DIRECT DEMOCRACY AND REGIME LEGITIMACY IN EUROPE

Sergiu Gherghina

Institute of Political Science

Goethe University Frankfurt

gherghina@soz.uni-frankfurt.de

Paper prepared for presentation at the ECPR Joint Session of Workshops

10-15 April 2014, Salamanca

Workshop: “Regime Legitimisation through Institutional Reform:

Analysing its Dimensions and Effectiveness”

1

Introduction

In the last two decades an increasing body of literature discusses the disaffection of

citizens with representative democracy. The existence of several problems in the process

of political representation led to a gradual abandon of the traditional modes of

involvement (e.g. electoral turnout, party membership). One solution to address this

“democratic malaise” has been the adoption of direct democracy procedures through

which citizens receive a direct say in political decision-making (Scarrow 2001; Smith 2009;

Geissel & Newton 2012). A different strand of literature emphasizes the decrease of

regime legitimacy across the world. The recent street protests in Eastern Europe are only

one indication of citizens’ attitudes and behaviors towards state authorities. While earlier

studies have focused on the causes and consequences of direct democracy and regime

legitimacy, little attention has been paid to the relationship between them.

This paper tries to fill this empirical void and aims to identify a relationship

between legislative provisions and use of direct democracy, on one hand, and regime

legitimacy, on the other hand. The cross-national analysis includes 39 European countries

ranging from transition countries to established democracies. It uses bivariate statistical

analysis and country-level data collected from legislation, secondary sources, and

aggregate surveys. The central argument of my exploratory study is that possibilities for

public’s direct involvement reflect high importance given to citizens, openness of the

regime towards different ways of making decisions and ways to avoid unpopular

institutions such as parties or the legislature. As a consequence of these benefits, citizens

are likely to accept and support the regime, improving or maintaining its legitimacy.

The first section reviews the existing approaches to legitimacy and outlines the

conceptualization and measurement used in this study. This is followed by a discussion

about direct democracy, its forms, and expectations regarding its relationship with

legitimacy. The third section presents the variable operationalization, data and research

method. The next section presents and interprets the empirical results of the analysis,

while the conclusions discuss potential implications and avenues for further research.

Conceptualizing and Measuring Regime Legitimacy

Starting with Weber’s distinction between authority and coercion, the concept of

legitimacy has become central to a broad array of political analyses because it pertains to

2

how power may be used by state institutions and what are citizens’ reactions to it. In

defining authority Weber (1978) refers to the legitimate use of power that citizens accept

and act upon orders given to them because they consider this to be appropriate

behaviors. Legitimacy can be defined in a broad sense as “the capacity of a political

system to engender and maintain the belief that existing institutions are the most

appropriate or proper ones for the society” (Lipset 1959, p.86). Following this line of

argument, Beetham (1991) defines legitimacy as a sum of three components: conformity

to established rules, justifiability of the rules by reference to shared beliefs, and the

expressed consent of the people.

These definitions point in the direction of two distinct approaches towards

legitimacy. The names of these perspectives vary across studies but they have similar

meanings. On one hand, there is the system level, normative or input-based legitimacy

relating to the working principles and functioning of state institutions with emphasis on

decision-making and exercise of power. On the other hand, we have a grassroots

subjective, descriptive, or output-based legitimacy referring to the public assessment of

the relevance and quality of institutional performance. According to the system level

perspective state institutions must be inclusive, responsive, efficient and fair in order to

be legitimate (Weatherford 1992). These attributes enhance high quality of governance

that, in turn, contributes to their acceptance by the population, i.e. comply with the rules

of society because they believe they should. This approach corresponds to what Blatter

(2007) calls normative legitimacy concerned with the extent to which political realities

comply with ideal settings about how state institutions should be. Since democratic

systems are the only ones in which institutions have such properties, liberal democracy is

considered the norm for political legitimacy. However, the latter is a matter of degree

rather than all or nothing concept (Beetham 1991) and earlier research showed that all

types of political regimes sought legitimacy (Englebert 2002; Gandhi 2008; Gilley 2009;

Levitsky & Way 2010).

On the other hand, legitimacy can be seen from the citizens’ point of view: a

regime is legitimate as long as the population perceives it as such. In the words of

Beetham, state institutions’ conformity to rules has no effect if citizens do not believe

those rules are legitimate. If they are not justified in the terms of shared beliefs, the

population is likely to reject institutions and withdraw its support to be ruled. In brief, this

3

type of legitimacy is evaluative in that citizens decide if political institutions are

acceptable (Diamond 1999; Hurrelmann et al. 2007). These value-based judgments of

obeying state institutions are then translated into behaviors of compliance with laws and

regulations (Gibson & Caldeira 2003; Estlund 2009; Levi et al. 2009). This descriptive

approach (Blatter 2007) towards legitimacy substantively covers the subject (population),

object (state institutions), and the relationship between the two (attitudes and behaviors

as effects of institutional features). In light of these conceptual advantages, this paper

uses the subjective perspective and defines regime legitimacy as citizen evaluations of

the state institutions’ ability to rightfully hold and exercise political power. Essentially, it

refers to the degree of popular acceptance, consent and support of a specific system of

governance.

The use of the descriptive definition does not ignore its complementarity with the

normative approach towards legitimacy (Norris 1999; Levi et al. 2009). On the contrary,

citizens’ evaluations have – at least to some extent – a normative component in that they

compare what institutions are with what they should be according to their views. Thus,

legitimacy is not only the reflection of willing obedience in front of authorities (Tyler

2006) but pertains also to the development and maintenance of supportive attitudes

towards the regime. This idea according to which output-based (subjective) legitimacy

cannot be achieved without input-based legitimacy lies at the core of my theoretical

argument from the following section.

With regard to operationalization, the use of a descriptive conceptualization

brings legitimacy quite close to notions such as political support or trust. Easton (1965,

p.278) explains that “legitimacy is a distinct form of political support that concerns

evaluations of the state from a public or ‘common good’ perspective”. This means that

citizens have the ability to identify their common interests and develop standards to

assess the performance of state institutions with regard to those interests. The degrees

to which citizens’ standards are achieved determine the degrees of legitimacy. This

complex process cuts across the forms of support for state institutions and elites

(specific) or for political community and regime (diffuse). Institutional trust is a useful

concept but too narrow to make it an equivalent of legitimacy. Earlier studies have shown

that trust is usually an assessment of authorities’ performance and/or competence

(Rothstein 2005) or a belief that institutions are motivated to rightfully and fairly deliver

4

what they promise (Levi et al. 2009). People are likely to support those state institutions

producing the goods and services expected of them. In addition, legitimacy refers to the

acceptance of general norms and principles of the political system. This is where the

notion of diffuse support, conceptualized as an established set of attitudes toward

politics and its functioning, comes into play (Dalton 2004).

The combination of these two types of political support leads to a multi-

dimensional structure of legitimacy. Gilley (2006b) has partly followed this path and

measured legitimacy through three components referring to citizens’ views and

acceptance of government legality, justification of the government, and consent to the

government. In this case, the multi-dimensionality refers to the types of attitudes. While

the three indicators are useful, the main shortcoming of his approach is the absence of

regime-oriented attitudes. His operationalization brings legitimacy to close to the array of

specific support. While his choice is understandable given the large number of analyzed

countries – where only few common indicators were available – it is less justifiable from a

theoretical perspective.

Booth and Seligson (2009) have solved this problem by looking at attitudes at a

second layer, i.e. the political system. Their measurement of legitimacy used six

dimensions: support for core regime principles, recognition of a political community,

evaluation of regime performance, support for political institutions, support for local

government, and support for political elites. Among these indicators, the support for

local government is not more than a proxy for legitimacy in the European context.1 This

type of support is embedded in broader attitudes towards core regime principles or

political institutions. Most European countries have effective local government for more

than half a century. Arguably, attitudes towards political elites are reflected in the

evaluation of regime performance or in the support for political institutions.

In measuring legitimacy this paper follows Booth and Seligson’s methodology

after applying the above mentioned modifications, i.e. the removal of two indicators.

Consequently, legitimacy is seen as a function of four types of attitudes: 1) acceptance of

core regime principles; 2) evaluation of regime performance; 3) support for political

1 Booth and Seligson conducted their research in Latin America and the political context may requiredifferent measurement. This is why my criticism should be regarded mostly in the framework of theEuropean analysis from this paper.

5

institutions and 4) support for institutions of authority. While details about

operationalization will be provided in the research design section, it is now time to

discuss about direct democracy as a potential source of regime legitimacy in Europe.

Direct Democracy and Regime Legitimacy

By direct democracy this paper understands four broad types: referendums (including

plebiscites), citizens’ initiatives, agenda initiatives, and recall. Under referendum the

public votes on a proposal put on the ballot by the state institutions (government,

legislature, president) or by themselves in those countries where citizen-initiated

referendums are possible. The citizens’ initiative refers to proposals submitted by citizens

to the legislature; if the latter does not act on the proposal the initiative may be

submitted to popular vote (referendum) alone or with a competing measure proposed by

state authorities. Agenda initiatives bypass the Parliament and citizens make a legislative

proposal on which they vote after circulating a petition. The recall procedure means that

citizens vote to remove from office a representative either at local or national level. One

way to decrease the disaffection with representative democracy is to provide citizens

these alternative opportunities for political involvement. In this sense, representative

mechanisms are not replaced but rather complemented by different means of political

engagement. In reality, not even in the small-size countries, we observe no pure forms of

direct democracy. Instead, there are different degrees to which representative and direct

democracy are. Accordingly, the arguments in the following lines refer to how increased

presence of direct democracy affects legitimacy.

The message conveyed in the previous section is that regime legitimacy depends

upon the popular acceptance of institutional procedures, behaviors, and outcomes. If the

public suspects the decision-making process in their country to be corrupted, legitimacy

suffers. If citizens consider that the degree of their influence over policy-making is low,

legitimacy is likely to be also low. In representative democracies elections are the crucial

mechanism that guarantees citizens’ involvement. However, elections rarely provide

effective opportunities for citizens to influence policy-making (Budge 1996). Elections

allow citizens to choose between several intermediaries who promise to aggregate and

represent their interests. Essentially, they are contests between political competitors –

parties or candidates – with general promises, often fairly similar, that are seldom

6

pursued after elections. Moreover, even when choices between competitors are

meaningful electoral results do not always reflect the popular will, being artificially

influenced by the rules of the game, i.e. electoral systems (Knight & Johnson 1994). As a

result, it is not surprising that there is growing public dissatisfaction with the system of

representative democracy (Norris 1999; Gray & Caul 2000; Pharr et al. 2000; Dalton 2004)

and citizens have become impatient with intermediaries between their opinions and

public policies (Morris 1999).

Direct democracy can compensate for these shortcomings and link citizens’

preferences to political decisions. To begin with, direct democracy can accommodate the

demands of an increasingly educated public. Over the last five decades the changing

social trends have fostered the development of a large category of critical citizens (Norris

1999; Newton & Norris 2000; Dalton 2004; Geissel & Newton 2012). This category rejects

the guidance of representative institutions and criticizes them either for deficiencies in

delivering the expected goods and services or for their redundant tasks. Let us take an

example to illustrate this perceived redundancy. Politics consists of complex and

sophisticated processes and the choice among initial alternatives is not often accessible

to ordinary voters. In this context, political parties simplify choices and generate symbols

of identification and loyalty (Borre & Katz 1973; Rosenstone & Hansen 1993; Dalton &

Wattenberg 2000). Educated voters do not need the simplification of initial alternatives,

while the rest of voters often figure out how to vote on issues of interest to them

(Nicholson 2005). Consequently, the possibility to express non-mediated preferences

allows voters to feel satisfied regarding their deeper involvement in decision-making.

Related to the previous argument, if representative institutions do not live up to

the public expectations of democratic citizenship (Pateman 1970; Budge 1996), direct

democracy may alter the degree of discontent by favoring the emergence of a (partial)

system of self-government. The latter allows the pursuit of interests and preferences that

may be (for various reasons) squeezed out of representative politics. One possible

consequence is a better and more responsible decision-making process from both

representative institutions – that will feel the pressure of alternative policy-making – and

citizens who will have no one to blame for poor decisions. In brief, the openness of the

regime towards direct democracy tools is likely to boost popular evaluations of

legitimacy.

7

Furthermore, representative democracy has been considered too hierarchical,

bureaucratic and partisan to be able to deal effectively with questions of popular

sovereignty or social exclusion. For example, some voters see their preferences discarded

when the parties for which they voted are not large enough to participate in policy-

making, i.e. fail to gain access to Parliament. The disappointment may end up in exclusion

from participation in the next elections. Direct democracy tools reduce the likelihood of

such situations by allowing citizens to express their preferences although they may be in

minority, e.g. through initiatives. Contrary to the case of elections presented above,

direct participation leads to cohesive society with limited exclusion (Barnes 1999).

On a pragmatic note, the use of direct democracy tools can enhance the popular

involvement in representative mechanisms. For example, the organization of a

referendum – on a sensitive or salient issue in society – at the same time with a candidate

race is likely to attract more people to the polls compared to regular elections. This

particular setting can diminish voter apathy because it combines two ways of policy

influence. Moreover, candidates can take stances on the matter debated in the

referendum and thus determine particular segments of the electorate to turn out and

vote for them.

All these arguments suggest that there are theoretical reasons to believe that

direct democracy tools can shape regime legitimacy in a country. Consequently, I

formulate three hypotheses that are tested in the empirical section of this paper:

H1: The existence of many direct democracy tools at national level is associated with high

regime legitimacy.

H2: The existence of many direct democracy tools at local level is associated with high

regime legitimacy.

H3: The extensive use of referendums at national level is associated with high regime

legitimacy.

The hypotheses refer to direct democracy at two different levels and cover both their

existence in the legislation and practical use. This approach is driven by theoretical and

8

empirical reasons. First, local level direct democracy tools are expected to have a

separate effect on legitimacy than the tools at national level because citizens behave

differently. They are concerned about problems in their community and know that direct

involvement in the decision-making process can bring change. At national level, only

salient issues attract the interest of citizens who are also aware that their involvement

has marginal influence due to the high number of voters. Moreover, as it will be

illustrated in the empirical section, some countries adopted local direct democracy in the

absence of such tools at national level (e.g. Belgium, Germany). Corollary, some countries

have national level direct democracy and are less concerned with what happens at local

level.

Second, H3 looks only at the use of national level referendums due to

quantification reasons. Referendums are organized either because they are a) mandatory

as it is the case with EU accession or constitutional changes in some countries, b) initiated

by state institutions (parliamentary majority, president) or c) initiated by citizens. They

are always organized and both successful and unsuccessful outcomes matter. Unlike

referendums, the other three types - citizen initiative, agenda initiative and recall – are

difficult to count. In their case, we can take into consideration only the successful

outcomes, e.g. when a citizens’ initiative is subjected to vote. None of the unsuccessful

attempt can be counted and this introduces bias. At the same time, local level

referendums are too numerous to count and there is no reliable information on their

numbers across countries and time.

The hypothesis testing will refer to the division between Western and Eastern

Europe. This is relevant mainly from the perspectives of democratic consolidation and

experience with democracy. Thus, in addition to the identification of relationships for all

39 countries, variations within the two regions will be also analyzed.

Research Design

The analysis is conducted at country level for a total number of 39 European countries.

This universe of cases resulted after excluding the authoritarian regimes2 (e.g. Azerbaijan,

2 The linkage between direct democracy and regime legitimacy requires the existence of basic rights andliberties such as freedom of expression and involvement. This is the reason for which I included in my studyonly democracies or democratizing countries.

9

Belarus), countries with different rules for segments of population (Bosnia and

Herzegovina), and cases for which (reliable) data were not available (e.g. Andorra,

Cyprus, Israel, Liechtenstein, Turkey). There is a geographical balance between Western

and Eastern Europe (19 vs. 20 cases) and countries of various sizes are included (Germany

and the UK at one extreme and Iceland and Luxembourg at the other).

Variable operationalization

One crucial variable of this study is regime legitimacy. Being about citizens’ evaluations of

the attributes of their political system, survey data can best capture it. I use data from the

4th wave of the European Values Study (2008). This survey is not randomly picked, being

from the last year before the financial crisis: the crisis is likely to influence regime

legitimacy especially that income levels (GDP/capita) are almost deterministic (Gilley

2006a). While survey data is collected at individual level, the aggregation at country level

required several steps that are briefly explained below. As explained at the end of the

section about conceptualization and measurement, regime legitimacy includes four types

of attitudes; for each of them I use two indicators. The acceptance of core regime

principles – basic rules of democracy – is operationalized as the rejection of a strong

leader who does not need the legislature to rule and the rejection of army rule. The

evaluation of regime performance uses two indicators: satisfaction with democracy and

an assessment of how good or bad the current government is. The support for political

institutions is measured as citizens’ support for parliament and government.3 The support

for institutions of authority is measured as people’s support for police and justice system.

Most of these eight indicators (two for every type of attitudes) were measured in

the survey on a four-point scale to capture the degree of support and acceptance. To

aggregate them, I recoded each dichotomously to have value 0 for no acceptance, no

support, or negative assessment and 1 for acceptance, support, or positive evaluation

(see Appendix 1). The next step was the creation of an index of legitimacy that included

the eight indicators, equally weighed (minimum value of 0 and a maximum value of 8).

Each respondent from the 39 analyzed countries had a score on this nine-point index. To

3 In theory, the support for government and the assessment of the government (good or bad) are differentevaluations. The low correlation coefficient between the two (0.39) shows that this is empirically the casein the used dataset.

10

measure legitimacy at country level, I took the average score of individuals in the same

country. The standard deviations of the averages are fairly similar and therefore it is not

necessary to standardize them. As a supplementary check, I have checked the correlation

between these averages and the lowest quartile of the index at country level. The

correlation coefficient is higher than 0.93, statistically significant at 0.01. Table 1 reflects

the distribution of average legitimacy between countries with a range between 2.19 in

Bulgaria and 6.2 in Denmark; the vertical axis presents the values of the country level

index.

Table 1: The Levels of Regime Legitimacy at Country Level

The first two hypotheses speak about the existence of direct democracy tools at national

and local levels. These refer to mentions in the legislation about referendums (mandatory

and optional type), citizens’ initiative, agenda initiative, and recall. Both variables are

count measures and they represent the sum of these types. For example, if a country

specifies in its legislation all direct democracy tools at national level it gets a score of 5

(referendum has two versions); the same applies at local level. Most data comes from the

Direct Democracy Database available on the website of International Institute for

Democracy and Electoral Assistance (IDEA). For the countries that are not included in the

IDEA Database (e.g. Armenia) I gathered data from either primary (legislation) or

secondary sources (reports, earlier studies, or websites dedicated to the topic such as

Navigator to Direct Democracy). Whenever possible, I have triangulated the information

11

for all countries – including those in the IDEA database – to make sure that the

information is accurate and complete.

The variable in the third hypothesis is the use of referendums at national level. In

addition to the number of times when a referendum is used, two further issues are

important (Setala 1999; Mendelsohn & Parkin 2001; LeDuc 2003): what happens after the

vote (binding or not) and how easy a referendum can succeed (required quorum).

Following these considerations, the variable in H3 is calculated with this formula:

Rc/2 x Q + Rb x Q

where Rc is the number of consultative referendums,Rb is the number of binding referendums,Q is the required quorum to pass the referendum.

Binding referendums are weighed twice as much as consultative because citizens have a

much stronger voice if their decisions are immediately implemented rather than re-

examined by an institution. In light of the arguments presented in the previous section,

this can have a higher impact on regime legitimacy. The quorum may be different for the

two types of referendums and has three values: 1 for both turnout and approval

quorums, 2 for turnout only, and 3 for no quorum required. The reasons behind this

ordinal ranking is that the very permissive legislation (no quorums) can make the

referendum institution much more effective than in the case where both approval and

turnout quorums are required to have a successful referendum. These scores are

calculated for the referendums organized between 1990 (to ensure comparability

between Western and Eastern Europe) and 2008.

The low number of cases does not allow sophisticated quantitative analyses. At

the same time, the goal of this paper is exploratory and intends to identify the existence

of relationships between direct democracy and regime legitimacy. This can be done with

bivariate analysis in the form of correlations presented in the following section.

Findings

General descriptive statistics sheds light on the extent to which direct democracy tools

are mentioned in the legislation of the 39 analyzed countries. To begin with the national

level, Figure 1 presents six categories according to the number of legislative provisions.

12

Belgium and Germany are the only countries that do not allow any direct involvement of

citizens at national level. The numbers indicate that among the countries with direct

democracy provisions at national level the smallest category is that of countries with one

tool and that is always the referendum. This category is not geographically clustered and

includes Czech Republic, Greece, Norway, Sweden, and the United Kingdom. The largest

categories are those of countries that have two and four direct democracy tools at

national level. Almost half of the countries fall in one of these two categories. Many of

the countries in the category with two types do not combine referendums with other

types of direct democracy but provide citizens the possibility to engage in both

mandatory and optional referendums. Examples of countries using this approach are

Armenia or Estonia I Eastern Europe, but also Denmark, France or Ireland in Western

Europe.

Figure 1: The Distribution of Direct Democracy Tools at National Level in Europe

The category of countries with inclusive legislation, i.e. mentioning all types of direct

democracy, has a clear geographic concentration. With the exception of Italy, all other

countries come from Eastern Europe. This is not surprising since many post-communist

countries have opted for participatory constitutions when they changed the regime in the

beginning of the 90s. This option for inclusive and direct citizen participation was partly a

reaction to the decades of authoritarianism. For some of them, the way in which

institutional change took place is illustrative for their options. For example, in Hungary,

13

the negotiations regarding the transfer of power between communists and opposition

took the form of Round Table Talks (Walsh 1994) where 50 delegates and 500 experts

formed committees and sub-committees. The whole process involved approximately

1,000 documented meetings for three months (Bruszt 1990). Another example is Slovakia

where the “velvet divorce” from the Czech Republic in 1993 has involved some

consultations with the masses.

The direct democracy at local level has more categories because I have counted

the regulations for both regional and community level. In some countries, in particular

the federations, the regional level is very important. Thus, it is not surprising the only two

countries with nine provisions (Figure 2) regarding direct democracy at local level are

federal: Germany and Switzerland. At the other extreme, Greece is the only country with

no direct democracy at local level. Since the institution of referendum exists at national

level, there is no possibility to use it at local level. Instead of direct democracy tools,

Greece gives preference to dialog-oriented procedures such as participation in the town

meetings or assemblies of neighborhood councils.

Figure 2: The Distribution of Direct Democracy Tools at Local Level in Europe

The category of countries with four direct democracy tools at local level is the largest,

being closely followed by the categories with one or two such tools. Although there are

more categories at the local than at national level, more than 80% of the countries have a

maximum number of four provisions in their legislation. Sometimes, the legislation does

14

not explicitly mention the possibility for direct involvement of citizens, but allows room

for manoeuver in practice. For example, in Belgium the constitution prohibits any regional

or local level referendums making it clear that state institutions at local level are

responsible for decision-making. According to this provision, local authorities cannot be

compelled by citizens in the exercise of their responsibilities. However, this does not

impede local authorities to organize consultative referendums on particular matters; such

consultative votes have been organized in practice. Similarly, in Denmark the local

authority can decide without the approval of a higher administrative institution to hold

consultative referendums although there is no provision in the legislation.

The correlation between the provisions for direct democracy at national and local

level is negative (-0.14, not statistically significant).4 This value indicates a weak tendency

of countries with no or few direct democracy provisions at national level to have

legislation on (more) direct democracy tools at local level. The most obvious example in

this case is Germany with no direct democracy tool at national level but with nine at local

level. France, Luxembourg and Poland are other examples of countries following this

pattern. The reciprocal is also valid: countries with extensive provisions at national level

provide fewer opportunities to their citizens to get involved at local level. An example is

Georgia that has five provisions at national level and only one for local level or Portugal

(fours vs. two). The low value of the coefficient is due to the fact that not many countries

have sharp differences between the national and local level; several countries have

exactly the same number e.g. Bulgaria, Estonia, or Iceland. The correlation has a different

sign and strength when we divide the universe of countries according to their

geographical (and political) location. For Western Europe the coefficient is very small

(0.05) indicating the absence of an empirical relationship. In Eastern Europe the

coefficient is negative (-0.26) indicating a medium tendency of countries with more

provisions for direct democracy at national level to allow less involvement at local level.

Empirically, this is the case in a series of countries such as Latvia, Lithuania, Macedonia,

Moldova, or Montenegro.

4 The statistical significance does not provide substantial information because the analysis takes intoaccoutn the universe of cases and not a sample from which we can generalize. At the same time, thenumber of cases is fairly limited and this can influence the level of significance.

15

Similarly to the legislative provisions for direct democracy, the use of referendums

has broad variations between the examined countries. The score obtained after using the

formula presented in the research design section ranges between 0 in countries like

Bulgaria, Greece or Iceland to more than 150 in Switzerland.

A Complex Picture

Table 1 summarizes the correlation coefficients for the three hypothesized relationships.

The bivariate statistical analyses for all European countries indicate empirical support for

H2 and H3. At the same time, evidence goes against H1: levels of regime legitimacy are

lower in countries with more provisions for direct democracy at national level. The

relationship is driven by East European countries that have legislation with many tools for

direct democracy and low levels of legitimacy. The adoption of inclusive legislation during

democratization was encouraged by the negative experiences with limited

representation in the previous regime, on one hand, and the European Union (EU) pre-

accession agreements, on the other hand.

Table 1: The Correlation Coefficients for Direct Democracy and Regime LegitimacyAll European Countries Western Europe Eastern Europe

Direct Democracy at National Level -0.40* -0.05 -0.20Direct Democracy at Local Level 0.18 0.17 -0.13Use of National Referendums 0.25 0.16 0.47** p > 0.05; ** p > 0.01.

These differences are not only relative to Western European countries but also within the

region. The coefficient calculated separately for Eastern Europe shows that even within

the region countries with lower legitimacy have more legislation about direct democracy

at national level. For illustrative purposes Figure 3 includes a comparison of direct

democracy provisions (gray dots) and regime legitimacy (black dots) in the EU New

Member States from Eastern Europe (these are 10 out of the 20 East European countries

in the analysis). The comparison between these 10 countries is relevant because their

democratic performances were officially acknowledged in 1999 when the EU started the

accession process; two of them – Bulgaria and Romania – were late joiners due to delayed

reforms. The negative relationship can be easily observed also within this small group:

countries with many provisions (Bulgaria, Hungary, Latvia and Lithuania) have low levels

16

of legitimacy, while countries with relatively high levels of legitimacy (Czech Republic,

Estonia, Poland and Slovenia) have few provisions.

Figure 3: National Level Direct Democracy and Regime Legitimacy in Eastern Europe (EU NMS)

Regarding H2, there is a positive relationship between direct democracy provisions at

local level and regime legitimacy (0.18). As the correlation between national and local

level provisions for direct democracy was negative (see the previous sub-section), it is

quite intuitive to have a different relationship with regime legitimacy. This time, West

European countries have more legislative provisions and proportionally higher levels of

regime legitimacy than countries in Eastern Europe. Within the regions, East European

countries follow the same pattern as in the case of H1: there is a negative correlation

indicating that countries with more provisions have lower levels of legitimacy.

In Western Europe the coefficient is fairly similar with what observed for the

pooled analysis. To better illustrate this relationship Figure 4 includes 10 countries

selected on the basis of their geographical positioning (three from North, four from

Central Europe, and three from South). The provisions for direct democracy at local level

(gray dots) are minimal in Norway and Portugal and quite extensive in Germany and

Switzerland. Regime legitimacy (black dots) is quite homogenous among these countries

especially when compared with the legitimacy for Eastern Europe in Figure 3. The

distribution of the 10 countries on these two variables indicates the existence of a

positive relationship. Countries with high number of provisions (Spain, France, Germany,

17

and Switzerland) have higher levels of legitimacy than countries with few provisions

(Portugal, Italy). The relationship would have been stronger if Norway or the Netherlands

– both with high levels of legitimacy – had more direct democracy provisions in their

legislation. This observation takes us back to the theoretical expectations and provides

one possible explanation for the few provisions in some West European countries. These

countries in which legitimacy is high – and has been high for decades –have little

incentives to promote direct democracy tools at national level.

Figure 4: Local Level Direct Democracy and Regime Legitimacy in Western Europe

Finally, there is empirical support for H3: in countries that use more referendums (binding

and easy to pass) the levels of regime legitimacy are higher than in the rest of cases. The

correlation coefficient has an average value (0.25) with consistent positive signs in both

regions. In Eastern Europe the coefficient is considerably higher (o.47 statistically

significant at 0.05) than in Western Europe (0.16). This result tells a complex story when

corroborated with the findings from H1. According to the latter, many legislative

provisions about direct democracy at national level correspond to low legitimacy. As the

use of referendums is positively associated with legitimacy, we may conclude that direct

democracy does not increase legitimacy if it is only on paper. The mere existence of

provisions without practical implementation may not be perceived as alternative means

of involvement. This is particularly observable in Eastern Europe where the high number

of provisions was not backed by practical use. The bivariate correlation between

18

provisions and national level and use of referendums in Eastern Europe shows no

relationship between the two; the value of the coefficient is 0.01.

The implementation of direct democracy is a possible explanation for the different

results provided by H1 and H2. Local level direct democracy rarely remains only on paper.

Although difficult to quantify the exact amount of tools used at local level there are many

examples showing that it is practiced in different communities across countries. As the

existence of regulations at the local level is often accompanied by implementation, it

differs significantly from what encountered at national level.

Conclusions

This paper tried to identify a relationship between direct democracy and regime

legitimacy in 39 European countries. The main results indicate that provisions for direct

democracy at local level and use of referendums at national level are positively associated

with legitimacy. The negative correlation between provisions for direct democracy at

national level and legitimacy may be explained through the absence of implementation.

The major conclusion of these findings is that direct democracy may have an impact only

if applied. When provisions remain solely on paper, direct democracy does not trigger the

expected attitudes. Equally important, this paper showed that the differences between

Eastern and Western Europe are smaller than initially considered.

This analysis is preliminary in its conclusions. Its exploratory goals were primarily

meant to identify an empirical relationship and to draw attention to a possible

explanation for the variation of regime legitimacy in Europe. Its major implication is

theoretical in the sense that these results give sufficient reason to include direct

democracy in future analyses of sources of legitimacy. On empirical grounds, this study

has produced evidence to account for synchronic differences between levels of

legitimacy. Moving beyond the relationships presented here, more empirical evidence is

required to determine how direct democracy tools offer citizens incentives to increase

their regime legitimacy. Further research can also investigate reverse causation with low

levels of legitimacy leading to the adoption of direct democracy tools. Regime legitimacy

can be better understood as soon as the correspondence between formal provisions of

direct democracy and practices is established.

19

Appendix 1: The Operationalization of Regime Legitimacy

Dimension Item Question Answers Code

Acceptance of core

regime principles

Attitude against

strong leader

Q66.A: Having a strong leader who does not have to bother with

parliament and elections

Very good

Fairly good

Fairly bad

Very bad

0

Attitude against army

rule

Q66.C Having the army rule the country 1

Evaluation of regime

performance

Satisfaction with

democracy

Q64: On the whole are you (answer options) with the way democracy is

developing in our country?

Very satisfied

Rather satisfied

Not very satisfied

Not at all satisfied

1

0

Good vs. Bad

Government

Q65: People have different views about the system for governing this

country. Here is a scale for rating how well things are going: 1 means

very bad; 10 means very good

1-5 0

6-10 1

Support for political

institutions

Support for

legislature

Q63: Please look at this card and tell me how much confidence you have

in Parliament.

A great deal

Quite a lot

Not very much

None at all

1

Support for

Government

Q63.R: Please look at this card and tell me how much confidence you

have in Government.

0

Support for

institutions of

authority

Support for police Q63.F: Please look at this card and tell me how much confidence you

have in the police.

A great deal

Quite a lot

Not very much

None at all

1

Support for justice Q63.N: Please look at this card and tell me how much confidence you

have in the justice system.

0

20

Bibliography:

Barnes, M., 1999. Researching Public Participation. Local Government Studies, 25(4),

pp.60–75.

Beetham, D., 1991. The Legitimation of Power, London: Macmillan Education Ltd.

Blatter, J., 2007. Legitimacy. In M. Bevir, ed. Encyclopedia of Governance. Thousand Oaks:

Sage Publications, pp. 519–523.

Booth, J.A. & Seligson, M.A., 2009. The Legitimacy Puzzle in Latin America: Political Support

and Democracy in Eight Nations, New York: Cambridge University Press.

Borre, O. & Katz, D., 1973. Party Identification and Its Motivational Base in a Multiparty

System: A Study of the Danish General Election of 1971. Scandinavian Political Studies,

8(1), pp.69–111.

Bruszt, L., 1990. 1989: The Negotiated Revolution in Hungary. Social Research, 57(2),

pp.365–387.

Budge, I., 1996. The New Challenge of Direct Democracy, Cambridge: Polity Press.

Dalton, R.J., 2004. Democratic Challenges, Democratic Choices: The Erosion of Political

Support in Advanced Industrial Democracies, Oxford: Oxford University Press.

Dalton, R.J. & Wattenberg, M.P. eds., 2000. Parties Without Partisans: Political Change in

Advanced Industrial Democracies, Oxford: Oxford University Press.

Diamond, L., 1999. Developing Democracy: Toward Consolidation, Baltimore: Johns

Hopkins University Press.

Easton, D., 1965. A Systems Analysis of Political Life, New York: Wiley.

Englebert, P., 2002. State Legitimacy and Development in Africa, Boulder: Lynne Rienner

Publishers.

Estlund, D.M., 2009. Democratic Authority: A Philosophical Framework, Princeton:

Princeton University Press.

Gandhi, J., 2008. Political Institutions Under Dictatorship, New York: Cambridge University

Press.

Geissel, B. & Newton, K. eds., 2012. Evaluating Democratic Innovations: Curing the

Democratic Malaise?, London: Routledge.

Gibson, J.L. & Caldeira, G.A., 2003. Defenders of Democracy? Legitimacy, Popular

Acceptance, and the South African Constitutional Court. The Journal of Politics,

65(01), pp.1–30.

21

Gilley, B., 2006a. The Determinants of State Legitimacy: Results for 72 Countries.

International Political Science Review, 27(1), pp.47–71.

Gilley, B., 2006b. The Meaning and Measure of State Legitimacy: Results for 72 countries.

European Journal of Political Research, 45(3), pp.499–525.

Gilley, B., 2009. The Right to Rule. How States Win and Lose Legitimacy, New York:

Columbia Press.

Gray, M. & Caul, M., 2000. Declining Voter Turnout in Advanced Industrial Democracies,

1950 to 1997: The Effects of Declining Group Mobilization. Comparative Political

Studies, 33(9), pp.1091–1122.

Hurrelmann, A., Schneider, S. & Steffek, J. eds., 2007. Legitimacy in an Age of Global

Politics, Basingstoke: Palgrave Macmillan.

IDEA, 2014. Direct Democracy Database. Available at:

www.idea.int/elections/dd/search.cfm [Accessed April 2, 2014].

Knight, J. & Johnson, J., 1994. Aggregation and Deliberation: On the Possibility of

Democratic Legitimacy. Political Theory, 22(2), pp.277–296.

LeDuc, L., 2003. The Politics of Direct Democracy: Referendums in Global Perspective,

Toronto: Broadview Press.

Levi, M., Sacks, A. & Tyler, T., 2009. Conceptualizing Legitimacy, Measuring Legitimating

Beliefs. American Behavioral Scientist, 53(3), pp.354–375.

Levitsky, S. & Way, L.A., 2010. Competitive Authoritarianism: Hybrid Regimes After the Cold

War, New York: Cambridge University Press.

Lipset, S.M., 1959. Some Social Requisites of Democracy: Economic Development and

Political Legitimacy. American Political Science Review, 53(1), pp.69–105.

Mendelsohn, M. & Parkin, A., 2001. Referendum Democracy Citizens, Elites and Deliberation

in Referendum Campaigns, Basingstoke: Palgrave Macmillan.

Morris, D., 1999. The New Prince, Los Angeles: Renaissance Books.

Newton, K. & Norris, P., 2000. Confidence in Public Institutions: Faith, Culture or

Performance? In S. J. Pharr & R. D. Putnam, eds. Disaffected Democracies: What’s

Troubling the Trilateral Countries? Princeton: Princeton University Press, pp. 52–73.

Nicholson, S.P., 2005. Voting the Agenda: Candidates, Elections, and Ballot Propositions,

Princeton: Princeton University Press.

22

Norris, P. ed., 1999. Critical Citizens: Global Support for Democratic Government, Oxford:

Oxford University Press.

Pateman, C., 1970. Participation and Democratic Theory, Cambridge: Cambridge University

Press.

Pharr, S.J., Putnam, R.D. & Dalton, R.J., 2000. Trouble in the Advanced Democracies? A

Quarter-Century of Declining Confidence. Journal of Democracy, 11(1), pp.7–25.

Rosenstone, S.J. & Hansen, J.M., 1993. Mobilization, Participation, and Democracy in

America, New York: Macmillan Publishing Company.

Rothstein, B., 2005. Social Traps and the Problem of Trust, New York: Cambridge

University Press.

Scarrow, S., 2001. Direct democracy and institutional change. A comparative

investigation. Comparative Political Studies, 34(6), pp.651–665.

Setala, M., 1999. Referendums in Western Europe - A Wave of Direct Democracy?

Scandinavian Political Studies, 22(4), pp.327–338.

Smith, G., 2009. Democratic Innovations. Designing Institutions for Citizen Participation,

Cambridge: Cambridge University Press.

Tyler, T.R., 2006. Psychological perspectives on legitimacy and legitimation. Annual

Review of Psychology, 57, pp.375–400.

Walsh, H., 1994. Political Transition Processes in Central and Eastern Europe. Comparative

Politics, 26(4), pp.379–394.

Weatherford, M.S., 1992. Measuring Political Legitimacy. American Political Science

Review, 86(1), pp.149–166.

Weber, M., 1978. Economy and Society: An Outline of Interpretive Sociology, Berkeley:

University of California Press.