SEARCH

SEARCH BY CITATION

Keywords:

  • government effectiveness;
  • diffuse support;
  • support for democracy;
  • democratic legitimacy

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

Diffuse support for democracy, as captured in mass surveys, tends to be treated as impervious to regime performance. Such a finding is often presented as confirmation of the basic distinction between ‘diffuse’ and ‘specific’ support as proposed by David Easton. This study argues that this line of argument stems from an incomplete reading of important aspects of Easton's theorisation about the relationship between system outputs and diffuse support. Using multilevel models, evidence from more than 100 surveys in close to 80 countries, and different measures of democratic support, it is shown that government effectiveness is the strongest macro-level predictor of such support. In democratic regimes, government effectiveness, understood as the quality of policy-making formulation and implementation, is linked to higher levels of support for democracy. Furthermore, in non-democracies, effectiveness and support for democracy are, under some model specifications, negatively related.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

Why are levels of popular support for democracy higher in some countries and for some people than for others? This question has implications for the crucial issue of regime stability. Democratic consolidation is thought to hinge on the popularly shared notion that democracy is ‘the only game in town’ (Linz & Stepan 1996: 15). And although autocracies are thought to rely more for survival on coercion, privilege and growth (Haggard & Kaufman 1995; Wintrobe 1998), their strenuous efforts to control the media and education indicate that popular support is crucial there too (Geddes & Zaller 1989; Kennedy 2008). Mass disloyalty, by leading to popular mobilisation and increasing divisions within ruling coalitions, may pose even greater threats to autocratic than to democratic survival (Magaloni & Wallace 2008).

In the research on regime support, increasingly based on cross-national surveys, two central ideas prevail. First, support for democracy seems today remarkably widespread, crossing borders defined by dominant religions, cultural heritages and even regime types: ‘[I]n the last decade, democracy has become virtually the only political model with global appeal, no matter what the culture’ (Inglehart & Norris 2003: 70). Second, beyond the long-term factors (development, culture, democratic experience) that determine whatever cross-national variations remain in the popular legitimacy of democracy, domestic political or economic performance seems to have little influence. This is often presented as a confirmation of David Easton's distinction between types of system support: while performance should clearly affect specific support, directed to ‘the perceived decisions, policies, actions, utterances or the general style of … authorities’ (Easton 1975: 437), diffuse support, ‘representing as it does attachment to political objects [such as regimes] for their own sake, will not be easily dislodged because of current dissatisfaction with what the government does’ (Easton 1975: 445).

This study suggests instead that fundamental preferences about regimes are, in fact, greatly affected by performance – particularly by the effectiveness of governments, understood as the quality of policy formulation and implementation. In light of the extant literature, the suggestion may seem to verge on the heretical. However, it does not stem from a refutation of Easton's treatment of the nature of political support. Instead, it precisely recovers the way in which he himself theorised about how attitudes and actions of support for a political regime should also be seen as affected by the political system's outputs, its public policies, by what governments do and how well they do it. Besides, Easton was not alone in this reasoning: three other prominent figures of twentieth-century political science – Lipset, Dahl and Linz – similarly argued that regime legitimacy should be seen as affected by the regime's ‘effectiveness’ (Lipset 1959; Dahl 1971; Linz 1978). This argument is presented and tested here employing multilevel models that bring together data from four waves of the World Values Survey (WVS).

Effectiveness and diffuse support

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

What defines ‘democracy’, and how it should be distinguished from other regime types, is a longstanding discussion in political theory. It is clear, nonetheless, that democracy and effective governance do not overlap conceptually or empirically: ‘[G]overnability is a challenge for all regimes, not just democratic ones’ (Schmitter & Karl 2008: 85–86). The internal heterogeneity of regimes from this point of view has been confirmed in several studies (Montinola & Jackman 2002; Bäck & Hadenius 2008; Charron & Lapuente 2010). In a review of the evidence, Holmberg et al. (2009: 138) conclude that, ‘empirically, there is no straightforward relationship between establishing electoral representative democracy and QoG [Quality of Government] in the exercise of public power’.

If democracies vary in terms of ‘how well’ they are governed, such variation should be reflected in citizens' satisfaction with the way they perform. This flows from Easton's theorisation about the nature of specific support, conceived as directed to ‘the perceived decisions, policies, actions, utterances or the general style of … authorities’ (Easton 1975: 437). If citizens are able to ‘perform a rational calculation of whether the authorities' actions address their needs and demands’ (Torcal & Moncagatta 2011: 2564), specific support (the evaluation of ‘the effectiveness of the political regime’, according to Klingemann 1999) should improve if they perceive those demands to be met. There is considerable empirical support for this conjecture. Popular satisfaction with the way democracy works has been related to several important features of political performance, including the real or perceived quality of the institutions of governance (Wagner et al. 2009; Curini et al. 2012; Linde & Erlingsson 2012) and, particularly, to government effectiveness (Dahlberg & Holmberg 2012).

However, should this also be true for democratic legitimacy? At first glance, there are reasons to believe not. Easton (1975: 444) defined ‘diffuse support’ as ‘evaluations of what an object is or represents – to the general meaning it has for a person – not of what it does’. Changes in diffuse support should occur slowly as its sources are found in social learning and socialisation (Easton 1957: 395–400; 1965: 125–127; 1975: 445). ‘Outputs and beneficial performance may rise and fall while this support, in the form of a generalized attachment, continues’ (Easton 1975: 444). Indeed, studies using survey data from the WVS (Klingemann 1999; Dalton 2004), the Comparative National Elections Project (Gunther et al. 2007), the Afrobarometer (Bratton et al. 2005), the Latinobarometer (Lagos 2003) or a combination of these (Chu et al. 2008) have repeatedly found that satisfaction with democratic performance and support for democracy have different etiologies. ‘Democracy is a stable cognitive value cultivated through the socialization process in the society [and] popular belief in the superiority of democracy is not susceptible to the ups-and-downs of government performance or the short-term economic fluctuation’ (Huang at al. 2008: 56–58, 58–59). While the perceived supply of democracy is ‘an instrumental, performance driven-attitude’, support for democracy ‘is largely a principled affair’ (Mattes & Bratton 2007: 201).

There is, however, a different point of view from which these conclusions must be seen as surprising. In an article revisiting the reception to his theories in the previous decades, Easton reminded us that diffuse support should also be related to the performance of governments, if properly understood, in terms of outputs (rather than outcomes):

Diffuse support may also, however, derive from experience. If only because this is a source usually associated with specific support, its significance for diffuse support may easily be overlooked or underemphasized. Members do not come to identify with basic political objects only because they have learned to do so through inducements offered by others – a critical aspect of socialization processes. If they did, diffuse support would have entirely the appearance of a non-rational phenomenon. Rather, on the basis of their own experiences, members may also judge the worth of supporting these objects for their own sake. Such attachment may be a product of spill-over effects from evaluations of a series of outputs and of performance over a long period of time. (Easton 1975: 446; see also Easton 1965: 119–120)

Easton was not alone in this reasoning. For Lipset (1959: 86–87), legitimacy – ‘the belief that existing political institutions are the most appropriate or proper ones for the society’ – is ‘more affective and evaluative’, while effectiveness – ‘the actual performance of a political system … marked by an efficient bureaucracy and decision-making system’ – is ‘primarily an instrumental dimension’. However, he also argued that ‘even in legitimate systems, a breakdown of effectiveness, repeatedly or for a long period, will endanger its stability’ (Lipset 1959: 89). Similarly, in his Polyarchy, Dahl (1971: 144–146) argued that although popular beliefs about regimes may be a ‘more or less fixed element in the political culture of a country … acquired through early socialization’, effectiveness also mattered: ‘[I]f a government is perceived as effective, its successes are likely to enhance the prestige of the authority patterns it embodies; the converse is true if it fails’ (Dahl 1971: 149). Finally, for Linz (1978: 54), although socialisation into particular ideals played a central role in fostering democratic legitimacy, regime performance – the ability to find and implement satisfactory solutions to basic policy problems – should be seen as being in constant interplay with legitimacy: ‘[T]he lack of effectiveness weakens state authority and, as a result, its legitimacy.’1

Although the notion that regime legitimacy should be affected by effectiveness has been around for a long time, empirical support for it has seldom been sought. A few studies do come close. Aspects of institutional performance, such as individual-level evaluations of levels of corruption (Mishler & Rose 2001) or levels of confidence in the legal system (Staton & Reenock 2010) have been related to democratic support. Even closer to our argument, Linde (2012) shows that, in ten new democracies, perceptions of fair and impartial treatment by public authorities are correlates of democratic support. However, these studies only indirectly address the relationship between effectiveness and support suggested by Easton. They test their hypotheses through contemporaneous correlations between attitudinal variables, raising problems of potential reciprocal causation and rationalisation (Bartels 2006: 147). Finally, they focus exclusively on democracies. Although there are good reasons to do so (more on this later), Easton's conjecture is more ambitious: effective governance should elicit greater diffuse support, regardless of the kind of regime. Is there a better way to test the relationship between effectiveness and regime support? I suggest that there is.

Hypotheses and data

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

The argument to be examined here is that effective governance increases diffuse regime support. However, we need to consider the fact that not all regimes are alike. Take, for example, democracies. If effectiveness increases support, this means that for people living under democratic regimes, effectiveness should increase support for democracy:

  • H1: In democratic regimes, greater levels of effectiveness produce greater diffuse support for democracy.

A second implication is that, in non-democratic regimes, greater effectiveness should increase support for whatever type of regime under which people happen to live. Testing that hypothesis is, however, more difficult. Questions about regime support in cross-national surveys tend to measure citizens' views about democracy. Some, as we are about to see, do look into attitudes vis-à-vis a limited number of non-democratic regime types, but this still is far from covering their bewildering variety, which includes military, personalistic, single-party and theocratic dictatorships, or even different sorts of hybrid regimes. Furthermore, in many if not most non-democracies, survey respondents are likely to face ‘pressures to limit their responses to opinions supportive of the regime’ (Scotto & Singer 2004: 479), raising additional measurement problems. This is probably why most studies examining the determinants of regime support tend to restrict their analysis to citizens living under democracy (which is also what is proposed here with H1). However, given that we do have measures of democratic support in non-democratic regimes, it is possible to test a second, more conditional and tentative hypothesis, flowing indirectly from the argument about effectiveness and regime support:

  • H2: In non-democratic regimes, greater levels of government effectiveness produce lower diffuse support for democracy.

Support for democracy

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

The dependent variable – diffuse support for democracy – has been measured in different ways. Most have involved asking survey questions about how respondents feel about ‘democracy’ or whether they generically prefer it to other regimes (see Rose 1997 for a review). However, given the positive image ‘democracy’ has today around the world, there is the danger that such questions end up capturing little more than mere ‘lip service’ (Inglehart 2003: 52). This problem has been addressed in the past by constructing indexes combining questions about ‘explicit’ support for democracy with questions capturing rejection of autocracy (see, e.g., Klingemann 1999; Inglehart & Welzel 2005; Mattes & Bratton 2007; Bratton et al. 2005; Dalton & Ong 2005; Tusicisny 2007). In the WVS studies, namely in the third/1994–1999, fourth/1999–2000, fifth/2005–2006 and sixth/2008–2010 waves of the WVS, four items are particularly designed for this purpose, with higher values in e114 to e116, (and lower values in e117) denoting greater support for democracy:

I'm going to describe various types of political systems and ask you about each one as a way of governing the country. For each one, would you say it is a very good [0], fairly good [1], fairly bad [2] or very bad [3] way of governing this country?

e114. ‘Having a strong leader who does not have to bother with parliament and elections?’

e115. ‘Having experts, not governments, make decisions according to what they think is best for the country.’

e116. ‘Having the army rule.’

e117. ‘Having a democratic political system.’

Alternatively, other questions have measured respondents' sentiments concerning trade-offs involved in democratic rule. People who may support democracy and reject autocracy ‘in principle’ may nevertheless see democracy in tension with widely shared goals along ‘valence’ dimensions. In the WVS (third, fourth and fifth waves), four items have been used to gauge the extent to which citizens see democracy as an impediment to political stability, good economic management or political decisiveness (see, e.g., Dalton & Ong 2005; Hofmann 2004). Higher values in items e120 to e122 and lower values in item e123 capture greater support for democracy.

I'm going to read off some things that people sometimes say about a democratic political system. Could you please tell me if you agree strongly [0], agree [1], disagree [2], or disagree strongly [3] after I read each of them?

e120. ‘In democracy, the economic system runs badly.’

e121. ‘Democracies are indecisive and have too much squabbling.’

e122. ‘Democracies aren't good at maintaining order.’

e123. ‘Democracy may have problems but it's better than any other form of government.’

Combining these items, although an improvement over using responses to single items, may still be inappropriate if those items are shown to be internally inconsistent, unreliable and lacking cross-cultural equivalence. Ariely and Davidov (2011), applying confirmatory factor analysis to the WVS data, find that, to the extent that items e114 to e117 are designed to capture a general democracy–autocracy preference (DAP), item e117 ends up either insignificantly or negatively loaded on the construct. Something similar happens when they look at items e120 to e123, supposed to capture a democratic performance evaluation (DPE). Here, it is item e123 that underperforms. They argue that the explicit mention of ‘democracy’ in e117 (in contrast with the remaining DAP items) and the solicitation of an opinion about ‘democracy’ without evoking trade-offs in e123 (in contrast with the remaining DPE items) is likely to cause the problem.

Our analysis of the WVS data supports this. Using the responses to all the surveys in the WVS longitudinal data file with these eight items, and inverting the coding for e117 and e123 (so that larger values mean greater democratic support), principal components analysis reveals the emergence of three different factors (see Table A3 in the online appendix): one formed by e114 to e116 (DAP); another formed by e120 to e122 (DPE); and a third factor formed by e117 and e123, which we will designate as capturing explicit democratic support (EDS). The factor loadings and the goodness of fit statistics from confirmatory factor analysis (see Table A4 in the online appendix) show that the three-factor model fits the data well, and that the construct of ‘support for democracy’ seems indeed to be comprised of three latent factors: EDS, DAP and DPE. On the basis of these findings, I built three additive indexes to capture each dimension of support: DAP and DPE both range from 0 to 9, while EDS ranges from 0 to 6, in all cases with higher values meaning greater support for democracy.2

Effectiveness

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

What about government effectiveness? Among the cross-national measures of ‘governance’ available, I employ the one that seems conceptually closest to the notions of ‘effectiveness’, ‘output efficiency’ and ‘quality of public policies and their implementation’ that derive from the discussions of Easton, Lipset, Dahl and Linz: the World Bank's ‘government effectiveness’ indicator (Effectiveness),3 which captures ‘perceptions of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government's commitment to such policies’ (Kaufmann et al. 2010: 4). The index is built upon 15 different data sources, mostly expert assessments or surveys of firms on the quality of the bureaucracy, the supply of basic public goods, policy stability and implementation, and the quality of budgetary and financial management.

In spite of some criticism to which the WGI indicators have been subjected, even the harshest critics concede that Effectiveness ‘clearly attempts to capture the ability of the state to formulate and implement its goals’ (Kurtz & Schrank 2007: 543). It has been used, and thus assessed in terms of predictive validity, in many studies. For example, it has been shown to be related with per capita income, infant mortality and literacy (Kaufmann et al. 1999), foreign investment flows (Globerman & Shapiro 2002), and several health and environment-related outcomes (Brooks et al. 2005; Holmberg et al. 2009). Closer to our purpose, others have treated it as a determinant of subjective attitudes, such as well-being (Helliwell & Huang 2008) and satisfaction with democracy (Dahlberg & Holmberg 2012).

Democracy

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

H1 is that government effectiveness increases support for democracy in democratic regimes, while H2 is that such a relationship is negative in non-democracies. I resort to two well-established data sources to distinguish democracies from other regimes. The first is the DD (Democracy and Dictatorship Revisited) dataset (Cheibub et al. 2010). The variable DDemocracy is coded, for each country-year, as 1 (if the regime qualifies as democratic) and 0 (if not). ‘Democracy’, from this point of view, is a polity where the chief executive and the legislature is chosen by popular election or by a popularly elected body, where more than one party competes in elections, and where alternation in power has taken place at least once. The second source of data about regime types used here is Freedom House. Since 1973, Freedom House has issued reports and surveys on the extent to which political rights and civil liberties are protected in the world. ‘Free’ countries are liberal democracies, with regular free and fair elections, multipartyism, universal suffrage, access of parties to the media and campaigning, and effective protection of political and civil rights. In contrast, ‘Not Free’ countries are typically governed by ‘military juntas, one-party dictatorships, religious hierarchies or autocrats’ and have ‘severely restricted rights of expression and association’, while ‘Partly Free’ designates situations where rulers, albeit allowing means of popular influence in government, manipulate elections and restrict the formation of political groups. These two ways of measuring ‘democracy’ – ‘Democracies’ (DD) or ‘Free/Liberal Democracies’ (FH) – and distinguishing them from other regimes have different normative and empirical underpinnings. Therefore, I will alternatively employ both in the analysis.

Controls

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

Several features of polities besides government effectiveness are likely to be related to attitudes towards democracy. Economic development is one of them. On the one hand, the relationship between high levels of income and high quality of governance is strong and basically ‘incontrovertible’ (Rodrik 2008). On the other hand, economic development is likely to endow citizens with the economic security, education and individual opportunities that socialise them into lower deference towards authority, greater demand for liberalisation and higher support for democratic rule (Inglehart 1997). Here, we measure economic development with (the natural log of) gross domestic product (GDP) per capita in each country-year (LnGDPpc).

A second control is Yearsdem, the number of years that, by the time of the survey, each country had held a democratic regime, capturing the possibility that, in countries where that experience with democracy has been longer, support for democracy may be stronger (Huang et al. 2008; Staton & Reenock 2010). I counted the number of years, since 1946 (at most) until the year of the survey, that DDemocracy was coded as 1 in the dataset. For the models where the Freedom House indicators are used, since they only exist since the 1970s, I also employ the measure based on the DD dataset.

A third contextual control is the level of income inequality (Gininet), measured by the Gini index of inequality of net household income for each country-year. Solt (2012) found that economic inequality tends to foster authoritarian attitudes, including citizens' views about ‘obedience’, ‘respect for authority’ and obeisance of workplace superiors. Thus, I take into account the possibility that inequality's effect in breeding authoritarian values may spill over to a more unfavourable attitude towards democracy.

Ethnic fractionalisation is thought to be related to a variety of aspects of relevance here, such as the extent to which a country is likely to become democratic (Welzel 2007), the quality of governance (Easterly & Levine 1997; La Porta et al. 1999), and several important political attitudes, including – especially for minorities – lower support for democracy (Dowley & Silver 2002). We rely on Alesina et al.'s (2003) measure of ethnic fractionalisation (Ethnicfrac) to code the different countries in our sample.

I also employ several of the individual-level controls. Previous findings indicate that richer and more educated individuals tend to exhibit greater support for democracy, while women display the opposite tendency (Hofmann 2004; Huang et al. 2008; Staton & Reenock 2010). Thus, I include Female, Education and Income in the models. The effects of age are somewhat less clear. As Solt (2012) notes, whatever the impact of age on democratic attitudes that has been found in many studies may be a reflection of other attributes correlated with age, such as marrying or having children. Therefore, following Solt, and to isolate the effects of Age, I also include Single and Children. Unemployed is a dummy variable measuring whether the respondent was unemployed at the time of the survey.

Finally, I include two attitudinal measures as individual-level controls. Good governance, to the extent it contributes to effectively enforce private agreements, seems to generate high levels of social trust (Herreros & Criado 2008). On the other hand, there is also evidence that, in turn, high levels of social trust tend to be associated with several relevant political attitudes, including confidence in government (Keele 2007) and more favourable attitudes towards democracy (Zmerli & Newton 2008). Therefore, there is a potential link between generic ‘quality of government’ and regime support that goes through social trust and general confidence in government. This is an important line of inquiry. However, the particular theoretical relationship we want to examine is a different one – the relationship between the effectiveness and political performance of governments and popular support for regimes. Thus, I control here for both interpersonal trust (social trust) and confidence in government.

Analysis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

One way to start making sense of the data is by focusing on aggregate-level patterns. How are average levels of support for democracy related to government effectiveness in different types of regimes? On the basis of the WVS longitudinal data file and the availability of items e114–e117 and e120–e123 in the different surveys and countries, we can estimate average levels of EDS, DPE and DAP in, respectively, 89, 92 and 142 surveys. I then matched these levels of democratic support per country-year with Effectiveness and regime type for the same years. This was possible for almost all cases, except those where measures of regime type were unavailable for the respective country-years and for surveys conducted before 1996, when the WGI indicators start being available.4

Figures 1, 2 and 3 plot our support for democracy indicators against Effectiveness in different regime types. Effectiveness and support for democracy are positively correlated in democracies, lending some preliminary support to H1. Countries like Norway, Sweden, Finland, Canada, Switzerland, New Zealand or Australia, which display the highest levels of government effectiveness in the sample (1.8 or more), also systematically appear among those where explicit support for democracy, rejection of autocratic solutions and better views of democracy's performance are more prevalent. Conversely, democracies like Romania, Mali, Indonesia, Bulgaria and El Salvador, where government effectiveness is lowest, are also among those where rejection of autocratic alternatives and positive views of democracy along valence dimensions are less prevalent.

figure

Figure 1. Government effectiveness and explicit democratic support (EDS) in different types of regimes.

Download figure to PowerPoint

figure

Figure 2. Government effectiveness and democratic performance evaluations (DPE) in different types of regimes.

Download figure to PowerPoint

figure

Figure 3. Government effectiveness and democracy-autocracy preference (DAP) in different types of regimes.

Download figure to PowerPoint

However, EDS has comparatively little aggregate-level variance in our cases: there is not a single country-year that falls on the bottom half of the scale. When questions elicit from respondents their explicit support for democracy (as a ‘good’ regime and as being ‘better’ than other forms of government), responses tend to be, on average, generally favourable and do not provide much discrimination between countries and surveys. The correlations between Effectiveness and EDS among democracies (0.17 among DDemoc = 1 and 0.41 among Free = 1) are weaker than those with DPE or DAP. In contrast, 32 and 41 per cent of the observations of, respectively, DAP and DPE are in the bottom half of the scale and their correlations with Effectiveness among democracies are stronger. Among ‘liberal democracies’ the relationship between effectiveness and support is always stronger than the same relationship within ‘democracies’ (reaching 0.81 with DAP). Finally, preliminary support for H2 – a negative relationship between effectiveness and democratic support in non-democracies – is flimsier. Correlations are mostly negative, as expected, but weaker (the strongest being –0.17 for the relationship between Effectiveness and DPE among non-democracies).

However, all this is still exploratory. First, we need multivariate analysis, rather than just looking at correlations between Effectiveness and the different measures of support. Second, we can take advantage of the fact that we have individual-level responses. Early on, we proposed three types of controls. Some vary across countries, like ethnic fractionalisation (Ethnicfrac). Others vary across countries and across time, such as lnGDPpc, Gininet and regime type (DDemocracy or Free), as well as Effectiveness, our main independent variable. Finally, EDS, DPE and DAP, as well as the individual-level controls, vary across survey respondents. Taking into account this three-level structure of the data – countries, country-years and individuals – we can estimate a multilevel model. Model 1, for individual i in country-year j and country k, treats EDS as a continuous variable, includes predictors at the three levels of analysis, as well as varying intercepts and error terms for country ( and year). Individuals are thus treated as being nested within country-years (surveys), which in turn are treated as nested within countries. This allows us, on the one hand, to take into account the strong possibility that observations about support for democracy taken from within the same contexts are not independent. On the other hand, average levels of support for democracy within these groups (countries and years) are allowed to vary reflecting factors that are not included in the model.

We include two interaction terms. The first is between Effectiveness and the dummy DDemocracy, allowing us to test H1 and H2. The second interaction is between DDemocracy and lnGDPpc. In other words, while we allow effectiveness to affect democratic support differently in democracies and other regimes, we do the same for economic development. On the one hand, we saw that, at the aggregate level, citizens living under more effective democracies are more supportive of democracy. On the other hand, we know that quality of governance and economic development are positively related (Rodrik 2008). Therefore, we need to take into account the possibility that what may be driving the correlation between effectiveness and support is that citizens in wealthier democracies make a better judgment about democracy, rather than that judgment being affected by political performance.5 Model 1 takes that into account, and will also be also estimated for DPEijk and DAPijk as dependent variables.

  • display math(1)

Model 2 is very much the same, with the single difference that, this time, we will be interacting Effectiveness and lnGDPpc with the Free (‘liberal democracy’) dummy variable, based on Freedom House's data.

  • display math(2)

Random effects ANOVA analyses allow us to determine the portion of the variance in each dependent variable due to country or country-year differences. In the case of EDS, about 11 per cent of the variance is due to differences across countries or across country-years. For DPE, the value is 10 per cent. Finally, for DAP, 10 per cent of variance is at the country level and 10 per cent at the country-year level. In the context of a study such as this, where the number of individuals in the survey samples is very large in comparison with the number of groups, the variance attributable to the grouping structure can be seen as reasonably large and to justify multilevel analysis (Hox 2010: 244).

Table 1 shows the results of the estimation of the multilevel linear models 1 and 2.6 Combinations between availability of all items in the scales and all individual-level and macro-level control variables vary between the dependent variable employed and the model that is estimated, leading to loss of observations in relation to the aggregate-level plots shown earlier. At a minimum, we are left with 55 countries and 72 surveys (for EDS).7

Table 1. Government effectiveness and support for democracy
Dependent variableEDSDPEDAP
Model121212
  1. Notes: * p < 0.10; ** p < 0.05; *** p < 0.001 (two-tailed tests). Standard errors adjusted for country clusters.

Effectiveness−0.02 (0.11)−0.24 (0.13)*0.03 (0.18)−0.27 (0.22)−0.02 (0.27)−0.28 (0.16)*
Effectiveness*DDemocracy−0.17 (0.15)0.47 (0.24)*0.58 (0.32)*
DDemocracy−2.51 (1.28)*0.32 (1.63)−3.33 (2.48)
Effectiveness*Free0.26 (0.15)*1.23 (0.28)***0.83 (0.23)***
Free−0.08 (0.78)2.77 (1.09)**−4.70 (0.30)**
Contextual controls      
lnGDPpc−0.16 (0.07)**−0.05 (0.08)−0.20 (0.10)**−0.08 (0.11)−0.39 (0.18)**−0.37 (0.15)**
lnGDPpc*DDemocracy0.29 (0.14)**−0.04 (0.19)0.40 (0.28)
lnGDPpc*Free0.02 (0.09)−0.35 (0.13)***0.59 (0.26)**
Yearsdem0.004 (0.003)0.003 (0.003)−0.006 (0.005)−0.01 (0.005)*0.007 (0.007)0.004 (0.005)
Gininet−0.01 (0.01)−0.005 (0.005)−0.01 (0.01)−0.00 (0.01)0.00 (0.01)0.00 (0.01)
Etnnicfrac−0.13 (0.27)−0.09 (0.27)−0.15 (0.41)−0.23 (0.37)0.13 (0.44)0.28 (0.32)
Individual controls      
Female−0.06 (0.01)***−0.06 (0.01)***−0.18 (0.03)***−0.18 (0.03)***−0.11 (0.02)***−0.11 (0.02)***
Age0.003 (0.001)***0.003 (0.001)***−0.001 (0.002)−0.001 (0.002)0.002 (0.001)**0.002 (0.001)**
Children0.007 (0.004)0.007 (0.004)−0.00 (0.01)−0.00 (0.01)−0.00 (0.01)−0.00 (0.01)
Single0.02 (0.02)0.02 (0.02)0.02 (0.03)0.02 (0.03)0.01 (0.02)0.01 (0.02)
Education0.05 (0.01)***0.05 (0.01)***0.10 (0.01)***0.10 (0.01)***0.11 (0.01)***0.11 (0.01)***
Income0.03 (0.01)*0.03 (0.01)*0.11 (0.02)***0.11 (0.02)***0.05 (0.01)***0.05 (0.01)***
Unemployed−0.01 (0.03)−0.01 (0.03)−0.08 (0.04)**−0.08 (0.04)**−0.06 (0.03)**−0.06 (0.03)**
Soctrust0.06 (0.03)**0.06 (0.03)**0.23 (0.05)***0.23 (0.05)***0.07 (0.03)**0.07 (0.03)**
ConfGov0.09 (0.03)***0.09 (0.03)***0.10 (0.03)***0.10 (0.03)***−0.07 (0.03)***−0.07 (0.03)***
Constant5.61 (0.77)***4.48 (0.81)***5.99 (1.05)***4.68 (1.06)***7.46 (1.74)***6.86 (1.38)***
Variance components      
Country-year intercept0.030.030.150.060.520.53
Country intercept0.120.130.190.230.090.01
Countries555676
Country-years7272116
Respondents74,12678,955132,005

First, almost all of the individual-level covariates behave similarly regardless of the dimension of support for democracy under examination and the model employed: males and individuals with higher levels of education, income and interpersonal trust tend to be more supportive of democracy. The coefficients for the unemployment condition are also negative in all models. The effects of age are not univocal: while it has no bearing on views about democratic performance, it does have a positive effect on both explicit democratic support (EDS) and democracy-autocracy preferences (DAP).

The findings concerning the contextual control variables are mostly non-findings: the signs for Yearsdem, Ethnicfrac and Gininet are different depending on the measure of support for democracy employed and are almost always far from conventional significance. We also find no support for the notion that economic development is positively related with support in either democracies or other regimes. In fact, once government effectiveness, its contingent effect on democratic support, and – quite importantly – individual-level measures of affluence and cognitive resources are taken into account, estimation of the marginal effects under the different conditions shows that the only significant effects of lnGDPpc to be found are negative.8

Our core concern in this study, however, is the impact of Effectiveness on democratic support. Figure 4 simplifies the reading of Table 1 in this regard by showing its marginal effects on EDS, DAP and DPE and the respective 90 per cent confidence intervals on the basis of models 1 and 2. We present those effects for the different values of the conditioning variables DDemocracy and Free.9

figure

Figure 4. Marginal effects of Effectiveness on democratic support in different types of regimes (90 per cent confidence intervals, models 1 and 2).

Download figure to PowerPoint

When support for democracy is measured by eliciting from respondents their explicit approval of democratic regimes, H1 is not supported: more effective democracies do not produce higher EDS levels. All effects are very close to zero and the one that is closest to statistical significance at conventional levels is actually negative. Citizens living in ‘democracies’ (or ‘liberal democracies’) that experience more effective governance seem no more or less likely to have better things to say about ‘democracy’ than those who live in democracies plagued by deeper failures in policy making and implementation. This absence of a relationship between this measure of governmental performance and EDS confirms the general tenor of the literature on democratic legitimacy where ‘explicit’ measures of support have typically been used.

However, the picture changes radically when, in order to gauge principled democratic support, we employ measures that evoke potential trade-offs in performance (DPE) or capture respondents' rejection of autocratic alternatives (DAP). There, H1 receives very clear support, regardless of whether we use DPE or DAP or whatever measure of regime is employed. In country-years identified with value 1 for DDemocracy, a one standard deviation increase in Effectiveness is expected to increase both DPE and DAP by about two-thirds of a standard deviation – a substantively important effect. When the regime citizens live under is a ‘liberal democracy’, the impact of Effectiveness emerges as even larger. This makes sense, as a regime typology like Freedom House's strengthens the contrast between genuinely liberal democracies and other sorts of regimes. Besides, it is relevant that H1 stands regardless of whether we use DAP or DPE as they both have important advantages and disadvantages. While DAP avoids the use of ‘agree–disagree’ formats, which we know to have several undesirable properties, particularly in the way they create incentives for ‘satisficing’ and generate an acquiescence bias (Krosnick 1991), DPE shows, unlike DAP, both metric and (partial) scalar invariance (Ariely & Davidov 2011) and also, as we showed earlier, greater reliability.

Support for H2 is clearly weaker, but nevertheless suggestive. Although the measures of regime support available are not specifically designed to measure it for the many different sorts of non-democratic regimes, and data limitations forced us to lump together many different types of ‘non-democracies’, H2 does receive partial support: the marginal effects of Effectiveness on democratic support in those cases are negative in four out of six estimations and statistically significant in two of them. In sum, the results also suggest the plausibility that effectiveness may drive down support for democracy in non-democratic regimes.

Implications

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

If ‘democracy’ was indeed ‘the only political model with global appeal’ and if democratic support was impervious to ‘performance’, democrats could rejoice. Policy failures in democracies might affect ‘instrumental’ dimensions of people's attitudes towards politics, but would leave the ‘principled’ support for the regime unscathed. However, if there are problems with the way democratic support has been measured, if such support is less widespread than previously thought, and if the effectiveness of governments affects it, the implications are somewhat less felicitous. In democracies that remain plagued by deep institutional failures in policy making and implementation, popular support for democracy may suffer.

Our results lend some credence to this latter less optimistic view. Although several established findings about individual-level predictors of support for democracy are confirmed, the results also show that, among the plausible macro-level predictors, government effectiveness emerges, by far, as the most important. This is particularly so when the measures of democratic support evoke less ‘explicit’ responses from interviewees and thus most clearly circumvent the risk that they may be paying ‘lip service’ to democratic principles (Inglehart 2003: 52). In sum, there are good reasons to believe we should seriously reconsider the notion that diffuse regime support is impervious to performance – at least if by performance we understand the quality of policy making and implementation.

The second implication of Easton's conjectures about the relationship between effectiveness and diffuse support is that the legitimacy on non-democratic regimes should also be enhanced by effectiveness. Unfortunately, we were not able to test that precise hypothesis here due to data limitations. Having said that, we found that, at least in those contexts where civil and political liberties fall short of qualifying countries as ‘liberal democracies’, Effectiveness and democratic legitimacy are negatively (rather than positively) related. That effect is statistically significant for two of our support measures and borderline (in)significant for the third. In other words, the usual practice in the study of popular support for regimes – to focus only on what occurs within democracies – may be neglecting the investigation of the sources of popular support for dictatorships (Geddes & Zaller 1989), with government effectiveness as one of them. This line of inquiry is even more important considering the growing literature on ‘autocratic stability’. The survival of autocracies seems to increase with the adoption of institutions that foster credible commitments, formalise interactions between ruling elites and their allies, and increase transparency in policy making (Gandhi & Przeworski 2007; Boix & Svolik 2011), and such features are, in turn, closely related to the quality of governance (Gehlbach & Keefer 2011; Charron & Lapuente 2011). This calls attention to a potentially important link between ‘autocratic institutionalisation’ and regime stability: in those ‘institutionalised’ autocracies, governance is likely to be more effective, recognised as such by citizens and converted into regime support – or, at least, into greater popular rejection of democracy.

Finally, there is still much to do in the investigation of the sources of regime support. First, we need greater caution about the available measures of regime support in existing surveys. Items that have often been combined in mostly intuitive ways seem in fact to tap different dimensions of support and, when used to build scales, do not necessarily travel in the same way, nor do they have the same correlates.10 Second, more and better data are needed. On the one hand, existing examinations of the dimensionality of items and the reliability of scales are constrained by the relatively small number of those items available in cross-national surveys. This is an empirical problem with theoretical implications. For example, dimensions such as DAP and DPE, built upon questions that ask citizens what kinds of regimes would be ‘good’ for their countries or whether they see democracy as compatible with ‘valence’ goals, seem potentially more oriented towards capturing something closer to ‘output-oriented’ legitimacy (Scharpf 1999) and thus more likely to be shaped by the effectiveness of governments. New items that could be designed to capture ‘input-oriented’ legitimacy – perceptions of whether regimes favour equal participation of all and proper aggregation of interests – might conceivably have different correlates. On the other hand, we seem to be particularly in need of more and better indicators of support for regimes other than democratic ones, and of more surveys conducted in such contexts. To be sure, the fact that we lack them is not just a matter of chance or neglect: the obstacles involved in obtaining good quality measures in mass surveys conducted in dictatorships and the ensuing case selection bias (since conducting quality mass surveys is not even a possibility in many autocracies) are very difficult to overcome. Nevertheless, the results of this study suggest the importance of devoting particularly strong efforts to that endeavour.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information

Earlier versions of this work were presented at the conference ‘The Quality of Democracy in Hard Times’, organised by the Barometer of the Quality of Democracy (Institute of Social Sciences of the University of Lisbon, Portugal) and at the Seminar in Comparative Politics and International Relations of CESNOVA at the Universidade Nova de Lisboa, Portugal. I'm grateful to the participants for the comments and suggestions received. I am also grateful to the editors of EJPR and to the four anonymous reviewers for their very helpful comments. All remaining errors remain the responsibility of the author. A substantial part of the work for this article was conducted while I was FLAD visiting professor at the Department of Government of Georgetown University. I'm grateful to the department and to the Luso-American Foundation for Development for their support.

Notes
  1. 1

    In a different context, related to the legitimisation of the European Union as a political system, ‘legitimacy’ has also been treated as having an ‘output-oriented’ component, linked not only to avoidance of abuses of public power but also to ‘effective problem solving’ (Scharpf 1999:13). I'm grateful to Yannis Papadopoulos for pointing this out to me.

  2. 2

    Cronbach's alpha for the items composing DPE, DAP and EDS are, respectively: 0.78, 0.53 and 0.56. The latter two values are below the ‘rule of thumb’ usually employed to determine ‘acceptable’ reliability. However, besides underestimating true reliability (Sijtsma 2009), alpha depends very much on the number of items, which in our cases is extremely small (three for DPE and DAP, two for EDS). For example, while the correlation between the two EDS items (0.39) yields an alpha of 0.56, a mean item intercorrelation of 0.39 for five items would already yield an alpha comfortably above 0.70 (Iacobucci & Duhacek 2003).

  3. 3

    See Tables A1 and A2 in the online appendix for descriptions and sources of all variables employed.

  4. 4

    For surveys in 1997, 1999 and 2001, I coded country-years with the average of the preceding and following year's measures of Effectiveness. Ideally, one would prefer to code each country-year with an average of past measurements of Effectiveness, reflecting a sustained high or low level of government effectiveness. However, using, for example, the average of four lagged observations would leave us with just half of the country-year observations. Levels of government effectiveness across countries seem to be rather stable through time. If we take all country-level measures of Effectiveness for all countries and territories between 1996 and 2010 and correlate those in a particular year with those of the preceding year, the lowest value obtained is 0.98. Furthermore, the correlation between our measure of Effectiveness for each country-year and the same variable measured in 1996 for all countries is 0.96.

  5. 5

    I am grateful to one of the reviewers for pointing out this possibility.

  6. 6

    The Stata 12.1 xtmixed command was used for this purpose. Standard errors of coefficients were adjusted for country clusters, using the vce (cluster Country) option. All models were estimated using unweighted survey data. Estimations using the sampling weights provided in the WVS dataset, s017 and s018, do not change the results in any relevant way.

  7. 7

    See Table A5 in the online appendix for a list of the surveys employed in the analysis.

  8. 8

    Results are available from the author upon request. One possible concern about the results of the macro variables would be multicollinearity. To address it, I ran simple OLS models including all main terms at all levels, and then estimated their respective variance inflation factors. The largest VIF obtained was 3.6, suggesting that multicollinearity is not a concern.

  9. 9

    The Stata 12.1 margins command was used for this purpose.

  10. 10

    For similar generic conclusions about the multidimensionality of ‘support for democracy’, this time in terms of different aspects – contestation, participation, limits on executive, and institutions and process – of the ‘polyarchy’ construct in 12 Latin American democracies, see Carlin and Singer (2011).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information
  • Alesina, A. et al. (2003). Fractionalization. Journal of Economic Growth 8(2): 155194.
  • Ariely, G. & Davidov, E. (2011). Can we rate public support for democracy in a comparable way? Cross-national equivalence of democratic attitudes in the World Values Survey. Social Indicators Research 104(2): 271286.
  • Bäck, H. & Hadenius, A. (2008). Democracy and state capacity: Exploring a J-shaped relationship. Governance 21(1): 124.
  • Bartels, L.M. (2006). Three virtues of panel data for the analysis of campaign effects. In R. G. Johnston & H.E. Brady (eds), Capturing campaign effects. Ann Arbor, MI: University of Michigan Press.
  • Boix, C. & Svolik, M. (2011). The Foundations of Limited Authoritarian Government: Institutions and Power-sharing in Dictatorships, 24 September. Available online at: http://ssrn.com/abstract=1352065
  • Bratton, M., Mattes, R.B. & Gyimah-Boadi, E. (2005). Public opinion, democracy and market reform in Africa. Cambridge: Cambridge University Press.
  • Brooks, N., Adger, W.N. & Kelly, P.M. (2005). The determinants of vulnerability and adaptive capacity at the national level and the implications for adaptation. Global Environmental Change 15(2): 151163.
  • Carlin, R.E. & Singer, M.M. (2011). Support for polyarchy in the Americas. Comparative Political Studies 44(11): 15001526.
  • Charron, N. & Lapuente, V. (2010). Does democracy produce quality of government? European Journal of Political Research 49(4): 443470.
  • Charron, N. & Lapuente, V. (2011). Which dictators produce quality of government? Studies in Comparative International Development 46(4): 397423.
  • Cheibub, J.A., Gandhi, J. & Vreeland, J.R. (2010). Democracy and dictatorship revisited. Public Choice 143(1): 67101.
  • Chu, Y. et al. (2008). Public opinion and democratic legitimacy. Journal of Democracy 19(2): 7487.
  • Curini, L., Jou, W. & Memoli, V. (2012). Satisfaction with democracy and the winner-loser debate: The role of policy preferences and past experience. British Journal of Political Science 42(2): 241261.
  • Dahl, R.A. (1971). Polyarchy: Participation and opposition. New Haven, CT: Yale University Press.
  • Dahlberg, S. & Holmberg, S. (2012). Understanding Satisfaction with the Way Democracy Works. QoG Working Paper 2012(8). Gothenburg: Quality of Government Institute.
  • Dalton, R.J. (2004). Democratic challenges, democratic choices: The erosion of political support in advanced industrial democracies. Oxford: Oxford University Press.
  • Dalton, R.J. & Ong, N.T. (2005). Authority orientations and democratic attitudes: A test of the ‘Asian values’ hypothesis. Japanese Journal of Political Science 6(2): 211231.
  • Dowley, K.M. & Silver, B.D. (2002). Social capital, ethnicity and support for democracy in the post-communist states. Europe-Asia Studies 54(4): 505527.
  • Easterly, W. & Levine, R. (1997). Africa's growth tragedy: Policies and ethnic divisions. Quarterly Journal of Economics 112(4): 12031250.
  • Easton, D. (1957). An approach to the analysis of political systems. World Politics 9(3): 383400.
  • Easton, D. (1965). A framework for political analysis. Englewood Cliffs, NJ: Prentice-Hall.
  • Easton, D. (1975). A re-assessment of the concept of political support. British Journal of Political Science 5(4): 435457.
  • Gandhi, J. & Przeworski, A. (2007). Authoritarian institutions and the survival of autocrats. Comparative Political Studies 40(11): 12791301.
  • Geddes, B. & Zaller, J. (1989). Sources of popular support for authoritarian regimes. American Journal of Political Science 33(2): 319347.
  • Gehlbach, S. & Keefer, P. (2011). Investment without democracy: Ruling-party institutionalization and credible commitment in autocracies. Journal of Comparative Economics 39(2): 123139.
  • Globerman, S. & Shapiro, D. (2002). Global foreign direct investment flows: The role of governance infrastructure. World Development 30(11): 18991919.
  • Gunther, R., Montero, J.R. & Torcal, M. (2007). Democracy and intermediation: Some attitudinal and behavioral dimensions. In R. Gunther , J.R. Montero & H.-J. Puhle (eds), Democracy, intermediation and voting on four continents. Oxford: Oxford University Press.
  • Haggard, S. & Kaufman, R.R. (1995). The political economy of democratic transitions. Princeton, NJ: Princeton University Press.
  • Helliwell, J.F. & Huang, W. (2008). How's your government? International evidence linking good government and well-being. British Journal of Political Science 38(4): 595619.
  • Herreros, F. & Criado, H. (2008). The state and the development of social trust. International Political Science Review 29(1): 5371.
  • Hofmann, S.R. (2004). Islam and democracy: Micro-level indications of compatibility. Comparative Political Studies 37(6):652676.
  • Holmberg, S., Rothstein, B. & Nasiritousi, N. (2009). Quality of government: What you get. Annual Review of Political Science 12: 135161.
  • Hox, J.J. (2010). Multilevel analysis: Techniques and applications. New York: Routledge.
  • Huang, M., Chang, Y. & Chu, Y. (2008). Identifying sources of democratic legitimacy: A multilevel analysis. Electoral Studies 27(1): 4562.
  • Iacobucci, D. & Duhacek, A. (2003). Advancing alpha: Measuring reliability with confidence. Journal of Consumer Psychology 13(4): 478487.
  • Inglehart, R. (1997). Modernization and postmodernization: Cultural, economic and political change in 43 societies. Princeton, NJ: Princeton University Press.
  • Inglehart, R. (2003). How solid is mass support for democracy – and how can we measure it? Political Science and Politics 36(1): 5157.
  • Inglehart, R. & Norris, P. (2003). The true clash of civilizations. Foreign Policy 135: 6370.
  • Inglehart, R. & Welzel, C. (2005). Modernization, cultural change and democracy: The human development sequence. Cambridge: Cambridge University Press.
  • Kaufmann, D., Kraay, A. & Mastruzzi, M. (1999). Governance matters. Washington, DC: World Bank.
  • Kaufmann, D., Kraay, A. & Mastruzzi, M. (2010). The worldwide governance indicators: Methodology and analytical issues. Washington, DC: World Bank.
  • Keele, L. (2007). Social capital and the dynamics of trust in government. American Journal of Political Science 51(2): 241254.
  • Kennedy, J.J. (2008). Maintaining popular support for the Chinese Communist Party: The influence of education and the state-controlled media. Political Studies 57(3): 517536.
  • Klingemann, H.D. (1999). Mapping political support in the 1990s: A global analysis. In P. Norris (ed.), Critical citizens: Global support for democratic government. Oxford: Oxford University Press.
  • Krosnick, J.A. (1991). Response strategies for coping with the cognitive demands of attitude measures in surveys. Applied Cognitive Psychology 5(3): 213236.
  • Kurtz, M.J. & Schrank, A. (2007). Growth and governance: Models, measures and mechanisms. Journal of Politics 69(2): 538554.
  • Lagos, M. (2003). Support for and satisfaction with democracy. International Journal of Public Opinion Research 15(4): 471487.
  • La Porta, R. et al. (1999). The quality of government. Journal of Law, Economics and Organization 15(1): 222279.
  • Linde, J. (2012). Why feed the hand that bites you? Perceptions of procedural fairness and system support in post-communist democracies. European Journal of Political Research 51(3): 410434.
  • Linde, J. & Erlingsson, G.Ó. (2012). The eroding effect of corruption on system support in Sweden. Governance. DOI: 10.1111/gove.12004.
  • Linz, J.J. (1978). Crisis, breakdown and reequilibration. In J.J. Linz & A. Stepan (eds), The breakdown of democratic regimes. Baltimore, MD: Johns Hopkins University Press.
  • Linz, J.J. & Stepan, A. (1996). Problems of democratic transition and consolidation: Southern Europe, South America and Post-Communist Europe. Baltimore, MD: Johns Hopkins University Press.
  • Lipset, S.M. (1959). Some social requisites of democracy: Economic development and political legitimacy. American Political Science Review 53(1): 69105.
  • Magaloni, B. & Wallace, J. (2008). Citizen Loyalty, Mass Protest and Authoritarian Survival. Paper presented at the ‘Dictatorships: Their Governance and Social Consequences’ conference, Princeton University.
  • Mattes, R. & Bratton, M. (2007). Learning about democracy in Africa: Awareness, performance and experience. American Journal of Political Science 51(1): 192217.
  • Mishler, W. & Rose, R. (2001). Political support for incomplete democracies: Realist vs idealist theories and measures. International Political Science Review 22(4): 303320.
  • Montinola, G.R. & Jackman, R.W. (2002). Sources of corruption: A cross-country study. British Journal of Political Science 32(1): 147170.
  • Rodrik, D. (2008). Thinking about governance. In World Bank (ed.), Governance, growth and development decision-making. Washington, DC: World Bank.
  • Rose, R. (1997). Survey Measures of Democracy. Working Paper 294. Centre for the Study of Public Policy, University of Strathclyde.
  • Scharpf, F.W. (1999) Governing in Europe: Effective and Democratic? New York: Oxford University Press.
  • Schmitter, P.C. & Karl, T.L. (2008). What democracy is … and is not. Journal of Democracy 2(3): 7588.
  • Scotto, T.J. & Singer, M.M. (2004). Beginning democracies. In J.G. Geer (ed.), Public opinion and polling around the world. Santa Barbara, CA: ABC-CLIO.
  • Sijtsma, K. (2009). On the use, misuse and the very limited usefulness of Cronbach's alpha. Psychometrika 74(1): 107120.
  • Solt, F. (2012). The social origins of authoritarianism. Political Research Quarterly 65(4): 703713.
  • Staton, J.K. & Reenock, C. (2010). Substitutable protections: Credible commitment devices and socio-economic insulation. Political Research Quarterly 63(1): 115128.
  • Torcal, M. & Moncagatta, P. (2011). Political support. In B. Badie , D. Berg-Schlosser & L. Morlino (eds), International Encyclopedia of Political Science. London: Sage.
  • Tusicisny, A. (2007). Security communities and their values: Taking masses seriously. International Political Science Review 28(4): 425449.
  • Wagner, A.F., Schneider, F. & Halla, M. (2009). The quality of institutions and satisfaction with democracy in Western Europe: A panel analysis. European Journal of Political Economy 25(1): 3041.
  • Welzel, C. (2007). Are levels of democracy affected by mass attitudes? Testing attainment and sustainment effects on democracy. International Political Science Review 28(4): 397424.
  • Wintrobe, R. (1998). The political economy of dictatorship. Cambridge: Cambridge University Press.
  • Zmerli, S. & Newton, K. (2008). Social trust and attitudes towards democracy. Public Opinion Quarterly 72(4): 707724.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Effectiveness and diffuse support
  5. Hypotheses and data
  6. Support for democracy
  7. Effectiveness
  8. Democracy
  9. Controls
  10. Analysis
  11. Implications
  12. Acknowledgements
  13. References
  14. Supporting Information
FilenameFormatSizeDescription
ejpr12024-sup-0001-si.pdf207K

Table A1.Macro-level variables

Table A2.Individual-level variables

Table A3.The dimensionality of the WVS democratic support items

Table A4.The dimensionality of the WVS democratic support items (CFA)

Table A5.Country-years/WVS surveys for which there are measures of Effectiveness, Free or DDemoc, and EDS, DPE or DAP

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.