Measuring Democratic Efficacy

doi.org/10.17203/kdk426 Democratic Efficacy and the Varieties of Populism in Europe WORKING PAPER Measuring Democratic Efficacy July 2020 doi.org/10.17203/kdk426 Author(s) Vaidas MORKEVIČIUS, Rimantas RAULECKAS, Giedrius ŽVALIAUSKAS, Eglė BUTKEVIČIENĖ (Kaunas University of Technology), Zsolt BODA, and Márton BENE (Centre for Social Sciences) Contact Information contact@demos.tk.mta.hu This publication has received funding from the European Union’s Horizon 2020 research and innovation programme under the grant agreement No 822590. Any dissemination of results here presented reflects only the authors’ view. The Agency is not responsible for any use that may be made of the information it contains. Copyright MORKEVIČIUS et al. (2019). Abstract This paper studies the state and trends of democratic efficacy in democratic polities around the world. First, it uses data of an on-line survey conducted by DEMOS in 15 European countries to provide descriptive data on democratic efficacy. A first step to test the association of democratic efficacy indicators with populism was also made, showing that most of the items composing democratic efficacy have a negative association with populist attitudes. Second, the paper considers problems of a conceptual definition of democratic efficacy and suggests that it could be analysed following two conceptualisations: 1) more extensive (holistic), coupling political efficacy with citizens’ democratic capacities, and 2) less extensive (narrow), coupling political efficacy only with citizens’ support for important democratic values. Then the paper discusses the availability of measures of democratic efficacy in existing international survey data sets (ESS, ISSP, EVS, CSES, EES, and Eurobarometer). Further, the report empirically assesses the trends of democratic efficacy in a temporal comparative perspective using data from two waves (2004 and 2014) of the ISSP module ‘Citizenship’. Results show that the majority of the populations in the studied countries score low on both internal (a little more than 50 %) and external (a little more than 70 %) political efficacy and, overall, changes over the last decades are negligible. However, levels of political efficacy are quite different across countries and there is substantial crosstime variation for at least half of the studied countries. According to our data, there is a general tendency of lower levels of political efficacy in the CEE countries. Also, in some (mostly Western) countries political efficacy increased from 2004 to 2014. With regard to the support for values of liberal democracy we found that it is moderate (overall, only half the studied populations showed strong support for these values) and rather stable. However, at the country level, we found quite a lot of variation both across countries and over time. These cross-country and cross-time differences need to be further studied with multilevel models including macro (country) level explanatory variables. 2 Copyright MORKEVIČIUS et al. (2019). 1. Introduction The first systematic studies of voting behaviour started in the early 1950s leading to interest in political efficacy mostly due to the fact that it was found to be one of the most important predictors of political participation. Since political participation was always considered an important feature of democratic government, high political efficacy of citizens was consequently deemed an important characteristic of good democratic governance. However, at the beginning of this century many studies of political behaviour found decreasing levels of political participation in the Western1 democracies during the twentieth century (see for a recent example Blais 2010), which was sometimes seen as an indicator of declining quality of democracy. More recently, the decline in political participation seems to have stabilized and in some countries even increasing levels of political participation have been detected. However, increasing numbers of people participating in politics started supporting populist and extremist political actors (for increasing populist voting on both left and right of the political spectrum see https://populismindex.com). Consequently, it has become clear that only the crude level of political participation is an insufficient (or even misleading) indicator of the quality of democratic governance, and democratic quality of political participation should be taken into consideration, too. The same seems to apply to the concept of political efficacy. Political efficacy is an important motivational background for political participation, but it can say nothing about its democratic quality. Therefore, to make the concept fit the contemporary challenges of political science and political realities, it seems reasonable to supplement it with a democratic component. The DEMOS project developed the concept of democratic efficacy for this purpose (see Bene – Boda, 2020). In this paper we use data of an on-line survey conducted by DEMOS in 15 European countries to provide descriptive data on democratic efficacy. We make the first step to test the association of democratic efficacy indicators with populism, showing that most of the items composing democratic efficacy have a negative association with populist attitudes. Then we come back to the problems of conceptual definition of democratic efficacy and propose broader as well as narrower (minimal) definitions of democratic efficacy. Then we discuss availability of measures of democratic efficacy in existing international survey data sets (ESS, ISSP, EVS, CSES, EES, and Eurobarometer). Further, we empirically assess the trends of 1 Importantly, it is even lower in the democratic countries of Central and Eastern Europe. 3 Copyright MORKEVIČIUS et al. (2019). democratic efficacy in a temporal comparative perspective using data from two waves (2004 and 2014) of the ISSP module “Citizenship”. We finalize our paper with conclusions. 2. Defining democratic efficacy The concept of political efficacy is quite well-researched and interest in it dates back to a classical political behaviour study of Campbell, Gurin and Miller (1954). The authors proposed that the “sense of political efficacy may be defined as the feeling that individual political action does have, or can have, an impact upon the political process, i.e., that it is worthwhile to perform one’s civic duties. It is the feeling that political and social change is possible, and that the individual citizen can play a part in bringing about this change” (Campbell et al. 1954: 187). As such political efficacy was hypothesised to strongly determine whether people would participate in the political processes and this emphasis in empirical studies continues to this day (see Valentino, Gregorowicz, Groenendyk 2009; Marx, Nguyen 2016; de Zúñiga, Diehl, Ardévol-Abreu 2017). Lane (1959) already distinguished two dimensions of political efficacy: a person’s image of the self and that of the government. Consequently, internal and external dimensions of political efficacy have been conceptualised and used in political studies (Balch 1974; Craig, Maggiotto 1982; Acock, Clarke, Stewart 1985; Craig, Niemi, Silver 1990; Niemi, Craig, Mattei 1991). Broadly speaking, internal political efficacy refers to citizens’ beliefs that they have competences to understand and effectively participate in politics (subjective competence), and external political efficacy is related to public perceptions of responsiveness of governmental institutions (perceived system responsiveness). Importantly, low political efficacy means that citizens distrust governmental institutions and do not believe that their actions will have an effect on the governance. Therefore, low political efficacy is related to political alienation, which is detrimental to the health of a democratic political system. However, in recent studies measurement and analysis of political efficacy as an important motivational background for political participation brought an important disconcerting insight to light. Namely, any increase in general political participation appeared to be related to populist voting and increasing support for populist attitudes and activism. Therefore, the standard causal logic relating higher levels of political efficacy with growing political participation and, consequently, with increased quality of democratic governance appeared to be tenuous. Thus, to make the concept of political efficacy fit the contemporary challenges of democratic 4 Copyright MORKEVIČIUS et al. (2019). governance, it seems reasonable to supplement it with a democratic component. We propose that democratic participation is enabled by higher level of political efficacy that is paired with certain democratic capacities. Bene and Boda (2020) discerned five groups of democratic capacities and values that have to be paired with political efficacy in order to derive a measure of democratic efficacy:  Factual political knowledge (citizens need to have some general political knowledge, but also keep up with the day-to-day political processes).  Political news consumption (citizens need to develop habits of using the media for political information in a reflective manner).  Political reflexivity (citizens have to be reflective on the political information, however, strong emotional attachments (political identities) may undermine this reflexivity and introduce serious biases in reasoning and evaluation of political information).  Core values of democracy (citizens are required to embrace certain values as prerequisites of democracy in order to make it work properly: respect for political and legal equality (equality of interests), capability of making free and autonomous decisions (political autonomy), tolerance and listening to opinions different from their own (reciprocity).  Political or civic skills (citizens have to be at ease with practising certain behaviours and attitudes: coping with plurality and conflicts in politics and policy; scrutinising leaders and their decisions; being able to express one’s own legitimate needs, aspirations and preferences). Based on this conceptualisation people with complete democratic capacities are those citizens who have (1) a certain level of factual political knowledge; (2) are regular and reflective news consumers; (3) are non-intensive partisans (as a proxy of reflexivity); (4) strongly identify with the core values of democracy, i.e. political and legal equality, tolerance towards dissenting opinion, and individual autonomy, and (5) have some involvement in political activities. In order to measure the levels of democratic efficacy Bene and Boda (2020) turned to survey data available in the major international academically driven surveys. They employed data from the European Social Survey (ESS) well-known for its rigorous cross-cultural design. The eighth round of the ESS conducted in 2016-2017 contained items for the measurement of both political 5 Copyright MORKEVIČIUS et al. (2019). efficacy and democratic capacities with the exception of political knowledge. In the ESS questionnaire political news consumption was measured by the question of how many minutes respondents spend consuming political news on a regular day. They consider regular news consumers those respondents who read, watch or listen to news at least 30 minutes on a regular day. Regarding partisanship, closeness to a political party was measured on a 4-point scale, and only the extreme value indicating ‘very close’ was regarded as a highly partisan answer. Participants were also asked to what extent they identify with certain character types and values on a 6-point scale. Three items of the ESS survey are closely related to the three core values of democracy discussed above. In detail, values of 1 and 2 indicate identification with these statements as they were labelled as ‘very much like me’ and ‘like me’ respectively. Our political activity measure was based on the ESS questionnaire items measuring involvement in different types of political activity. Table 1 shows the share of respondents who meet these criteria for each component of our democratic capacities concept. Table 1. The share of respondents of the eighth round of ESS based on available measures of democratic capacities Criteria Percent News consumption At least 30 min. per day 79.4 Partisanship Equality 1 – 3 (4point scale) 95.2 Tolerance Autonomy 1-2 (6-point scale) 71.5 Source: Bene and Boda (2020) 65.2 Political activity At least 1 68.5 52.0 The findings above suggest that all of these capacities are widely shared in the democratic countries under investigation. The large majority of respondents share the following democratic capacities: they consume news regularly, are non-intensive partisans, and they believe in the values of equality, tolerance and individual autonomy. More than half of them have been involved in at least one political activity during the preceding 12 months. However, Table 2 indicates that their combined presence is not as universal: only a fifth of the respondents have all these democratic capacities, while the large majority of respondents have incomplete capacities. These findings suggest that these capacities are suitable to let us categorise respondents: their validity is supported by their wide presence in democratic countries, but their combined occurrence can differentiate people with greater or lesser democratic capacities. Table 2. The share of respondents of the eighth round of ESS based on the mixture of their democratic capacities People with complete democratic capacities 19.5 % People with incomplete democratic capacities 80.5 % 6 Copyright MORKEVIČIUS et al. (2019). Source: Bene and Boda (2020) 3. Democratic efficacy and populism In 2019 an original on-line survey was undertaken in 15 European countries (Germany, the UK, Czechia, Hungary, Italy, the Netherlands, Poland, France, Slovakia, Lithuania, Denmark, Turkey, Spain, Greece, and Bosnia and Herzegovina). Our desired representative sample size amounted to approximately 500 respondents per country, while quotas based on current census data were set up for gender, age and geographical region. The fundamental Eligibility Criterion for respondents was having lived in their current country of residence for at least 10 years, which we consider a sufficient time frame to feel at home in the country of residence. The questionnaire of the survey was developed by the researchers of the DEMOS project and the survey was administered by the University of Amsterdam. Based on the data of this survey we make a first attempt to use the concept of democratic efficacy as well as to test the association between democratic efficacy and populist attitudes. In our analysis we addressed the following questions: Descriptive research questions:  RQ1. How are our respondents distributed among the categories of the typology of political efficacy (23.1 – 24.2) (i.e. high, low, paternalist, sceptics – see, Bene and Boda, 2020) by country.  RQ2. What share of our respondents does have complete democratic capacities (i.e. follow news at least some days [25.1]; not extremely partisan [mean above 5] [17.1 – 18.5]; identify with democratic values [equality, tolerance, autonomy] [at least somewhat agree] [20.1 – 20.3]) by country.  RQ3. How are our respondents distributed among the categories of the typology of democratic efficacy (i.e. political efficacy + democratic capacities, see, Bene and Boda, 2020). Inferential research questions: 7 Copyright MORKEVIČIUS et al. (2019).  RQ4. What is the connection between external political efficacy (23.1 – 23.2) and populist attitudes (12.1 – 16.5).  RQ5. What is the connection between internal political efficacy (24.1 – 24.2) and populist attitudes.  RQ7. What is the connection between democratic capacities (complete vs. incomplete) and populist attitudes.  RQ8. What is the connection between the typology of democratic efficacy and populist attitudes. Findings: RQ1. How are our respondents distributed among the categories of the typology of political efficacy (23.1 – 24.2) (i.e. high, low, paternalist, sceptics – see, Bene and Boda, 2020) by country. Note: Both EE and IE are constructed from two items respectively by taking their average. All four items are measured on a 7-point Likert scale where the larger values indicate higher level of efficacy. Mean values above 4 are considered ‘high’ level and under 4 are ‘low’ level of efficacy. Respondents who were placed at the middle value (4) in each of the constructed EE or IE variables were not considered in the typology (45.5% of the respondents). As indicated in Table 3, 41.8% of respondents have low levels of political efficacy, while 24.7% of them are efficacious. As for the mixed types, more than a fourth of citizens have high levels of internal and low levels of external political efficacy and can therefore be labelled as sceptics, while only an 8% of them are paternalist, i.e. have low levels of internal and high levels of external political efficacy. However, there are remarkable variations between countries. In CEE (except Lithuania) and some Mediterranean countries (Italy, Greece, France), the share of people with high levels of PE is smaller, while they have a larger share of people with low PE than in Western-European countries. The proportion of paternalists is higher in Lithuania, Czech Republic, Turkey, and Greece, while the most sceptical respondents were found in Bosnia, Slovakia, Denmark and the UK. Table 3. Share of the respondents based on the level of their political efficacy by country and mean values of external and internal efficacy by country (last two columns) Country Germany Political efficacy (%) High Low Paternalists Sceptics 35% 40% 5% 20% 8 Resp. on middle value on each variable 37% EE IE Mean 3.61 4.22 Copyright MORKEVIČIUS et al. (2019). Country UK Czechia Hungary Italy Netherlands Poland France Slovakia Lithuania Denmark Turkey Spain Greece BosniaHerzegovina Total Political efficacy (%) High Low Paternalists Sceptics 32.4% 11.6% 8.7% 23.4% 39.1% 21.8% 23.1% 14.5% 35.4% 44.3% 26.1% 32.5% 15.1% 8.6% 32.8% 56.4% 64.6% 43.7% 30.6% 48.7% 53.4% 44.9% 30.1% 18% 33.6% 34.4% 45.1% 51.7% 4.3% 12.2% 6.3% 8.3% 5.8% 8.2% 7.2% 9.5% 13% 7% 12.1% 9.8% 10.4% 2% 30.5% 19.8% 20.5% 24.6% 24.5% 21.2% 16.3% 31.1% 21.4% 30.6% 28.3% 23.3% 29.4% 37.7% Resp. on middle value on each variable 40.1% 38% 29.4% 37.1% 41.8% 38.5% 39.3% 37.9% 38% 34.1% 35.1% 36.5% 31.2% 30.8% 24.7% 41.8% 7.9% 25.6% 36.5% EE IE Mean 3.64 3.29 2.82 3.45 3.74 3.33 3.44 3.36 3.90 3.95 3.60 3.69 3.29 4.37 3.53 3.39 3.95 4.36 3.78 3.77 4.01 4.17 4.56 4.28 4.10 3.90 2.79 3.92 3.46 4.03 RQ2. What share of our respondents does have complete democratic capacities (i.e. follow news at least some days [25.1]; not extremely partisan [mean above 5] [17.1 – 18.5]; identify with democratic values [equality, tolerance, autonomy] [at least somewhat agree] [20.1 – 20.3]) by country. There are slightly more people in our sample who have incomplete democratic capacities, but more than 46.4% of our respondents have complete democratic capacities (see Table 4). It is difficult to find clear geographical patterns behind the country-level variations, but it is noticeable that in Mediterranean countries (Greece, Bosnia, Spain, Italy) more people have complete democratic capacities than in other countries. It is also interesting that in some Western-European countries the proportion of people with complete DC is rather low (Denmark, Netherlands, France). Table 4. Share of the respondents based on the level of their democratic capacities by country Country Germany UK Czech Rep. Hungary Italy Netherlands Poland France Incomplete 49.4% 50.5% 70.1% 57% 52.8% 63.4% 59.1% 58.1% 9 Democratic capacities Complete 50.6% 49.5% 29.9% 43% 47.2% 36.3% 40.9% 41.9% Copyright MORKEVIČIUS et al. (2019). 59.3% 38.7% 63.9% 56.2% 47.5% 37.6% 43.1 53.6% Slovakia Lithuania Denmark Turkey Spain Greece Bosnia-Herzegovina TOTAL 40.7% 61.3% 36.1% 43.8% 52.5% 62.4% 56.9% 46.4% RQ3. How are our respondents distributed among the categories of the typology of democratic efficacy (i.e. political efficacy + democratic capacities, see Bene and Boda, 2020)? Only one in ten people have both complete democratic capacities and high levels of political efficacy (see Table 5). At the other end of the typology, 21 % of the respondents have both incomplete democratic capacities and low level of political efficacy. Interestingly enough, a large share of respondents with low political efficacy have complete democratic capacities, while 13 % of the respondents have both high levels of political efficacy and incomplete capacities. Half of the sceptics have complete and the other half have incomplete democratic capacities. Table 5. The share of the respondents in the intersections of democratic capacities and political efficacy Democratic capacities Complete democratic capacities Incomplete capacities Total Political efficacy High Low 11.2% 20.6% 13.2% 21.2% 24.7% 41.8% Total Paternalist 4.2% 3.7% 7.9% Sceptics 12.8% 12.8% 25.6% 49.1% 50.9% 100% RQ4. What is the connection between external political efficacy (23.1 – 23.2) and populist attitudes (12.1 – 16.5)? The scale of populist attitude is constructed from 13 items by averaging them. External political efficacy and populist attitudes a significantly and negatively correlated and the effect size is fairly remarkable. People with low level of external political efficacy have more populist attitudes (p < 0.001; Pearson R = -0.375). RQ5. What is the connection between internal political efficacy (24.1 – 24.2) and populist attitudes? Internal political efficacy is also significantly and negatively related to populist attitudes, but here the effect size is much smaller, only a weak relationship exists between the variables. 10 Copyright MORKEVIČIUS et al. (2019). However, people with low level of internal political efficacy have rather more populist attitudes (p < 0.001; Pearson R= -0.046). RQ7. What is the connection between democratic capacities (complete vs. incomplete) and populist attitudes? A weak, but significant negative association exists between democratic capacities and populist attitudes, too. People with incomplete democratic capacities are more likely to have populist attitudes (p < 0.01; Pearson’s R = -0.032). RQ8. What is the connection between the typology of democratic efficacy and populist attitudes? Table 6 shows the mean values of populist attitudes for each category of democratic efficacy. It seems that the most populist subcategory is people who are sceptics with incomplete democratic capacities, followed by the ‘incomplete capacities with low PE’ and ‘complete capacities with low PE’ categories. The less populist respondents are those who have complete democratic capacities and high levels of political efficacy. Interestingly enough, paternalists are also less populist, moreover, paternalists with incomplete democratic capacities are even a little less populist than those with complete capacities. Table 6. Mean values of populist attitudes by the categories of democratic efficacy. (SDs are in parentheses) Democratic efficacy Populist attitudes Incomplete capacities with low PE 5.05 (.66) Complete capacities with low PE 5.02 (.65) Incomplete capacities with paternalist PE 4.48 (.74) Complete capacities with paternalist PE 4.56 (.64) Incomplete capacities with sceptics PE 5.11 (.67) Complete capacities with sceptics PE Incomplete capacities with high PE 4.69 (.76) Complete capacities with high PE 4.38 (.74) TOTAL 4.88 (.73) Additional analysis 11 Copyright MORKEVIČIUS et al. (2019). Table 7 shows that at least a two-third majority of the respondents meet our criteria for each democratic capacity in each country. Consequently, there are widely shared capacities in all European countries under investigation. In the case of news consumption, a remarkable gap exists between CEE and Western European countries. Except for Lithuania, the smallest percentages of regular news consumers can be found in CEE countries. In the case of partisanship, the variance is minor among the countries as most of them are between 75 % and 82 %, but there are some outlier cases such as the strongly partisan Turkey or Netherlands, or the least partisan Lithuania. Equality as a value is the least popular in Denmark (!) and the Czech Republic, while they are highly favoured in some South European countries such as Greece, Spain, Bosnia and Turkey. Smaller variances can be found in the case of the two other values. However, it is remarkable that all of these values are highly popular in Bosnia, Lithuania, Greece and Spain. Table 7. Share of respondents who meet our criteria for each democratic capacity Country Germany UK Czechia Hungary Italy Netherlands Poland France Slovakia Lithuania Denmark Turkey Spain Greece Bosnia-Her. TOTAL News cons. 89% 84% 74% 74% 89% 79% 77% 79% 74% 86% 86% 92% 91% 88% 75% 83% Non-partisan 81% 77% 76% 82% 75% 72% 80% 81% 80% 87% 78% 61% 76% 85% 87% 79% Equality 78% 86% 72% 84% 85% 79% 82% 80% 82% 89% 68% 91% 90% 93% 94% 84% Tolerance 85% 86% 83% 85% 87% 82% 85% 83% 83% 92% 82% 89% 88% 91% 94% 86% Autonomy 90% 90% 91% 91% 86% 85% 88% 85% 91% 95% 83% 92% 90% 92% 95% 90% Looking at the bivariate correlation between different components of democratic capacities, it seems that partisanship is an outlier component (see Table 8). While there is a significant positive relationship between news consumption and the values of equality, tolerance and autonomy, non-partisanship is significantly and negatively correlated with each of them. Those who consume news at least some days are more likely to agree with the values of equality, tolerance and autonomy. At the same time, less partisan people consume news infrequently and they are more likely to be neutral or negative with these values. The strongest correlation is found between the three democratic values. 12 Copyright MORKEVIČIUS et al. (2019). Table 8. Bivariate correlations between the components of democratic capacities News cons. News cons. Non-partisan Equality Tolerance Autonomy -.132** .075** .111** .096** -.064** -.075** -.058** .430** .427** Non-partisan Equality Tolerance .481** Note: * p < .05, ** p < .01, *** p < .001 Table 9. Bivariate correlations between the components of democratic capacities and populism Democratic capacities Populism News consumption .009 Non-partisan -.165** Equality .136** Tolerance .121** Autonomy .222** Interestingly, democratic values are significantly and positively correlated with populist attitudes (see Table 9). It seems that democratic values and populist attitudes are not mutually exclusive. The only democratic capacity that significantly decreases populist attitudes is nonpartisanship, the component that is significantly negatively associated with the other democratic components. Non-extremely partisan citizens are less likely to identify with populist attitudes. News consumption is not significantly associated with populism, but when its interval variety is considered there is a small (R = 0.051), but significant (p < 0.001) relationship between the two variables showing that more frequent news consumption is correlated with less populist attitudes. 4. Measuring democratic efficacy across time and countries As our survey above demonstrates, the measurement of the full set of democratic capacities coupled with political efficacy is empirically possible and provides interesting results, but only with data from quite recent surveys (ESS, Comparative Studies of Electoral Systems, 13 Copyright MORKEVIČIUS et al. (2019). International Social Survey Programme). Unfortunately, analysis of time trends seems to be hardly available, since either political efficacy or democratic capacities are not measured in the same way in repeated rounds of major international social surveys. Therefore, we simultaneously propose a different — narrower — conceptualisation of democratic political efficacy, which basically remains on the attitudinal level. Thus, if we remain on the attitudinal level, the most important addition to the concept of political efficacy is that of certain democratic attitudes. Thus, we propose a narrow definition of democratic efficacy conceptualised as a personal sense of political efficacy (either internal, or external, or both) coupled with certain core democratic values. Then a person that could be described as democratically efficacious should feel confident about his/her ability to effectively participate in politics, feel that government is responsive to his/her needs and also subscribe to certain core democratic values. As a conceptualisation (and, to a certain extent, a measurement) of political efficacy (both internal and external) is already established, clearly defining which democratic values should be included into the concept of democratic efficacy remains a remarkable challenge. Indeed, this may be a daunting task, as the vast literature on theories of democracy, quality of democracy and its measurement abound. However, here we follow definitions of democracy that may be called sophisticated procedural or procedural with prerequisites (see general characterisation in Brettschneider 2006: 262-266). Importantly, every definition of democracy starts from the most important features — elections and voting. Every regime that wants to be called a democracy must be a political system where rulers are selected by competitive elections, that is, by citizens voting in free and fair elections for candidates and their supported policies. Procedural (or electoral, minimal) definitions of democracy are inclined to stop here as advocated by Schumpeter (1942) or Przeworski (1999). However, the majority of other theorists of democracy proceed further and suggest various additional features that are necessary for a truly democratic regime. The largest part of them subscribe to an adjective ‘liberal’ and propose definitions of ‘liberal democracy’. This group of scholars follows the tradition of Robert Dahl (1956, 1971) and his conception of pluralist democracy or polyarchy. Dahl (1989) formulated five prerequisites for democracy:  Inclusiveness, or equality of all citizens within the state and ability to get involved in the political process. 14 Copyright MORKEVIČIUS et al. (2019).  Voting equality at the decisive stage, or weight equality of each citizen’s vote when decisions are taken (voted upon).  Enlightened understanding, or ample and equal opportunities to discover information about political matters and base their decisions on this information.  Control of the agenda, or the opportunity for citizens to decide which political matters are important and which of them should be discussed and decided upon.  Effective participation, or adequate and equal opportunities of citizens to form their preferences and express them on the public agenda. While the first three criteria are rather procedural and every scholar advocating a minimal definition of democracy could subscribe to them, the two remaining would be problematic, as they require that citizens (not politicians) control the substance of politics and be free in formulating and expressing their policy preferences. The last requirement most directly points in the direction of liberal democracy as certain liberties are indicated as prerequisites of democratic government. All in all, definitions of liberal democracy formulate certain individual rights and principles of governance that are crucial prerequisites of a democratic regime. One of the most comprehensive attempts to formulate the prerequisites (as well as a definition) of liberal democracy2 was developed by Diamond and Morlino (2005). These authors (together with a team of fellow researchers) identified five procedural, two substantive and one resultsoriented dimension of (representative) democracy (see pages x-xxxi):  Procedural dimensions: o Rule of law, which means that all citizens are equal before the law, and that the laws themselves are clear, publicly known, universal, relatively stable, non-retroactive, and fairly and consistently applied to all citizens by an independent judiciary. Importantly, the legal system defends democratic procedures, upholds citizens’ civil and political rights, and reinforces the authority of other agencies of horizontal accountability. o Participation, which implies that all adult citizens have formal rights of political participation, including the right to vote, and are able to make use of these formal rights by being able to organise, assemble, protest, lobby for 2 Other versions of similar definitions abound (see, for example, Schmitter, Karl 1991; O’Donnell, Cullell, Iazzetta 2004; Bühlmann, Kriesi 2013). 15 Copyright MORKEVIČIUS et al. (2019). their interests, and otherwise influence the decision-making processes. Importantly, effective participation requires political as well as basic socioeconomic equality, tolerance of political and social differences and effective implementation of the rule of law. o Competition, which requires that political systems have regular, free, and fair electoral competition among various political actors. It also implies openness of access to the electoral arena for new political forces, the ease with which incumbents can be defeated, and equality of access to the mass media and campaign funding for the competing political actors. Importantly, effective competition requires functioning horizontal accountability (an independent electoral commission) and effective implementation of the rule of law. o Vertical accountability, which obliges elected political leaders to answer for their political decisions when asked by citizens or other constitutional bodies. Importantly, vertical accountability may be extended beyond elections and encompass also efforts of civic associations, NGOs, social movements, think tanks, and mass media to hold governments accountable in between elections. Effective competition and participation as well as the rule of law are necessary for effective functioning of vertical accountability. o Horizontal accountability, which requires that officeholders are answerable to other institutional actors that have the expertise and legal authority to control and sanction their behaviour. Governmental agencies and institutions must have their sphere of competence and yet be accountable to some supervising or controlling body. The vitality of horizontal accountability very much depends on a legal system that enables the exertion of checks and balances by public entities that are independent of the government.  Substantive dimensions: o Freedom, which consists of three types of rights: political, civil, and social or socioeconomic. Political rights include rights to vote, to run for office, to campaign, and to organise political parties. They make possible effective political participation and competition as well as vertical accountability. Essential civil rights include personal liberty, security, and privacy; freedom 16 Copyright MORKEVIČIUS et al. (2019). of thought, expression, and information; freedom of religion; freedom of assembly, association, and organisation; freedom of movement and residence; the right to legal defence and due process. However, rights and freedoms are subject to exceptions and qualifications, that is, some liberties may encroach on other social values, such as preventing disorder or protecting public health and safety. o Equality, which means, first of all, formal political equality of citizens requiring the same rights and legal protections for every citizen, access to justice and power as well as the prohibition of discrimination on various grounds. Importantly, political equality should not be decoupled from social and economic correlates, as at least some basic social and economic resources are required in order to effectively exercise one’s political rights.  Results dimension: o Responsiveness, which requires that governments, parties and politicians respond to the expectations, interests, needs, and demands of citizens. This implies that from the point of view of good governance, democratic institutions have to formulate and implement policies that citizens want. Thus, responsiveness is closely related to vertical accountability. It is clear from the definition and conceptualisation provided above that there are certain basic procedural criteria of democracy, which are enhanced by additional substantive and results oriented features3. Therefore, respect for core procedural attributes of ‘good’ democracy should be the building blocks onto which we could base our definition of democratic political efficacy. However, for a well-functioning democracy they are not sufficient and at least a certain level of other criteria should be realised4. All in all, concept and measures of democratic efficacy should include traditional definitions and indicators of the sense of political efficacy together with conceptual and operational definitions of respect for core procedural democratic criteria and at least some substantive and/or results oriented attributes of democracy. In the next section 3 Cumulative nature of the features of democracy is clearly stated in the conceptualisation itself: procedural criteria are called ‘basic’ and the remaining criteria are identified as ‘goals of ideal democracy’ and ‘broader standards of good governance’ (Diamond, Morlino 2005: x-xi). 4 The idea that substantive and results oriented criteria should be treated not as absolute necessary norms of democracy, but as to a certain degree attainable desirable attributes of democracy is supported by Coppedge (2004). 17 Copyright MORKEVIČIUS et al. (2019). we proceed with an overview of data sources that could be used for the measurement of democratic efficacy across time and countries. The components of democratic efficacy in international surveys In order to be able to study the state and trends of democratic efficacy in democratic polities around the world, we need to find appropriate indicators in the existing data sources collected by longitudinal international surveys. We performed a search among the variables of the most important socio-political surveys: European Social Survey (ESS), International Social Survey Programme (ISSP), European Election Studies (EES), European Values Study (EVS), Comparative Study of Electoral Systems (CSES) and Eurobarometer. Two criteria for evaluation were employed: availability and consistency of measurement of political efficacy, core values of democracy, political knowledge, news consumption, political attachment (party or ideological identification), and political skills (political participation besides voting). Results of the survey are presented in Table 10. Table 10. Indicator availability in major longitudinal international socio-political surveys Concepts ESS ISSP EES EVS Political efficacy Measured, but Measured in Not measured Not measured inconsistently some waves Core values of democracy Consistent measurement Political knowledge Not measured Not measured News consumption Political attachment Political skills CSES Measured in some waves Measured Measured Measured in inconsistently Not measured inconsistently some waves in some waves in some waves Measured in Measured in Not measured only one wave only one wave Measured Measured, but Measured, but inconsistently inconsistently inconsistently in some waves Consistent Consistent Consistent measurement measurement measurement Consistent Measured in Not measured measurement some waves Measured, but Not measured inconsistently Eurobarometer External measured in some waves Measured inconsistently in some waves Measured in some waves, (only EU) Measured in some waves Consistent Consistent Measured in measurement measurement most waves Consistent Not measured Not measured measurement Since indicators of political efficacy are most important when attempting to measure democratic efficacy, we first studied their availability. It appeared that the most systematic measurement was available in the ISSP modules ‘Role of Government’ and ‘Citizenship’. These two modules include both internal and external efficacy measures and span at least 10 years. The ESS measures political efficacy inconsistently: items were excluded in some waves and 18 Copyright MORKEVIČIUS et al. (2019). measurement differs between the earlier and more recent waves. It would only be possible to consistently study internal and external efficacy for the span of the last five to six years. Unfortunately, EES, EVS and Eurobarometer do not include any consistent longitudinal measurement of political efficacy at all. The same is mostly true for the surveys included in CSES (measures of political efficacy were included only in recent waves). All in all, the only two major longitudinal international socio-political surveys that could be used for measurement of political efficacy are the ISSP and the ESS. After further consideration, however, the ESS was also eliminated as its measurement of political efficacy spans only the last three waves (five to six years) and another very important measure for evaluating democratic efficacy — values of liberal democracy — do not include indicators of support for procedural democracy5. Further, looking at the questions in the ISSP module ‘Role of Government’ we could see that political efficacy is only measured by single items and measures of liberal democracy are almost absent. Therefore, for our empirical study we selected items form the ISSP module ‘Citizenship’. This module was conducted in 2004 and in 2014. Thus, it provides a sufficiently long perspective in order to be able to discern certain trends of stability or change. More importantly, this module included two items measuring both internal and external efficacy, and many items measuring public support for values of liberal democracy:  External political efficacy: o To what extent do you agree or disagree with the following statements?  People like me don't have any say about what the government does (Q33 2004; Q37 2014).  I don't think the government cares much what people like me think (Q34 2004; Q38 2014).  Internal political efficacy: o To what extent do you agree or disagree with the following statements?  I feel I have a pretty good understanding of the important political issues facing [COUNTRY] (Q35 2004; Q39 2014). 5 The ideal case here would have been the ESS module on ‘Democracy perceptions and evaluations’. However, it does not include any items measuring political efficacy. 19 Copyright MORKEVIČIUS et al. (2019).  I think most people in [COUNTRY] are better informed about politics and government than I am (Q36 2004; Q40 2014).  Support for values of liberal democracy: o There are different opinions as to what it takes to be a good citizen. As far as you are concerned personally on a scale of 1 to 7, where 1 is not at all important and 7 is very important, how important is it:  Always to vote in elections (Q1 2004; Q1 2014).  Always to obey laws and regulations (Q3 2004; Q3 2014).  To keep watch on the actions of government (Q4 2004; Q4 2014).  To be active in social or political associations (Q5 2004; Q5 2014).  To try to understand the reasoning of people with other opinions (Q6 2004; Q6 2014). o There are different opinions about people's rights in a democracy. On a scale of 1 to 7, where 1 is not at all important and 7 is very important, how important is it:  That government authorities respect and protect the rights of minorities (Q28 2004; Q29 2014). The measures of political efficacy are rather traditional and after performing principal components analysis on them (separately for 2004 and 2014) we discovered that they do distinguish into two scales: internal and external efficacy (see Table 11). However, for our further analysis we constructed three separate scales of: internal efficacy (averaging two PC1 items), external efficacy (averaging two PC2 items) and political efficacy (averaging all items). Even though the principal components analysis showed differentiation of public evaluations of internal and external efficacy items, we also merged them in order to construct a common scale of political efficacy and explore trends of more general democratic political efficacy. Table 11. Principal components analysis of political efficacy items in the ISSP module ‘Citizenship’: varimax rotated solutions 2004 2014 Items PC1 PC2 PC1 PC2 People like me don't have any say about 0.6902 0.6830 what the government does 20 Copyright MORKEVIČIUS et al. (2019). 2004 Items PC1 I don't think the government cares much what people like me think I feel I have a pretty good understanding of the important political issues facing country I think most people in country are better informed about politics and government than I am Eigenvalues (only > 1) Variance explained (%) Total variance explained (%) 2014 PC2 0.6929 PC1 PC2 0.6889 1.58 39.39 0.7716 0.7869 0.6358 0.6160 1.28 32.06 71.45 Data source: the ISSP modules ‘Citizenship’. Notes: Kaiser normalisation applied; loadings < 0.3 suppressed. 1.58 39.55 1.24 31.12 70.67 As was discussed previously, the ISSP module ‘Citizenship’ contains at least six items measuring public support of values of liberal democracy. They (in totality) reflect (at least partially) all the important characteristics or criteria of a liberal democracy presented in the theoretical part:  There are different opinions as to what it takes to be a good citizen. As far as you are concerned personally on a scale of 1 to 7, where 1 is not at all important and 7 is very important, how important is it:  Always to vote in elections (Participation, Vertical accountability, Responsiveness).  Always to obey laws and regulations (Rule of law, Horizontal accountability).  To keep watch on the actions of government (Vertical accountability, Responsiveness).  To be active in social or political associations (Participation, Competition, Freedom).  To try to understand the reasoning of people with other opinions (Equality, Competition).  There are different opinions about people's rights in a democracy. On a scale of 1 to 7, where 1 is not at all important and 7 is very important, how important is it: 21 Copyright MORKEVIČIUS et al. (2019).  That government authorities respect and protect the rights of minorities (Freedom, Equality). In order to investigate, whether these items could be subsumed into a single scale of support for liberal democratic values, we performed principal components analysis (separately for 2004 and 2014). The results (see Table 12) showed that the items could be considered as loading on a single component as there is only one eigenvalue larger than 1. However, the single component solution explains only about 40 % of variance of the items and the last item from the different question battery loads poorly on the first component. On the other hand, Cronbach’s alpha calculated for the scale including all the six items approaches 0.7 (2004=0.6845 and 2014=0.6875), which we consider to be an acceptable indicator of consistency. All in all, we constructed a scale of support for values of liberal democracy averaging all the six items. Table 12. Principal components analysis of items corresponding to values of liberal democracy in the ISSP module ‘Citizenship’: unrotated solutions 2004 2014 Items PC1 PC1 Always to vote in elections 0.4407 0.4311 Always to obey laws and regulations 0.3460 0.3548 To keep watch on the actions of government 0.4934 0.4916 To be active in social or political associations 0.4416 0.4373 To try to understand the reasoning of people with other opinions That government authorities respect and protect the rights of minorities Eigenvalues (only > 1) Variance explained (%) 0.4133 0.4231 0.2772 0.2763 2.35 39.17 2.37 39.48 Data source: the ISSP modules ‘Citizenship’. The ISSP module also has many other items suitable for the analysis of the democratic efficacy of citizens. These include multiple items of participation (proxy indicator of political skills), an item on political attachment (voting for a party in the last general election, containing categories for far-right and far-left party voting), and an item measuring frequency of using media to get political news or information (however, included only in the 2014 wave). Finally, the ISSP data also includes many socio-demographic characteristics: gender, age, education, household income, socio-economic class and place of residence (urban vs. rural). 22 Copyright MORKEVIČIUS et al. (2019). For the analysis of the state and trends of democratic efficacy in democratic polities around the world we integrated the scales of political efficacy and support for values of democracy, thus constructing three measures of: democratic political efficacy, democratic internal efficacy and democratic external efficacy. Before merging the items we dichotomised them in order to get fourfold classification of respondents according to all the three scales of democratic efficacy6:  Scoring low on both political efficacy and support for values of liberal democracy.  Scoring low on political efficacy, but high on support for values of liberal democracy.  Scoring high on political efficacy, but low on support for values of liberal democracy.  Scoring high on both political efficacy and support for values of liberal democracy. The thresholds of dichotomising the scales were used the following:  Political (internal and external) efficacy: low = lowest through 3, high = more than 3).  Support for values of liberal democracy: low = lowest through 5.5, high = more than 5.5). These thresholds are somewhat unbalanced towards the higher values end, since we wanted to identify as highly efficacious or strong supporters of liberal democracy only those respondents who were clear and not ‘middle-of the-road’ (‘average’) cases. 5. Results of the longitudinal cross-country analysis In this section, we present our findings of the secondary data analysis of two the waves (2004 and 2014) of the ISSP module ‘Citizenship’. On the most general level, in 2004 almost twothirds (64.6 %) of respondents scored low on political efficacy, while in 2014 this share dropped somewhat to 61.4 %). Low levels of political efficacy are more widespread in Central and Eastern European (CEE) countries. The share of respondents who scored high on both political efficacy and support for values of liberal democracy increased somewhat (from 21.5 % to 6 Unfortunately, these groups do not reflect the classification of respondents into four groups of political efficacy as suggested previously: highly efficacious, highly non-efficacious, paternalists, and sceptics. 23 Copyright MORKEVIČIUS et al. (2019). 23.5 %) between 2004 and 2014. However, the share of respondents who score high on political efficacy and low on support for values of liberal democracy also increased from 13.9 % to 15.2 % between 2004 and 2014. Increase in both groups was mostly due to an increase in political efficacy in both groups as the share of people strongly supporting values of liberal democracy remained almost unchanged (52.3 % in 2004 and 51.9 % in 2014). By and large, the increase in political efficacy was more pronounced in the Western countries. The scores on the scales of political efficacy and values of liberal democracy in 2004 reveal that the highest scores of the LE+LL indicator (LE – low score on political efficacy + LL – low score on support for values of liberal democracy) were in Czechia (64.7 %), Latvia (60.1 %), Hungary (53.5 %), Slovenia (52.0 %) and Belgium: Flanders (51.7 %) (see Table 13). The lowest scores were in France (17.0 %), United States (17.5 %), Denmark (17.7 %), Canada (20.0 %) and Israel (20.4 %) (see Table 13). In the second wave of the 2014 round the highest scores of the LE+LL indicator were again mainly among formerly communist Eastern European countries, i.e. in Slovakia (56.1 %), Czechia (54.1 %), Belgium: Flanders (49.3 %), Poland (49.2 %) and Slovenia (47.2 %). On the other hand, the lowest scores were in Turkey (13.4 %), Iceland (16.4 %), Norway (17.2 %), France (18.4 %) and Sweden (19.5 %). Table 13. Cross-country and temporal distributions of respondents in four groups according to scores on the scales of political efficacy and values of liberal democracy, percentages Country Australia Austria Belgium: Flanders Bulgaria Canada Croatia Cyprus Czechia Denmark Finland France Georgia Germany: East Germany: West Great Britain Hungary Iceland Ireland Israel: Arabs + Jews LE+LL 20.5 41.2 49.3 2014 LE+HL HE+LL 27.7 14.4 24.4 15.0 19.5 15.3 HE+HL 37.4 19.3 15.8 42.8 37.5 6.0 13.6 54.1 22.0 38.2 18.4 26.6 40.6 30.9 28.2 45.3 16.4 21.4 22.6 18.8 14.9 45.5 15.7 16.7 29.3 25.2 17.8 14.7 22.5 23.0 29.1 7.9 21.2 24.8 13.7 13.5 20.1 9.9 33.0 20.0 37.5 20.0 22.5 27.6 28.8 16.1 45.6 29.9 38.9 11.3 19.9 24 LE+LL 21.2 35.8 51.7 2004 LE+HL HE+LL 37.6 12.4 33.4 10.3 20.0 16.1 HE+HL 28.8 20.6 12.1 50.5 20.0 27.8 42.4 10.2 7.6 11.5 30.0 25.1 64.7 17.7 44.3 17.0 18.7 14.6 23.1 17.6 15.0 19.1 15.2 22.0 23.4 25.3 37.0 5.5 37.2 14.7 42.7 50.1 37.7 39.6 53.5 22.6 26.8 26.0 23.6 16.4 16.9 16.4 12.5 10.9 18.6 18.0 10.5 25.1 20.4 40.3 41.8 10.8 8.7 23.8 29.1 Copyright MORKEVIČIUS et al. (2019). Country LE+LL 34.9 38.9 2014 LE+HL HE+LL 19.3 23.4 38.5 9.1 HE+HL 22.4 13.6 LE+LL 35.5 39.9 60.1 2004 LE+HL HE+LL 19.2 22.6 42.2 8.2 26.3 6.9 HE+HL Japan 22.7 Korea (South) 9.7 Latvia 6.7 Lithuania 42.7 40.3 6.7 10.3 Netherlands 22.7 23.1 18.4 35.8 27.1 19.3 23.1 30.6 New Zealand 30.2 23.9 19.3 26.6 Norway 17.2 17.2 23.7 42.0 23.8 21.4 19.8 35.0 Poland 49.2 32.7 7.4 10.7 36.1 52.7 3.2 8.1 Portugal 26.0 51.3 5.5 17.2 Russia 44.2 25.0 17.7 13.1 43.9 40.2 9.2 6.7 Slovakia 56.1 23.4 14.3 6.2 50.7 32.0 9.5 7.8 Slovenia 47.2 41.2 3.8 7.8 52 29.4 11.7 6.9 South Africa 34.8 41.1 8.6 15.5 24.3 45.3 7.4 23.1 Spain 27.8 40.2 9.0 23.0 28.2 37.1 12.0 22.7 Sweden 19.5 24.1 19.3 37.0 29.1 33.9 14.9 22.1 Switzerland 25.6 12.4 29.7 32.2 33.5 15.2 27.0 24.2 Turkey 13.4 38.6 12.9 35.0 United States 22.0 34.0 9.1 34.9 17.5 25.5 11.5 45.5 Data source: the ISSP modules ‘Citizenship’. Notes: LE – low score on political efficacy; HE – high score on political efficacy; LL – low score on support for values of liberal democracy; HL – high score on support for values of liberal democracy. The scores on the scales of political efficacy and values of liberal democracy in 2004 indicate that the highest scores of the HE+HL indicator (HE – high score on political efficacy + HL – high score on support for values of liberal democracy) were in the United States (45.5 %), France (42.7 %), Denmark (37.2 %), Cyprus (37.0 %) and Norway (35.0 %) (see Table 13). On the opposite side of the HE+HL indicator scale are the former post-communist countries Czechia (5.5 %), Latvia (6.7 %), Russia (6.7 %), Slovenia (6.9%) and Slovakia (7.8 %). In 2014 the highest scores of the HE+HL indicator were in Iceland (45.6 %), Norway (42.0 %), France (37.5 %), Australia (37.4 %) and Sweden (37.0 %). On the other hand, Slovakia (6.2 %), Slovenia (7.8 %), Czechia (9.9 %), Lithuania (10.3 %) and Poland (10.7 %) compose a group of counties, where the score is the lowest. Regarding the score of the LE+HL indicator (LE – low score on political efficacy + HL – high score on support for values of liberal democracy), in 2004 the highest score is in Poland (52.7 %), Portugal (51.3 %), South Africa (45.3 %), Canada (42.4 %) and South Korea (42.2 %) (see Table 13). On the other hand, in Czechia (14.6 %), France (15.0 %), Switzerland (15.2 %), Finland (17.6 %) and Cyprus (18.7 %) these scores were the lowest. In 2014 the highest scores of the LE+HL indicator were in Georgia (45.5 %), Slovenia (41.2 %), South Africa (41.1 %), Lithuania (40.3 %) and Spain (40.2 %). These are the five countries with the lowest scores: Switzerland (12.4 %), France (14.9 %), Germany: East (15.7 %), Germany: West (16.7 %) and Norway (17.2 %). 25 Copyright MORKEVIČIUS et al. (2019). Finally, the scores on the scales of political efficacy and values of liberal democracy in 2004 reveal that the highest scores of the HE+LL indicator (HE – high score on political efficacy + LL – low score on support for values of liberal democracy) is in Switzerland (27.0 %), France (25.3 %), Finland (23.4 %), the Netherlands (23.1%) and Japan (22.6 %) (see Table 13). On the opposite side on the HE+LL indicator scale is Poland (3.2 %), Portugal (5.5 %), Latvia (6.9 %), South Africa (7.4 %) and Canada (7.6 %). In 2014 the highest scores of the HE+LL indicator is in Switzerland (29.7 %), France (29.1 %), Germany: West (24.8 %), Norway (23.7 %) and Japan (23.4 %). On the other hand, the lowest score is in Slovenia (3.8 %), Croatia (6.0 %), Lithuania (6.7 %), Poland (7.4 %) and Georgia (7.9 %). Table 14. Temporal differences (2014-2004) between shares of groups according to levels of democratic political efficacy Country Sweden Germany: East Difference (LE+LL) -9.6 -9.5 Difference (LE+HL) -9.8 -6.9 Difference (HE+LL) 4.4 4.8 Difference (HE+HL) 14.9 11.6 Great Britain Germany: West Australia Switzerland Norway Russia -11.4 -6.8 -0.7 -7.9 -6.6 0.3 3.3 -10.1 -9.9 -2.8 -4.2 -15.2 -2.7 7.9 2.0 2.7 3.9 8.5 10.8 9.0 8.6 8.0 7.0 6.4 Hungary Finland Netherlands Czechia Korea (South) Belgium: Flanders -8.2 -6.1 -4.4 -10.6 -1.0 -2.4 1.6 1.2 3.8 6.8 -3.7 -0.5 1.0 -0.4 -4.7 -0.5 0.9 -0.8 5.6 5.3 5.2 4.4 3.9 3.7 Poland Slovenia Spain Japan Austria Slovakia 13.1 -4.8 -0.4 -0.6 5.4 5.4 -20.0 11.8 3.1 0.1 -9.0 -8.6 4.2 -7.9 -3.0 0.8 4.7 4.8 2.6 0.9 0.3 -0.3 -1.3 -1.6 Denmark 4.3 -0.5 0.5 -4.2 France 1.4 -0.1 3.8 -5.2 South Africa 10.5 -4.2 1.2 -7.6 Israel: Arabs + Jews 9.5 -2.9 2.6 -9.2 United States 4.5 8.5 -2.4 -10.6 Data source: the ISSP modules ‘Citizenship’. Notes: LE – low score on political efficacy; HE – high score on political efficacy; LL – low score on support for values of liberal democracy; HL – high score on support for values of liberal democracy. 26 Copyright MORKEVIČIUS et al. (2019). From Table 14 we see that the increase in groups of HE+LL and HE+HL was mostly due to increasing levels of political efficacy in some democratic countries7. Figures 1 and 2 show country differences in profiles of distributions of separate groups. In 2004 high democratic efficacy was relatively more pronounced in only two countries: the United States and Denmark. Fig.1. Relative cross-country distributions of respondents into four groups according to scores on the scales of political efficacy and values of liberal democracy in 2004. In 2014 the ‘picture’ changed and we clearly see that many more countries may be identified as having a relatively sizeable share of population scoring high on political efficacy and strongly supporting values of liberal democracy: Iceland, Australia, Sweden, the Netherlands, Denmark and Norway. 7 However, in some Western countries, such as, Denmark, France, Israel and United States efficacy decreased substantially. 27 Copyright MORKEVIČIUS et al. (2019). Fig.2. Relative cross-country distributions of respondents into four groups according to scores on the scales of political efficacy and values of liberal democracy in 2014. Further, we present results of analysis of tendencies with regard to external democratic efficacy. On the most general level, in 2004 almost three-fourths (74.3 %) of respondents scored low on external political efficacy, while in 2014 this share dropped slightly to 70.5 %. Low levels of external political efficacy again appeared to be more widespread in the CEE countries. The share of respondents who scored high on both external political efficacy and support for values of liberal democracy increased only slightly from 15.5 % to 17.7 % between 2004 and 2014. The share of respondents who scored high on external political efficacy and low on support for values of liberal democracy remained relatively stable (10.3 % in 2004 and 11.8 % in 2014). Thus, external democratic efficacy remained rather stable during the studied period. The scores on the scales of external efficacy and values of liberal democracy in 2004 reveal that the highest scores of the LE+LL indicator (LE – low score on external political efficacy + LL – low score on support for values of liberal democracy) were in Czechia (68.3 %), Latvia 28 Copyright MORKEVIČIUS et al. (2019). (61.7 %), Belgium: Flanders (59.2 %), Hungary (58.3 %) and Germany: East (57.8 %) (see Table 15). The lowest scores of the LE+LL indicator were in the United States (20.7 %), France (21.6 %), Canada (22.2 %), Denmark (23.5 %) and Israel (24.4 %). In 2014, we have another group of countries where the highest scores of the LE+LL indicator are mainly in Eastern Europe, i.e. Czechia (58.4 %), Slovakia (57.0 %), Belgium: Flanders (54.0 %), Poland (51.2 %) and Slovenia (48.1 %). On the opposite side of the LE+LL indicator scale are Turkey (14.0 %), Iceland (20.8 %), Norway (22.1 %), Sweden (23.7 %) and the United States (24.5 %). Regarding the score of the HE+HL indicator (HE – high score on external political efficacy + HL – high score on support for values of liberal democracy), in 2004 the highest scores were in France (34.9 %), the United States (34.6 %), Cyprus (27.7 %), Denmark (27.6 %) and Norway (27.5 %) (see Table 15). On the other hand, this score was the lowest in countries, which had communist regimes after WWII, i.e. Czechia (3.2 %), Slovakia (3.4 %), Latvia (5.0 %), Germany: East (5.8 %) and Slovenia (5.8 %). Table 15. Cross-country and temporal distributions of respondents in four groups according to scores on the scales of external efficacy and values of liberal democracy, percentages Country Australia Austria Belgium: Flanders Bulgaria Canada Croatia Cyprus Czechia Denmark Finland France Georgia Germany: East Germany: West Great Britain Hungary Iceland Ireland Israel: Arabs + Jews Japan Korea (South) Latvia Lithuania Netherlands New Zealand LE+LL 25.8 47.1 54.0 2014 LE+HL HE+LL 41.6 9.2 31.9 9.1 25.5 10.7 HE+HL 23.4 11.9 9.8 45.7 45.4 3.1 5.7 58.4 29.7 43.2 25.6 27.3 46.9 36.3 33.0 47.3 20.8 24.4 32.7 23.6 25.2 49.8 22.3 21.2 38.9 28.2 26.4 10.5 14.9 18.2 22.0 7.3 14.9 19.4 9.1 11.5 15.8 6.7 22.7 15.0 27.2 15.6 15.8 23.1 19.1 13.0 37.1 33.9 44.2 7.3 14.6 36.1 40.4 20.5 41.1 22.3 7.5 21.1 11.0 44.8 28.4 42.2 31.9 4.5 12.7 8.5 27.0 29 LE+LL 26.1 38.0 59.2 2004 LE+HL HE+LL 49.0 7.5 42.7 7.9 26.0 8.6 HE+HL 17.4 11.4 6.2 54.2 22.2 31.9 51.9 6.6 5.5 7.3 20.4 27.8 68.3 23.5 49.5 21.6 28.2 16.9 32.6 21.3 22.6 16.3 11.6 16.2 18.5 20.9 27.7 3.2 27.6 10.8 34.9 57.8 43.1 45.0 58.3 28.0 34.8 32.0 26.4 8.4 11.7 11.2 7.7 5.8 10.4 11.9 7.6 27.5 24.4 50.1 53.0 8.3 4.7 14.1 17.9 39 40.8 61.7 23.4 43.5 27.9 19.3 7.4 5.4 18.3 8.3 5.0 33.6 37 27.6 32.6 16.5 12.6 22.3 17.8 Copyright MORKEVIČIUS et al. (2019). Country LE+LL 22.1 51.2 2014 LE+HL HE+LL 23.5 18.8 36.0 5.5 HE+HL 35.6 7.3 LE+LL 27.8 37.4 27.6 45.2 54.9 55.6 25.4 31.9 33.3 36.1 2004 LE+HL HE+LL 28.7 15.9 54.6 1.9 56.8 3.6 40.9 7.9 36.4 5.3 30.3 8.2 48.6 6.4 44.5 8.3 40.1 10.6 19.9 24.3 HE+HL Norway 27.5 Poland 6.0 Portugal 12.0 Russia 45.7 25.4 16.3 12.7 6.0 Slovakia 57.0 22.5 13.7 6.8 3.4 Slovenia 48.1 45.3 3.0 3.7 5.8 South Africa 35.6 42.9 7.8 13.6 19.6 Spain 32.5 51.5 4.4 11.5 15.3 Sweden 23.7 32.1 15.3 29.0 16.0 Switzerland 29.6 17.7 25.8 26.9 19.6 Turkey 14.0 41.8 12.2 32.0 United States 24.5 45.7 6.6 23.2 20.7 36.5 8.3 34.6 Data source: the ISSP modules ‘Citizenship’. Notes: LE – low score on external political efficacy; HE – high score on external political efficacy; LL – low score on support for values of liberal democracy; HL – high score on support for values of liberal democracy. In 2014 the highest scores of the HE+HL indicator were in Iceland (37.1 %), Norway (35.6 %), Turkey (32.0 %), Sweden (29.0 %) and France (27.2 %). Five former post-communist countries had the lowest scores: Slovenia (3.7 %), Croatia (5.7 %), Czechia (6.7 %), Slovakia (6.8 %) and Poland (7.3 %). The scores on the scales of political efficacy and values of liberal democracy in 2004 indicate that the highest scores of the LE+HL indicator (LE – low score on external political efficacy + HL – high score on support for values of liberal democracy) were in Portugal (56.8 %), Poland (54.6 %), Israel (53.0 %), Canada (51.9 %) and Ireland (50.1 %) (see Table 15). On the opposite side of the LE+HL indicator scale are these countries: Czechia (16.9 %), Switzerland (19.9 %), Finland (21.3 %), France (22.6 %) and Japan (23.4 %). In 2014 the highest scores of the LE+HL indicator were in Spain (51.5 %), Georgia (49.8 %), United States (45.7 %), Croatia (45.4 %) and Slovenia (45.3 %). Switzerland (17.7 %), Japan (20.5 %), Germany: West (21.2 %), Germany: East (22.3 %) and Slovakia (22.5 %) compose a group of countries, where the scores were the lowest. Finally, the scores on the scales of political efficacy and values of liberal democracy in 2004 reveal that the highest scores of the HE+LL indicator (HE – high score on external political efficacy + LL – low score on support for values of liberal democracy) were in Switzerland (24.3 %), France (20.9 %), Japan (19.3 %), Finland (18.5 %) and the Netherlands (16.5 %) (see Table 15). On the opposite side of the HE+LL indicator scale are Poland (1.9 %), Portugal (3.6 %), Israel (4.7 %), Slovakia (5.3 %) and Latvia (5.4 %). In 2014 the highest scores of the HE+LL indicator one might see in Switzerland (25.8 %), Japan (22.3 %), France (22.0 %), Germany: West (19.4 %) and Norway (18.8 %). Again, the lowest 30 Copyright MORKEVIČIUS et al. (2019). scores were mainly in Central and Eastern European countries, i.e. Slovenia (3.0 %), Croatia (3.1 %), Spain (4.4 %), Lithuania (4.5 %) and Poland (5.5 %). In Table 16 we see that temporal changes with regard to external democratic efficacy are very similar to the case of the general democratic efficacy: an increase in HE+HL was most pronounced among the Western countries8. Table 16. Temporal differences (2014-2004) between shares of groups according to levels of external democratic efficacy Country Sweden Germany: West Difference (LE+LL) -9.6 -6.8 Difference (LE+HL) -8.0 -13.6 Difference (HE+LL) 4.7 7.7 Difference (HE+HL) 13.0 12.7 Germany: East Norway Switzerland Great Britain Russia Australia -10.9 -5.7 -6.5 -12.0 0.5 -0.3 -5.7 -5.2 -2.2 6.9 -15.5 -7.4 6.5 2.9 1.5 -2.1 8.4 1.7 10.0 8.1 7.3 7.2 6.7 6.0 Hungary Netherlands Finland Belgium: Flanders Czechia Slovakia -11.0 -5.2 -6.3 -5.2 -9.9 2.1 1.8 4.3 2.3 -0.5 7.5 -13.9 3.8 -3.8 -0.3 2.1 -1.1 8.4 5.4 4.7 4.2 3.6 3.5 3.4 Japan Korea (South) Poland Austria Slovenia Israel: Arabs + Jews -2.9 -0.4 13.8 9.1 -7.5 9.5 -2.9 -2.4 -18.6 -10.8 15 -8.8 3 0.1 3.6 1.2 -5.2 2.6 2.8 2.7 1.3 0.5 -2.1 -3.3 Spain 0.6 7.0 -3.9 -3.8 Denmark 6.2 0.1 -1.3 -4.9 South Africa 10.2 -5.7 1.4 -6.0 France 4.0 2.6 1.1 -7.7 United States 3.8 9.2 -1.7 -11.4 Data source: the ISSP modules ‘Citizenship’. Notes: LE – low score on political efficacy; HE – high score on political efficacy; LL – low score on support for values of liberal democracy; HL – high score on support for values of liberal democracy. Figures 3 and 4 show country differences in profiles of distributions of separate groups. In 2004 high external political efficiency was relatively more pronounced in Cyprus, United States, Denmark, and Norway. 8 However, in some Western countries, such as, Denmark, France and United States external democratic efficacy decreased substantially. 31 Copyright MORKEVIČIUS et al. (2019). Fig.3. Relative cross-country distributions of respondents into four groups according to scores on the scales of external efficacy and values of liberal democracy in 2004. In 2014 the ‘picture’ changed only moderately as the Netherlands, Norway and Iceland (not studied in 2004) joined the club of countries with a relatively large share of highly effective citizens who are strong supporters of liberal democracy. 32 Copyright MORKEVIČIUS et al. (2019). Fig.4. Relative cross-country distributions of respondents into four groups according to scores on the scales of external efficacy and values of liberal democracy in 2014. Finally, we present results of analysis of tendencies with regard to democratic internal efficacy. On the most general level, in 2004 about half (50.6 %) of respondents scored low on internal political efficacy, and in 2014 this share remained almost the same (51.7 %). Low levels of internal political efficacy are again more widespread in CEE countries. The share of respondents who scored high on both internal political efficacy and support for values of liberal democracy remained almost unchanged (28.8 % in 2004 and 28.2 % in 2014). The same was true about the share of respondents who score high on internal political efficacy and low on support for values of liberal democracy (20.5 % in 2004 and 20.2 % in 2014). Thus, internal democratic efficacy remained rather stable during the studied period (similar to the trends with regard to external democratic efficacy). The scores on the scales of internal efficacy and values of liberal democracy in 2004 reveal that the highest scores of the LE+LL indicator (LE – low score on internal political efficacy + LL – 33 Copyright MORKEVIČIUS et al. (2019). low score on support for values of liberal democracy) were in former post-communist countries, i.e. Czechia (53.5%), Latvia (51.4%), Hungary (46.0%), Japan (42.0 %) and Slovenia (41.5 %) (see Table 17). The lowest scores were in old democracies: Australia (11.7 %), Canada (13.1 %), United States (13.5 %), Denmark (13.8 %) and Israel (13.9 %). In the second wave of the ISSP module ‘Citizenship’ (2014) the highest scores of the LE+LL indicator were in Slovakia (49.3 %), Russia (45.6 %), Hungary (42.8 %), Japan (42.4 %) and Czechia (41.3 %). On the other hand, the lowest scores were again in the group of the old democracies, i.e. Australia (12.5 %), the Netherlands (13.7 %), Iceland (13.9 %), Denmark (16.0 %) and Norway (16.2 %). Table 17. Cross-country and temporal distributions of respondents in four groups according to scores on the scales of internal efficacy and values of liberal democracy, percentages. Country Australia Austria Belgium: Flanders Bulgaria Canada Croatia Cyprus Czechia Denmark Finland France Georgia Germany: East Germany: West Great Britain Hungary Iceland Ireland Israel: Arabs + Jews Japan Korea (South) Latvia Lithuania Netherlands New Zealand Norway Poland Portugal Russia Slovakia Slovenia South Africa Spain LE+LL 12.5 33.4 31.8 2014 LE+HL HE+LL 15.0 22.5 19.6 22.8 12.6 32.3 HE+HL 50.0 24.2 23.3 28.2 20.7 20.6 30.4 41.3 16.0 30.0 22.0 23.0 27.3 25.3 20.9 42.8 13.9 16.1 14.2 14.3 20.0 39.7 13.7 13.5 20.0 24.6 14.6 27.4 28.5 31.1 25.6 11.4 34.5 30.0 21.1 15.9 22.6 15.1 41.3 24.7 32.4 25.9 24.5 31.3 38.0 16.8 48.8 24.5 32.0 15.9 27.6 42.4 36.4 26.4 34.6 16.5 11.5 14.8 17.5 37.1 13.7 36.7 13.6 12.0 27.1 14.2 45.6 16.2 39.8 16.2 25.5 24.7 16.6 42.9 18.1 45.6 49.3 32.8 32.6 18.0 28.2 20.7 24.9 37.9 22.7 16.3 20.9 17.9 10.8 18.7 9.9 9.1 24.3 18.7 40.6 34 LE+LL 11.7 25.4 28.3 2004 LE+HL HE+LL 19.6 21.8 20.7 20.6 9.6 39.1 HE+HL 47.0 33.3 23.0 41.3 13.1 21.5 24.4 19.4 14.2 17.8 48.4 25.4 53.5 13.8 36.3 19.6 17.6 9.7 18.2 13.2 22.6 18.6 26.2 25.9 31.5 22.3 38.3 10.6 42.2 19.0 35.5 31.4 27.5 29.3 46.0 9.8 16.8 15.1 20.6 35.1 26.9 27.0 19.7 23.7 28.8 28.6 13.7 13.9 13.9 18.8 31.1 22.0 15.1 45.3 39.9 42.0 37.6 51.4 24.3 39.6 22.4 16.1 10.5 15.4 17.7 12.2 10.8 18.1 18.6 17.4 29.5 20.6 40.5 37.2 41.5 23.1 22.5 12.4 11.8 15.5 43.3 35.4 36.0 21.1 23.0 45.7 28.1 31.9 30.9 26.2 9.8 10.9 12.3 23.1 22.5 8.4 17.5 37.6 38.6 41.0 17.5 33.1 11.2 18.7 13.0 22.8 31.9 Copyright MORKEVIČIUS et al. (2019). Country 2014 2004 LE+LL LE+HL HE+LL HE+HL LE+LL LE+HL HE+LL HE+HL Sweden 17.7 21.4 21.0 40.0 23.3 25.5 20.6 30.6 Switzerland 25.9 13.3 29.5 31.3 32.9 14.2 27.7 25.2 Turkey 18.6 52.5 7.8 21.0 United States 17.1 25.5 14.0 43.5 13.5 19.9 15.5 51.1 Data source: the ISSP modules ‘Citizenship’. Notes: LE – low score on internal political efficacy; HE – high score on internal political efficacy; LL – low score on support for values of liberal democracy; HL – high score on support for values of liberal democracy. The scores on the scales of internal efficacy and values of liberal democracy in 2004 indicate that the highest scores of the HE+HL indicator (HE – high score on internal political efficacy + HL – high score on support for values of liberal democracy) were in the United States (51.1 %), Canada (48.4 %), Australia (47.0 %), Ireland (45.3 %) and Denmark (42.2%) (see Table 17). On the opposite side of the HE+HL indicator scale were these countries: Czechia (10.6 %), Latvia (10.8 %), Russia (11.2 %), South Korea (12.2 %) and Slovenia (13.0 %). In 2014 the highest scores of the HE+HL indicator were in Australia (50.0 %), Iceland (48.8 %), the Netherlands (45.6 %), the United States (43.5 %) and Norway (42.9 %). On the other hand, Slovakia (9.1 %), Russia (9.9 %), Lithuania (14.2 %), Japan (14.8 %) and Czechia (15.1 %) composed a group of counties, where the scores were the lowest. Regarding the score of the LE+HL indicator (LE – low score on internal political efficacy + HL – high score on support for values of liberal democracy), in 2004 the highest scores were in South Africa (45.7 %), Poland (43.3 %), South Korea (39.6 %) Russia (36.0 %) and Portugal (35.4 %) (see Table 17). On the other hand, in Belgium: Flanders (9.6 %), Czechia (9.7 %), Germany: East (9.8 %), New Zealand (11.8 %) and the Netherlands (12.4 %) these scores were the lowest. In 2014 the highest scores of the LE+HL indicator were in Turkey (52.5 %), Georgia (39.7 %), South Africa (37.9 %), Lithuania (36.7 %) and South Korea (34.6 %). And these were the five countries with the lowest scores: Belgium: Flanders (12.6 %), Switzerland (13.3 %), Germany: West (13.5 %), the Netherlands (13.6 %) and Germany: East (13.7 %). Finally, the scales of internal efficacy and values of liberal democracy in 2004 reveal that the highest scores of the HE+LL indicator (HE – high score on internal political efficacy + LL – low score on support for values of liberal democracy) were in Belgium: Flanders (39.1 %), Germany: East (35.1 %), the Netherlands (31.9 %), Finland (31.5 %), and New Zealand (30.9 %) (see Table 17). On the opposite side of the HE+LL indicator scale were South Africa (8.4 %), Poland (9.8 %), South Korea (10.5 %), Portugal (10.9 %) and Russia (12.3 %). In 2014 the highest scores of the HE+LL indicator were again in countries of Western Europe: Germany: East (34.5 %), Belgium: Flanders (32.3 %), Finland (31.1 %), Germany: West 35 Copyright MORKEVIČIUS et al. (2019). (30.0 %) and Switzerland (29.5 %). And the lowest scores were these countries: Turkey (7.8 %), South Africa (10.8 %), Georgia (11.4 %), South Korea (11.5 %) and Lithuania (12.0 %). Table 18. Temporal differences (2014-2004) between shares of groups according to levels of internal democratic efficacy Country Slovenia Great Britain Sweden Spain Difference (LE+LL) -8.7 -8.4 -5.6 -4.5 Difference (LE+HL) 1.9 4.9 -4.1 -5.4 Difference (HE+LL) -4.6 -5.9 0.4 1.2 Difference (HE+HL) 11.3 9.4 9.4 8.7 Netherlands Switzerland Finland Korea (South) Czechia Hungary Australia -4.4 -7 -6.3 -1.2 -12.2 -3.2 0.8 1.2 -0.9 1.1 -5 6.4 4 -4.6 -4.8 1.8 -0.4 1 1.2 -3.8 0.7 8 6.1 5.7 5.3 4.5 3.1 3 Germany: West Norway Germany: East Poland Belgium: Flanders Denmark -2.2 -1.2 -4.1 10.3 3.5 2.2 -3.3 0.7 3.9 -17.8 3 -4 3.1 -1.5 -0.6 6.8 -6.8 2.6 2.5 1.9 0.8 0.6 0.3 -0.9 Russia Japan France South Africa United States 5.1 0.4 2.4 9.5 3.6 -7.8 2.1 -2.6 -7.8 5.6 4 0.4 3.3 2.4 -1.5 -1.3 -2.9 -3.1 -4.1 -7.6 Austria 8 -1.1 2.2 -9.1 Slovakia 12.1 -0.4 -2.2 -9.6 Israel: Arabs + Jews 10.6 0.9 0.8 -12.3 Data source: the ISSP modules ‘Citizenship’. Notes: LE – low score on political efficacy; HE – high score on political efficacy; LL – low score on support for values of liberal democracy; HL – high score on support for values of liberal democracy. In Table 18 we see that temporal changes with regard to internal democratic efficacy are again similar to other forms of democratic efficacy: an increase in HE+HL was most pronounced among the Western countries (with the exception of the United States, Austria and Israel). 36 Copyright MORKEVIČIUS et al. (2019). Fig.5. Relative cross-country distributions of respondents into four groups according to scores on the scales of internal efficacy and values of liberal democracy in 2004. Figures 5 and 6 show country differences in profiles of distributions of separate groups. In 2004 high internal democratic efficiency was relatively more pronounced in a number of Western countries: The United States, Australia, Ireland, Denmark, Norway, France, Cyprus and Austria. 37 Copyright MORKEVIČIUS et al. (2019). Fig.6. Relative cross-country distributions of respondents into four groups according to scores on the scales of internal efficacy and values of liberal democracy in 2014. In 2014 the ‘picture’ changed somewhat as high internal political efficiency was relatively more pronounced in Australia, Iceland, Spain, Sweden, Norway, the Netherlands and Great Britain. To summarise, a majority of the populations in the studied countries scored low on both scales (internal, a little more than 50 %, and external, a little more than 70 %) of political efficacy. Low levels of political efficacy are more widespread in CEE countries. Moreover, these low levels of political efficacy are rather stable over time if looked at the aggregate level. However, at the country level there were certain changes, as in some (mostly Western) countries political efficacy increased from 2004 to 2014. Finally, support for values of democracy is also rather stable over time. However, it is quite low, as only about half of the populations in the studied countries are strong supporters of the values of liberal democracy. 6. Conclusions 38 Copyright MORKEVIČIUS et al. (2019). This paper aimed at defining and studying general trends of democratic efficacy around the world. Regarding the definition of the concept of democratic efficacy we presented two conceptualisations: 1) more extensive (holistic) coupling political efficacy with citizens’ democratic capacities and values, and 2) less extensive (narrow, minimal) coupling political efficacy with only citizens’ support for important democratic values. The first conceptualisation of democratic efficacy complements political efficacy with five types of democratic capacities and values: factual political knowledge of citizens; habits of political news consumption; citizen’s political reflexivity; support for core values of democracy (equality of interests, political autonomy and reciprocity); political or civic skills. However, even though this conceptualisation is more encompassing and fine-grained, it does not allow us to study the trends of democratic efficacy in the historical-temporal perspective, as there is no data that could be employed for this type of study. Therefore, we introduced a less extensive conceptualisation of democratic efficacy that includes only attitudinal aspects. We believe that this might be considered a minimal definition of democratic efficacy. It couples political efficacy with support for important values of democracy conceptualised according to Diamond and Morlino (2005) and including eight dimensions: five procedural (rule of law, competition, participation, horizontal and vertical accountability), two substantive (freedom and equality), and one results oriented (responsiveness). Although this definition only captures attitudinal aspect of democratic efficacy neglecting behavioural and reflexive capacities, it is nonetheless useful in cross-cultural and historical analysis employing survey data. After analysing data from the two modules of ISSP (Citizenship I conducted in 2004 and Citizenship II conducted in 2014) we found that, overall, changes over the last decades in democratic efficacy are negligible. However, there are important differences with regard to levels of internal and external political efficacy and their trends in separate countries and regions. First of all, we found that the majority of the populations in the studied countries scored low on both internal (a little more than 50 %) and external (a little more than 70 %) political efficacy. However, the difference between the two dimensions is substantial and should not be neglected in future studies of political efficacy. It is quite a usual practice in political research to study external dimension of political efficacy (beliefs about responsiveness of the political system) without including internal aspect of political efficacy (beliefs in competence to understand and participate in politics). Results of our study indicate that this might involve a risk to overlook much higher levels of internal compared to external political efficacy. Moreover, it seems that the most appropriate way of studying trends of political efficacy is to 39 Copyright MORKEVIČIUS et al. (2019). analyse the two dimensions separately, especially when political efficacy is integrated with other types of political capacities and values. With regard to support for the values of liberal democracy we found that it is moderate (only half of the studied populations showed strong support for these values) and rather stable, at least when looking at the general trends. At the country level, we see quite a lot of variation both across countries and over time. For example, support for values of liberal democracy was very high both in 2004 and 2014 in countries such as the US (71.0 % and 69.0 %, respectively) and Australia (66.7 % and 65.0 %, respectively). However, in other countries it decreased rather substantially, as for example, in Poland (from 60.7 % in 2004 to 43.5 % in 2014) and Israel (from 70.8 % in 2004 to 58.5 % in 2014). And yet in some other countries it increased substantially, as for example, in Britain (from 43.9 % in 2004 to 58.0 % in 2014) and the Netherlands (from 49.9 % in 2004 to 58.9 % in 2014). Similarly, levels of political efficacy are quite different across countries and there is substantial cross-time variation for at least half of the countries. According to our data, there is a general tendency of lower levels of political efficacy in the CEE countries. Also, in some (mostly Western) countries political efficacy increased from 2004 to 2014. These temporal changes and cross-country variation of both levels of political efficacy and support for values of liberal democracy produce yet other constellations of cross-time and cross-country differences in democratic political efficacy. For example, levels of democratic efficacy are markedly lower than those of democratic internal efficacy (around 1/6 compared to more than 1/4). Thus, these differences need to be further studied with multilevel models including different macro (country) level explanatory variables. Among such variables, influence of type of political, educational and media institutions, socio-economic development and cultural values should be studied. 40 Copyright MORKEVIČIUS et al. (2019). References Acock, A.; Clarke, H.D., Stewart, M.C. (1985). A New Model for Old Measures: A Covariance Structure Analysis of Political Efficacy. Journal of Politics, 47(4): 1062-1084. Arendt, P.H.; Kohn, J. (1968). Between Past and Future: Eight Exercises in Political Thought. New York: Viking Press. Balch, G.I. (1974). Multiple Indicators in Survey Research: The Concept ‘Sense of Political Efficacy’. Political Methodology, 1(2): 1-43. Bene, M; Boda, Z. (2020). Conceptualizing Democratic Efficacy. DEMOS Working Paper, https://openarchive.tk.mta.hu/419/1/Democratic%20Efficacy_31-03-2020.pdf Blais, A. (2010). Political Participation. In LeDuc, L.; Niemi, R.G.; Norris, P. (eds.). Comparing Democracies 3. London: Sage, pp. 165-183. Brettschneider, C. (2006). The Value Theory of Democracy. Politics, Philosophy & Economics, 5(3): 259-278. Bühlmann, M.; Kriesi, H. (2013). Models for Democracy. In Kriesi, H.; Bochsler, D.; Matthes, J.; Lavenex, S.; Bühlmann, M.; Esser, F. (eds.). Democracy in the Age of Globalization and Mediatization. Basingstoke: Palgrave Macmillan, pp. 44-68. Campbell, A.; Gurin, G.; Miller, W.E. (1954). The Voter Decides. Evanston: Row Peterson. Coppedge, M. (2004). Quality of Democracy and Its Measurement. In O’Donnell, G.; Cullell, J.V.; Iazzetta, O.M. (eds.). The Quality of Democracy: Theory and Applications. Notre Dame: University of Notre Dame Press, pp. 239-248. Craig, S.C.; Maggiotto, M.A. (1982). Measuring Political Efficacy. Political Methodology, 8(3): 85-109. Craig, S.C.; Niemi, R.G.; Silver, G.E. (1990). Political Efficacy and Trust: A Report on the NES Pilot Study Items. Political Behavior, 12(3), 289-314. Dahl, R.A. (1956). A Preface to Democratic Theory. Chicago: University of Chicago Press. Dahl, R.A. (1971). Polyarchy: Participation and Opposition. New Haven: Yale University Press. Dahl, R.A. (1989). Democracy and Its Critics. New Haven: Yale University Press. Delli Carpini, M.X.; Keeter, S. (1996). What Americans Know about Politics and why it Matters. New Haven: Yale University Press. Dewey, J. (1954). The Public and Its Problems. Athens: Swallow Press. Diamond, L.; Morlino, L. (2005). Introduction. In Diamond, L.; Morlino, L. (eds.). Assessing the Quality of Democracy. Baltimore: Johns Hopkins University Press, pp. ix-xliii. 41 Copyright MORKEVIČIUS et al. (2019). Lakoff, S. (1990). Autonomy and Liberal Democracy. The Review of Politics, 52(3): 378-396. Lane, R.E. (1959). Political Life: Why People Get Involved in Politics. Glencoe: The Free Press. Marx, P.; Nguyen, C. (2016). Are the Unemployed Less Politically Involved? A Comparative Study of Internal Political Efficacy. European Sociological Review, 32(5): 634-648. Mutz, D.C. (2006). Hearing the Other Side: Deliberative versus Participatory Democracy. Cambridge: Cambridge University Press. Niemi, R.G.; Craig, S.C.; Mattei, F. (1991). Measuring Internal Political Efficacy in the 1988 National Election Study. American Political Science Review, 85(4): 1407-1413. O’Donnell, G.; Cullell, J.V.; Iazzetta, O.M. (eds.) (2004). The Quality of Democracy: Theory and Applications. Notre Dame: University of Notre Dame Press. Petersen, M.B.; Skov, M.; Serritzlew, S.; Ramsøy, T. (2013). Motivated Reasoning and Political Parties: Evidence for Increased Processing in the Face of Party Cues. Political Behavior, 35(4): 831-854. Przeworski, A. (1999). Minimalist Conception of Democracy: A Defense. In Shapiro, I.; Hacker-Cordon, C. (eds.). Democracy’s Value. Cambridge: Cambridge University Press, pp. 23-55. Schmitter, P.C.; Karl, T.L. (1991). What Democracy Is... and Is Not. Journal of Democracy, 2(3): 75-88. Schumpeter, J.A. (1942). Capitalism, Socialism and Democracy. New York: Harper & Brothers. Shapiro, R.Y.; Bloch-Elkon, Y. (2008). Do the Facts Speak for Themselves? Partisan Disagreement as a Challenge to Democratic Competence. Critical Review, 20(1–2): 115139. Valentino, N.; Gregorowicz, K.; Groenendyk, E. (2009). Efficacy, Emotions and the Habit of Participation. Political Behavior, 31: 307-330. Verba, S.; Schlozman, K.L.; Brady, H. (1995). Voice and Equality: Civic Voluntarism in American Politics. Cambridge: Harvard University Press. de Zúñiga, H.G.; Diehl, T.; Ardévol-Abreu, A. (2017). Internal, External, and Government Political Efficacy: Effects on News Use, Discussion, and Political Participation. Journal of Broadcasting & Electronic Media, 61(3): 574-596. 42