Original Question

I’m currently working with Timothy Sisk on an assignment from the UNDP Oslo Governance Centre to write a guideline to UNDP offices to better understand the linkages between election violence and governance, the consequences of election violence on governance and development, and a need for a better appreciation of the ways in which international electoral assistance programming can be improved to prevent and manage conflict in the future.

We are particularly interested in case studies or descriptions (stories) of particularly effective programs or interventions. We would like to invite you to share with us your reflections, lessons, practices, ideas and literature on the following questions:

3. How can electoral assistance "embrace the underlying political, social, and economic conditions in a society that give rise to and perpetuate conflict" and become more of a "process-oriented approach to peace building and nation building, development, and democracy support, which in turn will serve the cause of conflict prevention" as some scholars are advocating?

4. What kinds of approaches to evaluations of previous electoral processes are the ones which will help with positive inputs into new initiatives?

Summary of the Responses

Distinguishing Between Different Types of Electoral Violence

As most practitioners who participated in the discussion point out, violence is one of the greatest threats to free and fair elections. This may be especially true if related to new and emerging democracies. However, Ilona Tip also notes that elections are adversarial in their nature as they represent a contest for power. Yet, as highlighted in the ACE Focus on Elections and Security, “the emergence of electoral violence is not a result of the process being followed, but signals a critical departure from the accepted rules that govern the process.”

Discussing prevention strategies that could be implemented in order to mitigate electoral violence firstly requires the identification of the different types of electoral violence. Carl Dundas suggests distinguishing between violence aimed at disrupting elections by actors who do not want the elections to take place at all, and violence triggered by the rivalry between the contesting candidates/parties.  

If the former type of violence occurs, national security forces can be mobilized and target the obstructionists. Dr Noor Mohammed also brings to attention the aforementioned type of conflict where some parties do not agree on whether or when to hold elections, and likewise suggests using force on one hand and educating the public on the other hand. Additionally, Ilona Tip considers active disruptions of the voting process by spoiler groups and refusals to participate in elections either due to the shape of earlier negotiations, which were perceived as biased for instance due to such a measure as constituency demarcation or due to fears of power loss.

The latter type of violence mentioned by Dundas, i.e. violence triggered by rivalry between the contestants, can be controlled by other measures such as: codes of conduct for political parties with strong sanctions penalizing electoral violence, intensive voter/civic education programmes concerned with election campaigns and polling, and also special programmes/ activities aimed at preparing parties and candidates for election loss. Furthermore, community based organizations and citizens can help in preventing both types of violence and should have a more prominent role in strategies and programmes aimed at the prevention of electoral violence. Similarly, EMBs have a role in establishing structures for all stakeholders taking part in elections, especially political parties and civil society organizations, which could also be used as a platform for discussions about preventive measures. All of this should be done within a transparent system. Bringing in specific examples of South Africa and Zambia, Ilona Tip also underscores the need for Codes of Conduct legislated or informally signed by parties and other key stakeholders, as they have the potential to serve as conflict managing mechanisms.  

Dr. Noor Mohammed brings in yet another distinction of electoral violence/conflict, based on the different times during the electoral cycle when violence takes place. Hence conflicts before, during and after elections. The pre-electoral period conflicts are often connected to voter registration processes and their perception as unfair or biased. Such conflicts could be mitigated by transparent registration accompanied by voter education. Dr. Mohammed also suggests compulsory registration as opposed to voluntary, since in the case of the latter, eligible voters may be prevented from registering, which could in turn lead to escalating conflict situations. Furthermore, he claims that attention should be given to boundary delimitation and polling station locations, so that the needs of voters in rural and remote areas are catered to. At the same time, he argues that if the voters are attached to particular polling stations, and if any fraud allegations arise during or after elections, re-polling could be done with the same voters, at the same station.

 Focusing on post-electoral issues, Michael Meadowcroft favours the declaration of results in each polling station accompanied with an invitation to the party representatives to sign the results sheets. A signed copy, posted at the polling station can in Meadowcroft’s opinion be crucial in preventing higher party officials from disowning the results. Yet another trigger according to Tip is also the legitimacy and impartiality of the EMBs, monitoring bodies, but also courts and dispute resolution bodies. Dr. Mohammed suggests that transparency and various initiatives aimed at the confidence of different stakeholders (candidates/political parties, civil society, election watch groups etc.) should be put forth. An efficient election dispute mechanism also has to be in effect. 

Ilona Tip additionally emphasizes the significance of placing elections in a political context and historical perspective, so that some issues could be pre-empted and responded to pro-actively. In many cases, election related violence is indeed contextual and can be anticipated – as for instance in countries which have experienced long history of conflict and armed violence or countries where tensions and conflict are part of the political climate; sometimes due to unresolved issues from the previous elections. In addition to the above mentioned, she also points out some other possible triggers for electoral violence such as: restrictive legislation, only state owned media available to the contesting parties, allegations of fraud and rigged polls as well as the harassment of voters by police/ armed forces. 

Programs/ strategies aimed at preventing electoral violence

Therese Laanela suggests three strategies/ programmes to be implemented on all levels, (from a national level down to the polling station level), that should lead to decreased electoral violence. She specifically emphasizes the significance of the district level electoral officials who can act as mobile trouble shooting units on the polling day and who possess the best knowledge of the ongoing events in the field before and after elections.

 These three areas which could be incorporated in guidelines against electoral violence include:

1. training programmes: polling station staff training has to be more interactive (role play activities, scenario-building) in order to increase staff capacity to handle election day complaints effectively before they escalate to violence 

2. dialogue among electoral stakeholders (political parties, EMBs, human rights organizations, security forces etc.) to create an agreement on the rules of the game, but also the build-up the commitment and relationships (exchange of information, regular meetings, contingency planning). Dialogue efforts are essential not only centrally but also on lower levels, ensuring that district/village level conflicts are resolved and that all actors respect the codes of conduct

3. efficient and credible complaints processes: unresolved complaints are one of the main triggers for electoral violence (especially as they relate for instance to party registration), so the mechanisms to deal with them have to be efficient  

Ilona Tip also dwells on the importance of efficient electoral dispute mechanism bodies, mentioning some cases such as South Africa, where special permanent electoral tribunals/courts have been established to review decisions of the Independent Electoral Commission. Electoral matters are heard as they appear, so the aim of the court is to address the need for a speedy electoral justice.

Aleida Ferreyra brings to attention the successfully conducted programme addressing violence and elections in Guyana where the Social Cohesion Programme was created as part of the electoral assistance, integrating several areas and stakeholders relevant for the prevention of electoral violence, such as: assistance to political parties, EMB, ethnic relations commission, the police etc.

Tip also brings to attention the political party liaison committees introduced in various places, such as South Africa at different levels from local to national, in order to minimize electoral conflict by providing political parties the opportunity to liaise with the commission on electoral processes. The committees serve as consultative forums that encourage and promote transparency and accountability of the work of the EMBs. By discussing issues regarding other parties, electoral legislation and other contentious topics, these forums serve as means to resolve conflicts before the parties resort to violence. 

Another model, developed by EISA and the Independent Electoral Commission of South Africa establishes a Conflict Management Panel, in which mediators are identified by stakeholders and drawn from respected community and religious leaders, lawyers etc. that belong to different ethnic and religious groups, as well as youth, women, civil society and academics. The mediators are presented to and agreed upon with political parties and trained to deal with pre-, during and post-election violence at all levels. The Conflict Management Model was used in different places such as South Africa (1999), DRC (1996), Lesotho (2002) and Zambia (2006). The DRC case was especially successful as the trained mediators have been used even afterwards in resolving community disputes including household conflicts, land disputes and other problems at grassroots level.

Different methodologies are also mentioned by Debashis Sen, based on the Indian example. These include vulnerability maps where areas that have history of violence are identified and security forces  are deployed in these areas; phased election schedule if adequate security forces are not available; intensive observation by independent observers so that each polling station is covered at least once on the polling day; automatic re-poll if violence occurs and rule of law enforcement weeks ahead of the polls by organizing Flag Marches, rounding up criminals and executing arrest warrants. 

There are several organizations that have already developed specific tools and methodologies to monitor and eliminate electoral violence. Tihana Blanc points to the IFES Election Violence Education and Resolution (EVER) methodology, which records incidents of electoral violence throughout electoral process. These incidents are broken down into specific components such as motives, victims, perpetrators, responses and impacts of violence. After thorough analysis, patterns are established and findings are presented to electoral commissions, government agencies, security forces and other relevant stakeholders, who are thought to craft responses that will ultimately prevent or mitigate further electoral violence. Part of the methodology is also connected to monitoring and reporting on the responses. The EVER methodology was applied in several cases such as Guyana (2006), Ghana and Kyrgyzstan (2004). As Blanc points out though, several criteria have to be met in order for EVER to be successful.  

As several experts point out, any programme designed to prevent electoral violence would benefit substantially from targeting the entire electoral cycle. It is rarely if ever sufficient to introduce a variety of measures shortly before the Election Day. Although electoral violence is not condoned by any of the practitioners, Michael Meadowcroft poses a question on the actual effect of the violence on the results of the polls, especially if the level of violence is low.

 Replies were received with thanks from:

Therese Pearce Laanela

Carl Dundas

Aleida Ferreyra

Ilona Tip

Jeff Fischer

Dr. Noor Mohammad

Tihana Bartulac Blanc

Debashis Sen

Michael Meadowcroft

Other Sources

ACE Focus on Elections and Security

Fischer, Jeff (2002) IFES White Paper: Electoral Conflict and Violence: A Strategy for Study and Prevention

IFES Elections Today (2004): Focus on Elections During Conflict

Sisk, Timothy D. (2008): Elections and Fragile States: Between Voice and Violence. Graduate School of International Studies, University of Denver

UNDP Crisis Prevention and Recovery Newsletter (2009): Prevention of Electoral Violence: Experiences in Ghana, Kenya and Sierra Leone

UNDP Guyana

The best approach to electoral violence is Avoiding It!

I would want to see programmes that included a) training programmes b) dialogue between electoral stakeholders c) efficient and credible complaints processes. 

Each should be on levels from national down to polling station level, and the first two should occur well before election day.

Polling station staff training, in general, needs to be much more activity and roleplay oriented if staff/chairpersons are going to be able to handle election day complaints (before they escalate to violence!).

Perhaps the most important level to prepare/train in dispute resolution (and often the least invested in) are the District Level electoral officials, who can act as mobile troubleshooting units on election day. Also, they are in the best position before and after election day to feel the pulse of the district, including possible trouble areas.

Dialogue processes, well in advance of elections, are critical - between parties, human rights organisations,  Electoral management bodies, possibly security forces - to agree on rules of the game, to build relationships and commitments to each other (i.e. on exchanging information, regular meetings), and contingency planning. Too often these dialogue processes occur too centrally, (province and national level) - while on the district/village level low level conflicts are left unresolved. Lower level party members may not behave according to the codes of conduct that their senior colleagues have signed...

Complaints left unresolved are the most potent fuel for any violent flame, in particular perhaps those to do with party registration, etc. The mechanisms to deal with them - throughout! - must be efficient and credible.

Sounds like a terrific assignment, best of luck to both of you. ACE has a new Focus-On on security and elections; and I think the BRIDGE Project version 2 will have a module on the same. Your report would be a great resource.

Violence is the greatest threat to free and fair elections in new and emerging democracies. Perhaps a useful distinction should be made between violence aimed at disrupting an election by those who have no interest in a particular election taking place, and election violence triggered by the rivalry between contesting political parties and or candidates. In the former, the might of the national security services can be mobilized against the obstructionists. Good local surveillance by citizens and community-based organizations (CBOs) can help to prevent such mischief from taking root by reporting suspicious behaviour in the neighbourhood in a timely manner.

Violence instigated or openly encouraged by candidates or political party functionaries can be checked or controlled by a number of measures, such as codes of conduct for political parties and their supporters backed by strong sanctions. Intensive voter/civic education programmes with respect to election campaigns and polling should be undertaken on a timely basis. Special attention should be given to preparing political parties and candidate for election losses.

A prominent role should be reserved for community-based organizations which are interested in election conflict prevention locally. Electoral management bodies also have a role to play with respect to the establishment of structures for stakeholders, particularly political parties and civil society organizations, within which they can discuss preventive measures against election violence.

The electoral legislative scheme should pronounce itself clearly and firmly by way of stiff penalties for election violence.

There should be a transparent regime for the speedy redress of threats of violence and of all types of election disputes.

One of the most successful programmes relating issues of violence and elections is Guyana.  In particular, I suggest looking at the Social Cohesion Programme created as part of the electoral assistance in the country.  The programme integrated several areas such as assistance to political parties, the electoral commission, the ethnic relations commission, the police, etc. For more information, click here to see their website. 

Elections are by their nature adversarial as they are a contest for power. In many cases it can be constructive as it provides a useful context within which voters and members of the public are able to question the current leadership, challenge progress, or the lack thereof, and make a decision about their own vote. Placing elections in the political and historical context in which they take place will provide a useful framework to assess the issues that may arise and need to be analysed, pre-empted and responded to pro-actively. In a multi-party election registered political parties are trying to persuade voters to support their party rather than another. In attempting to garner support from voters contesting parties/candidates will criticise each others programmes and manifestos, question their track records etc on the one hand to resorting to physical violence, intimidation and threats.
Election-related violence is contextual and can be anticipated within a country specific context. Where a country is emerging from a long history of armed conflict or political violence, or the current political climate is one of tension and conflict, or the events surrounding the previous election are unresolved, and whether the electorate and public at large has between one election to another, continued to express its dissatisfaction about the outcome, or restrictive legislation or only state owned media is available to contesting parties, or misuse of public funds, property by the ruling party will provide a framework as to the nature and level of election related conflict is likely to occur. Depending on the nature of the conflict so the level of violence will arise.
Aside from the above many other issues may generate conflict in the election process itself:

• allegations of fraud and rigged polls;
• refusals to participate owing to unhappiness with the shape of earlier negotiation processes, perceptions of being sidelined through such measures as constituency demarcations or simply fears of loss of power (UDPS boycott – Kasais and Kinshasa);
• active disruption of the voting process by spoiler groups;
• disruption owing to problems in voter rolls – especially in situations where many have been displaced owing to violence in their home regions;
• challenges to the legitimacy or impartiality of electoral commissions, monitoring bodies, courts or other decision-making or dispute resolution bodies;

• harassment of voters by police or armed forces.

• review any decision of the Electoral Commission relating to an electoral matter;
• hear and determine appeals against any decision of the Commission;
• determine which courts shall have jurisdiction to hear particular disputes and complaints about infringements of the Electoral Code of Conduct;
• hear and determine any matter relating to the interpretation of any law referred to it by the Commission; and
• investigate any allegation of misconduct, incapacity or incompetence of a member of the Commission.

Political Party Liaison Committees have been introduced as a mechanism for minimizing conflict. In South Africa for example, Party Liaison Committees are established in terms of the Party Liaison Regulations and provide represented political parties with an opportunity to liaise closely with the commission on the electoral process. The Committees are in place at national, provincial and local level and are a consultative forum that encourages and promotes accountability and transparency of the work of the EMB. Parties are able to utilise these forums to also table issues about other parties, discuss proposed electoral legislation, etc with the intention of resolving these issues before they become so conflictual that parties resort to violence as a means of “resolving” them.
From the EISA experience first introduced in South Africa in preparation of the 1999 national and provincial elections, EISA in collaboration with the IEC put in place the Conflict Management Panel. This model has been used extensively in the 1996 DRC elections and in different formations in other SADC Countries, notably Lesotho in 2002 and local government elections in 2005 and Zambia in 2006.
Aside from some of the possible intervention processes outlined, the Conflict Management Model has been particularly effective. Conflict mediators were identified by stakeholders and drawn from respected, trusted community leaders, religious leaders and to a large extent lawyers. The potential mediators were presented to and agreed with political parties and trained in a specifically designed course to deal with pre, during and post election conflict. The model includes guidelines for mediators and administration guidelines for the coordinators of the panels.
The DRC experience in particular is a noteworthy case study as a conflict resolving and reducing process. A particular example is when, during the registration process, an angry mob outside a registration station became extremely aggressive towards five people trying to register for the Democratic Republic of the Congo’s (DRC) first democratic elections in 41 years. The scared victims were locked up in a room with election officials. EISA-trained mediators arrived quickly with loudhailers, and stood between the angry mob and the group, speaking to all stakeholders. Within thirty minutes they were able to calm the crowd and registration continued. The cause of the conflict was that the five people were mistaken for Rwandans, and therefore thought to be trying to register illegally. Through the EISA programme about 3000 similar incidents were recorded by the Conflict Panel Mediators countrywide. The mediation programme located within, coordinated and supervised by EISA, which included designing training material, and training 197 core trainers who in turn trained over 3500 mediators countrywide, was a partnership with the CEI (the electoral authority). (EISA is currently analysing the 3000 reported interventions.)
Mediators were drawn from different ethnic and religious groups, including the youth, women, civil society, and academics.  This diverse cross section of the population was trained to interact at all levels of election-related conflict. This was an integral part of achieving a climate of stability in order to uphold one of the pillars of democracy: elections. The mediators successfully intervened in cases of political and ethnic intolerance, political violence, misunderstandings, police repression, as well as murder allegations and unlawful arrests.
Given the infrastructural challenges in the DRC the United Nations supported the mediators by providing logistical support and often called on them to intervene in conflicts. In addition, The European Union Election Observer Mission as well as other international observer missions also praised the mediators work in their reports. Although the project has ended, for the mediators are using their skills to play other roles in our communities, like preventing household conflicts, land disputes, and any other social or political conflict at grassroots levels.
The design of the programme and the way it is implemented is critical to its effectiveness. It is important to design shape of the programme to fit the particular transformation process underway.
Professor Mark Anstey from the Nelson Mandela Metropolitan University has written two excellent papers on the issue of conflict management as a tool for responding to electoral conflict on Zimbabwe and the DRC. They are as yet unpublished and may be useful.
In his paper presented to a conference in the DRC on conflict and elections he noted that:
The moment of election may convince some parties that a continuation of conflict or at least the space provided by continuing ambiguity is preferable. Motives which inform such preferences include:

• beliefs that processes, negotiations or structures are illegitimate or flawed;
• loss of status and influence when the rules of the game change (RDP);
• larger gains possible through coercion than concession-making or problem-solving;
• limited repertoire of skills – only know how to fight;
• risks of being held to account for atrocities or war crimes.

(If you wish to contact Mark Anstey, please send an e-mail to the ace facilitators.)

A methodology for assessing electoral threats and formulating responses to those threats can be found in a White Paper that I wrote for IFES, Electoral Conflict and Violence, a Strategy for Prevention and Management with a subsequent article expanding on that research in an Elections Today edition on "Elections and Conflict". Both are on the IFES web site.

I would like to categorize election conflict as follows:

1. Conflicts on whether and when to hold elections – It may be seen in post conflict situations where parties to the conflict don’t agree on whether or when to hold elections.  Talibans in Afghanistan would not like elections at all as that will lead to a legitimate government. Only way is to weaken such powers by force on one hand and try to take public along by educating the public on the other. Election administration is not directly concerned and does no have to deal with such conflict.  However this does have an impact on the capacity of the electoral body to do its own job i.e. to conduct elections

2. Conflicts before elections – such conflicts largely arise when the voter registration process is unfair or is perceived to be unfair. This is under the purview of the election administration. A transparent registration process accompanied by voter education may help resolve most of these. To my mind a compulsory registration which involves the local elders in the process may be better than voluntary registration in which a number of eligible voters may be prevented from registering. Any lapse at this stage may lead to conflict situations in future.

The electoral process should be chosen carefully and the constituencies may be carved out without prejudice and polling station locations may be located to facilitate access by the voters particularly in rural and remote areas. Some times it may be difficult to provide such access because of security situation or accessibility.  But a compromise has to be made and as the security situation eases the polling stations should be located nearer to the voters.

The voters may be attached to polling stations so that if on the Election Day there is some fraud done in a particular polling station, there may be a re-poll in that PS in which the voters registered in that PS may vote again.

3. Conflicts during poll preparations, poll, counting and declaration of results – This is completely in the jurisdiction of the electoral body. A transparent approach that enjoys the confidence of the stakeholders like the candidates, political parties, civil society and election watch groups and observers would go a long way.  There is ample scope for innovation by the EMB to take initiatives to achieve this. In fact the EMB should be judged by its success mainly in dealing with this aspect.

In addition, a suitable grievance redressal mechanism has to be put in place to resolve grievances of the stake holders.

4. Conflicts post elections – In my view the EMB has no role to deal with this stage.

IFES has used Election Violence Education and Resolution (EVER) methodology in several countries to monitor and mitigate electoral violence. The methodology aims to record incidents of electoral violence (usually by relying on a network of local civil society groups) throughout the electoral process. The methodology breaks incidents down into components such as motives, victims, perpetrators, responses, and impacts of violence. It further analyzes them, establishes patterns where possible, and presents findings to electoral commissions, government agencies, security forces, political parties and other stakeholders. Ideally, those stakeholders are able to craft responses that prevent or mitigate further electoral violence. Responses are also monitored and reported on as a part of the methodology.

There are a number of useful case studies, a particularly successful one being that of Guyana in the fall of 2006, where the project has strengthened local institutions and significantly contributed to a national movement towards peaceful elections and resulted in the first in a series of elections not marred by serious violence. As a direct response to the project goals in Guyana, political parties issued peace messages in the crucial days leading to the election inviting their supporters in no uncertain terms to resist from using violence and settling their accounts in the streets. Political party manifestos were issued outlining the main policies. The government publicly endorsed the project even when it privately disagreed with reports of ruling party involvement in violence. Serious efforts by the government to use security forces more effectively were enhanced by EVER briefs to entire military, police and fire fighter contingents. The project was greatly if unexpectedly assisted by a formation of a third major political party with inter-ethnic focus. This party created a new cause for concern in the leadership of the two major, ethnically identified parties. They worried for the first time that voters dissatisfied with party violence would be able to express that dissatisfaction in their votes. Another successful example is Ghana, where electoral violence remained confined to the northern regions in part as a result of the presence of a widespread network of civil society representatives specifically monitoring for electoral violence.  In Kyrgyzstan, an early warning mechanism devised as a part of EVER methodology served to present and mitigate violence to stakeholders at regular roundtables and greatly contributed to violence-free presidential election in 2004.

From my experience in all but two countries (Iraq 2005 and Bangladesh 2006/7) where the EVER project has been implemented, it can be most useful when certain conditions are met. First, it is an electoral violence project. When violence is the result of a broader social breakdown, either through the collapse of the rule of law, violent conflict over natural resources, or outright civil war, EVER may be less effective. Second, smaller political parties, electoral management bodies, security services, and other stakeholders in smaller transitional democracies seem to be the most susceptible to being forced into action by clear public reporting. In larger countries, or where illiberal election institutions have already established ways of controlling popular discontent, EVER faces more serious challenges. This does not mean that EVER does not provide a useful base for how to address violence in these contexts as well, but more substantial adaptation of the methodology may be required.

I will indicate few very brief methodologies that have actually been adopted in recent elections in India to avoid electoral violence:-

1. Vulnerability Maps: Identify on GIS map areas having history of violence. Deploy security forces based on vulnerabilty mapping.

2. Phased Election Schedule: If adequate security forces are not available, hold elections in phases.

3. Intensive Observation: Deploy independent observers intensively so that each polling station is covered by her at least once, and in no particular order, on the day of poll.

4. Automatic Repoll: Enact rules laying down automatic repoll if violence occurs.

5. Create Ambience: Enforce "Rule of Law" weeks ahead of the polls by organising Flag Marches, rounding up criminals, executing arrest warrants etc.

-- There were no acts of violence in the 2006 General Assembly Elections; there were no repolls either.

Debashis Sen

Reply uploaded on behalf of Michael Meadowcroft:

No-one, particularly in our line of work, would ever condone violence but there is an important aspect in relation to its impact on the election process itself: the question needs to be asked - did the violence have any effect on the voting on polling day?

In my experience I doubt whether, in itself, low level violence has much effect on an individual elector's intention to vote or on his or her voting preference. Voters are remarkably resilient and tend to be entrenched in their views by violence or threats of violence. A difficult judgement needs therefore to be made as to whether violence has affected the result(s), whilst still persevering with efforts to end violence.

There is one further aspect of electoral practice that impinges on threats and electoral practice. I am very much in favour of the declaration of results in each polling station, coupled with an invitation to the party representatives to sign the results sheet, a copy of which should be posted at the polling station. I can think of a number of elections in which the existence of PS results sheets, signed by party representatives, was crucial in preventing party officials higher up disowning the results.

The problem, however, is that whereas a well run election will protect the secrecy of an individual's vote, a single polling station may serve one relatively small village whose vote is consequently public and which is therefore vulnerable to threats of collective violence.

Difficult to solve!
Michael