The Impact of Democratic Value Orientations on Regime Support in Democratic, Hybrid and Authoritarian Regimes

Transcription

1 The Impact of Democratic Value Orientations on Regime Support in Democratic, Hybrid and Authoritarian Regimes Marlene Mauk Department of Political Science, Johannes Gutenberg University, Mainz, Germany Jakob-Welder-Weg 12, Mainz, Germany Paper prepared for the panel Effects of Values and Context on Political Attitudes, section Public Opinion, Political Attitudes and Values, ECPR General Conference, Montréal, August 26-29, *Draft. Please do not quote without author s permission*

2 The Impact of Democratic Value Orientations on Regime Support in Democratic, Hybrid and Authoritarian Regimes Political values are often conceptualized as pre-determining more specific political attitudes such as regime support. Based on the assumption of internal attitudinal coherence, democratic value orientations should have a decidedly negative influence on regime support in non-democratic contexts, while a positive relationship is more likely in democratic settings. This paper systematically explores the relationship between democratic value orientations and regime support in a total of 45 democratic, hybrid, and authoritarian regimes based on the latest available round of the Afrobarometer ( ) and Asian Barometer ( ). On the theoretical level, individual regime perceptions, rather than actual regime context, are identified as a moderating variable. Empirically, a consistently negative impact of democratic value orientations on regime support is found. Conforming to the expectation, the strength of this relationship appears to be dependent of regime perception: It becomes bigger the less democratic the regime is perceived as. However, this interaction turns statistically insignificant as soon as procedural performance evaluations are controlled for. Keywords: Africa, Asia, Comparative Politics, Political Culture, Quantitative, Values Acknowledgements: I would like to thank Ayse Gün, who has been a great help in preparing the merged data file and editing the results of the empirical analyses. Introduction Political culture research has long conceptualized support for the political system into several distinct, yet interrelated dimensions that can broadly be categorized as support for political values, regime structures, and political personnel. Of these dimensions, political value orientations are seen as the most enduring concept, pre-determining more specific political attitudes such as support for the political regime and its authorities. Political value orientations and attitudes towards the political regime should thus be 1

3 internally coherent. Based on this proposition, democratic value orientations should have a decidedly negative influence on regime and authorities support in nondemocratic regime contexts, while a positive relationship is more likely although, as the critical-citizens literature suggests, not granted in democratic settings. However, a growing body of research reveals both rising support for the value of democracy as well as continuingly high support for incumbent non-democratic regimes and authorities, challenging this very proposition of internal attitudinal coherence. This paper thus seeks to put this proposition to a direct empirical test, analysing the relationship between political value orientations and regime support in democratic, hybrid, and authoritarian regimes. Using comparative public opinion data collected by the Afrobarometer ( ) and Asian Barometer ( ) 1 survey projects, the relationship between orientations towards democratic values and support for the actual incumbent regime is systematically explored in a total of 45 countries, covering a variety of regime contexts from liberal democracies to closed authoritarian regimes. The analysis focusses on the following main question: Are the direction and strength of the relationship between political value orientations and regime support dependent on (perceived) regime contexts? 1 Data analysed in this paper were collected by the Asian Barometer Project ( ), which was co-directed by Professors Fu Hu and Yun-han Chu and received major funding support from Taiwan s Ministry of Education, Academia Sinica and National Taiwan University. The Asian Barometer Project Office ( is solely responsible for the data distribution. The author appreciates the assistance in providing data by the institutes and individuals aforementioned. The views expressed herein are the author s own. 2

4 Theoretical framework A generalized model of political support In line with the classical works of Gabriel Almond and Sidney Verba (1963) as well as David Easton (1965, 1975), political support is defined here as citizens attitudes towards the political system. Following Dieter Fuchs (2002, ), the political system is conceptualized in three hierarchically structured levels (Figure 1): values, structure, and process. While the values level contains the basic political values and principles a political regime may be built upon, 3 the structure level encompasses only the actual institutional structure implemented in a given country, and the process level is comprised of the political authorities and their actions. In accordance with the definition of political support as citizens attitudes towards the political system, specific attitudes are then related to these three levels of the political system (Figure 1): On the values level, citizens are committed to certain political values which may or may not be incorporated in their political regime s institutional structure ( political value orientations ). 4 On the structure level, citizens attitudes towards this actual institutional structure are reflected as support for the incumbent political regime ; likewise, on the process level, support for the political authorities results from citizens evaluations of 2 Even though Fuchs himself limits the scope of his model to democratic political systems only, the author argues that vital parts of it can be generalized to apply in both democratic and nondemocratic contexts. 3 As these values are not actual manifest objects, but abstract reference points, all political values are always present in any kind of political system they may just not be incorporated in its institutional structure. This sets them apart from the actual institutional structure of a regime and the political authorities filling its roles. 4 This means that even in authoritarian regimes, citizens can hold democratic value orientations, and vice versa. 3

5 the political authorities and their actions on the process level. In this model, then, our dependent variable of regime support is situated on the structure level, while our central independent variable of (democratic) value orientations can be located on the values level and thus accurately separated from the dependent variable. 5 As has been implicitly theorized by Easton (1965: ; 1975: ) and explicitly by Fuchs (2002: 37-38; 2009: 165), the different levels of political support are not independent of but influence each other (Figure 1). On the one hand, an overflow of values can be expected from the upper to the hierarchically lower levels: People will more likely support the incumbent political regime if its institutional structure conforms to their value orientations; likewise, if the values held by the population are not reflected in the institutional structure of the regime, support for it should dwindle. Similarly, the political structure of the regime as well as the dominant value orientations determines what citizens expect from the political authorities, and incumbents acting against the norms set by the political values and the regime structure will rarely find the support of citizens committed to them. On the other hand, the opposite effect, a generalization of experiences from the lower to the upper levels, can also occur: As people accumulate experiences with the political authorities, support for the political regime these authorities represent can be 5 Strictly speaking, our construct of political value orientations can have varying values which could be placed on a continuous scale ranging from clearly democratic to clearly authoritarian, brushing various stages of hybrid or ambiguous orientations. In the same way, however, we could conceptualize our construct as expressing varying degrees of democratic value orientations, with the values ranging from very high to very low/none. Political value orientations and democratic value orientations are therefore conceptually the same. 4

6 inferred, and sustained positive experiences with the regime itself will ultimately lead to the adoption of the political values this regime is built upon. Figure 1: A generalized model of political support levels attitudinal constructs values political value orientations structure support for the incumbent political regime process support for the political authorities overflow of values generalization of experiences Democratic value orientations and regime support in different contexts A causal relation between political value orientations and regime support can thus be directly derived from the model of political support. In general, as long as the institutional structure of the regime conforms to the basic political values held by its population, support for this regime should be high as well regardless of context, i.e. regime type (similarly, Peralta 2013: 481). This, however, makes it impossible to specify a direction of this relation: Political value orientations cannot take high or low values, but only democratic or authoritarian ones (or anything in between, of course), and so their relation to regime support cannot be described as positive or negative. To allow for this expression, therefore, we need to specify political value 5

7 orientations as democratic value orientations 6 with regard to the formulation of hypotheses and the subsequent empirical analysis. This specification, however, does make the direction of the causal relation context-dependent: In hybrid and authoritarian regimes, i.e. a nondemocratic context, the apparent violation of democratic principles by the incumbent regime should result in a decidedly negative relation between democratic value orientations and regime support. On the contrary, in regimes whose institutional structure is based upon democratic principles i.e., democratic contexts a positive relation between democratic value orientations and support for the incumbent regime could be assumed. Yet if, as I have argued above, we theorize political value orientations as shaping the expectations citizens have of their political regime, the opposite effect could also occur: Since no actual political regime is ever capable of meeting ideal democratic standards (cf. Dahl 1989), even the most democratic regime may fall short of the standards derived from its citizens democratic value orientations and hence evoke more critical evaluations and consequently less regime support. 7 Following this line of argument, we could therefore expect a negative effect of democratic value orientations on regime support in all types of political regime contexts. There is yet another caveat to take into account: The whole proposition outlined in the preceding paragraph rests on the assumption that citizens do actually perceive their regime as conforming or objecting to their basic political values correctly, i.e. have a regime perception that is true to reality. This supposition, however, is highly questionable as evidenced by a number of studies especially in non-democratic regimes 6 Of course, a specification as authoritarian value orientations would be just as possible. 7 This line of argument is prominent in the critical-citizens literature (see, e.g., Dalton 2000, 2004: 67-68; Norris 1999a: 24-25, 2011: 31). 6

8 (e.g., Shin 2013: 23-27), but also in established democracies (cf. Norris 1999b). To account for the possibility of misperceptions, we need to slightly reformulate the basic supposition as follows: The relationship between democratic value orientations and regime support is not directly dependent upon the type of incumbent regime, but upon the perception citizens have of their incumbent regime. 8 We would therefore expect an interactive effect between democratic value orientations and regime perception (instead of actual regime type) when trying to explain regime support (Figure 2). Figure 2: The relationship between democratic values, regime perception, and regime support regime perception democratic value orientations support for the incumbent regime Alternative sources of regime support Apart from democratic value orientations, prior literature lists a variety of other potential sources of regime support, which can be categorized into two main strands of argument: the culturalist or socialization approach and the performance approach. Culturalists argue that enduring values, which are formed mainly during the socialization process and are dependent on the macro-cultural and -social context surrounding the individual, influence citizens attitudes towards their political regime 8 Of course, assuming at least a loose relationship between actual regime type and regime perceptions, the relationship between democratic value orientations and regime support will not be completely independent of regime type (for empirical evidence, see Ariely 2015 (who does, however, show that the relationship is in fact rather loose!); Bratton, Mattes, and Gyimah-Boadi 2005: 94-96) 7

9 (Eckstein 1988; Mishler and Rose 2002: 7; Peralta 2013: ; Wang/Tan 2013: 200). 9 Opinions diverge on whether these cultural values are subject to change over time or not: While the culturist school (e.g., Pye 1985) posits a lasting and basically unchangeable influence of certain cultural traditions such as Asian values on political values, the universalist school (e.g., Inglehart 1997; Inkeles/Smith 1974) claims that socioeconomic modernization will lead to value change and ultimately erode any type of undemocratic culture. Integrating this strain of thought into our model of political support, cultural values can be thought of as pre-determining political value orientations which in turn pre-determine support for more specific political objects such as the incumbent regime and thus regime support (see also Chu/Welsh/Chang 2013: ). The second main strand of literature follows another line of argumentation, viewing the performance of a political regime as the decisive factor for its support among the population. The classical democratic-legitimacy or institutionalist approach locates this performance at the input side of the political system, claiming that the institutional structure of a political regime determines the extent to which it can perform well on criteria like accountability and that, by definition, democratic regimes are inherently superior to nondemocratic institutional arrangements in this respect (Dahl 2006; Fukuyama 1992; Shapiro 2003). 10 More recent approaches instead focus on the 9 The notions that values are an enduring concept, are formed mainly during the socialization process, and ultimately influence regime support do also hold true for our central independent variable, democratic value orientations. In this way, our argument can also be characterized as standing in the tradition of the culturalist approach. 10 This argument can be spun further in distinguishing not only between regime types, but also sub-types such as majoritarian versus consensus democracies: Basically, it rests on the assumption that citizens will support their regime more strongly if they are part of the majority, i.e. in democracies on the side of the election winners. Since institutional structures such as electoral rules determine how many losers and winners there will be and 8

10 output side of the political system, positing that regime support is much more dependent upon economic performance (substantive performance) and good governance (procedural performance) than on classical legitimacy derived from democratic institutional structures (Dalton 2004: 63-67; McAllister 1999: ; Norris 2011: ; Rothstein 2009). In this view, results matter more than procedures and thus nondemocratic regimes may in fact generate more support than democratic ones if they manage to provide either economic well-being or/and high-quality political output and administrative services 11 to their citizens (Mishler and Rose 2002: 8; Wang/Tan 2013: 206). Figure 3 provides an overview of these main approaches and how different sources may influence regime support as well as interact with each other. In addition, a myriad of other potential sources and influences such as social capital or media effects has been discussed in the literature; for a very cursory overview, see Gilley 2006: how freely the winners will be able to rule over the losers, they make a difference for the overall level of regime support (Anderson and Guillory 1997: 66-69; Norris 1999b: ; Peralta 2013: 481, ). 11 E.g., transparent and responsive government, but also more substantial outcomes such as the protection of the environment or a working social security system (Wang/Tan 2013: 206). 9

11 Figure 3: A (simplified) theoretical model of regime support cultural context cultural value orientations (culturalist) democratic value orientations political context regime perception (input performance) good governance (output performance) regime support economic context economic performance (output performance) 10

12 Prior research Levels of regime support and its sources have been a central topic in political culture research for many decades and the body of prior research is far too vast to discuss here in its entirety. I will therefore concentrate solely on studies examining the relation between political value orientations and regime support, focusing on works of a comparative nature and especially those that distinguish between different regime types. Even though a wide variety of individual- and country-level determinants of regime support has been examined empirically, e.g. economic and political performance (e.g., Armingeon and Guthmann 2014; Dalton 2004; Gibson 1996; McAllister 1999; Miller and Listhaug 1999; Norris 2011), socioeconomic modernization (e.g., Bratton, Mattes, and Gyimah-Boadi 2005; Mattes and Bratton 2007), social capital (e.g., Dalton 2004; Newton and Norris 2000), corruption (e.g., Chang 2013), media consumption (e.g., Booth and Seligson 2009; Geddes and Zaller 1989), institutional characteristics (e.g., Anderson and Guillory 1997), policy preferences (e.g., Citrin, Levy, and Wright 2014), generational differences (e.g., Mujani and Liddle 2013), migration experiences (Careja and Emmenegger 2012), political participation (e.g., Quintelier and van Deth 2014), ethnic group membership (Ruiz-Rufino 2013), or political knowledge (e.g., Norris 2011), surprisingly few studies exists that analyze the relationship between political value orientations and regime support. Those analyses that do incorporate political value orientations as an independent or at least a control variable almost unanimously find negative effects of democratic value orientations on regime support (Chang, Chu, and Welsh 2013; Chen and Dickson 2008; Chu 2011; Chu, Welsh, and Chang 2013; Huang, Chu, and Chang 2013; Ikeda 2013; Ma and Yang 2014; Park 2013; Park and Chang 2013; Tang and Huhe 2014; Wu 11

13 and Chu 2007). 12 As prior research shows, this negative relation is present under both non-democratic and democratic regime contexts (e.g., Chang, Chu, and Welsh 2013; Chu, Welsh, and Chang 2013; Ma and Yang 2014; Park 2013; Park and Chang 2013; Wu and Chu 2007). This corroborates the argument made above that stronger support for democratic values leads to higher expectations and consequently lower evaluations of the incumbent regime even if it is in fact a democratic one. Even though the majority of prior studies is not confined to either democratic or non-democratic regimes, only a handful differentiate between different regime contexts. These first results indicate that the impact of democratic value orientations on regime support is comparatively stronger in hybrid and authoritarian regimes than in democratic ones (Chang, Chu, and Welsh 2013; Chu, Welsh, and Chang 2013; Park 2013), hinting at a possible interaction effect between regime type and democratic value orientations. None of the existing research, however, examines the more direct relation between regime perception, democratic value orientations, and regime support proposed in the theoretical model above. The comparative strength of the impact of democratic values on regime support varies from non-significant and weak effects (e.g., Mattes and Shin 2005; Park and Chang 2013) to strong and dominant effects (e.g., Chen and Dickson 2008; Huang, Chu, and Chang 2013), depending on the countries surveyed, the operationalization of the dependent and independent variables, and the other variables included in the multivariate models. Yet, most studies agree that indicators of economic and political performance exert a stronger influence on regime support than political value 12 The only deviating results show positive, but very weak effects in South Africa (Mattes and Shin 2005) and Haiti and the Dominican Republic, respectively (Stoyan et al. 2014). 12

14 orientations (e.g., Chang, Chu, and Welsh 2013; Mattes and Shin 2005; Park 2013; Stoyan et al. 2014). To summarize, prior research on the relationship of democratic value orientations and support for the incumbent regime contains several key findings: First, democratic value orientations appear to be negatively correlated with regime support regardless of regime context; second, the relative explanatory strength of democratic value orientations is higher in non-democratic than in democratic regime contexts; and third, compared to political and economic performance, democratic value orientations exert a rather weak influence on regime support. This research, however, is severely limited and thus inconclusive with regard to the research question of how democratic value orientations influence regime support: On the one hand, almost all empirical analyses are confined to the region of East Asia (the only exceptions being Mattes and Shin 2005 and Stoyan et al. 2014) and hence their results cannot simply be generalized to other parts of the world. On the other hand, the moderating effect of regime perception is neglected completely: At the most, prior analyses compare democratic value orientations impact on regime support across different groups of countries (democratic, hybrid, and authoritarian) which is both theoretically as well as methodologically inadequate. Hypotheses As has been outlined above, a negative effect of democratic value orientations on regime support can definitely be assumed in nondemocratic regime contexts, and is likely even in democratic ones. Prior empirical analyses indeed confirm this proposition, so that a negative relationship can be expected in all regime contexts. 13

15 Hypothesis 1: Democratic value orientations have a negative influence on regime support. Building on the theoretical model, we can further assume that this negative influence increases when citizens perceive a larger discrepancy between their value orientations and their regime s institutional structure. We would therefore expect that the strength of the effect of democratic value orientations is dependent upon the perception citizens have of their incumbent regime and, more specifically, that the effect size will increase when citizens perceive their regime as less democratic. Hypothesis 2: The strength of democratic value orientations influence on regime support is dependent upon the perception of the incumbent regime. Hypothesis 2a: The more democratic the incumbent regime is perceived as, the less pronounced is the negative influence of democratic value orientations on regime support. As far as the relative strength of the correlation between democratic value orientations and regime support is concerned, the theoretical model does not allow for any specifications. However, prior research has shown that, while the absolute influence of democratic value orientations can vary depending on the research design, it is usually a weaker predictor of regime support than economic and political performance variables. In addition, prior analyses suggest that this rank order is fairly constant across regime types and thus most likely independent of regime perceptions, too. Hypothesis 3: The relative influence of democratic value orientations on regime support is smaller than the influence of economic and political performance variables. Hypothesis 4: The relative strength of democratic value orientations influence of regime support is independent of the perception of the incumbent regime. 14

16 Data and measurement This article employs data from the third wave of the Asian Barometer Survey ( ) and from the fifth wave of the Afrobarometer Survey ( ). Both of these surveys are regional, cross-national projects, and face-to-face interviews have been conducted on representative samples in almost fifty countries across Africa and East Asia, covering regimes ranging from consolidated liberal democracies to closed authoritarian one-party states. 13 In total, the sample contains 21 democratic, 19 hybrid and five authoritarian regimes. Countries of the Arab Spring (Egypt and Tunisia) are excluded from the sample due to their drastic and sometimes frequent changes in regime type during the last few years, which make it difficult for respondents to unambiguously identify the reference object for regime support. For the same reason, Madagascar was excluded as during 2013, it was in the transition phase from a military dictatorship/transitional government to a multiparty hybrid regime. Both the Afrobarometer as well as the Asian Barometer Survey include a large variety of items tapping into respondents political attitudes and values. For the dependent variable of regime support, two established measures are available from both surveys: on the one hand, the often used, even though highly disputed (cf. Canache, Mondak and Seligson 2001; Linde and Ekman 2003) satisfaction with democracy indicator, and on the other hand several items gauging institutional trust. As regime support has been conceptualized above as support for the institutional structure of the incumbent political regime, the latter measure is chosen. Regime support is thus 13 Regimes are classified according to Larry Diamond s regime typology (Diamond 2002). Employing Diamond s qualitative criteria, the political regimes under analysis are classified as democratic (21 countries), hybrid (19 countries), or authoritarian (5 countries). A complete list of countries and their classification can be found in the appendix (Table ). 15

17 measured by a total of six questions asking for respondents confidence in the president or prime minister (trustpm), the national government or ruling party 14 (trustgov), the parliament (trustparl), the courts of law (trustcourts), the police (trustpolice), and the armed forces (trustmilitary) (Table 1). These six items are aggregated into two indices measuring trust in more politicized institutions (parliament, president, government) and less politicized institutions (courts, police, military), respectively, as well as a composite index of institutional trust. This dimensional structure is confirmed by CFA. Democratic value orientations are also measured in both surveys, although not with identical questions (Table 1). However, central dimensions of (liberal) democratic values are covered in both surveys: vertical accountability, political equality, political pluralism, and the rule of law. Table 1 lists the respective items. Dimensionality analyses (CFA) confirm that all democratic values items can be conceptualized as measuring a single latent construct and thus a composite index is constructed. The intervening variable of regime perception is measured by an index comprised of three items asking respondents whether they perceived their regime as democratic, to place their current system of government on a 10-point scale ranging from completely undemocratic to completely democratic, and how free and fair the last national elections were. To control for alternative sources of regime support, traditional cultural values are captured by a question tapping into the preference of male versus female children. Output or systemic performance of the political system is measured by two indices: one 14 While the Asian Barometer asks directly for trust in the national government, the Afrobarometer asks for trust in the ruling party. Even though not exactly the same, both items are seen as interchangeable since it is the ruling party which forms the government and it is most likely the government which is thought of by respondents when asked about the ruling party. 16

18 for procedural performance (good governance), aimed at the perceived adherence of government officials to the law, and one for substantial (economic) performance. In addition, sociodemographic standard variables such as age, gender, and education, are controlled for. Table 1: Operationalisation of dependent and central independent variables Asian Barometer Afrobarometer item question wording scale question wording scale dependent variable: regime support trustpm I m going to name a number of institutions. For each one, please tell me how much trust do you have in them? the president/prime a great deal of minister trust - quite a lot of trust - not very much trust - no trust at all trustgov the national government the ruling party trustparl parliament parliament trustcourts the courts courts of law trustpolice the police the police trustmilitary the military the army independent variable: political values vertical2 equality1 pluralism3 Government is our employee, the people should tell government what needs to be done vs. The government is like a parent, it should decide what is good for us Women should not be involved in politics as much as men Multiple parties compete to represent political interests vs. One party represents the interests of all the people intervening variable: regime perception rp1 In your opinion, how much of a democracy is [country]? agree strongly with s1 - agree with s1 - agree with s2 - agree strongly with s2 strongly agree - somewhat agree - somewhat disagree - strongly disagree agree strongly with s1 - agree with s1 - agree with s2 - agree strongly with s2 a full democracy a democracy, but with minor problems a How much do you trust each of the following, or haven t you heard enough about them to say? the president not at all - just a little - somewhat - a lot The government is like a parent. It should decide what is good for us vs. The government is our employee. We are the bosses and should tell government what to do Men make better political leaders than women, and should be elected rather than women vs. Women should have the same chance of being elected to political office as men Political parties create division and confusion; it is therefore unnecessary to have many political parties in [country] vs. Many political parties are needed to make sure that [citizens] have real choices in who governs them In your opinion, how much of a democracy is [country] today? agree very strongly with s1 - agree with s1 - agree with s2 - agree very strongly with s2 agree very strongly with s1 - agree with s1 - agree with s2 - agree very strongly with s2 agree very strongly with s1 - agree with s1 - agree with s2 - agree very strongly with s2 not a democracy a democracy, with major problems a 17

19 rp2 rp4 control variables cultural1 lawfulness1 lawfulness2 corruption1 eco1 eco3 Where would you place our country under the present government? On the whole, how free and fair would you say the last national election was? If you could only have one child, a boy is more preferable Do officials who commit crimes go unpunished? How often do you think government leaders break the law or abuse their power? How widespread do you think corruption and bribe-taking are in the national government? How would you rate the overall economic condition of our country today? What do you think will be the state of our country s economic condition a few years from now? democracy, with major problems not a democracy completely undemocratic (1) completely democratic (10) completely free and fair free and fair, but with minor problemsfree and fair, with major problems not free and fair strongly agree strongly disagree always most of the time sometimes rarely always most of the time sometimes rarely hardly anyone is involved not a lot of officials are corrupt most officials are corrupt almost everyone is corrupt very good good soso bad very bad much better a little better about the same a little worse much worse Where would you place our country today? On the whole, how would you rate the freeness and fairness of the last national election? If funds for schooling are limited, a boy should always receive an education before a girl vs. If funds for schooling are limited, a family should send the child with the greatest ability to learn Do officials who commit crimes go unpunished? How often does the president ignore the courts and laws of this country? How many of the president and her officials are involved in corruption? In general, how would you describe the present economic condition of this country? Looking ahead, do you expect the economic conditions in this country to be better or worse in twelve months time? democracy, but with minor problems a full democracy completely undemocratic (0) completely democratic (10) not free and fair free and fair, with major problems free and fair, but with minor problems completely free and fair agree very strongly with s1 - agree with s1 - agree with s2 - agree very strongly with s2 never rarely often always never rarely often always none some of them most of them all of them very bad fairly bad neither good nor bad fairly good very good much worse worse same better much better All variables and indices were recoded so that 0 represents low values and 1 high values on the respective construct. 18

20 Results Democratic value orientations and regime support in Africa and Asia Taking a look first at a few descriptive statistics (Table 2, Table 3), we can observe that regime support appears to be fairly high across all regions and all types of regimes. In fact, regime support seems to be higher in less democratic regimes than in more democratic ones, reaching its lowest value in liberal democracies and its highest one in authoritarian regimes. Democratic value orientations, on the other hand, are distributed in accordance with the basic theorem of political culture, being more prominent among citizens of more democratic regimes. These simple descriptives also show that citizens regime perceptions are not completely independent of the actual institutional structure of the incumbent political regime, with more democratic regimes being perceived as somewhat more democratic than more authoritarian regimes; however, with the differences in regime perceptions being fairly small (0.57 in authoritarian regimes compared to 0.67 in liberal democracies), they cannot be deemed an accurate representtation of the actual level of democracy in a country. In fact, regime perceptions are only weakly correlated with actual regime types (measured as reversed FH scores; r =.18). Table 2: Descriptive analysis, split by regions Asia Africa pooled data trust in politicized institutions (18588) (45267) (63855) trust in unpoliticized institutions (18822) (46834) (65656) institutional trust (total) (18355) (44722) (63077) democratic value orientations (17567) (47236) (64803) regime perception (18349) (45501) (63850) Notes: Means. Weighted data (combinedwt). Number of cases in parentheses. Source: Asian Barometer Survey 3, ; Afro Barometer 5,

21 Table 3: Descriptive analysis, split by regime type authoritarian hybrid electoral liberal regimes regimes democracies democracies pooled data trust in politicized institutions (6571) (26706) (15393) (15185) (63855) trust in unpoliticized institutions (7997) (26886) (15629) (15144) (65656) institutional trust (total) (6521) (26309) (15253) (14994) (63077) democratic value orientations (7225) (26743) (15637) (15198) (64803) regime perception (7391) (26004) (15264) (15191) (63850) Notes: Means. Weighted data (combinedwt). Number of cases in parentheses. Source: Asian Barometer Survey 3, ; Afro Barometer 5, On a bivariate level, the expected negative correlation between democratic value orientations and regime support can be observed both in the pooled data as well as in all regime contexts (Table 4). However, the strength of this correlation varies greatly across regime types: While it is strongest in hybrid regimes, it is weakest in electoral democracies and authoritarian regimes, and somewhere in the middle in liberal democracies. This lends support to the argument made above that the effect of democratic value orientations is in fact largely independent of the actual type of regime the respondents live in and much more dependent of the perception respondents have of their regime. Table 4: Bivariate correlations between democratic values and regime support, split by regime type authoritarian regimes hybrid regimes electoral democracies liberal democracies pooled data trust in politicized institutions -.06 ** -.17 ** -.03 ** -.10 ** -.11 ** trust in unpoliticized institutions -.04 ** -.14 ** -.03 ** -.04 ** -.08 ** institutional trust (total) -.06 ** -.17 ** -.03 ** -.08 ** -.11 ** N Notes: Pearson s correlation coefficients. Weighted data (combinedwt). ** p<0.01 (two-tailed). Source: Asian Barometer Survey 3, ; Afro Barometer 5, The second part of this argument is further corroborated by the results reported in Table 20

22 5. These show that the strength of the correlation between democratic value orientations and regime support increases the less democratic the incumbent regime is perceived as. Table 5: Bivariate correlations between democratic values and regime support, split by regime perception not a democracy a democracy with major problems a democracy with minor problems a full democracy pooled data trust in politicized institutions -.12 ** -.10 ** -.09 ** -.07 ** -.11 ** trust in unpoliticized institutions -.10 ** -.06 ** -.06 ** -.06 ** -.08 ** institutional trust (total) -.13 ** -.09 ** -.08 ** -.07 ** -.11 ** N Notes: Pearson s correlation coefficients. Weighted data (combinedwt). ** p<0.01 (two-tailed). Source: Asian Barometer Survey 3, ; Afro Barometer 5, However, it needs to be noted that the overall correlations between democratic value orientations and regime support are rather small. Yet, the relationship does remain significant as well as negative when alternative explanatory factor such as performance evaluations or education are controlled for (Table 6), corroborating hypothesis 1. This holds true both in Asia as well as in Africa, although the strength of the relationship is considerably bigger among the Asian subsample (Table 7). Looking at different kinds of regimes (Table 8), it again becomes apparent that the effect of democratic value orientations on regime support is biggest in hybrid regimes and much smaller, verging on insignificance, in authoritarian regimes and liberal democracies. In electoral democracies, there seems to be no connection between democratic value orientations and regime support at all as soon as other explanatory factors are controlled for. These results also support hypothesis 3, showing that democratic value orientations have a comparatively smaller impact on regime support than both substantive as well as procedural performance evaluations (Tables 6-8). 21

23 Table 6: Multivariate regression, pooled data dependent variable: politiziced unpoliticized institutional trust B beta B beta B beta democratic values ** ** ** procedural performance ** ** ** substantive performance ** ** ** traditional values (0.00) ** 0.00 (0.00) (0.00) * age 0.00 (0.00) 0.02 ** 0.00 (0.00) 0.03 ** 0.00 (0.00) ** gender (female) (0.00) ** (0.00) ** (0.00) ** edu_primary (0.00) ** 0.00 (0.00) (0.00) * edu_secondary (0.00) ** (0.00) ** (0.00) ** edu_tertiary ** ** (0.00) ** (constant) r² N Notes: Multivariate linear regression. Weighted data (acrosswt). ** p<0.01; * p<0.05. Source: Asian Barometer Survey 3, ; Afro Barometer 5, Table 7: Multivariate regression, split by region dependent variable: institutional trust Asia Africa B beta B beta democratic values ** ** procedural performance ** ** substantive performance ** ** traditional values 0.00 (0.00) (0.00) ** age 0.00 (0.00) (0.00) 0.03 ** gender (female) 0.00 (0.00) (0.00) ** edu_primary (0.00) ** edu_secondary ** (0.00) ** edu_tertiary ** ** (constant) 0.30 (0.02) 0.28 r² N Notes: Multivariate linear regression. Weighted data (acrosswt). ** p<0.01; * p<0.05. Source: Asian Barometer Survey 3, ; Afro Barometer 5,

24 Table 8: Multivariate regression, split by regime type dependent variable: authoritarian electoral liberal hybrid regimes institutional trust regimes democracies democracies B beta B beta B beta B beta democratic values * ** (0.02) * procedural ** ** ** 0.33 performance (0.02) 0.31 ** substantive ** ** ** 0.28 performance (0.02) 0.26 ** traditional values ** (0.00) (0.00) (0.00) (0.00) * age ** ** ** 0.00 (0.00) (0.00) (0.00) (0.00) gender (female) ** ** (0.00) (0.00) (0.00) ** edu_primary ** * ** edu_secondary ** ** ** ** edu_tertiary ** ** ** ** (constant) (0.02) (0.02) r² N Notes: Multivariate linear regression. Weighted data (acrosswt). ** p<0.01; * p<0.05. Source: Asian Barometer Survey 3, ; Afro Barometer 5, A moderating effect of regime perception? Yet, despite the implications of the bivariate correlations (Table 5), hypothesis 2 is not supported by the multivariate analyses: The interaction term of democratic values x regime perception does not exert a statistically significant influence on regime support and thus a moderating effect of regime perception on the relationship between democratic value orientations and regime support cannot be deduced from the empirical data (Table 9). It does, however, point in the expected direction, indicating that at lower levels of regime perception (i.e. people thinking their regime is less democratic) the negative effect of democratic value orientations on regime support is indeed stronger than at higher levels of regime perception (i.e. people thinking their regime is more democratic). In addition, this effect becomes statistically insignificant only when 23

25 procedural performance evaluations are included as a control variable; in all other models, it does in fact show statistical significance. Table 9: Multivariate regression including interaction term dependent variable: politiziced unpoliticized institutional trust B beta B beta B beta democratic values ** ** ** procedural performance ** ** ** substantive performance ** ** ** traditional values (0.00) ** 0.00 (0.00) (0.00) * age 0.00 (0.00) (0.00) 0.02 ** 0.00 (0.00) 0.01 ** gender (female) (0.00) ** (0.00) ** (0.00) ** edu_primary (0.00) ** 0.00 (0.00) (0.00) edu_secondary (0.00) ** (0.00) ** (0.00) ** edu_tertiary (0.00) ** ** (0.00) ** regime perception ** ** ** interaction c_rp1*c_demval (0.03) (0.02) 0.00 (constant) r² N Notes: Multivariate linear regression. Weighted data (acrosswt). ** p<0.01; * p<0.05. Source: Asian Barometer Survey 3, ; Afro Barometer 5, Conclusion This paper set out to systematically examine the relationship between political value orientations and regime support across different regime contexts in order to test the proposition that political value orientations pre-determine regime support and the two attitudinal dimensions should thus be internally coherent. Building mainly on the works of Dieter Fuchs, political values were conceptualized as shaping the expectations citizens have of their incumbent regime and thus setting standards against which the actual institutional structure is judged. It was then deduced that, when specifying political value orientations as democratic value orientations, a negative relationship with regime support should be present across all regime contexts since all existing regimes were bound to fall short of the standards of a perfect democracy. As the 24

26 discrepancy between actual regime structures and democratic standards by definition grows larger the less democratic a political regime is, the strength of this effect could be thought to be dependent upon regime context. Yet, to account for the possibility that citizens perceive their political regime inaccurately, i.e. as more or less democratic than it actually is, citizens regime perceptions were identified as the moderating variable rather than actual regime type. The main hypotheses derived from the theoretical argument were subsequently tested empirically based on recent survey data from the Afrobarometer ( ) and Asian Barometer ( ). It was found that democratic value orientations are in fact negatively, albeit fairly weakly correlated to regime support regardless of the type of regime people live in or how they perceive this regime. In addition, the strength of this negative relationship appears to be indeed dependent of regime perception: It is strongest when the regime is perceived as non-democratic, and weakest when it is perceived as a full democracy. This moderation effect is also found initially in a multivariate setting, where the interaction term of democratic value orientations and regime support turns out to be positive and statistically significant. However, the interaction turns statistically insignificant as soon as procedural performance evaluations are added as a control variable. Overall, this study could therefore enhance existing research in three main respects: Firstly, on the theoretical level, it added to conceptual clarity by identifying regime perceptions instead of regime context as the decisive moderating variable. Secondly, on the empirical level, it confirmed prior results that democratic value orientations and regime support are negatively correlated even in democratic settings, and that the explanatory strength of political value orientations is weaker than that of performance evaluations. In doing so, it took an important step towards generalizability 25

27 of these results by analysing both East Asian as well as African countries, being the first study not confined to a single geographical region or just a handful of countries. Thirdly, it augmented the empirical body of research by for the first time examining the interaction effect of regime perceptions and democratic value orientations on regime support, whereas prior studies merely compared effect strengths across different groups of countries. Yet, more research needs to be done to fully understand the complex relationship between political value orientations and regime support. For one, more survey data on these attitudes is needed from other regions of the world to find out whether the results can truly be generalized to, for example, the advanced industrial democracies of Western Europe or North America. Additionally, with the data available at the moment, the key concepts of democratic value orientations and regime support could only be operationalized rather crudely; this might be one of the reasons the effect found empirically turned out so small. Theoretically, further discussion is warranted, especially with regard to the potential moderating effect of other variables such as evaluations of different types of regime performance. This could also help answer the question why the interaction between democratic value orientations and regime perception vanished after adding procedural performance evaluations as a control. Nevertheless, even at this early stage of research, implications can be drawn pertaining to both the persistence of non-democratic as well as democratic regimes. On the one hand, while the good news is that support for non-democratic regimes is negatively affected by the growing support for democratic values, this effect unfortunately is only noticeably strong if citizens actually perceive their regimes as being non-democratic. Therefore, as long as hybrid and authoritarian regimes continue to successfully trick their citizens into believing they are being governed 26