05:32
Unknown
Patterns of Electoral Malpractice*
Electoral malpractice and manipulation are among the principal obstacles to full democratization in states that have made an initial transition from authoritarianism. The manipulation of elections is not uncommonly used as a means of maintaining authoritarian and semi-authoritarian regimes while paying lip-service to the demands of the international community for competitive politics. Developing our understanding of this phenomenon is thus important to our appreciation of the various factors that contribute to the quality of democracy.
There is a small but growing comparative literature on electoral malpractice in the contemporary world (see Birch, 2008a for an overview). This paper aims to develop this field further by presenting a new cross-national dataset of electoral malpractice that allows researchers to differentiate among different forms of abuse. Specifically, this paper is designed to describe overall patterns in electoral malpractice across the world, and also to explore variations in sub-types of malpractice.
Why disaggregate electoral malpractice into sub-types?
Several studies have conceptualized the variety of electoral malpractice (Schedler, 2002; Elklit and Reynolds, 2005; Calingaert, 2006). At the same time, relatively little progress has been made in theorizing or measuring the relative prevalence of different forms of manipulation. It is the contention of this paper that differentiating sub-types of electoral malpractice is crucial to an understanding of this topic. Let us pause for a moment to consider why this might be so.
Elite decisions on whether to adopt a strategy of electoral manipulation involve a trade-off between power and legitimacy. On the one hand, incumbent leaders have an incentive to manipulate elections so as to reduce the inherent uncertainty in the electoral process and to ensure their continued hold on power. Under the right conditions, they can be expected to use all means at their disposal to achieve these ends, including blatant fraud. On the other hand, electoral manipulation is a strategy that entails risks to both domestic and international standing. Rigging elections lowers the legitimacy of leaders, which can affect their ability to rule without using coercive means, can cost them genuine electoral support, and can lead to international sanctions. Elites must therefore select their manipulation strategies carefully, and they can be expected to opt for those forms of manipulation that carry the least risk in the contexts in which they are operating.
For the purposes of analysis, electoral malpractice can be understood as operating along three principal dimensions: manipulation of the rules governing elections, manipulation of vote choice, and manipulation of voting (electoral administration).[1] Though risks will necessarily vary with contexts, some general hypotheses about choice of strategy can nevertheless be made. Elites can be expected to employ the manipulation of rules and institutions where possible, as this strategy carries a relatively low level of risk. Rules can be altered through legal means, and though international legal standards have been developed in this sphere, state actors nevertheless enjoy considerable leeway in the interpretation of international norms. Crucially, it is difficult for international observers to label the manipulation of rules ‘corrupt’ or ‘fraudulent’; leaders that employ this strategy as their principal form of manipulation can therefore expect to be spared condemnation on the grounds that their elections were ‘stolen’.
Yet there are limits to the extent to which the manipulation of rules alone can be counted on to ensure the desired outcome. Even when institutions are designed so as to favour one outcome over another, the very availability of genuine electoral competition creates a risk of electoral defeat. Moreover, heavily manipulated rules tend to attract criticism from international observers, who may even refuse to observe an election altogether if the basic pre-conditions for electoral competition are absent.
Cautious electoral manipulators will therefore often seek to employ other methods to secure the desired outcome. In many contexts, the next-most-attractive option is the manipulation of vote choice. Subtle pressure and selective incentives can often be used quite effectively to sway key sectors of the electorate with relatively little risk of international or domestic outcry, especially if the populace is accustomed to this type of activity, and especially if it takes place through informal channels. The disadvantage of this form of manipulation is that it may be quite costly, especially if an adequate infrastructure is not in place to sway vote choice through intimidation or particularistic rewards.
Only if rulers lack confidence in their ability successfully to manipulate rules and vote choice can they be expected to rely heavily on the manipulation of actual electoral administration, as this is the aspect of the electoral process that is easiest for domestic and international actors to monitor. Electoral administration is generally (expected to be) carried out in public and documented. Blatant forms of fraud such as the stuffing of ballot boxes and the falsification of results can therefore be expected to be a strategy of last resort.
In conclusion, we can expect manipulation of rules to be most common, followed by manipulation of vote choice, and lastly by manipulation of voting (electoral administration).
What general patterns should we expect to observe in the distribution of electoral malpractice?
As a first stab at assessing cross-national patterns of electoral malpractice, this paper assesses differences in overall levels of manipulation and abuse across regions of the world, over time, and across election type (parliamentary, presidential or combined).
General knowledge of experience with democracy suggests that we should expect there to be variations in the prevalence of electoral malpractice across regions of the world. The data used in this study are from three regions: Latin America , Eastern Europe and the former Soviet Union , and Sub-Saharan Africa (see below). We can expect the level of electoral malpractice to be correlated with how well entrenched democracy is in a region. Of the three regions analysed here, Latin America has the longest experience of democracy, while this regime type only took hold in many African states in the mid-to-late 1990s. We would therefore expect to see corresponding variations in the level of electoral wrong-doing. Following the same logic, we should also expect to see a gradual improvement over time in the quality of elections, given that the world has since the mid 1970s been experiencing a ‘wave’ of democratization.
It is more difficult to predict whether presidential, parliamentary or combined elections should attract malpractice. The concentration of power in the presidency suggests that this office is one that aspiring leaders might well go to considerable lengths to win; on the other hand, the larger number of beneficiaries in a parliamentary election might be anticipated to entail a multiplication of abuse. Finally, it is unclear whether the combination of presidential and parliamentary polls would be expected further to augment the extent of malpractice, or whether actors seeking the different offices would have an incentive to check each other in order to enhance the legitimacy of their own election.
Data Sources
Development of the study of electoral malpractice has been hindered by lack of comparable cross-national data, yet recent years have witnessed the generation of several relevant datasets by academic and non-academic organizations.
One such is the ‘fraud’ indicator in the World Bank Database of Political Institutions (Beck et al., 2001). This dummy variable represents answers to the question: ‘Were vote fraud or candidate intimidation serious enough to affect the outcome of elections?’ The variable definition document accompanying the dataset notes that ‘there may have been instances of fraud/violence that were not reported, thus resulting in false negatives’ (Keefer, 2005). In practice, the variable is coded ‘1’ in only 8.7 per cent of the cases included in this dataset, making fraud a relatively rare event. The binary approach taken to the assessment of election and the possibility of relatively large numbers of false negatives thus make this indicator less than satisfactory from the point of view of cross-national analysis of electoral malpractice.
More recently, Freedom House began in 2006 to release the sub-category scores that contribute to the construction of their Freedom in the World dataset. One of the sub-categories is ‘electoral process’ which rates ‘an individual's ability to […] vote freely in legitimate elections’ (Freedom House, 2007). To date, country scores have been released for 2006 and 2007, which in most cases represents a single election. These data are therefore of limited value in conducting longitudinal analysis of election quality.
A final source of cross-national data is survey data. Both the Comparative Study of Electoral Systems survey project and the International Social Survey Project have included questions on perceptions of election quality.[2] These data have proved useful in analysing the micro-level foundations of electoral malpractice (see Birch 2005; 2008b), but they are of less use in undertaking broad comparative analysis due to limited country coverage.
A common disadvantage of all the aforementioned data sources is that they fail to provide indicators for the wide range of sub-types of electoral malpractice. The World Bank ‘fraud’ indicator singles out specific aspects of electoral quality, while the Freedom House indicator is more general. The survey questions tap the overall quality of elections indirectly by measuring popular perceptions.
Given the limitations of existing datasets, the Project on Electoral Malpractice in New and Semi-Democracies initiated the construction of an Index of Electoral Malpractice on the basis of election observation reports. The reports included were restricted to those produced by international organizations that carried out large-scale election observation missions between 1995 and 2006 in three regions of the world - Latin America , Eastern Europe and the former Soviet Union And Sub-Saharan Africa -according to a standardized methodology in accordance with internationally-recognized best practice. These criteria are fulfilled by missions carried out by the European Union (EU), the Organization of American States (OAS), and the Organization for Security and Co-operation in Europe (OSCE).
The choice of regions was determined in part on conceptual grounds and in part on grounds of data availability. Recent democratization processes in all three regions have spawned rich comparative literatures on the electoral process; electoral mechanisms in each region have enough similarities that scholars have been able to identify common trends and regional specificities, providing each of the three areas with a conceptual coherence that has heretofore been somewhat lacking vis-à-vis other regions such as Asia and the Middle East. Furthermore, existing work by inter-governmental, non-governmental and academic researchers has generated substantial corpuses of electoral law and electoral result databases for each area. Finally, electoral observation has in each area been conducted on a relatively systematic basis by regional the inter-governmental organizations mentioned.
Each report was coded independently by three coders according to a schema that distinguishes between the targets of various manipulative strategies: the legislative framework, campaign practices, and various aspects of the electoral administrative process. The resulting dataset includes indicators of 15 different forms of electoral malpractice in 136 elections held between 1995 and 2006 in 57 new and semi-democracies.[3]
As data sources, election observation reports have advantages as well as disadvantages. Their advantages include the fact that those who compile them are election experts who set out systematically to assess electoral equality, and as accredited observers, they in theory have good access to key aspects of the electoral process. Since the mid 1990, the methodology subtending electoral observation has improved considerably, and the inter-governmental organizations whose reports are used to compile this database all employ scientific techniques following rigorous methodological frameworks. Observers form major inter-governmental organizations are undoubtedly better placed than any other group to assess the quality of many elections.
At the same time, election observation reports also have drawbacks from the point of view of the aims of this study. Like other forms of corruption, electoral malpractice is an activity that is difficult to observe directly due to its illegal nature. Moreover, some aspects of the phenomenon are easier to observe than others. The media, by their public nature, are highly susceptible to monitoring and evaluation, while at the other end of the spectrum, vote-buying is an activity that is often very difficult for election observers to identify and measure, due to its particularistic and covert nature, and also due to the fact that both the givers and the receivers of electoral bribes have an incentive to collude in preventing disclosure of the transaction. Other forms of electoral malpractice mostly fall in between these extremes. For example, ballot-box stuffing is often noted by attentive observers at polling stations, but observers are not present in all polling stations all the time, and the stuffing of mobile ballot boxes and those in closed institutions such as hospitals and prisons is often virtually impossible to monitor.
Another oft-noted concern with this type of data is the possibility of politically-motivated bias in reports (Carothers, 1997). Inter-governmental organizations may be under pressure from a variety of sources to deliver a certain verdict, be it to further the interests of a particularly influential member state, or to prevent an outbreak of post-electoral violence in the state whose election is being observed. This suggests that the overall assessment of the election may in some cases be less informative than the sub-components whose evaluation is typically embedded in the reports’ details.
These concerns warrant circumspection when employing election observation reports as a data source. At the same time, there may in many cases be no superior source of data with which to address specific research questions. Where this is the case, the careful use of election observation reports makes good sense. Their deficiencies are likely at worst to reduce the quality of the data to the point where actually occurring patterns and causal relationships cannot be discerned. The results of such analyses should therefore be considered alongside research based on other sources, such as qualitative studies and studies that use data derived from other data collection exercises (surveys, for example).
Construction of the Index of Electoral Malpractice
The Index of Electoral Malpractice includes 15 separate indicators of various aspects of electoral manipulation. Within the three broad dimensions of malpractice referred to above – the manipulation of rules, of voters and of voting - it is possible to distinguish a number of specific sub-types. In each case, coders were advised to code the evaluations of the severity of malpractice as indicated in the observation mission reports, not their own evaluations of the factual information contained in the reports. The resulting indicators thus measure observation mission judgements of electoral quality.
Legal framework. For the purposes of this coding project, the manipulation of rules category was not broken down into sub-types; rather an overall evaluation was provided as to whether the legal framework governing the elections conforms to international standards. The OSCE and EU reports provide detailed evaluations of the legal framework governing the election on which they are reporting, whereas the OAS reports tend to restrict themselves to descriptive accounts of such provisions. It was therefore not possible to code this item for all cases.
The remaining two categories – electoral choice and electoral administration, were broken down into sub-types as follows:
The manipulation of electoral administration
Electoral authority independence: the term ‘electoral authorities’ refers to those bodies and/or or agencies involved in running the election. This typically includes the election management body (electoral commission or election department), as well as the judiciary, in as much as dispute adjudication often involves the courts. Other regulatory and administrative agencies may also be involved in certain aspects of the election. For the purposes of the construction of this indicator, all such bodies are considered electoral authorities when exercising a designated role in the election. Electoral authorities are meant to operate in an impartial and transparent manner, regardless of how they are selected. This item indicates whether the electoral authorities in practice exhibit independence and impartiality. It is not an indicator of the composition of the authority. The presence of party members on electoral commissions is not a cause for concern, provided they perform their role in a professional and impartial manner. Noteworthy violations of impartiality include instances where (one or more of) the electoral authorities are clearly under the control of (or pressure from) the incumbent government or some other agent (the military, a foreign power, etc). Lesser violations include noted bias in favour of one or more electoral contestants. Though observation reports do not all address this question explicitly, they generally give some indication of the degree of impartiality of the electoral authorities.
Contestation: This item is designed to indicate whether major contestants were banned or had their applications to contest the elections refused/obstructed in (evident) violation of the law. Open contestation is crucial to electoral integrity. When eligible candidates and/or parties are denied access to the ballot paper, this constitutes a fundamental irregularity. Noteworthy violations include the arbitrary and/or spurious denial of candidate/party registration, the removal of candidates/parties from contestation on spurious or arbitrary grounds, or the obstruction of candidate/party nomination/registration. This question does not refer to the campaign activities of contestants, but only to ballot paper access.
Voter registration: This item measures the extent to which sections of the citizenry were refused registration or mis-registered in (evident) violation of the law. Accurate and comprehensive voter registration is another key feature of free and fair elections, though it must be recognized that the maintenance of a full and accurate electoral roll is a difficult and resource-intensive task. Noteworthy violations include the systematic failure to register certain categories of people, the existence of numerous non-eligibles (‘dead souls’) on the register, gross inaccuracies in the data maintained in the register, widespread failure to correct known inaccuracies, and refusal to allow public inspection of the register. Significant logistical problems affecting the accuracy and/or completeness of the electoral register are also grounds for concern, though in coding greater weight was given to evidence of systematic attempts to manipulate the register than to insufficient logistical capacity (in as much as it is possible to distinguish between these sources of inaccuracy).
Polling arrangements: This indicator taps whether polling arrangements were adequate and equitable. Polling places should be provided in sufficient number that all voters have a reasonable opportunity to exercise their franchise. They should be located relatively conveniently, and the facilities provided should enable voters to cast a ballot without waiting an excessive amount of time. Ideally, special provision should be made for certain categories of voters, such as the disabled/infirm, and those living in remote locations. The ability of the electoral authorities to facilitate voting will to some extent depend on their own resources, as well as on geographical considerations. Noteworthy violations include polling arrangements that are manifestly inadequate (too few, located too far from where many people live, or too poorly resourced - such that voters have to wait for long periods of time to vote), or provisions that are clearly biased against particular regions or groups (for example, too few polling stations in opposition strongholds). Failure to open polling stations on time or to keep them open for the designated period is also a common problem. Greater weight was given to evidence of systematic bias than to logistical problems, though logistical problems were also considered grounds for concern.
Voting: This indicator reflects the extent to which voting proceeded in accordance with the law. ‘Free and fair’ voting entails unhindered and equal access to voting for all electors, and the safeguarding of ballot secrecy. Noteworthy violations include denial of access to voting opportunities in polling places, ballot box stuffing, violations of ballot secrecy (including open voting, family voting, illegal proxy voting, or voting under the observation of unauthorized persons), and lack of transparency in the organization of voting. In coding, greater weight was given to evidence of systematic bias than to logistical problems, though logistical problems were also considered to be grounds for concern.
Counting, tabulation and the reporting of results: This item captures the linked processes of counting ballots, tabulating and reporting results. Once votes are cast, they must then be counted, tabulated (aggregated) and reported accurately. Many of the most traditional forms of electoral fraud occur at this stage in the process. Noteworthy violations include miscounting, inaccurate tabulation, and the mis-reporting of results; failure to report results in a timely manner is also a cause for concern, as is lack of transparency in the process of counting, tabulation and/or reporting. In coding, greater weight was given to evidence of systematic bias than to logistical problems.
Dispute adjudication: This indicator describes the quality of the dispute adjudication process, and specifically whether this process was conducted in a timely, comprehensive and impartial manner. Following an election, there are virtually always legal disputes over some aspect of the process. The timely and impartial adjudication of these disputes can in many cases be crucial to the overall quality of electoral conduct. Noteworthy violations in this domain include failure to consider significant numbers of claims, undue delay in processing claims, and clear bias in deciding individual cases.
Observer access: This indicator designates the access of international and domestic observers to all relevant aspects of the electoral process. The elections included in this study are by definition ones in which international election observation has played a role. Nevertheless, election observation missions are commonly impeded in a variety of subtle and not-so-subtle ways. Noteworthy problems in this area include failure to grant observers access to aspects of the electoral process they are entitled to observe, intimidation of observers, and obstruction of their activities. Note that this question refers to both domestic observer groups (where they exist) and to international observation missions (though the absence of domestic observer groups is not a cause for concern per se).
The manipulation of vote choice
Media coverage: This variable captures the extent to which the media provided balanced coverage of the electoral campaign. Balanced media coverage is an important component of any fair election. Though fully equitable media reporting is virtually unachievable, blatant biases in campaign reporting represent a serious obstacle to the supply of effective voter information. This item is designed to tap the degree of bias in media coverage, and, in particular, the extent to which the media provided more coverage of, or was more favourable to, particular contestants. Particular attention was given to the state media.
Misuse of resources: This indicator describes the extent to which regulations governing the use of campaign resources were observed. Virtually all states regulate the use of campaign resources in some way, though the nature and extent of regulation varies considerably. This question is intended as a means of assessing the extent to which parties and candidates complied with the resource regulations in force in the state in question. Noteworthy violations include the abuse of state resources (i.e. the use of state resources for partisan campaigning purposes), known over-spending (campaign spending in excess of a limit established in the law), fundraising from illegal sources, and failure to supply required information on use of resources/campaign finance.
Vote-buying: This indicator for vote-buying designates the extent to which there were allegations of vote-buying noted in the report. The purchase of votes is common in many states, though electoral observers are not often well placed to monitor this sort of activity. Nevertheless, if vote-buying is widespread, it is often commented upon in election observation reports. Of relevance is the extent and magnitude of this practice, as well as its likely impact on the outcome of the election.
Voter intimidation and/or obstruction: This item taps the extent to which voters were reported as having been intimidated, coerced or obstructed. The above comments on vote-buying are largely relevant to intimidation also (as the intimidation of voters is an alternative to the provision of particularistic rewards through vote-buying). This item refers to the intimidation and/or obstruction of voters by any agent. Inappropriate involvement of the military or police in the electoral campaign was normally coded as a form of intimidation; this includes cases where military officers instruct their subordinates how to vote (bearing in mind that the proportion of the electorate serving in the military is typically small).
Candidate intimidation and/or obstruction: This item measures the extent to which candidates or party activists were reported as having been intimidated or coerced, or having their campaign activities unlawfully obstructed. Candidates and party activists should be able to carry out lawful campaign activities unhindered by state authorities, by the representatives of other parties, or by other agents. If candidates and activists have their campaigning unlawfully obstructed, or if there is evidence that they have suffered personal intimidation or coercion (including attacks on campaign offices or equipment), this is a serious violation of campaign standards. When there are reports of intimidation or obstruction, it is often unclear who is responsible for these actions (particularly when they involve violence). This item designates intimidation, coercion, and/or the unlawful obstruction of campaigns through violent or non-violent means by any agent – be it a state body, another political party, groups of voters, or unknown persons. Such events were coded only if the evidence suggests intent to affect campaigning capacity; mishaps such as accidents and robberies that do not appear to have a political motive were disregarded (though interpretations of such events often vary – coders were advised to code the interpretation given by the report). The obstruction of campaigns by electoral authorities can include the unequal application of certain aspects of the law, including authorization for campaign meetings and rallies. If there was evidence that the meetings and rallies of one or more parties are systematically being denied authorization on spurious grounds, this was coded as obstruction.
Overall assessment of the election: This indicator is designed as a summary measure of the overall quality of the election, as indicated in the report. Coders were alerted to the fact that this assessment could differ from that which they might reach on the basis of an evaluation of the detail presented in the report. They were advised to code this item with reference to the overall assessment of the election that is typically provided in the executive summary and/or the conclusion of the report.
Table 1 presents the mean scores for the 15 different forms of electoral malpractice derived from the Index of Electoral Malpractice. These data show that the media manipulation and the misuse of state resources are the areas in which there is the greatest amount of overall abuse. This suggests that political actors are more likely to seek to manipulate the process of vote choice than the process of voting itself. It is also noteworthy that efforts to manipulate vote choice through the media and the abuse of state resources are more common than attempts to intimidate voters or candidates; the carrot appears to be a more attractive strategy than the stick. In this context it is worth pausing to consider the particularly low score for vote-buying - another form of ‘carrot’. As noted above, this aspect of the electoral process is one that is particularly difficult for electoral observation missions to monitor with any degree of accuracy. Not only are votes typically bought in private, but the buyer and the seller both have an incentive to collude to keep the transaction secret. Anecdotal evidence of vote-buying does in some cases attract the attention of election observation missions, but the data presented here undoubtedly under-estimate the prevalence of this practice.
Among the various aspects of electoral administration, voter registration and vote counting, tabulation and reporting are activities that are particularly susceptible to manipulation. It is noteworthy that the actual casting of ballots is less prone to malpractice than the processing of ballots following the close of voting. This may well be due to the presence of observers at polling stations. By contrast, the obstruction of contestation and of election observation are aspects of the electoral process that are relatively immune to abuse.
According to the schema elaborated above, the various forms of electoral abuse can be divided into the following types: manipulation of the rules governing elections (No 1 in Table 1), manipulation of vote choice (Nos 10-14), and manipulation of the vote through efforts to alter electoral administration (Nos 2-9). When the 14 sub-types of malpractice listed above are divided into these types and mean scores are calculated for each category, we see that, as expected, manipulation of the rules governing the electoral process is the area where we observe the greatest amount of abuse, followed by the manipulation of vote choice. Abuse of the electoral administrative process is least susceptible to manipulation (see Table 2).[4] These results provide support for our hypotheses regarding the hierarchy of malpractice strategies based on risk of disclosure.
-- Table 2 about here --
* I am grateful to the University of Essex British Academy
0 comments:
Post a Comment