and generally agreed measures of media framing are still lacking and require extensive manual coding or sophisticated Natural Language Processing (NLP) algorithms applied to Big Data scraped from the web.

• Another missing element is the change over time in the frequency of certain narratives and topics, which could require either manual coding or web scraping and the use of NLP algorithms.

Furthermore, there is a relative paucity of measures of media consumption:

• Traditional ones, as featured in the ESS, ask individuals to choose between categories of time spent on different media (e.g. television, newspapers...) and different topics (distinction between overall time spent and time spent specifically on news and politics-related programs), ranging from “no time at all” to “more than 3 hours” (Héricourt and Spielvogel, 2013[73]).

• A type of information that is usually missing and poorly evaluated in large-scale surveys is the general orientation or slant of the content consumed by individuals: newspapers, social media feeds, and TV channels engage differently with the same topics. However, gathering internationally comparable data of this kind presents significant challenges, as it would probably require at a minimum creating harmonised political coding of content providers from many different countries and for different media.

Looking at OECD surveys and tools:

• The Drivers of Trust in Public Institutions survey includes a question on respondents’ regular sources of information (“From which of the following sources do you get information about politics and current affairs at least once per week”), with different possible sources listed.

Social dimensions

Individuals’ economic position and rank in the income distribution can be elicited and characterised in two broad manners:

• First, researchers can directly ask respondents to give an estimate of their income, whether in the form of a precise number or by selecting the income bracket they belong to. Similarly, respondents can be assigned to groupings by educational and occupational classes, using the CASMIN and ISCED classification schemes based on respondents’ self-reported education and occupation, as in Bucca (2016[74]).

• Another important and complementary approach is to ask respondents where they believe they are situated in the social hierarchy. This second measure tends to be more predictive of individuals’ actual attitudes, decisions, and support for specific policies, as it reflects their subjective lived experience. For an example of this kind of measure, Bucca (2016[74]) and Mijs (2021[66]) use a 1 to 10 discrete scale with the following text: “In our society, there are groups which tend to be towards the top and groups which tend to be towards the bottom. Below is a scale that runs from top to bottom. Where would you put yourself now on this scale?”

• Building on this approach, measures of subjective social mobility can be constructed by comparing individuals’ perception of the position of their family growing up (“social origin”) and perception of their position at present (“social destination”). For example, Mijs et al. (2022[75]) present respondents with the image of a seven rung “social ladder” (“In our society there are groups which tend to be towards the top and groups which tend to be towards the bottom”) and ask them first to place themselves on the ladder and second, thinking back on their childhood, to place the family in which they grew up.

Looking at OECD surveys and tools:

• Compare Your Income features questions about the perceived rank of the individual in the income distribution. This yields information on income position bias across countries (for example, do a majority of respondents in each country place themselves around, above or below the median?). It also asks respondents about their perceived economic opportunities in the future. Finally, the survey features an informational experiment by showing individuals their actual place in the income distribution.

• Trustlab features classic socio-demographic variables, as well as questions about financial security. Specifically, it asks whether respondents believe their financial situation will remain the same, improve or worsen in the coming 12 months.

Another factor to consider when trying to understand an individual’s perceptions of inequality is a respondent’s ethnicity or ethnic identity:

• Alesina, Stantcheva and Teso (2018[76]) find evidence in their survey of perceptions of economic mobility that African-Americans are more optimistic regarding their own prospects. The authors suggest two possible explanations for this finding. First, respondents may exhibit proof of System Justification Theory, meaning that individuals try to reduce cognitive dissonance when faced with an unjust and disadvantageous system. Second, they may compare themselves to members of their own communities. Hence, for individuals in statistically disadvantaged communities, individuals may have lower points of reference when assessing inequality and prospects. Ethnicity is typically measured in international surveys with simple questions where respondents self-assess their ethnicity and choose between multiple options, more or less granular.

Looking at OECD surveys and tools:

• The Opportunities module in the 2022 Risks that Matter survey includes a series of questions relating to discrimination. These questions cover perceived (“Recently, have you witnessed discrimination or harassment?”) and experienced (“Have you ever felt discriminated against or harassed?”) discrimination, as well as the prevalence (“Thinking about your personal experiences over the past year, how often have you felt discriminated against or harassed?”) and effects of discrimination.

Political orientation or political identity is another factor that acts as a potent filter in the information 🡺 perception 🡺 attitude pipeline:

• Alesina, Stantcheva and Teso (2018[76]) suggest that conservative-leaning individuals are on average less confident in governments’ competence and more optimistic about social mobility. The authors ask respondents to position themselves on the political spectrum with regard to economic policy, with answers on a 5-point scale ranging from “very liberal” to “very conservative”. Other surveys, such as the ESS, use a numeric discrete 11-point Likert Scale that is general and not domain-specific.

Looking at OECD surveys and tools:

• The Drivers of Trust in Public Institutions survey asks respondents about political attitudes and participation.¹ It uses 11-point scales to measure political efficacy through questions about individuals’ general level of interest in politics, confidence in their ability to participate in politics, and perception of the degree to which people like them have a say in determining policy. It also includes items designed to measure individuals’ actual political participation: voting, interaction with government, attendance of trade union meetings, signature of petitions… The survey does not include a direct question on political identity but does ask respondents whether the party they voted for at the last national election is currently in government or not.

• Trustlab uses a simple question asking respondents to position themselves on the political spectrum using an 11-point Likert-scale (0-10), ranging from Left to Right.

• The Survey on International Attitudes toward Climate Policies includes a series of questions on interest in politics, political participation and preferences (relating to candidates and parties), and positioning on economic policy with candidates asked to situate themselves on a scale from 1 to 5, where 1 is Left and 5 is Right.

• The Risks that Matter survey asks respondents to position themselves within their national political spectrum (“If a national election were held tomorrow, for which party would you vote?”).

Block 2: Policy

Support for policies

Support for a specific policy can be measured fairly simply, with either a binary question or, better still, a question featuring a Likert-scale allowing individuals to signal the extent of their agreement or disagreement:

• Support for a specific policy can be measured fairly simply, with either a binary question or, better still, a question featuring a Likert-scale allowing individuals to signal the extent of their agreement or disagreement. For instance, de Groot and Schuitema (2012[77]) ask respondents to rate policy using a discrete scale ranging from very unacceptable to very acceptable. It is important to directly measure the support for a policy rather than assume support for a policy if the individual generally agrees with its goal as “translation issues” are frequent.

• Another under-developed element in the literature is the comparison of perceived support by others for certain policies. This would require asking individuals to assess support for given policies by several reference groups and key figures.

Looking at OECD surveys and tools:

• The Risks that Matter survey features general questions designed to elicit respondents’ preferences for redistribution (“Should the government tax the rich more than they currently do in order to support the poor?”) and social protection (“Do you think the government should be doing less, about the same, or more to ensure your economic and social security and well-being?”). The Opportunities module included in the 2022 wave adds a similar question focusing on equality of opportunity, using the same scale.

• Risks that Matter also explores priorities and support for public spending across different areas of social policy (“Would you like to see the government spend less, spend the same, or spend more in each of the following areas?”). Interestingly, the survey seeks to assess willingness to pay: respondents are first asked whether they would want more provision given the costs involved (”bearing in mind the taxes paid and the benefits received”) and then asked the same question with a concrete price-tag (“would you be willing to pay an additional 2% in taxes and social security contributions?”).

Effectiveness of policies

The perceived effectiveness of policies can be measured as follows:

• The most straightforward way is to ask respondents in surveys to give their impression on how effective a policy would be at tackling a specific problem.

• A more complex perspective on the topic is described in Bolderdijk et al. (2017[78]), as the authors argue for a reverse causality where respondents, finding a policy unattractive or unfair, would in turn doubt its potential effectiveness. Their results suggest that “offering optimistic effectiveness estimates” may not bolster support for a policy. Hence, measuring the perceived effectiveness of a policy precisely might be either impossible or yield little information.

Looking at OECD surveys and tools:

• In the Drivers of Trust in Public Institutions survey, the questionnaire explores respondents’ perceptions of policies and public services by asking them to rate their satisfaction on a 11-point Likert-scale. The areas covered include education, healthcare and the quality of administrative services. The survey also asks respondents to rate public institutions on the five dimensions of the OECD Framework on Drivers of Public Trust in Institutions: Integrity, Responsiveness, Reliability, Openness, and Fairness (Brezzi et al., 2021[79]). The questions on Responsiveness, Reliability and Openness cover perceptions relating to the quality, ease of access and effectiveness of both policies and public services. Finally, it includes questions on the government’s perceived ability to address long-term and global challenges (including reducing inequality and discrimination, combatting climate change and managing migration flows).

• The Risks that Matter survey includes questions on satisfaction with the quality of and access to public services (including more specific questions on administrative burden) and on satisfaction with the level of protection provided against different sources of income loss. These are measured using 5-point scales.

• In Trustlab, the experimental setup includes a questionnaire featuring questions about respondents’ degree of satisfaction with multiple public services, including the education system, the health care system, public transports, childcare services, welfare benefits, public housing, security and crime prevention, environmental services and cultural facilities. A question to raise however concerns the endogeneity of these measures: are individuals reporting their satisfaction with the efficiency or perceived legitimacy of the services? Also, one may ask whether respondents include collective, individual or a mix of considerations when answering the questions.

• The Survey on International Attitudes toward Climate Policies assesses respondents’ views on the effectiveness of climate policies by asking them whether they agree or disagree with a series of policy goals and outcomes. It also asks them to rate the availability (ease of access and frequency) of public transportation where they live.

Fairness of policies

Maestre-Andrés, Drews and Van Den Bergh (2019[80]) offer an interesting framework for measuring the perceived fairness of a policy by respondents in their literature review on the evaluation of carbon-pricing schemes. According to the researchers, a policy’s fairness can be assessed through four clusters of measures:

• Perceived distributional impact: Respondents in the different studies are asked to estimate how different publics would be affected, which group would suffer the most or the least (rural vs urban households for instance). Also, some studies elicited opinions from respondents on whether the burden would be fairly shared between companies and households.

• Perceived procedural fairness: This cluster featured general questions about the trust of respondents in government, their level of information on the policies, their trust in the government’s use of the funds raised by the carbon tax, knowledge about the actors involved in designing the policy...

• Perceived impact on the respondents themselves: Individual studies ask respondents to evaluate for instance whether the policies would endanger their jobs, reduce their purchasing power or well-being, limit their choices...

• Perceived ends: This includes questions about a government’s intent and goal when implementing a policy. Respondents in studies are asked whether they think the carbon-pricing schemes are a means to increase tax revenues or mitigate climate change for instance.

Other considerations can be included to augment this framework: in particular, specific questions can be asked of respondents to understand their conceptions of fairness and their preferred redistributive impact. Pontusson et al. (2020[40]) ask respondents to choose for multiple policies, like unemployment benefits, one of three distributional impacts inspired by the types of fairness distinguished by Pontusson et al. (2020[40]): proportional to one’s contribution (equity principle), equal for all (equality principle), targeted to the poorest households (needs principle).

Looking at OECD surveys and tools:

• In the Risks That Matter survey, respondents are asked whether they feel they receive “a fair share of public benefits” given their contributions. This question evaluates the proportionality dimension of fairness, as well as the “pocketbook” considerations that individuals may have. The survey also explores the reciprocity dimension of fairness (by measuring people’s degree of agreement with the statement “Many people receive public benefits without deserving them”) and procedural aspects (through the degree of agreement with the statement “Government incorporates the views of people like me when designing or reforming public benefits and services”).

• The Survey on International Attitudes toward Climate Policies includes a question on respondents’ degree of agreement with the statement that “[Carbon tax with revenue recycling] is fair”. It also assesses respondents’ views on the distributional impact of a policy of carbon tax with revenue recycling by asking them whether they believe that their own household would win or lose financially under such a policy, and also whether different socio-demographic groups (Low-income earners/The middle class/High-income earners/People living in rural areas) would win or lose.

• The Drivers of Trust in Public Institutions survey asks respondents to rate public institutions on the five dimensions of the OECD Framework on Drivers of Public Trust in Institutions (see above). The questions on Integrity, Openness and Fairness cover a number of relevant perceptions relating to equal treatment, transparency, corruption, voice…

Role of the State

Individuals’ perspectives on the extent of the role of the state, and more specifically on whether tackling income inequality and other forms of inequality is part of governments’ responsibilities, tend to be evaluated in a simple and straightforward manner.

• For instance, in the ESS and the Inequality & Politics Survey, respondents are asked to rate the responsibility of multiple entities for economic inequalities, including the State and the European Union, on a scale ranging from 0 (no responsibility) to 10 (full responsibility) (Pontusson et al., 2020[40]).

• Mijs, de Koster and van der Waal (2022[81]) ask respondents if they think that "It is the responsibility of the government to reduce the differences in income between people with high incomes and those with low incomes", while Mijs, Herrera Huang and Regan (2023[82]) ask respondents whether they think "It is the government’s responsibility to combat racial and ethnic discrimination”.

Looking at OECD surveys and tools:

• The Opportunities Module included in the 2022 wave of Risks that Matter asks respondents to rate the role and responsibility of a range of different actors in ensuring economic inequality remains at an acceptable level. These actors include different levels of government, the private sector and financial institutions, civil society organisations and individuals (both wealthy and ordinary). Respondents are then asked to identify priority actions for both government and the private sector (“Which of the following do you think are the most important in order to reduce economic inequality and/or foster equal opportunities in your country?”).

• The Survey on International Attitudes toward Climate Policies asks a general question on the preferred scope for government intervention in terms of the balance between public and private action (“Some people think the government is trying to do too many things that should be left to individuals and businesses. Others think that the government should do more to solve our country’s problems. Which come closer to your own view?”).

• The Drivers of Trust in Public Institutions survey asks respondents about priorities and preferred modes of government action for addressing global challenges (including unilateral and multilateral modes of action, greater international cooperation, engagement with the private sector and civil society…).

Trust in institutions

Trust in government may be too vague or insufficient to capture individuals’ perceptions of political authorities. In response to this limitation, the ESS provides an informative battery of questions to measure trust in institutions at multiple levels: the questionnaire asks respondents to rate their trust, on a scale between 0 and 10, in parliament, the legal system, the police, politicians, the European Parliament, and the United Nations. Other surveys, such as Eurobarometer, sometimes ask similar questions but distinguish between local and national government, which matters given that individuals tend to trust local authorities more (Fitzgerald and Wolak, 2016[83]).

Looking at OECD surveys and tools:

• The Drivers of Trust in Public Institutions survey measures institutional trust using an 11-point Likert-scale. The institutions covered include national government, local government, the national parliament, police, the civil service, news media, the legal system and international organisations. As mentioned above, the survey includes questions aimed at understanding the determinants of trust in institutions from different perspectives: first, it asks them to rate governments and public institutions on the five dimensions of the OECD Framework on Drivers of Public Trust in Institutions: Responsiveness, Reliability, Integrity, Openness, and Fairness. Second, the survey includes questions asking respondents to rate their satisfaction with services and the governments’ efforts to tackle long-term global challenges.

Identity and groups

Again, individual identity tends to interact with many of the metrics cited above. Alesina, Stantcheva and Teso (2018[76]) provide an example illustrating this. Results from their survey suggest for instance that liberals and conservatives differ strongly in their views on government efficiency, favoured policies, preferences for the extent of government intervention...

Another aspect currently lacking in the literature is the heterogeneity in perceived group-specific impacts of policies. As noted earlier, individuals’ perception of a policy’s efficiency is strongly associated with their support for a policy and the fairness they attribute to that policy. Hence, support for a policy seems generally to align with one’s own perceived interest. The development of “representative personas” (e.g. archetypes of the most relevant groups) can constitute a useful tool for taking account of the group-specific impacts of policy at the design stage and for communicating on these impacts (Schäfer et al., 2019[84]).

• Little is known about the trade-offs individuals are willing to make between their own interest and that of other groups. This latter question remains underdeveloped and could provide a link between measures of perceived policy trade-offs (like the one between inequalities and economic growth Pontusson et al. (2020[40]) ask respondents about) and perceived differences in fairness in the way institutions treat groups. For example, Buckler and Higgins (2016[85]) ask respondents whether they believe the police acts fairly towards different ethnic groups, before measuring their support for “Stop & Frisk” policies). An ideal survey could fill this gap with questions measuring the perceived fairness or benefits of a policy for different sub-groups and a final measure of general policy support.

Looking at OECD surveys and tools:

• The Opportunities module in the 2022 wave of Risks that Matter introduces questions relating to the broader theme of discrimination, including support for efforts to combat different grounds of discrimination (“Do you think that efforts made to fight discrimination against the following grounds have gone too far or not far enough?”) and support for specific measures designed to foster diversity and combat discrimination.

Block 3: Society

Characteristics of society

The demographic structure of a society, and in particular its ethnic composition, appear to be meaningful factors in determining policy preferences.

• Several studies for example show a negative relationship between ethnic diversity at a local level and support for redistributive policies (Alesina, Stantcheva and Teso, 2018[76]; Algan, Hémet and Laitin, 2016[86]). An interesting metric to measure this phenomenon is the “fractionalisation” described by Alesina et al. (2003[87]), which can be used for ethnicity, language and religion.

• Other measures, like the Multigroup Theil Index used in Alesina, Stantcheva and Teso (2018[76]), also exist and are described in Reardon and Firebaugh (2002[88]). These metrics are computed using demographic data, often drawn from census sources or household surveys.

Expectations of others

Trust in others is a major predictor of functioning societies:

• Pitlik and Rode (2021[89]) report analyses of large international surveys, suggesting that generalised social trust predicts moderate political attitudes regarding government intervention and redistribution. This generalised measure of social trust is usually a self-reported question, such as the one from the ESS and GSS: “Generally speaking, would you say that most people can be trusted or that you can’t be too careful in life”. Answers include “yes”, “no” and “depends” (Reeskens and Hooghe, 2008[90]). Other variants include more granular Likert-scales.

• However, one needs to be careful regarding the validity of self-reported measures of generalised trust: Murtin et al. (2018[13]) conclude from the Trustlab experiments that “Self-reported measures of trust in others capture a belief about trustworthiness (as well as altruistic preferences), whereas experimental measures [sending money to other participants in the hope of a higher payoff in a trust game] instead capture willingness to cooperate and one’s own trustworthiness. Therefore, both measures are loosely related and should be considered complementary rather than substitutes”. Hence, mixed methodologies (experiments and surveys) might ultimately be more appropriate to measure trust, although more expensive and time-consuming.

Looking at OECD surveys and tools:

• In practice, Trustlab combines questionnaires with economic games, allowing researchers to obtain self-reported as well as behavioural measures for various traits. With regards to trust in others, Trustlab includes multiple measures, both self-reported and experimental/behavioural. Self-reported measures capture individuals’ expectations of cooperation from others, with some of the traditional questions, as well as a new experimental one: “[Participants start with an endowment of 10 Euro]. Imagine you sent 5 Euro, so Participant B receives 15 Euro, making his or her total budget 25 Euro. Participant B has no information about your identity. What amount would you expect Participant B to return to you?” The behavioural measure of trust on the other hand is determined by the amount of money participants transfer to the other players during a trust game. The 2020 experiment investigates heterogeneity in trust by giving participants information about the other players, including their ethnicity as well as their income. Hence, trust games can allow for the observation of differences between general trust and group-specific trust.

• As with institutional trust, the Drivers of Trust in Public Institutions survey includes a simple measure of interpersonal trust using an 11-point Likert-scale and the following question: “On a scale from zero to ten, where zero is not at all and ten is completely, in general how much do you trust most people?” As explained by Murtin et al. (2018[13]), this indicator measures individuals’ expectations vis-à-vis other members of society but may be an inappropriate proxy for behavioural manifestations of trust in others.

Group-specific trust also plays a role in predicting support for redistribution and other policies:

• Cetre et al. (2020[39]) also use Trustlab to measure class and ethnic biases in behavioural trust. Although results suggest the existence of discrimination in cooperative games, and thus lack of trust, towards certain specific minorities, full results also suggest intra-ethnic distrust based on class differences and potentially class resentment. Hence, regardless of ethnicity, wealthier individuals are more trusting towards each other. Such a target-specific measure of trust could also be adapted with self-reported measures, although there would be a high-risk of social desirability bias in the self-reported answers.

Norms of reciprocity (and fairness more generally)

More than just the degree of reciprocity, societies differ in terms of the type and prevalence of social norms. This can be explored for example by looking at support for the different types of fairness principles described by Konow (2001[91]). Pontusson et al. (2020[40]) offer a visually striking illustration of these differences by plotting support for equity over need, against support for equity over equality, averaged by country (see Annex Figure 1.B1).

Annex Figure 1.B1.  Relative support for equity, equality and needs-based fairness, by country

Note: To assess their support for the needs principle relative to the equity principle and support for the equality principle relative to the equity principle, respondents were asked: (i) to indicate the extent of their disagreement or agreement with the statement that “a society is fair when it takes care of those who are poor and in need regardless of what they give back to society,” with response categories ranging from (1) “strongly disagree” to (5) “strongly agree”; and (ii) to place their view of a “fair society” on a scale from 0 to 10, where 0 means that there are no income differences as rewards for individual efforts and 10 means that such differences are large. The figure maps answers to these two questions by country in a two-dimensional space with higher values corresponding to preferences for the equity principle over the equality principle (on the x-axis) and over the needs principle (on the y-axis). Answers to the question about needs have been inverted such that 1 stands for “strongly agree” and 5 stands for ”strongly disagree”.

Source: Pontusson et al. (2020[40]), “Introducing the Inequality and Politics Survey: Preliminary findings”, Unequal Democracies Working Paper N. 16, University of Geneva, https://archive-ouverte.unige.ch/unige:135683.

In the literature, reciprocity and trust are found to be closely linked. When authors are unable to use experimental methodologies, they tend to use self-reported measures of generalised trust as proxies for willingness to reciprocate:

• In the ESS for example, this includes questions on the extent of agreement (on an 11-point Likert-scale) with statements such as “Most people can be trusted”, “Most people would try to take advantage of me”, “Most people would try to be fair” (León, 2012[92]).

• In Trustlab, the Trust games record the amount that players are transferring to other players, as well as the amount sent back by their partners. The amount sent back by partners provides a behavioural measure of reciprocity. A similar assessment of reciprocity can be made experimentally in Trustlab using the amount players put in common during Public Goods games. Moreover, given that participants in the 2020 experiment have information about each other, one can observe whether reciprocity is general or group-specific. With regards to self-reported measures, Trustlab features two different questions in order to measure both positive reciprocity and negative reciprocity (e.g. propensity to punish non-cooperators).

Proportionality or equity tend to be measured in large surveys by asking respondents to choose what society is the fairest, between a society with no income differences (0) and a society with large differences in income (10).

• The survey used in the Program for Applied Research on Climate Action (PARCA Canada) includes a question that seeks to measure mutual expectations and the perceptual aspect of norms by asking “People who are close to me generally expect that I do my part to help limit climate change” with responses on a 5-point scale going from “strongly agree” to “strongly disagree”.

• The emergence, evolution and dynamics of social norms can also be studied. Young (2015[93]), for example, uses experimental measures and field data to test the predictions of evolutionary game-theoretic models in this area.

Similarly to the precedent item, preference for equality is easily measured in large surveys with answers bound on a Likert-scale. Pontusson et al. (2020[40]) measure it by reverse-coding the previous question about the type of society that is considered fairest.

Pontusson et al. (2020[40]) provide an easily replicable example of metric for need that can be used in large surveys. The authors ask respondents the extent of their agreement, between 1 and 5, with the statement “A society is fair when it takes care of those who are poor and in need regardless of what they give back to society”.

Note

1. The questionnaire has been updated for the 2^nd wave of the survey conducted in 2023, with results published in (OECD, 2024[15]). The updated questionnaire was not available at the time this review was conducted. For the purposes of the review, we are referring to the version used in the previous wave of the survey conducted in 2021. Information on the survey design and questionnaire can be found in (Nguyen et al., 2022[94]).

Annex 1.C. Assessing the data landscape: Preliminary review of the quality and reliability of indicators

Annex Table 1.C1. Overview of possible indicators for the key factors identified in the conceptual framework

Concept	Measurement	Type	Precision	Internal Reliability	External Reliability	Source/Reference	Other
Perceived Magnitude of Inequality	Post-computed ratio of self-reported estimated earnings of low and high SES jobs (e.g. workers & CEOs).	Composite indicator created from self-reported measures	Ratios made from items rated on continuous scales -> Higher levels of sensitivity compared to indicators using only Likert-scales	N/A	Correlation with support for redistribution	International Social Survey Programme (ISSP) (Choi, 2019[95])
	Estimated ratios between bottom 10% and 50% percentiles, and between 50% and top 90% percentiles of the income distribution			N/A		Inequality & Politics Survey (Pontusson et al., 2020[40])
Concerns / Tolerance for Inequality	Agreement, measured on a Likert-scale, with the statement “Income differences [in my country] are too high"	Self-reported	Single-item measure -> Potentially high rates of measurement error, low sensitivity and granularity	N/A	Negative association with actual country-level income inequalities	ISSP (Mijs, 2021[66])
	Agreements, measured on a Likert-scale, with 5 statements on economic inequality, including: "Economic inequality is not a problem" and "I am very disturbed by the amount of economic inequality in the world today"	Self-reported	Measure created from answers to multiple questions rated on Likert-scales -> Higher precision & sensitivity	Internal consistency validated -> unit-weighted linear composite (ICC) = 0.94	Correlation with attitudinal (support for redistribution) and behavioural outcomes (donations) + Correlation with similar indicators (e.g. support for inequality indicators from World Values Survey [WVS] & ISSP)	(Wiwad et al., 2019[96])
Justification of Inequality	Choosing amongst multiple factors (e.g. family, hard work etc.) the one that best explains why "the rich are rich" and "the poor are poor”	Self-reported	Low-precision binary measure	N/A	N/A	(Mijs and Hoy, 2022[68])
	Composite measure of procedural justice, from the ratings between 0 and 100 (or on a 1-to-5 Likert scale) of the importance of multiple factors (meritocratic and structural) in explaining economic success	Composite indicator created from self-reported measures	Composite measure with high granularity and sensitivity	Cronbach's Alpha > 0.6	Correlation with tolerance of current levels of economic inequality, call for governmental action, and support for redistribution	ISSP (Larsen, 2016[97]) (Becker, 2021[98])
	Perceived Distributive Justice -> Measured through self-centered and other-centered aggregated measures, based on answers to 4 questions on equality of opportunity, rated on 0-to-10 Likert scales.	Self-reported	Composite measures each based on two 11-points Likert scales for higher precision and sensitivity than single-items	Cronbach's Alpha between 0.5 and 0.6 for the two aggregated measures	Correlation with political and institutional trust	European Social Survey (ESS) (Schnaudt, Hahn and Heppner, 2021[99])
	Subjective probability of intergenerational social mobility -> Individuals are invited to give their subjective probability that individuals in an income quintile will transition to another higher quintile (e.g. in the article, Q1 to Q1, Q1 to Q4, Q1 to Q5)	Self-reported	Single items measured on granular continuous scales, thus with higher sensitivity	N/A	Correlation with support for many conservative and liberal social policies	(Alesina, Stantcheva and Teso, 2018[76])
Economic Position	Subjective social status -> Individuals positioning themselves in the social hierarchy on a scale between 1 and 10 with the question "In our society, there are groups which tend to be towards the top and groups which tend to be towards the bottom. Below is a scale that runs from top to bottom. Where would you put yourself now on this scale?”. A similar question also exists for respondents to situate their families.	Self-reported	Single-item measure and a low-granularity scale, thus with lower precision and sensitivity	N/A	Correlation with tolerance for inequality & belief in meritocracy	ISSP (Larsen, 2016[97]) (Mijs, 2021[66])	The difference between respondents' subjective social status and the subjective social status of families could be used to create a subjective indicator of social mobility
Financial Security	Rating of financial expectations for one's own situation in the next 12 months, using a 3-point Likert scale (worse, the same, better)	Self-reported	Single-item measure and a low-granularity scale, thus with lower precision and sensitivity	N/A	Non-significant, but positive association with trust in government in (Murtin et al., 2018[13])	(Murtin et al., 2018[13])
	The Economic Security Index corresponds to the difference between two time periods of the gross household income minus non-discretionary spending like housing, divided / weighted by the household equivalence scale / cost of living. This indicator can be used to identify households experiencing a shock beyond a given threshold (e.g. a 25% drop) or a more general loss of income.	Composite indicator created from self-reported and economic data	Single-item measure, either binary (yes/no), discretised or continuous. It is based on factual economic data.	N/A	Negatively associated with declining mental health as per the GHQ-12 and political trust	Economic Security Index (Hacker et al., 2014[100]) (Kopasker, Montagna and Bender, 2018[101]) (Wroe, 2016[102])
Political Identity	Measured using Likert scales, usually between 0 and 10	Self-reported	Single-item discrete measures with rather low sensibility	N/A	Correlated with effectiveness of informational treatments, attitudes towards government, views on inequality and support for numerous policies	(Alesina, Stantcheva and Teso, 2018[76])
Perceptions of and Attitudes towards Government	Satisfaction with different government services and institutions, rated on 11-point Likert scales	Self-reported	Single-item discrete measures with rather low sensibility	N/A	Correlated for certain services like security and crime prevention, with self-reported generalised social trust and trust in government	(Murtin et al., 2018[13])
	Self-reported trust in government -> multiple instances of this single-item measure exist: they typically assess trust using a single question, with answers rated on a discrete scale and not necessarily using Likert scaling	Self-reported	Single-item discrete measures with sensitivity ranging from low to very low based on the response range available to respondents	N/A	Evidence that this measure is associated with support for spending in multiple policy domains like defense and welfare	(Rudolph and Evans, 2005[103])
	Composite political trust -> This indicator is a composite of answers to multiple questions asking respondents to rate multiple institutions on discrete Likert scales. The EVS version uses 4 questions for parliament, the justice system, the armed forces & the police. The ESS uses 5 questions for the parliament, politicians, political parties, the legal system and the police.	Composite indicator created from self-reported measures	Measure created from answers to multiple questions rated on Likert-scales -> Higher precision and sensitivity	ESS version: Cronbach's Alpha > 0.8 EVS version: Cronbach's Alpha ~ 0.7	Association with legal permissiveness and political participation	European Value Survey (EVS) and European Social Survey (ESS) (Hooghe and Marien, 2013[104]) (Marien and Hooghe, 2011[105])
Societal Heterogeneity	Racial Segregation through the Multi-Group Theil Index / Fractionalization -> These indicators are derived from census-type data, including the proportion in areas of foreign-born individuals in regions of interest	Measure derived from hard census-type data	Based on continuous percentages and hence granular and sensitive	N/A	Associated with optimistic perceptions of social mobility, decreased social trust, and decreased support for social spending and government intervention for respondents to the centre and right of the political spectrum	(Alesina, Stantcheva and Teso, 2018[76]) (Alesina, Murard and Rapoport, 2021[106]) (Stolle, Soroka and Johnston, 2008[107])	No significant relationship between diversity and support for the welfare state according to alternative method from (Sumino, 2014[108])
Social Trust	Self-reported social trust -> single item question asking respondents how much they trust others, with answers on a Likert-scale, more or less granular. Questions include: "Generally speaking, would you say that most people can be trusted or that you can’t be too careful in life"	Self-reported	N/A	N/A	Associated with support for the welfare state, social spending, and trust in government	(Murtin et al., 2018[13]) (Arikan and Ben-Nun Bloom, 2015[109]) (Daniele and Geys, 2015[110])
	Behavioural trust -> This indicator measures, during trust games, the amount / share of the endowment participants are willing to send their partners in the hopes of a higher payoff	Behavioural measure	Given that the measure is a continuous function of the amount transferred, the measure should be granular, precise and sensitive	N/A	Associated with other-regarding preferences, willingness to cooperate, self-reported trust, institutional and political trust	(Murtin et al., 2018[13])
	Behavioural In-group Bias -> Whether normalised or not, it essentially captures the difference between the average in-group transfer and the average out-group transfer in a series of economic trust games. The measure can be used to consider intersectional groups (ethnicity + income for instance)	Derived from behavioural measures	Given that the measure is a continuous function of the amount transferred, the measure should be granular, precise and sensitive	N/A		(Cetre et al., 2020[39])
Fairness Norms	Preference for equality / Need over equity -> These two self-reported measures are obtained from single-item questions where respondents rate their agreement to statements on Likert scales. The statements are respectively 1°) extent of income differences between individuals, on a scale spanning 0 and 10, and 2°) "A society is fair when it takes care of those who are poor and in need regardless of what they give back to society"	Self-reported measures	These two measures are derived from single item questions. They will thus tend to be less precise and reliable than composite measures. The granularity will depend upon the granularity of the Likert scales.	N/A		Inequality & Politics Survey (Pontusson et al., 2020[40])
	Behavioural preference for equality over equity -> The behavioural preferences of individuals are determined by comparing the amounts transferred by players to the ones they would transfer if they followed an equality-based, equity-based, or needs-based decision function, during economic games with production phases - typically Dictator Games. It is a function of a participant's performance during the production phase, as well as the performances of his/her partners.	Derived from behavioural measure	The measure can either be a continuous coefficient weighting the individual's own performance in the decision function, or a simple binary indicator indicating the closest pure decision function. Depending upon the design choice, the measure is more or less granular, sensitive and precise.	N/A	N/A	(Balafoutas et al., 2013[111])

Notes

← 1.  The OECD’s work on inequality has been underpinned by the collection of data and development of internationally comparable indicators through the OECD Income Distribution Database and Wealth Distribution Database. Building on that evidence base, the OECD has provided analysis on key aspects of inequality and recommendations on how to address the challenges they raise. Most notably, it has done so through a series of landmark reports covering (i) the trends in social mobility as well as in inequality in income and wealth (OECD, 2008[154]; 2011[155]; 2018[59]); (ii) the causes of inequality and the policies that can contribute to promote inclusive growth (OECD, 2015[58]; 2018[56]); and (iii) the consequences of inequality for society as a whole and for the middle class in OECD countries (OECD, 2019[156]). Released in 2021, Does Inequality Matter? added to this series of OECD flagship publications on inequality (OECD, 2021[1]). A follow up report to OECD (2018[59]) is planned for release in 2025, focusing on the level, trends and determinants of inequality of opportunity in OECD countries and on policies to address them.

← 2.  In addition to the perceived level of inequality throughout society as a whole, other relevant perceptions notably include people’s views on their own position within the income distribution (Cruces, Pérez Truglia and Tetaz, 2013[47]; Karadja, Möllerström and Seim, 2017[48]; Hvidberg, Kreiner and Stantcheva, 2023[49]) and on their prospects for social mobility (Piketty, 1995[148]; Benabou and Ok, 2001[112]; Alesina, Stantcheva and Teso, 2018[76]).

← 3.  Dechezleprêtre et al. (2022[3]) observes a relative lack of explanatory power of individual socio-economic characteristics in predicting perceptions of and attitudes towards climate policies. See for example p.8: “It is difficult to predict either beliefs or policy views based on socio-economic and lifestyle characteristics only. Put differently, it is not the case that we are easily able to infer people’s policy views or beliefs based on their age, country, gender, education, income, political leanings or how much they rely on polluting sources of energy”. Similarly, p.29: “Overall, 70% of [policy support is] explained by [key] beliefs and socio-economic and lifestyle characteristics, compared to only 24% explained by individual characteristics alone”. This conclusion is confirmed more broadly by evidence showing that cultural heterogeneity has increased more within groups than between groups (Desmet, Ortuño-Ortín and Wacziarg, 2017[113]).

← 4.  Here again, Dechezleprêtre et al. (2022, p. 25[3]) reaches a similar conclusion: “It is difficult to delineate specific groups for or against climate policies.”

← 5.  For key references, see for example Piketty (1995[148]), Fong (2001[25]), Alesina and Angeletos (2005[149]), Alesina and La Ferrara (2005[114]). See Alesina, Stantcheva and Teso (2018[76]) for an overview of the economic literature.

← 6.  For example, the linear approach to the formation of policy preferences used in Does Inequality Matter? leaves important questions open regarding the way in which the different factors interact. This problem becomes more salient as the scope of the analysis expands and a wider range of perceptions and attitudes are considered. Similarly, the fact that the model is largely premised on individual rational choice makes it well adapted for the analysis of “vertical” forms of inequality, such as income disparities, but also means that it is less well equipped to take account of the “horizontal” aspects of inequality and the role of group identity. On the role of group identity and “group threat” in shaping attitudes towards inequality, see for example a recent report by the European Commission (Cassio, 2024[150]).

← 7.  In identifying the key subjective factors for analysing policy preferences, the experts considered 3 main questions: (i) How do people understand issues relating to inequality?; (ii) How do people assess policies designed to promote inclusion and equality of opportunities?; and (iii) What expectations do people have of each other in the context of welfare and redistribution?

← 8.  For further reference on people’s views about inequality and their subjective determinants, see for example Kluegel and Smith (1986[70]), Janmaat (2013[7]), Kraus, Rucker and Richeson (2017[115]), Howarth et al. (2019[116]), Kraus et al. (2019[117]), Trump (2020[118]), Benson et al. (2021[119]), Easterbrook (2021[120]), Davidai (2022[121]).

← 9.  For evidence confirming this effect, see OECD (2021, pp. 62-67[1]). See also Karadja, Möllerström and Seim (2017[48]) and Mijs (2018[2]; 2021[66]).

← 10.  See Ciani, Fréget and Manfredi (2021[152]) for a meta-analysis of the literature on information treatments in survey experiments on economic inequality. As noted, while information treatments tend to raise awareness and concern about inequality, this does not necessarily translate into increased support for inequality-reducing policies. In this respect, the effects of the information on different types of belief may offset one another or the information may trigger cognitive dissonance reduction mechanisms (motivated reasoning, adaptive preferences…). See for example Kuziemko et al. (2015[123]), Alesina, Stantcheva and Teso (2018[76]), Alesina, Miano and Stantcheva (2020[124]), Hoy and Mager (2021[125]). For similar evidence on carbon taxation, see for example Douenne and Fabre (2022[126]).

← 11.  For example, in contrast to Dechezleprêtre et al. (2022[3]), a similar survey experiment focusing on tax policy finds that informing people about distributional impacts and implied trade-offs between equity and efficiency increases public support for progressive income and estate taxes, but providing information on their efficiency costs has no effect on its own (Stantcheva, 2021[127]).

← 12.  The OECD Framework to Decarbonise the Economy provides an overview of the trade-offs between policy objectives for the main climate policy instruments by characterising them in terms of their cost effectiveness, fiscal implications, distributional impacts and the political economy challenges they are likely to raise (D’Arcangelo et al., 2022[128]). The OECD has also developed a complementary Framework for Assessing and Addressing Adaptation Needs and Priorities to help guide the design and implementation of climate change adaptation policies (OECD, 2024[129]). Taken together, these two frameworks aim to support strategies for addressing climate change that are more effective and more likely to receive public support by ensuring greater integration between mitigation policies, adaptation policies and broader economic policy.

← 13.  On principles of fairness and how they are applied in the context of resource allocation and distributive justice, see for example Konow (2001[91]), Pontusson et al. (2020[40]) and Trump (2020[118]).

← 14.  In this case, fairness evaluations would typically be of the following kind: “How fair is it that a person’s income reflect their luck/hard work/talents?”

← 15.  In this case, fairness evaluations would typically be of the following kind: “How fair is it for some individuals to receive more in social benefits than they contribute?”

← 16.  See for example Hobbes (1974 [1650][151]), Hume (1985 [1740][130]), Rawls (1971[131]) and Nozick (1974[132]). For commentary highlighting this point in Hobbesian thought, see MacPherson (1964[133]) and Oakeshott (1975[134]). For a discussion on how to define and operationalise the notion of social contract in a policy context, see Bussolo et al. (2019[146]) and OECD et al. (2021[153]). For qualitative evidence on citizens’ perceptions of the social contract in France and the United Kingdom, see IDDRI / Hot or Cool Institute (2024[147]).

← 17.  See for example Lewis (1969[135]), Gauthier (1986[136]), Sugden (1986[32]), Young (2001[35]) and Kimbrough, Smith and Wilson (2008[137]).

← 18.  For an overview of the experimental literature on information search in economics, see Capozza et al. (2021[138]).

← 19.  Here, the OECD is supporting efforts to strengthen the integrity of information and implement policies that promote the transparency, accountability and plurality of information sources through its Reinforcing Democracy Initiative. See notably OECD (2022[139]; 2024[140]) and the OECD DIS/MIS Resource Hub.

← 20.  Schelling (1960[141]) offers a classical example by showing how the provision of additional relevant information (on household size, expenditure…) can reduce players’ ability to coordinate on a fair income-proportionate share of a joint surplus in an experimental bargaining game. Camerer and Loewenstein (1993[142]) generalises this point through a study of the relation between information, fairness and efficiency in the context of bargaining. The practical implications of this point are illustrated in Marino, Iacono and Möllerström (2024[143]) which reviews the literature on misperceptions of inequality and finds that the provision of information is more likely to increase political polarisation than decrease it, as different societal groups respond by strengthening (rather than adapting) their existing preferences.

← 21.  On the informational and cognitive limits of human decision-making and how they can be addressed in the context of “bounded rationality” models, see for example (Todd, 2001[144]).

← 22.  Almås et al. (2017[145]), Karadja, Möllerström and Seim (2017[48]) and Hvidberg, Kreiner and Stantcheva (2023[49]) provide notable illustrations of the insights that can be drawn by linking survey data on perceptions of inequality with administrative data on real-life outcomes.