Why malpractices generate pressures for electoral reform: An agenda setting model.

NORRIS: MALPRACTICES AND ELECTORAL REFORM www.electoralintegrityproject.com Why malpractices generate pressures for electoral reform: An agenda setting model. Pippa Norris Harvard University and the University of Sydney Department of Government and international Relations 443, Merewether Building (HO4) The University of Sydney NSW 2006 Australia Cell: (US)+1 857 445 9105 Email: pippa.norris@sydney.edu.au John F. Kennedy School of Government 79 JFK Street Harvard University Cambridge, MA 02138, USA Cell: (US)+1 857 445 9105 Email: Pippa_Norris@Harvard.edu Synopsis: Do electoral malpractices strengthen public pressures for reform? The study develops an agenda setting model postulating that where flawed elections fail to meet international standards of electoral integrity, this colors public perceptions of electoral malpractices, undermines feelings of political legitimacy, dampens voter turnout, and encourages protest politics. To support these claims, Part I sets out the theoretical framework and core propositions. Part II outlines the research design, data, and indices. Data is drawn from a new battery of survey evidence measuring perceptions of electoral integrity and malpractices, included for the first time in the 6 th wave of the World Values Survey 2010 2012 (WVS). The results in Part III confirm that public perceptions of electoral integrity closely reflect independent assessments of the quality of elections in the countries under comparison. Moreover, lack of confidence in elections undermines feelings of political legitimacy and encourages protest activism. The conclusion in Part IV theorizes that this process strengthens the likelihood that elites will respond to popular pressures, although the outcome may be democratic concessions, or state repression, or regime transitions, depending upon the type of regime. Keywords: elections, political legitimacy, protest politics, electoral reform. 1

NORRIS: MALPRACTICES AND ELECTORAL REFORM Countries around the world face major challenges in meeting international standards of electoral integrity. The most overt malpractices used by rulers include imprisoning dissidents, harassing adversaries, coercing voters, vote rigging, ballot stuffing and finally, if losing, blatantly disregarding the people s choice. 1 Serious violations of human rights, undermining electoral credibility, are widely condemned by domestic observers and the international community. But in many countries, minor irregularities are more common, exemplified by inaccurate voter registers, maladministration of polling, petty corruption, pro government media bias, lack of transparent campaign finance regulation, tabulation miscounts, and excessively high vote seat thresholds. Even long established democracies are not immune, as exemplified by the notorious hanging chads in Florida in 2000, more recent accusations of voter suppression through over zealous identification requirements during the Obama Romney contest, 2 and security vulnerabilities in UK postal ballots. 3 Even where electoral officials prove impartial, professional and independent, new challenges arise in maintaining standards, exemplified by contemporary European debates about the most appropriate regulation of campaign finance, the rules governing political broadcasting, and the deployment of new voting technologies. 4 A growing literature derived from several sub fields, developed by scholars and applied policy analysts, has started to study these issues by conceptualizing the normative standards of electoral integrity, 5 gathering systematic evidence from electoral forensics, content analysis of international observer election reports, and mass surveys, 6 examining techniques commonly used to manipulate electoral processes, 7 and identifying the effects of deploying observers to monitor malpractices. 8 Building upon this emerging literature, this study seeks to determine whether, and under what conditions, electoral malpractices strengthen public pressures for reform. The agenda setting model advanced in this study postulates that where elections are unfair, corrupt, or flawed, this strengthens public perceptions of electoral malpractices, undermines feelings of political legitimacy, dampens voter turnout, and encourages protest politics. By contrast, where elections meet international standards of electoral integrity, so that contests are widely regarded as legitimate, this reduces mass pressures for reform. To buttress this argument, Part I sets out the theoretical framework and core propositions. Part II outlines the research design, data, and indices. Problems of electoral integrity and malpractices are gauged from Susan Hyde and Nikolay Marinov s dataset National Elections across Democracy and Autocracy (NELDA). 9 To monitor popular perceptions, this study presents the initial results of a new battery of survey evidence measuring perceptions of electoral integrity and malpractices, included for the first time in the 6 th wave of the World Values Survey 2010 2012 (WVS). Surveys currently remain in the field in many countries but at present ten societies can be compared with diverse elections, ranging from Uruguay, Estonia and Mexico to Zimbabwe, Azerbaijan, and Kazakhstan. The WVS survey also includes a wide range of indicators of system support, to measure feelings of political legitimacy, as well as measuring voting participation and protest activism. Part III analyzes the empirical evidence for the core propositions. Part IV summarizes the key findings, and considers their broader implications. The conclusion theorizes that this process increases the likelihood that elites will respond to mass pressures and the loss of legitimacy, although, as Acemoglu and Robinson suggest, the outcome may be democratic concessions, or state repression, or regime transitions. 10 The analysis supports the core claims. Concerning the first proposition, the evidence strongly confirms that public perceptions of electoral integrity in the ten countries under comparison closely reflect the actual quality of elections (monitored in the NELDA dataset) and the quality of liberal democracy (measured by Freedom House), suggesting that the public is capable of making a rational and informed assessment. Many political institutions are far removed from ordinary people s daily lives, making it difficult to judge directly how well they are working, but elections provide unique 2

NORRIS: MALPRACTICES AND ELECTORAL REFORM opportunities to engage every citizen. Moreover the public does not just make blanket yes/no judgments about the quality of these contests; instead different problems with elections are distinguished and highlighted by people living in each country, for example highlighting the problems of electoral violence in Nigeria and Zimbabwe, lack of fairness among electoral officials in Kyrgyzstan and Ukraine, and the role of money politics in Ghana and Mexico. These nuanced evaluations suggest that the public identifies specific problems of elections, probably fairly accurately according to observer reports, so that judgments are not simply reflecting broader assessments about the performance of democracy. Moreover, public perceptions of electoral integrity and malpractices have a direct effect upon feelings of political legitimacy (monitored by satisfaction with the performance of democracy, satisfaction with respect for human rights, and confidence in elected institutions), as well as direct and indirect effects upon political participation (voter turnout and protest activism). The evidence therefore demonstrates that where elections are flawed, public concern is likely to increase mass pressure from below for reform of electoral institutions and processes, highlighting the importance of this issue on the policy agenda. Whether and how elites choose to respond, when seeking to restore legitimacy and contain popular protests, however, is theorized to differ in democratic and autocratic regimes. I: Theoretical framework The concept of electoral integrity is gaining in popular usage as an all encompassing way to conceptualize many related issues. 11 Similar notions include a thicket of negative phrases such as electoral malpractice, flawed elections, manipulated contests, and electoral fraud, as well as positive terms, where elections are described as credible, acceptable, genuine, reflecting the will of the people, or the standard diplomatic rhetoric of free and fair. 12 In this study, which is part of The Electoral Integrity Project (www.electoralintegrityproject.com), it is proposed to ground the overarching concept of electoral integrity in terms of international commitments, endorsed in a series of conventions, treaties, protocols, and guidelines, with agreed principles applying universally to all countries worldwide throughout the electoral cycle, including during the pre electoral period, the campaign, on polling day, and its aftermath. Conversely, electoral malpractice is used to refer to violations of electoral integrity, whether intentional or unintentional, legal or illegal. Flaws may arise from the failure to observe democratic rights, or lack of effective electoral governance capacity, or indeed some mix of both. There is no question that electoral malpractices are widely regarded as intrinsically important where they constitute violations of basic democratic principles: Article 21 in the 1948 Universal Declaration of Human Rights declares that the will of the people shall be the basis of the authority of government; this will be expressed in periodic and genuine elections which shall be by universal and equal suffrage and shall be held by secret vote or by equivalent free voting procedures guaranteeing everyone the right to take part in the government of his country, directly or through freely chosen representatives. 13 These principles have been elaborated and endorsed in many subsequent international treaties and instruments, notably the International Covenant on Civil and Political Rights, and the International Covenant on Economic, Social and Cultural Rights which entered into force in 1976. 14 But what are the instrumental consequences of electoral integrity? In long established democracies like the UK and US, it is commonly believed (although not well established) that even minor irregularities, such as slipshod clerical errors in out of date registration rolls, have the capacity to corrode citizen s confidence and trust in electoral institutions and processes. It is also feared that this could have more serious consequences, if voting fraud and errors undermine political legitimacy, depress voter turnout, and catalyze demonstrations. 15 Public discontent, reinforced by opposition 3

NORRIS: MALPRACTICES AND ELECTORAL REFORM parties, the news media and watchdog NGOs, is expected to strengthen calls for electoral management bodies to rectify these problems in subsequent contests. Problems in established democracies, however, pale in significance compared with flawed contests held in many fragile states around the world, which may potentially trigger fatal consequences. Popular elections have now become a standard part of peace building initiatives in deeply divided societies racked by decades of conflict, exemplified by Afghanistan, Nepal, and the Democratic Republic of Congo, as part of regime change, such as in post Mubarak Egypt and post Gaddafi Libya, as well as in newly independent nations which are still building state capacity, such as Timor Leste and South Sudan. As illustrated by Kenya in 2007, claims of even minor electoral irregularities may be used by demagogues to trigger social instability, riots, and deadly violence, undermining fragile gains in democratization and development. 16 Following the Kenyan elections, the UN High Commissioner for Human Rights estimated that more than 1,200 were killed and thousands injured, including cases of sexual violence, about 42,000 houses and many businesses were looted or destroyed, with more than 300,000 people displaced, costing the country more than one billion dollars and deterring potential investors. 17 Kutz and Thomson argue that in cases such as Serbia, Georgia and Ukraine, stolen elections have been revolutionary triggers, catalyzing regime change through mobilizing ordinary citizens, strengthening the opposition, and dividing the regime. 18 It is therefore vitally important to understand what strengthens electoral integrity. This raises a large research agenda which includes establishing whether and under what conditions the public comes to believe that elections are unfair, fraudulent, or manipulated, and whether these perceptions undermine feelings of political legitimacy, strengthen protests, and thereby stoke mass pressures for strengthening electoral institutions and processes. To help understand these issues, in earlier work, I developed a model which depicts the policymaking process for changes to electoral institutions and processes as a series of cyclical stages involving agenda setting, policymaking, implementation, and feedback evaluation. 19 Building upon this idea, here the agenda setting stage can be broken down further into a sequential process, illustrated schematically in Figure 1. [Figure 1 about here] The first step in the model (H 1 ) suggests that events in the real world shape public concerns. In particular, through the feedback learning loop, malpractices which come to light before, during, or after any election, as well as broader flaws in the workings of democratic institutions and processes, are likely to strengthen public beliefs that electoral processes are fraudulent, dishonest, or unfair, undermining confidence in the electoral process. This proposition is not as straightforward as it might appear: the public is often regarded are being poorly informed about many basic factual matters about politics, such as public policies, political institutions, and events in their own country, at least in the United States. 20 Knowledge of democracy, such as the ability to identify the classic characteristics of liberal democracy, also varies substantially worldwide, strengthening with historical experience of democratization. 21 When asked to express satisfaction with the working of democracy, for example, citizens in Vietnam and China awarded their system higher marks than those living in Italy and the United States. 22 Nevertheless unlike other political institutions, popular elections provide unique opportunities to engage all citizens, allowing informed judgments to be made on the basis of direct experience at the ballot box, as well as from critical cues provided by the reaction of the news media, political parties, watch dog NGOs, official election management bodies, and the international community. Common malpractices are widely evident to many citizens; in Nigeria, for example, according to Afro barometer estimates, one out of five Nigerians was personally exposed to vote buying and around one in ten experienced threats of electoral violence. 23 It follows that the public can be expected to offer more informed and rational judgments 4

NORRIS: MALPRACTICES AND ELECTORAL REFORM about the working of elections than other more abstract political institutions, such as the courts or legislature. Integrity and political legitimacy Moreover the second proposition (H 2 ) predicts that public confidence in electoral integrity will influence broader feelings of political legitimacy. The concept of legitimacy represents acceptance of the underlying rules of the game, so that all actors, even electoral losers, willingly consent to rule by the regime, without the sanction of force. Legitimacy, in Seymour Martin Lipset words, "involves the capacity of a political system to engender and maintain the belief that existing political institutions are the most appropriate and proper ones for the society." 24 For this reason alone, citizens feelings of political legitimacy should be regarded as a critical indicator of the underlying stability of the regime. Feelings of legitimacy are typically measured by indicators of support for the political system. As in previous work, following the classic Eastonian framework 25, the concept of system support is understood as a multidimensional phenomenon which can be broken down into its separate parts. 26 This includes five components, ranging from diffuse to specific levels: (i) (ii) (iii) (iv) The most general and fundamental feelings of citizens towards belonging to the national community, exemplified by feelings of national pride and identity; Support for general regime principles, including approval of democratic and autocratic attitudes and values; Evaluations of the overall performance of the regime, exemplified by satisfaction with the workings of democracy; Confidence in state institutions, notably government, parliaments, parties, the civil service, the courts, and the security forces; and (v) Trust in elected and appointed office holders, including politicians and leaders. Confidence in electoral integrity is predicted to be especially important at the most specific levels for the popular acceptance of the outcome and thus trust in the legitimate authority of elected officials and elected institutions. Conversely public perceptions of electoral malpractices (whether based on true or false claims) are likely to weaken confidence in elected authorities. Long after Florida in 2000, for example, many Democrats continued to question the legitimacy of President George W. Bush, due to the way that he ascended to office, following a close result and the bitter battle of the lawyers, with the final outcome determined by the Supreme Court. 27 Perceptions that elections are won fair and square should therefore strengthen confidence and trust in regime institutions which are most closely related to these contests, including political parties, parliaments, and elected presidents, with only weaker diffuse effects on other state institutions, such as the judiciary, civil service, and military. A long series of flawed and fraudulent contests may also corrode general satisfaction with the performance of democracy in each country. 28 Some empirical evidence from previous studies supports these propositions. In Russia, for example, Rose and Mishler found that perceiving the Duma elections as unfair made people feel less proud of their country, as well as reducing trust in state institutions such as parties, parliament and the army, and it also made them less likely to endorse the regime. 29 Similarly McAllister and White found that perceptions of electoral fairness in Russia proved significant predictors of satisfaction with democracy and a sense that the system is responsive to citizen s concerns (external political efficacy). 30 Elsewhere, as well, Afrobarometer data suggests that Africans consider elections to be the best means 5

NORRIS: MALPRACTICES AND ELECTORAL REFORM of forming a government; and they judge the quality of democracy accordingly. 31 High quality elections in the region are found to strengthen feelings of legitimacy. Moreover in my last book, comparing countries worldwide based on the 5 th wave of the World Values Survey (2005 7), citizen satisfaction with democracy in each society proved to be closely related to the procedural performance of democratic governance, measured at macro level by objective macrolevel indices for levels of democratization, good governance, and human rights. By contrast, satisfaction with democracy was usually unrelated to most macro level indicators of policy performance, reflecting the capacity of governments to manage the delivery of basic public goods and services, such as levels of economic growth, human development and social welfare. Based on this evidence, the previous study concluded that the public was capable of making fairly rational evaluations about the performance of democracy, based on how well democratic procedures actually worked. This relationship strengthened with each country s historical experience of liberal democracy, suggesting a learning process. Moreover, when people were asked in the World Values Survey about how they understood the notion of democracy, a procedural conceptualization also proved by far the most common in every type of society. Thus eight out of ten people saw free elections in which people chose their leaders as an essential characteristic of democracy. 32 For all these reasons, public perceptions of the quality of elections are expected to shape feelings of political legitimacy. Electoral integrity and political behavior In turn, the third proposition claims that lack of confidence in election integrity is likely to affect political participation, by depressing voter turnout and also simultaneously strengthening the propensity for ordinary people to engage in protest activism. The model in Figure 1 also suggests that there could also be an indirect effect of perceptions on behavior, which is mediated by political legitimacy. The reasoning is straightforward. Where an election is widely regarded as rigged, corrupt, or unfair, this is likely to depress conventional forms of civic participation, especially voting turnout among those backing losing parties. After all, confidence that the political system is responsive to citizen s concerns (external political efficacy) has long been established as strongly associated with political participation. 33 Several studies support the idea that lack of confidence in electoral processes reduces turnout. For example McCann and Dominguez examined a series of public opinion polls in Mexico in the mid 1980 to the mid 1990s, during the time of the predominance of the PRI and prior to more competitive and cleaner contests. They found that those citizens expecting electoral fraud were more likely to stay home on election day, and this group was also more likely to support the opposition. 34 In the United States, Alvarez et al found that American who were more confident about the integrity of electoral procedures were more likely to cast a ballot, with malpractices having particular strong impacts by depressing African American turnout. 35 Birch s study, utilizing data from Module I of the Comparative Study of Electoral Systems, confirmed that perceptions of electoral fairness help to account for cross national patterns of voter turnout. 36 At the same time, the impact may not be so straightforward, as different types of malpractices may well have contrary effects. For example, widespread threats of violence at polling places may plausibly depress the willingness of citizens to risk casting a ballot, but in contrast patronage may induce voting; vote buying has been found to provide material incentives which encourage participation. 37 These cross cutting pressures may cancel each other out. Feeling that the election outcome was fraudulent, rigged, or stolen can also be expected to encourage direct actions by anti government forces, notably opposition boycotts, political strikes, and mass demonstrations protesting against the process and outcome. Large scale protests against Putin s return to the Kremlin were mobilized following the 2012 Russian presidential contest, despite the crackdown by security forces and new laws restricting demonstrations. 38 Similar movements were evident in the so called colored revolutions in Eastern Europe, where mass mobilization reportedly led 6

NORRIS: MALPRACTICES AND ELECTORAL REFORM to the reversal of the results of several fraudulent elections. 39 More recently, widespread mass demonstrations were also triggered in Mexico City by claims of vote buying and fraud, following victory for Institutional Revolutionary Party (PRI) candidate Enrique Peña Nieto in the 2012 presidential contest. Similarly, mass protests also erupted in Tahrir Square following the constitutional coup d état suspending the newly elected parliament in Egypt. Accordingly the direct and indirect effects of perceptions of electoral integrity on mass political participation can be examined. II Research Design, Data and Evidence Empirical evidence to test each of these theoretical propositions requires independent expert indices of the quality of electoral integrity and liberal democracy in each country, as well as measures of mass perceptions of electoral integrity and malpractices, political legitimacy, voting turnout, and protest activism. Indices of electoral integrity and democracy Independent evidence about common problems in elections is available from upon Susan Hyde and Nikolay Marinov s coding of National Elections across Democracy and Autocracy (NELDA). 40 The dataset has the advantage of covering all national legislative and presidential election events in independent nation states with populations over half a million during the period from 1945 to 2010. Moreover the dataset is derived from an extensive list of secondary sources, including academic election handbooks, online resources, news media, and official reports, in addition to observer reports. Compared with other available indices, NELDA was selected as it provides more comprehensive coverage of countries worldwide, and it is also derived from multiple sources of evidence. A measure of electoral integrity was developed using the NELDA dataset by combining four items: Nelda11 Before elections, are there significant concerns that elections will not be free and fair?, Nelda13 Were opposition leaders prevented from running, Nelda15 Is there evidence that the government harassed the opposition? and, Nelda16 In the run up to the election, were there allegations of media bias in favor of the incumbent. The items do not tap into all dimensions of the broader concepts of electoral integrity and malpractices, by any means, nevertheless they do monitor several common problems, especially those found in many electoral autocracies. The items were tested using principal component factor analysis with varimax rotation and they were found to form a single dimensional. Reliability checks further confirmed that the four items formed a consistent scale (Cronbach s Alpha 0.756). The items were summed and the score was standardized, for ease of comparison. Mean scores on this index were summarized for each nation for national legislative and executive elections held during the decade from 2000 2010. In addition, public perceptions of the quality of elections are likely to be colored by cumulative experience of broader political processes and institutions, including the role of political parties, elected officials, and parliaments, as well as by the outcome of particular contests. For example, the Russian presidential election in March 2012, while reported to be flawed, were only part of a series of similar contests during the previous decade, and assessments of the outcome are likely to be framed within the context of the gradual but sustained erosion of human rights under the leadership of Vladimir Putin. 41 Accordingly confidence in electoral integrity is also expected to be shaped at a more diffuse level by broader indices of liberal democracy, measured in this study from the Freedom House index of political rights and civil liberties, providing the broadest coverage of countries worldwide, with the 14 point score standardized to 100 points for ease of interpretation. Measuring public perceptions of electoral integrity Evidence of public opinion is derived from the 6 th wave (2010 2012) World Values Survey. 42 Previous public opinion surveys have occasionally been used to gauge various components of the quality 7

NORRIS: MALPRACTICES AND ELECTORAL REFORM of elections, for example, the annual Gallup World Poll has asked survey respondents in over one hundred countries about their confidence in the honesty of election in their own country. 43 Unfortunately single dichotomous (yes/no) items, reflecting thin notions, are unable to capture the multidimensional and thicker concept of electoral integrity; for example, in some elections, problems could arise from vote fraud and dishonest practices at the ballot box and vote count. But elsewhere types of malpractices unrelated to honesty per se may be serious, such as pro government media bias, restrictions on ballot access for opposition parties, or threats of violent intimidation used against opponents. Along similar lines, survey items monitoring fraudulent activities, such as manipulation of the ballot box, can overlook malpractices arising from incompetence or maladministration. To improve measurement for thicker concepts, the 6 th wave of the World Values Survey (2010 2012) asked respondents the following battery of questions: In your view, how often do the following things occur in this country s elections? V228A. Votes are counted fairly V228B. Opposition candidates are prevented from running V228C. TV news favors the governing party V228D. Voters are bribed V228E. Journalists provide fair coverage of elections V228F. Election officials are fair V228G. Rich people buy elections V228H. Voters are threatened with violence at the polls V228I. Voters are offered a genuine choice in the elections Since these are new items, it is worth discussing their design and construction in some detail, before examining the evidence. The questions were framed to capture perceptions of this country s elections in general, at whatever level, rather than to gauge reactions to a specific contest for local, legislative or executive office. Clearly controversial results in recent high profile contests may have been at the forefront of respondents minds when answering, such as accusations of voter fraud in Ukraine or scandals over corruption in Italy, but the study sought to monitor perceptions of the usual quality of elections in each country over successive contests. Careful measurement is also important. Previous indices of electoral integrity have often been coded using binary categories or ordinal level measures using limited scales, denoting summary pass/fail judgments. But continuous or internal level measures are arguably preferable, providing finer grained responses, as perceived problems arise with different degrees of severity. Respondents in the WVS were thus asked to respond to each item using a 4 point ordinal scale ranging from very often to not at all often. Items on the list were also ranked to intersperse positive and negative questions, to encourage respondents to consider each item separately rather than passing summary judgments. Conceptual validity When selecting questions, it is important that they meet the requirements of conceptual validity, reflecting the underlying notions which we seek to measure. The items selected for inclusion in the survey are based upon the definition of electoral integrity advanced earlier, as they reflect widelyaccepted international principles and standards, such as the requirement that ballots should be fairly tabulated, opposition candidates should be free to run for elected office, and citizens should not face intimidation or violence at the polls. These basic and minimal requirements are firmly grounded in international conventions on universal human rights. The list of items did not monitor dimensions of electoral integrity where international agreement has not yet been established, such as asking whether there should be limits on campaign spending or donor contributions. These questions were also designed to reflect universal issues in elections, described in everyday language, avoiding more technical 8

NORRIS: MALPRACTICES AND ELECTORAL REFORM matters which may be poorly understood, such as attitudes towards specific laws, regulations or electoral procedures. The questions also ask respondents about their perceptions, rather than their direct experience. One reason is that respondents may be reluctant to acknowledge first hand involvement in some problems, such as vote bribery or violence, due to fear of retribution, legal penalties, or moral norms. The significance and severity of any violations for electoral integrity also varies across the items, as does their meaning and interpretation in different contexts, for example in judging what counts as genuine choice or pro government bias in TV news. [Figure 2 about here] The electoral cycle approach adopted in this study assumes that the most overt techniques which are intentionally designed to distort the genuine will of the people arise from ballot stuffing, vote rigging and fraud on polling day and its aftermath. But many more subtle and legal techniques which also violate international standards are often employed well away from the public spotlight to limit party and candidate competition before a single ballot is cast, during the pre election and campaigns periods, such as through overly restrictive voter registration processes, ballot access requirements, and electoral thresholds. As shown in Figure 2, the items were therefore designed to reflect different stages in the electoral cycle, not simply problems occurring on polling day. The conceptualization used in this study also assumes that no country is wholly immune from problems of electoral administration and challenges of human rights. To reflect this understanding, the questions were designed to reflect universal issues which can potentially undermine electoral integrity in every country. Thus although electoral violence and bribery of voters are exceptional today in long established democracies, the fairness of election officials and pro government bias in TV news are common challenges for the quality of elections in many societies. Summary scales The individual items are examined separately, to identify specific problems in each election. For an overview, however, the Perceptions of Electoral Integrity scale and the Perceptions of Electoral Malpractice scale were both constructed by summing the separate positive and negative item responses, each with equal weight, and standardizing the scale to 100 points, for ease of comparison with the expert indices. Principal Component Factor Analysis with varimax rotation, shown in Table 1, confirmed that the items fell into two distinct dimensions, measuring positive and negative evaluations. Country coverage [Table 1 about here] Data collection for the 6 th wave currently remains in progress, but nevertheless the initial results which are already available can be analyzed to compare responses in ten countries. Once fieldwork is completed, the final results of the 6 th wave WVS will expand the analysis of these items in around thirty societies. 44 The cases used for analysis in this study include diverse types of regimes, as well as countries in sub Saharan Africa, post communist Europe, and Latin America. This includes three electoral autocracies holding contests which are seriously flawed for different reasons, as exemplified by violent intimidation and state repression of opposition activists in Zimbabwe, flawed parliamentary elections in Azerbaijan, and non competitive contests in Kazakhstan, where President Nursultan Nazarbayev won an estimated 95.5% of the vote. Elections are also analyzed where there are less serious violations reported, including in Nigeria which experienced post election inter communal violence, Ukraine where opposition parties alleged electoral fraud, and Kyrgyzstan with a series of administrative irregularities in recent contests. Public perceptions are also compared with several elections which observers agree display relatively higher standard of integrity, including Mexico (prior to the controversy over vote buying in the 2012 presidential context), Ghana, Estonia, and Uruguay. The 9

NORRIS: MALPRACTICES AND ELECTORAL REFORM Technical Appendix summarizes background information about the election to the lower house of the national legislature or the presidential election which was immediately prior to the date of the survey fieldwork in each country. [Table 2 about here] Feelings of political legitimacy are measured and analyzed using several indicators in the WVS survey, tapping into system support. This included overall satisfaction with the performance of democracy and respect for human rights in each county, as well as confidence in elected political institutions (parties, parliaments and governments). Models in this paper control for many micro level social and demographic characteristics which have commonly been found to shape political orientations. This includes age (in years), sex (male), education (educational qualification 9 point standardized scale, household income scale, and urbanization (an 8 point scale for the size of urban rural communities). 45 All models were checked through tolerance statistics to confirm that there were free from problems of multicollinearity. III: Findings and analysis Do real world events shape public perceptions? The first proposition (H 1 in Figure 1) suggests that public confidence in elections is acquired from direct experience at the ballot box and of the overall democratic quality of processes and institutions. If so, then evaluations of electoral integrity by the general public within each country should reflect independent indices of electoral integrity and of liberal democracy. Although there are only ten cases for comparison, the results of the analysis do indeed confirm a remarkably strong and significant correlation between the public perceptions of electoral integrity and contemporary levels of democratization, as measured by Freedom House (R=.784, Sig. p=.007), as well as with the NELDA index estimating electoral integrity during the previous decade (R=.654, Sig. p=.055). Table 2, Figure 3 and Figure 4 illustrates these patterns. Hence citizens proved most confident about elections in Uruguay, Estonia and Ghana, and to a lesser extent in Mexico, all countries which were ranked highly in electoral integrity by NELDA and in liberal democracy by Freedom House. By contrast, people were most critical of elections in Zimbabwe, Kyrgyzstan and Kazakhstan, countries rated poorly by the two independent indices. This correlation suggests that standards of electoral integrity are not simply the product of expert assessments; instead global norms are widely shared by the general population in each society as well. The main outlier to these patterns was Ukraine, where relatively poor public perceptions can be attributed to the recent history of elections in this country and the heated controversy triggered by opposition allegations of large scale fraud occurring during the 2010 presidential elections, although OSCE observer mission reported that this particular contest largely met international standards. 46 [Table 2, Figure 3 and Figure 4 about here] To get a better sense of the results, responses in the WVS survey can be broken down into the component parts, as summarized in Table 3, to throw see whether the public distinguishes across the range of items in different countries, or whether they provided an overall blanket assessment. The results suggest that the public identified some specific problems in the elections under comparison and these can be plausibly related to the specific type of regime and the particular context of each country s political background and electoral experiences, as summarized in Technical Appendix A. Cases show that the public does distinguish among different problems and the flaws they emphasize are commonly also highlighted by observer reports. 10

NORRIS: MALPRACTICES AND ELECTORAL REFORM In the two African states, for example, citizens highlight the problem of electoral violence, a topic which has aroused considerable concern in the region. 47 In Zimbabwe, during the first round presidential elections in March 2008, the opposition leader, Morgan Tsvangirai (MDC) gained more votes than President Robert Mugabe (ZANU PF), although no candidate was declared to have achieved the 50% threshold required for victory. A second round run off election was held in June, but Mr Tsvangirai subsequently withdrew following numerous reports of systematic attacks against MDC activists, including killings and abductions. Citizens were coerced to vote in the one candidate second ballot. The Secretary General of the UN criticized the violence and intimidation surrounding the elections. The violence only ceased following a brokered and uneasy power sharing agreement between the major leaders. In the case of Nigeria, the April 2001 Presidential and parliamentary elections were deemed credible and largely successful by African Union observers, with high turnout and with praise for the role of the Independent National Electoral Commission. After voting ended, the public could observe the vote count in each polling station. President Goodluck Johnson from the governing People's Democratic Party (PDP) was declared the winner with 57% of the vote, avoiding the need for a run off second ballot. Yet his victory was quickly followed by major outbreaks of post election inter communal violence between Muslim and Christian communities in several northern cities. Thousands of people have died over the past few years in communal rivalry and separatist aspirations have been growing, with the north lagging behind in education, infrastructure, and poverty. The results of the WVS survey show that a bare majority of Nigerians thought that elections were fair, but nevertheless, they expressed concern about electoral violence. By contrast in Ukraine and Kyrgyzstan, the public is far more concerned about problems of bribery and corruption, as well as pro government bias in broadcasting, and lack of fairness in electoral administration. This largely reflects the reports of electoral observer from the international community. Thus in Kyrgyzstan, following the 2011 presidential elections, the OSCE observer mission report noted problems in the integrity of voting, counting and tabulation with cases of ballot box stuffing, multiple and family voting, vote buying, and bussing of voters. 48 In the case of the Ukrainian 2010 elections, the country's first presidential contest since the 2004 Orange Revolution, the OSCE observer mission reported that elections largely met international standards, the process was transparent, and voters were offered a genuine choice of candidates, although the legal framework was deficient and unsubstantiated allegations of electoral fraud negatively affected the election atmosphere and voter confidence in the process. The opposition leader, Tymoshenko, who lost with 45% of the vote in the second round, alleged "large scale fraud", before subsequently withdrawing her charges, and then being sentenced to 7 years imprisonment for abuse of office. It appears that these accusations were widely believed, since the survey showed exceptionally widespread perceptions of electoral malpractices in Ukraine, largely through problems of corruption rather than violence. [Table 3 about here] Lastly, by contrast, Ghana exemplifies a case where the general public proved relatively positive in their perceptions of electoral integrity, which largely reflects expert assessments. Among West African nations, Ghana is widely regarded today as one of the most successful contemporary democracies on the continent. During the late 1980s the country began to move towards economic stability and the transition towards democracy. In April 1992, a constitution allowing a multi party system was approved in a referendum, ushering in a sustained period of democracy under the fourth republic. Since then, Ghana has experienced periodic multiparty contests which international observers have regarded as free and fair. 49 In the presidential elections in December 2008, the results which were extremely close, with less than 0.5% of the vote separating the top two presidential candidates, but nevertheless the outcome was largely peaceful. The European Union was among a host of institutions 11

NORRIS: MALPRACTICES AND ELECTORAL REFORM observing these contests, including the Carter Centre, ECOWAS, African Union, and over 4000 representatives from the Coalition of Domestic Observers (CODEO). The EU reported that the Electoral Commission who administered the contests proved impartial, professional, and independent, ensuring the transparency of the process. Electoral observers and party agents were able to observe all stages of polling, vote counting, and aggregation. There were clear legal channels for complaints and appeals challenging the results and the Commission worked to ensure conciliation and acceptance of the process among the major stakeholders. There were some minor administrative irregularities experienced on polling day, but the Electoral Commission apologized for these and sought to rectify the situation. Disputes followed the second round of presidential voting, and tensions rose in the tight contest with some slight delays in announcing the vote, but these were eventually resolved without conflict. The perceptions of electoral integrity items show that most people in Ghana believe that their elections are fair, and that there is free competition for opposition candidates, but nevertheless they remain more critical of voter bribery and suspicious of the power of rich people to buy elections. Therefore overall public evaluations are strongly correlated with independent expert indices of electoral quality and liberal democracy. The fact that public evaluations in each country also vary across different items increases confidence in the first proposition (H 1 ), suggesting that experience of elections in each society shapes public concern about their integrity. Indeed Hyde and Marinov argue that there is a more direct process, as international observer reports provide a credible cue to citizens about the quality of elections, especially where other reliable sources of domestic information are unavailable, and critical reports help to trigger popular protests. 50 Nevertheless irrespective of whether the source of information is direct experience, domestic cues, or international observers, as predicted, problems occurring during any election, and general weaknesses in democratic institutions and processes, strengthen the perception that electoral processes are flawed, undermining public confidence in the electoral process. The impact on feelings of political legitimacy (H 2 ) The second proposition (H 2 ) suggests that perceptions of electoral quality are likely to influence feelings of political legitimacy. Table 4 presents a series of regression models analyzing the effects of perceptions of electoral integrity and malpractices on three indicators of legitimacy, confidence in elected institutions (parties, parliaments and governments), satisfaction with the performance of democracy, and respect for human rights controlling for other related political attitudes and social characteristics which have commonly been found to shape political orientations. The results confirm the significant and consistent impact of confidence in elections on indicators of political legitimacy. Thus perceptions of electoral integrity were positively related to satisfaction with democracy, respect for human rights and confidence in electoral institutions. By contrast, perceptions of electoral malpractice were consistently related to each of these indices. Figures 5 and 6 illustrate these patterns visually, suggesting a linear relationship. The fact that similar patterns were observed across three different indices, as well as other measures not presented here, increases confidence in the robustness of the results. 51 These effects persisted as relatively strong and significant despite the battery of attitudinal and social controls which are incorporated into the models. Moreover comparison of the standardized beta coefficients can be used to compare the relative importance of each of these factors. The analysis demonstrates that in each of the models, electoral integrity and malpractice were consistently stronger predictors of each of the indices of political legitimacy than the effects of either sex, age income, education, or urbanization. The other attitudinal controls behaved much as expected, for example confidence in elected institutions was strongly associated with overall satisfaction with democracy and respect for human rights, although with cross sectional data it remains impossible to determine the direction of causality in this association. The evidence therefore strongly confirms the second 12

NORRIS: MALPRACTICES AND ELECTORAL REFORM proposition (H 2 ): public perceptions of electoral integrity and malpractices strongly shape feelings of political legitimacy. [Table 4, Figures 5 and 6 about here] The impact on political participation (H 3 and H 4 ) Moreover in Figure 1, confidence in the quality of elections is expected to have a direct effect by increasing voting participation (H 3 ), while perceptions that elections are flawed are simultaneously predicted to encourage protest activism (H 4 ). Indirect effects are also predicted, mediated by feelings of political legitimacy (H 5 and H 6 ). Table 5 presents the results of the models predicting political participation, with multiple controls. Figures 7 and 8 illustrate the bivariate relationships. [Table 5, Figures 7 and 8 about here] The result of the multivariate models confirm, as expected, that perceptions of electoral malpractices proved significant predictors of protest activism. Indeed malpractices had a stronger effect on protest than other attitudes, such as satisfaction with democracy or institutional confidence, as well as most of the standard demographic variables such as age, sex and education. The only stronger predictors of protest activism (measured by the size of the standardized beta coefficients) were for political interest and urbanization. At the same time, however, the picture was not wholly clear cut, since perceptions of electoral integrity also mobilized protest activism. This remains something of a puzzle. One possible reason, which remains to be determined, could be that anti government protests among opposition movements are often met by counter reactions mobilizing supporters of the winning parties. In Florida in 2000, for example, both Republicans backing Bush and Democratic supporters of Gore mobilized to carry placards and demonstrate vocally for and against the decisions of the electoral officials and courts. Electoral winners and losers have been found to react very differently to the outcome of any contest, viewing the results through a partisan prism. For example, in comparative European studies, losers have been proved far less satisfied with the way that democracy works, especially in newer democracies. 52 In Russia, as well, trust in elections was observed to vary among winners and losers. 53 This explanation will be explored in subsequent research. In terms of voting participation, feelings of electoral integrity were significantly related to turnout; those with more faith in the process proved more willing to cast a ballot. 54 Again the effects (in the standardized betas) outweighed all other control variables except for political interest and age. This is exactly as predicted: belief that the system is fair is more likely to encourage citizens to think that it is worthwhile to take part at the polls. Perceptions of electoral malpractice were also positive but not statistically significant. IV: Conclusions and implications Therefore the initial evidence compared in this research has amassed considerable empirical support for the agenda setting model, although further work remains. Once fieldwork for the World Values Survey is completed, then additional analysis can be tested across a broader range of countries and institutional arrangements. To recapitulate the key findings, however, firstly, public perceptions of the quality of elections in each country were closely related at macro level to independent expert assessments about both the quality of elections which have occurred during the last decade (NELDA) and contemporary levels of political rights, civil liberties, and democratization (Freedom House). The specific problems which were 13

NORRIS: MALPRACTICES AND ELECTORAL REFORM emphasized by the mass public, such as violence, corruption, and pro government media bias, also closely reflected issues highlighted by international observers and media coverage of these contests. Moreover public perceptions of electoral integrity and malpractices proved to be strong and consistent predictors of feelings of political legitimacy, including satisfaction with the way that democracy works, trust in electoral institutions such as parties and parliaments, and assessment of how far countries respect human rights. In understanding the phenomenon of critical citizens, it seems that the quality of elections is a previously largely over looked missing link, as the public often assesses democratic performance by their experience of the way that elections work or fail to do so. Lastly, there are behavioral consequences as well: perceptions of electoral flaws, fraud or manipulation do encourage popular protest, while by contrast belief in the fairness of elections encourages voter turnout. The behavioral impacts need to be further analyzed in future research, however, to distinguish the reactions of election winners and losers, as well as controlling for institutional contexts, such as the type of electoral system, the use of compulsory voting, and polling facilities, all of which have often been found to influence turnout. 55 So far, so good. But does the agenda setting process matter for changing electoral institutions? Evidence demonstrating the impact of agenda setting upon public policymaking, and thus the outcome of electoral reform debates, is well beyond the scope of this study. Nevertheless drawing upon the research literature, as Acemoglu and Robinson theorize, the response of elites to any public concerns about electoral processes seems likely to diverge sharply under democratic and autocratic regimes. 56 To understand the impact of the agenda setting process, we can utilize a policy cycle model of electoral reform developed in previous work. 57 This identifies multiple actors as players in a sequential policymaking process, involving: i. The agenda setting stage in the public sphere, engaging the public, political parties, the media and NGOs, which heightens the salience of institutional reforms as a key problem to be addressed on the policy agenda; ii. iii. iv. The policy making stage in the state, where policy options are formulated, coalitions are built, and regulatory policies are adopted to address these perceived problems, directly engaging decisions by political parties in the legislature and the executive, as well as, indirectly, the influence of public referendums and international assistance; The implementation stage, where the revised regulatory framework is put into practice in subsequent elections, involving election management bodies and the courts; and finally, The feedback evaluation loop, when learning about the consequences of the new regulatory framework shapes either satisfaction with the status quo or further demands for subsequent revisions. All this activity is occurs within a broader environment in each country, including the role of pathdependent historical democratic cultures, the type of past and present regime, and the socio economic structure of development. 58 Conventional explanations of changes to electoral institutions have focused upon the policymaking arena in established democracies, notably rational choice theories of stable equilibrium among parliamentary parties. 59 From this perspective, the key decision makers in electoral reform debates are actors within the legislature and executive who seek to maintain or revise constitutional, legal and regulatory frameworks when calculating the optimal rewards in elected office anticipated from any rule changes. 60 In these accounts, the concept of partisan interest can be framed in terms of policy outcomes, office seeking, and/or personal gain for key actors. 14

NORRIS: MALPRACTICES AND ELECTORAL REFORM In a widely cited article, for example, Carles Boix argued that the choice of electoral systems derives from the decisions ruling parties make to maximize their representation. As long as the electoral arena does not change and the current electoral regime benefits the ruling parties, the electoral system is not altered. As the electoral arena changes (due to the entry of new voters or a change in voters' preferences), the ruling parties modify the electoral system, depending on the emergence of new parties and the coordinating capacities of the old parties. Similarly, Kenneth Benoit posits that electoral laws change when two conditions are fulfilled: (i) when a party coalition forms with the power to alter electoral rules; and (ii) where each party in the coalition expects to benefit by gaining more seats under alternative electoral institutions. 61 Echoing these views, Josep Colomer argues that existing parties in assemblies and governments tend to prefer the electoral formulas that reinforce their power, with party fragmentation in many countries gradually encouraging the adoption of more inclusive formulas. 62 Rational choice explanations of changes to electoral institutions and processes have become conventional in the research literature but this approach is silent about why reform debates arise on the political agenda in the first place. As Leijenaar and Hazan argue, attention needs to shift towards explaining what takes place before the passage of any changes to electoral laws. 63 Moreover the newer body of research has focused on multiple dimensions of reform, not simply changes to the basic type of electoral system. 64 From this viewpoint, like other areas of public policy, the puzzle is to understand when, where and why some constitutional, legal and procedural electoral amendments become passed and implemented, while others fall by the wayside. Attention has also shifted from understanding the path dependent constraints which limit institutional change in long established democracies towards explaining more flexible institutional choices in transitional states, especially those like Afghanistan, Egypt, and Libya which are in the process of establishing new constitutions and electoral arrangements. My previous research demonstrated that democratic aspirations (support for democracy as an ideal) proved to be a strong, significant, and robust predictor of the occurrence of subsequent electoral reforms. 65 But attitudes towards democracy are relatively abstract and perceptions of electoral integrity can be expected be more closely linked with the outcome of reform processes. Therefore we can theorize that the outcome of popular pressures may well differ according to the type of regime. As Acemoglu and Robinson suggest, the outcome of public discontent and protest uprisings may be democratic concessions, or state repression, or regime transitions. 66 In democratic states, public concern about the quality of elections, low or declining levels of voter turnout, and popular protests, should all heighten pressures for the introduction of legal and regulatory institutional changes designed to ameliorate electoral flaws. In democracies, public concerns can be freely expressed without fear of reprisal or repression and electoral malpractices are open to criticism by losing parties, the news media and NGOs. Studies of political representation in established democracies have reported that policymakers often prove responsive to public concerns, although the strength of the congruence is contingent upon multiple institutional conditions. Hence in a cross national study, Soroka and Wlezien develop a thermostatic model of democratic responsiveness where the strength of any relationship connecting public preferences with government spending is expected to be contingent upon the saliency of the policy issues and several institutional features which determine the policy process in each regime, including the allocation of responsibilities in federal systems and in presidential, mixed, or parliamentary executives. 67 Others within country comparisons have linked the predominant ideology or general attitudes within U.S. states with public policies or policy outcomes in each state, such as the congruence between liberal attitudes toward homosexuals or feminism and related policies gay marriage bans or the number of abortions. 68 In more democratic regimes, therefore, sustained popular protests demonstrating against the legitimacy of electoral processes and outcomes can be expected to make it more likely, although far 15

NORRIS: MALPRACTICES AND ELECTORAL REFORM from certain, that policymakers will seek to adopt legal or procedural institutional changes designed to allay public concern. After widespread and noisy partisan protests following the events in the Gore v. Bush race in Florida in 2000, for example, Congress passed the Help America Vote Act (HAVA) in 2002, designed to replace mechanical punch card voting machines, to create an Electoral Assistance Commission, and to establish minimum electoral standards. 69 Nevertheless policies may fail to achieve their intended objectives; for example tracking surveys since 2004 suggest that controversy surrounding new voter registration requirements implemented by many American states may have slightly weakened, not strengthened, voter confidence in the electoral process. 70 In general, evidence derived from several Anglo American democracies suggests that changes to electoral institutions, such as the adoption of proportional representation electoral systems or legislative term limits, have little impact, at least in the short term, upon strengthening feelings of political efficacy and patterns of civic engagement. 71 There is a different logic in repressive states, however, as state control of the news media means that the public may not be aware of electoral malpractices or, if aware, they may not necessarily feel free to express their concerns or to mobilize dissent. Any mass protests and popular uprisings which do occur may potentially trigger either regime concessions, leading ultimately to more competitive multiparty contests, strengthening processes of democratization. This process is exemplified by developments in Mexico during the 1990s, were domestic and international pressures encouraged the ruling party to implement new measures gradually cleaning up elections. 72 Staffan Lindberg has emphasized that a series of elections in Africa, even flawed contests, provide platforms for the opposition which gradually facilitates the process of democratization. 73 Yet instead of progressive outcomes, alternatively popular protests may lead towards further state repression, as occurred recently following the 2012 Russian presidential elections. The contest was criticized by OSCE observers. 74 Once Putin s victory was announced, demonstrators immediately rallied and called for the election results to be annulled and for vote rigging to be investigated. Far from capitulating to public concerns, however, the Duma cracked down after the contest by passing new penalties for unsanctioned protests. Police raided the homes of several prominent activists and some independent websites went down, although this did not deter tens of thousands of citizens from engaging in antigovernment demonstrations on Russia Day, demanding fresh elections. 75 Or alternatively, in the case of several color revolutions, such as Georgia in 2003 or Ukraine in 2004, popular discontent with elections may embolden the opposition and trigger international pressures, forcing regime transition from below. 76 Thus there is not simple and direct link between public concern and electoral reform; instead the overall impact of flawed contests is expected to be conditioned by the broader context in which these problems come to light, especially the opportunities for public concerns to influence the policymaking agenda within democratic and autocratic regimes. 16

NORRIS: MALPRACTICES AND ELECTORAL REFORM Table 1: Factor analysis of survey items How often in this country's elections : Electoral malpractice Voters are bribed.801 Rich people buy elections.788 Voters are threatened with violence at the polls.691 TV news favors the governing party.620 Opposition candidates are prevented from running.505 Electoral integrity Election officials are fair.763 Votes are counted fairly.696 Journalists provide fair coverage of elections.692 Voters are offered a genuine choice in the elections.665 Notes: Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Source: World Values Survey 6 th wave 2010 2012, 10 nations 17

NORRIS: MALPRACTICES AND ELECTORAL REFORM Table 2: Evaluations of electoral integrity, malpractices and liberal democracy Measure Electoral Integrity Electoral malpractice Electoral integrity Liberal democracy Data Source and period WVS 2011 12 WVS 2011 12 NELDA Hyde and Marinov 2000 2010 Scales 100pts 100pts 100pts Freedom House 2011 100pts Uruguay 75 46 100 100 Estonia 75 54 100 100 Ghana 70 63 85 93 Mexico 68 70 100 71 Kazakhstan 67 52 12 36 Nigeria 65 72 50 57 Azerbaijan 64 46 43 36 Kyrgyzstan 63 67 40 43 Zimbabwe 62 66 20 28 Ukraine 59 65 68 64 Notes: See Tables 1 for the source of the WVS indices and the technical appendix for details about their construction. N/a Not available Source: World Values Survey 6 th wave 2010 2012, 10 nations 18

Table 3: Perceptions of electoral integrity and malpractice Country Code Votes are counted fairly Electoral integrity Election officials are fair Voters are offered a genuine choice in the elections Journalists provide fair coverage of elections Opposition candidates are prevented from running TV news favors the governing party Electoral malpractice Voters are bribed Rich people buy elections Voters are threatened with violence at the polls Uruguay 91 81 66 73 62 41 15 24 4 Estonia 81 71 78 62 44 61 41 35 5 Ghana 61 64 72 65 23 47 67 60 49 Kazakhstan 62 47 63 60 31 61 29 39 10 Azerbaijan 54 52 64 57 28 44 14 16 7 Mexico 52 50 72 69 49 64 75 72 54 Nigeria 54 49 53 62 53 65 72 70 66 Zimbabwe 52 53 54 47 38 68 47 52 54 Kyrgyzstan 53 34 61 56 56 70 68 63 36 Ukraine 38 25 47 56 59 72 71 79 19 Total 60 52 62 60 45 62 55 56 34 Notes: Q: In your view, how often do the following things occur in this country s elections? The above gives the proportion in each nation responding Very often and fairly often. The electoral integrity and electoral malpractice scales provide a mean score across the items. Source: World Values Survey 6 th wave 2010 2012, 10 nations. 19

Table 4: Electoral quality strengthens political legitimacy PERCEPTIONS OF ELECTION QUALITY Satisfaction with democracy Respect for human rights Confidence in elected institutions B S.E. Beta P B S.E. Beta P B S.E. Beta P Electoral integrity.024.001.164 ***.004.001.079 ***.131.012.111 *** Electoral malpractice.023.001.169 ***.002.001.049 ***.160.012.146 *** ATTITUDINAL CONTROLS National pride.299.036.085 ***.128.013.103 *** 3.37.295.121 *** Importance of democracy.212.012.168 ***.025.004.056 ***.391.104.039 *** Confidence in elected institutions.023.001.181 ***.007.000.169 *** Interest in politics.077.025.030 ***.041.009.046 *** 2.39.207.119 *** Life Satisfaction.112.012.101 ***.001.004.002 N/s.395.096.045 *** Satisfaction with democracy 1.56.090.196 *** SOCIAL CONTROLS Sex (male).101.048.020 *.008.017.005 N/s.128.401.003 N/s Age (years).010.002.066 ***.002.001.038 ***.032.013.025 * Income scale.135.013.107 ***.023.005.052 ***.487.112.048 *** Education.082.012.072 ***.014.004.034 ***.203.096.022 * Urbanization.022.010.022 *.016.004.044 ***.556.083.069 *** Constant 1.526 1.53 44.3 N. observations 8227 8218 8238 N. nations 10 10 10 Adjusted R 2.247.233.178 Note: OLS regression models monitoring satisfaction with the performance of democracy, respect for human rights and confidence in elected institutions as indicators of legitimacy. See the technical appendix for the definition and construction of all variables. The columns include the unstandardized beta (B), the standard error (SE), the standardized beta coefficient (b) and the significance. *=.05 **=.01 ***=.001. Source: World Values Survey 6 th wave 2010 2012, 10 nations 20

Table 5: Electoral quality predicts political participation Protest activism Voting participation B S.E. Beta P B S.E. Beta P PERCEPTIONS OF ELECTION QUALITY Electoral integrity.006.001.069 ***.008.001.099 *** Electoral malpractice.011.001.132 ***.002.001.032 ** ATTITUDINAL CONTROLS National pride.080.024.037 ***.134.022.068 *** Importance of democracy.016.008.020 *.026.007.037 *** Satisfaction with the performance of democracy.013.007.021 *.039.007.070 *** Confidence in elected institutions.005.001.061 ***.000.001.006 N/s Interest in politics.318.017.205 ***.239.015.168 *** Life Satisfaction.022.008.033 ***.013.007.020 N/s SOCIAL CONTROLS Sex (male).296.032.097 ***.044.029.016 N/s Age (years).009.001.089 ***.023.001.257 *** Income scale.034.009.044 ***.024.008.033 ** Education.009.008.013.024.007.038 *** Urbanization.083.007.133 ***.037.006.066 *** Constant 4.26 2.87 N. observations 8203 8227 N. nations 10 10 Adjusted R 2.109.247 Note: OLS regression models where the 9 point scale monitoring protest activism as the dependent variable. See the technical appendix for the definition and construction of all variables. The columns include the unstandardized beta (B), the standard error (SE), the standardized beta coefficient (b) and the significance. *=.05 **=.01 ***=.001. Source: World Values Survey 6 th wave 2010 2012, 10 nations 21

Figure 1: Agenda setting model of electoral integrity Electoral participation H3 H5 Quality of elections and democracy H1 Public perceptions of electoral integrity and malpractices H2 Political legitimacy H6 Elite response: Concessions, repression, or regime transition H4 Protest activism 22

Figure 2: The WVS survey items measuring components in the cycle of electoral integrity Pre electoral period Campaign period Polling day and its aftermath Opposition candidates are prevented from running TV news favors the governing party Voters are threatened with violence at the polls Rich people buy elections Journalists provide unfair coverage of elections Voters are bribed Election officials are unfair Voters are not offered a genuine choice in the elections Votes are counted unfairly Source: World Values Survey 6 th wave 2010 2012 23

Figure 3: Correlations of public perceptions of electoral integrity with liberal democracy Note: For the construction of the electoral integrity scale, see Table 1 and the technical appendix. Source: World Values Survey 2010 2012 24

Figure 4: Correlations of public perceptions of electoral integrity with expert assessments Note: For the construction of the electoral integrity scale, see Table 1 and the technical appendix. Sources: National Elections across Democracy and Autocracy (NELDA) Susan D. Hyde and Nikolay Marinov http://hyde.research.yale.edu/nelda/; World Values Survey 6 th wave (2010 2012). 25

Figure 5: Electoral integrity strengthen confidence in elected institutions Note: I am going to name a number of organizations. For each one, could you tell me how much confidence you have in them: is it a great deal of confidence, quite a lot of confidence, not very much confidence or none at all?... The government (in your nation s capital), Political parties, Parliament. Source: World Values Survey 2010 2012 26

Figure 6: Electoral integrity increases satisfaction with the performance of democracy Note: And how democratically is this country being governed today? Again using a scale from 1 to 10, where 1 means that it is not at all democratic and 10 means that it is completely democratic, what position would you choose? Source: World Values Survey 2010 2012 27

Figure 7: Electoral integrity strengthens voting participation Note: When elections take place, do you vote always, usually or never? Please tell me separately for each of the following levels local level, national level. Source: World Values Survey 2010 2012 28

Figure 8: Electoral malpractices strengthen protest activism Note: Now I d like you to look at this card. I m going to read out some forms of political action that people can take, and I d like you to tell me, for each one, whether you have done any of these things, whether you might do it, or would never under any circumstances do it.. Attending peaceful demonstrations, Joining political strikes. Source: World Values Survey 2010 2012 29