34990 Paper No. 31 / December 2005 Consolidating Indonesia's Democracy: Conflict, Institutions and the "Local" in the 2004 Legislative Elections Patrick Barron Melina Nathan Bridget Welsh Summary Findings Despite widespread fears of potential violent unresolved conflicts during the electoral period may conflict, Indonesia's April 5, 2004 legislative potentially feed into existing tensions and lead to future elections saw limited incidents of violence and outbreaks of conflict. represented a significant step in Indonesia's democratization process. In order to evaluate Second, these deeper more structural efforts to reduce conflict and conflict prevention, this study conflict can only be undertaken by strengthening analyzed four diverse provinces--Bali, East Java, institutions that can positively channel tensions. One of Maluku and NTT--from March 15 to May 15, the major study findings was that the performance of 2004 covering the election campaign, polling day institutions in Indonesia (both those focused on and the immediate aftermath. Studying the managing the electoral process, as well as state legislative elections provided a window through institutions more broadly) was mixed: capacity, which to examine issues related to local conflict professionalism, and constructive societal engagements and the role institutions play in managing it. were lacking in many localities. These institutional Given the extent to which decentralization has shortcomings resulted in high levels of protest and, in devolved power to the local level, examining the some cases, attacks against state institutions. process of local competition helps shed light on the nature and state of Indonesia's democracy and The 2004 legislative elections were the first time that the institutions that underlie it. the police was solely in charge of internal security and it successfully adopted conflict prevention measures Overall, the elections were peaceful. In the four that resulted in peaceful elections. However, the police provinces there were a limited number of performed best at deterrence when its presence by election-related conflicts, of which just over 10 sheer force of numbers reduced the likelihood of percent were violent. This is an impressive clashes. This study emphasizes, however, that achievement. The study shows that the low level successful conflict reduction occurs when police adopt of election-related conflict was the product of a community policing approaches which involve number of factors, but especially the lack of consulting and working with a broad range of actors interaction between pre-existing forms of conflict from local government and the military, to community and the local electoral processes, as well as a actors and community-based organizations. This range of conflict prevention measures. Local underscores the need to deepen and broaden police elites and grassroots initiatives, especially in high reform. At the same time, by adopting a supporting role conflict areas, played a pivotal role in minimizing and remaining neutral, the military contributed tensions. positively to the ongoing security reforms required to consolidate democracy in Indonesia. Although the elections were relatively peaceful, the study highlights institutional weaknesses and Finally, the study's most significant finding was the the need for increased attention to three broad pivotal role local actors and institutions played in areas. First, the study found that one of the major preventing and managing conflict. While national factors underpinning whether election-related agencies provided important policy guidelines, the conflict was likely to arise/escalate was prior success of conflict reduction measures depended on history of local conflict. While conflict prevention local knowledge, ownership and implementation at the was by and large successful in reducing electoral regional levels of state institutions, the quality of cross- violence, in many cases the measures tended to institutional coordination and cooperation, and the role address symptoms rather than causes. This of other local actors. The importance of local highlights the need to rethink the way in which knowledge and ownership underscores the need to we conceptualize conflict and conflict resolution. recognize and support community-driven peace- Conflict in Indonesia is not just an episodic, building efforts where they are successful and build event-driven occurrence, but is a symptom of capacity at this grassroots level where efforts are less broader problems, cleavages and interests. Thus, developed. SOCIAL DEVELOPMENT PAPERS Conflict Prevention & Reconstruction Paper No. 31 / December 2005 Consolidating Indonesia's Democracy: Conflict, Institutions and the "Local" in the 2004 Legislative Elections Patrick Barron Melina Nathan Bridget Welsh This Working Papers Series disseminates the findings of work in progress to encourage discussion and exchange of ideas on conflict and development issues. Papers in this series are not formal publications of the World Bank. This paper has not undergone the review accorded to official World Bank publications. The findings, interpretations and conclusions herein are those of the author(s) and do not necessarily reflect the views of the International Bank for Reconstruction and Development/The World Bank and its affiliated organizations, or its Executive Directors, or the governments they represent. The papers carry the names of the authors and should be cited accordingly. The series is edited by the Conflict Prevention and Reconstruction Unit in the Social Development Department of the Environmentally and Socially Sustainable Development Network of the World Bank. To request copies of the paper or for more information on the series, please contact the CPR Unit. Papers are also available on the CPR Unit's website. For additional copies of this paper, please contact: Conflict Prevention & Reconstruction Social Development Department The World Bank 1818 H Street, NW Washington, DC 20433 Fax: 202-522-3247 Web : http://www.worldbank.org/conflict (see "Publications" in the navigation menu) E-mail: cpr@worldbank.org Printed on Recycled Paper Table of Contents Acknowledgements.......................................................................................................................................ii I. Introduction ...............................................................................................................................................1 II. Overview of Election-Related Conflict....................................................................................................4 1. The Distribution of Election-related Conflict .......................................................................................4 2. Forms of Election-related Conflict........................................................................................................5 3. The Timing of Conflict..........................................................................................................................7 4. Urban/Rural Distribution.......................................................................................................................7 III. Causes of Conflict or Lack of Conflict ...................................................................................................9 1. Institutional Strength/Weakness..........................................................................................................10 2. Local Actors and Histories: The Mobilization of Pre-existing Tensions............................................14 3. The Nature of the Party System ..........................................................................................................17 4. Legislative Elections: A Less Tense Contest ......................................................................................18 IV. Limiting Violence: Successful Conflict Prevention .............................................................................19 1. Campaigning Guidelines and Implementation....................................................................................20 2. Proactive Policing ...............................................................................................................................21 3. Local Leaders and Grassroots Initiatives ............................................................................................24 4. Institutional Coordination and Cooperation........................................................................................25 V. Managing and Resolving Conflict: Mixed Success ...............................................................................26 1. Local Intervention ...............................................................................................................................27 2. Limited Police Role in Conflict Management.....................................................................................28 3. Varied Strength of Election Officials..................................................................................................30 VI. Reflections: Future Democratic Consolidation.....................................................................................32 References...................................................................................................................................................35 Appendix A: Conflict Incidents by Area ......................................................................................................1 Appendix B: Election-related Conflict Incidents by Type............................................................................2 Tables Table 1: Election-related conflict in case study provinces (March 15-May 15).................................4 Table 2: Forms of election-related conflict (March 15-May 15)...................................................6 Figures Figure 1: Election-related conflict incidents (March 15-May 15).................................................5 Figure 2: Type of dispute by province (March 15-May 15)........................................................7 Figure 3: The timing of election-related conflicts (March 15-May 15)...........................................8 Box Box 1: Overview of the local dynamics of conflict in the four provinces.......................................11 ii Acknowledgements Special thanks to Rachael Diprose (formerly of the World Bank in Jakarta) who conducted the field research in East Java, contributed to the paper, and participated in early discussions of the findings. Thanks are also due to the five provincial researchers: Heru Gutomo (Bali), Cici Novia Anggraini and Endro Probo (East Java), Jefferson O. Tasik (Maluku), and Stanis Didakus (NTT). Thanks to Elisabeth Dua Soru, Getrudis Erni, Budi Santoso, and Mirnawati Ladongga who provided assistance in compiling the newspaper dataset, and Samuel Clark who coordinated the newspaper research. The field research was supported by DfID. The study would not have been possible without the cooperation of a wide range of Indonesian election officials, political candidates, police officers, reporters and other members of the communities who kindly gave us their time and assistance during our fieldwork. We are extremely grateful to them all, although space does not permit us to name them individually. The views in this paper are those of the authors and do not reflect those of the institutions they are affiliated with. The authors, Patrick Barron (World Bank), Melina Nathan (World Bank/UNDP), and Bridget Welsh (Johns Hopkins University, SAIS), can be contacted at: pbarron@worldbank.org, melina.nathan@undp.org and bbwelsh@jhu.edu. Patrick Barron Melina Nathan Bridget Welsh CONSOLIDATING INDONESIA'S DEMOCRACY: CONFLICT, INSTITUTIONS AND THE "LOCAL" IN THE 2004 LEGISLATIVE ELECTIONS I. Introduction Despite widespread fears of the potential for conflict, the April 5th 2004 Indonesian legislative elections had a limited number of incidents of violence and were an impressive illustration of deepening conflict prevention and management efforts on the part of national and local actors. As such, the April elections represented a significant step in Indonesia's democratization. The elections, however, were not problem- free. This study of conflict and conflict prevention and management before and after the legislative elections points to varying experiences across Indonesia, shaped by local contexts. It also highlights common shortcomings in election processes as well as successful mechanisms for preventing and reducing conflict. Even as a non-violent competitive political process has become increasingly accepted, attention to conflict prevention and on strengthening local institutions must remain a priority, and in some areas improve, if Indonesia's democracy is to consolidate in the coming years. Indonesia's political framework has radically transformed since 1999. The country's new electoral system has created new procedural uncertainties, increasing the need for voter education, training of officials and cooperation among candidates/parties. At the same time, the process of decentralization has increased the stakes of elections at the local level. For the first time in April 2004, Indonesians voted for individual legislative candidates at the national, provincial, and district levels, increasing local competition. Four elections ran simultaneously, with voters choosing candidates and parties for the district, provincial and national parliaments as well as a new regional chamber, the Dewan Perwakilan Daerah (DPD), comprising independent (i.e., non-party) deputies. These elections were a precursor to the two rounds (in July and September) of Indonesia's first direct presidential election, and the first direct election of District Heads (Bupati) which followed in mid-2005. As the first major election in 2004, the legislative elections provide a valuable lens to understand the relationship between the electoral process and conflict, particularly at the local level, and to assess the state of Indonesia's democracy today. These electoral changes have occurred in a climate of institutional reform. The parliament has become a stronger institution (King 2004). Political parties have become more prominent and competitive. Yet, at the same time, they remain internally fragmented, lack uniform strength across Indonesia and, in some areas, rely heavily on paramilitary (satgas) and criminal preman groups for mobilization. The formation of national election implementation and monitoring agencies, the Komisi Pemilihan Umum (National Elections Commission or KPU) and Panitia Pengawasan (Committee for Election Supervision or Panwaslu) has created a new electoral bureaucracy with institutional interests, significantly restructuring electoral administration within the state. These institutions extend to the provincial and district levels. Security sector reforms have made the police rather than the armed forces (Tentara Nasional Indonesia-- TNI) primarily responsible for internal security and for ensuring an orderly and peaceful campaign (Riefqi 2004).1 Despite these significant reforms, an overall climate of weak institutions in Indonesia continues to underscore the potential for civil society to resolve conflicts through violent means. Attempts to reform 1Security sector reform has been a key focus of the international aid community. See the Partnership for Governance and Reform in Indonesia (http://www.partnership.or.id/sector.php?sub=18&lang_id=1) and the Asia Foundation's website (http://www.asisafoundation.org) 2 the judiciary and foster respect for the rule of law have occurred slowly and unevenly (Asia Foundation 2001).2 A history of violence associated with the 1999 elections, and high levels of ongoing local conflict in many areas,3 the transformed political landscape, and the increased competition it brings, led some to fear further violent conflict in the April 2004 legislative elections, and highlighted the need for effective conflict prevention and management mechanisms. In order to investigate conflict and conflict prevention, we examined four diverse provinces--Bali, East Java, Maluku and Nusa Tenggara Timur (NTT)--from March 15th through May 15th 2004, covering the election campaign, polling day, and the immediate aftermath.4 Studying the legislative elections provided a window through which to examine issues related to local conflict and the role institutions play in managing it. Given the extent to which decentralization has devolved power to the local level, examining processes of local competition helps shed light on the nature and state of Indonesia's democracy and the institutions that underlie it. Our approach stands in contrast to most of the early analyses of the 2004 Indonesian elections, which have focused primarily on the role of elites at the national level (see Aris et al. 2005; Emmerson 2004; NDI 2004; Rieffel 2004; Saiful and Liddle 2004 and 2005; Sherlock 2004; Singh 2003) or the implementation of multiple elections for the first time (see Sissener 2004). This is in part a function of the focus on the presidential elections, where national elites were prominent, and an interest in the future of particular parties, especially Islamic parties (Chanintira 2005; The Economist 2004; Saiful and Liddle 2005), as well as being symptomatic of the general tendency of the policy community to focus on macro dynamics at the national level.5 In contrast, we focus on local level case studies of four diverse provinces' experience with the legislative elections, where much of the competition was between local candidates and elites.6 This allows us to examine more closely how reform processes have played out locally in Indonesia's post-1998 democratic era. Careful attention was paid to distinguishing the April 2004 legislative elections from the July and September presidential elections, although there was overlap in May when the presidential election campaign began. In each province we selected a minimum of two districts for in-depth analysis based on variations in levels of violence and conflict resolution mechanisms. As a whole, the four provinces (and eight plus districts) had experienced a wide range of levels and forms of conflict, from ethnic tensions to 2For an assessment of Indonesian perceptions of judicial reform, see http://www.asiafoundation.org/pdf/IndoLaw.pdf 3The exact level of local conflict in Indonesia is unknown. Most reports on conflict in post-New Order Indonesia have focused on separatist violence in Aceh and Papua and ethnic violence in Maluku and Poso (e.g., Wilson 2001 and Klinken 2001) as well as West and Central Kalimantan (Bertrand 2004; Davidson 2003; Smith 2005). Focus on local conflicts has, however, been limited. According to the official government PODES 2003 survey, (mostly violent) conflict was present in over 7 percent of Indonesian villages in 2002, spread across every Indonesian province, resulting in 4,869 deaths (Barron, Kaiser and Pradhan 2004). Other research drawn from newspaper accounts, including Tadjoeddin (2002), Varshney, Rizal and Tadjoeddin (2004), Barron and Sharpe (2005), and work on vigilante violence (Welsh 2005) in four Indonesian provinces, suggests higher numbers of deaths. 4While the selected provinces cannot be said to be fully representative of Indonesia, they are a diverse sample. They vary in wealth, geographic location, population size, level of urbanization and development, level of past conflict, etc. As such, we can be confident that some of the diversity of experience with the 2004 Indonesian legislative elections is reflected within the sample. 5See, for example, Harris and Todaro (1970) on the incentives that tend to bias policy making, and social analysis toward urban areas. 6This reflects an increasing attention to local politics in Indonesia. This has included, more recently, a focus by anthropologists and others doing village level fieldwork on local political dynamics, including in analyses of the 2003-2005 district and mayoral elections. Some of this work was presented at the 4th International Symposium of Journal Anthropologi Indonesia at the University of Indonesia, July 12-15, 2005. See Amalinda (2005), Honna (2005), Kimura (2005), Morishita (2005) and Sakurai (2005). 3 disputes over resources. They were also familiar ground; we had previously conducted extensive research on local conflict and conflict management in these areas before the election period. Thus, the case studies allowed us to examine the relationships between the election and preexisting tensions. The study was developed through intensive research in the field sites as well as analysis of secondary data including the creation of a newspaper archive of conflict incidents.7 The research teams spent two-four week periods in each province, during and/or after the election, and conducted a wide range of interviews with local officials, candidates, media representatives, police, victims of conflict and other community members.8 Although admittedly limited by the time constraints of the research and inherent bias of newspaper accounts, the combination of interviews and newspaper analysis provided a broad set of research methods to help capture and provide an understanding of conflict. The research found that conflict management in the April elections was successful and violence was not significant. While election-related conflict incidents did arise, most were small and did not escalate, reflecting both a desire and an ability to resolve disputes in non-violent ways. The research did, however, pinpoint significant institutional weaknesses related to both the electoral bureaucracy and the security sector. The majority of conflicts that took place involved individuals and groups protesting against problems with the electoral process administered by the state. This is significant: it is in stark contrast to previous elections, where violence has tended to flow in the opposite direction, being administered by the state (or actors allied with the state) against the population. While this reversal of conflict dynamics is symbolic of growing political freedoms in Indonesia, it also highlights problems that, if unaddressed, will significantly impact on democratic consolidation. Strengthening the institutions that manage and channel conflict at the local level and devolving more authority to them to mediate disputes is of critical importance for Indonesia. We also found that local conflict dynamics, shaped by historical patterns of conflict, the nature of local party systems, and elite power relationships, determined whether election-related conflicts escalated. The form of preexisting conflict in local communities mattered more than the level of preexisting conflict, highlighting the critical dimension of the "local" in impacting conflict patterns within Indonesia. Furthermore, while national conflict prevention measures were important, the most significant conflict management mechanisms were initiated by local actors (individuals, independent local institutions, and local state and nationally-mandated or affiliated non-state institutions). This finding highlights the continued need to build capacity and support effective efforts at the local level, which has gained more importance as the decentralization process has evolved since 1999.9 The paper outlines the extent of and variations in levels of conflict and forms of conflict management, during the legislative elections in the four studied provinces. Section 2 provides an overview of election- related conflict in the research locations. Section 3 considers the causes of conflict, and of variation of 7The newspaper archive was based on local newspapers which included Bali Pos, Denpasar Pos and Nusa Pos in Bali; Radar Madura, Jawa Pos and Radar Bromo in East Java; Ambon Express, Dewa, Info Baru and Suara Maluku in Maluku; and Flores Pos, Pos Kupang and Nusa in NTT. The archive was created by clipping and then coding all reported incidents of conflict during the period studied. Incidents were then entered into a master dataset, and checked to ensure there was no double counting of incidents reported across sources. 8The eruption of sectarian violence in Ambon from April 25th 2004 effectively blocked access to the province for World Bank staff during the period under study. The findings of this case study were based on close collaboration with a local Ambon-based researcher, as well as extensive interviews with key informants conducted outside Maluku. 9The decentralization process began in 1999. Laws 22/1999 and 25/1999, and implementing regulation 25/2000, devolved significant fiscal authority and decision-making responsibility to rural districts and urban cities and, to a lesser extent, to villages (see ICG 2003a). This process has continued through 2004 as greater authority was granted to districts and local accountability has been enhanced with the onset of local elections. See Bell (2003), Crouch (2000), Klinken (2003b), Malley (2002; 2003). 4 incidence across and within the four provinces. Sections 4 and 5 examine, respectively, the conflict prevention and management mechanisms (both formal and informal) that were in place, and reasons for their success/failure. The paper concludes with a summary of the lessons we can learn from the April 2004 elections and provides some recommendations for preventing conflict, strengthening conflict management mechanisms, and improving future elections processes. The findings clearly illustrate that conflict prevention and management efforts need to move beyond episodic, event-driven thinking; effective local actors need to be empowered; and the institutions managing conflict and governing elections need to be strengthened considerably. Our understanding of conflict in democratizing Indonesia in particular has to move below the national level. This is vital if the democratic consolidation process is to be strengthened. II. Overview of Election-Related Conflict 1. The Distribution of Election-related Conflict Conflict during the legislative election period was assessed as being either directly related to the election or not; within each of these categories, we differentiated violent conflict from non-violent conflict. In this paper, we focus specifically on election-related incidents.10 This includes conflict over party symbols/attributes, over election results, and other forms of election-related mobilization, each with potentially violent or non-violent outcomes. Events where physical force was used against individuals or property are defined as violent. This definition excludes threats of force or intimidation, which were deemed non-violent when reported. The analysis of the newspaper archives and interviews during the two-month period found that there were 224 election-related conflict incidents in the four provinces. Due to the limited scope of the newspapers, the data is drawn from a narrow area in two of the provinces studied, East Java and NTT.11 As shown in Table 1 and Figure 1, Bali had by far the largest number of election-related conflicts (102) of the four provinces, followed by Maluku (49), East Java (45) and NTT (28). Overall, the number of incidents is low on a per capita basis considering the 11.4 million population of the area studied.12 Given the large population of the districts covered in East Java, the relative lack of conflict (violent and non-violent) is even more marked. Table 1: Election-related conflict in case study provinces (March 15-May 15) Number of Number of Violent Election Province Election Incidents Incidents Bali 101 9 (8%) East Java 45 2 (4%) Maluku 49 10 (20%) NTT 28 4 (14%) Total 223 25 (11%) 10During the March 15-May 15 period, we documented 105 cases of non-election-related violent conflict in the research areas. See Appendix A. 11This is because we need to use sub-provincial news sources to get an accurate picture of local level conflict (Barron and Sharpe 2005). Thus for East Java and NTT, we picked clusters of neighboring districts, allowing for the use of a limited number of newspapers. In Bali and Maluku, all districts were covered. In East Java, only six districts (out of 29) were covered. In NTT, the seven districts on Flores island were covered, while those in West Timor and Sumba were not. 12The population of the six districts covered in East Java is 5.6 million (of a provincial total of 32 million); Bali's population is 3.1 million; the population of the covered districts in Flores is 1.6 million; the population of Maluku is 1.1 million. 5 Most election-related conflicts were minor disputes, and election-related violence was relatively scarce, further illustrating the limited scope and impacts of conflict in the April elections. Of the election-related conflicts that took place, just over 10 percent were violent (25 out of 224). Significantly, there were no reported election conflict-related deaths in our research sites during the period of study.13 Maluku had the highest number of violent election-related conflicts (10), followed by Bali (9) and NTT (4). There were only two incidents of election-related violence in East Java. The levels were less than expected given the pre-existing tensions in each province and, comparatively, were much lower than for the 1999 legislative election with the exception of Maluku.14 Importantly, the sectarian violence that erupted in Ambon on April 25, 2004 was not linked to the legislative elections and thus not included in this study.15 Election- related violence was also relatively rare compared to other forms of violent conflict during the period of the elections, accounting for under 20 percent of all recorded violent conflict during the two month period. Figure 1: Election-related conflict incidents (March 15-May 15) 120 100 stn 80 dei ncItcil Violent 60 Non-violent onfC# 40 20 0 Bali East Java Maluku NTT Province 2. Forms of Election-related Conflict 13One legislative candidate was killed in West Java over a month before the elections. Another serious incident occurred in Bali in October 2003 before the elections where two people were killed, one supporting PDI-P and the other supporting Golkar. The research traced the incident back to residual tensions from the 1999 elections. 14While we do not have quantitative figures for the 1999 legislative election for three of the four research areas provinces, interviews reported a drop in the level of conflict in 2004, especially violent conflict. This drop is even more significant if one includes the violence around the election of the President by the parliament (MPR) in 1999, manifested most violently in Bali. The one exception is Maluku, where the period around the 1999 legislative elections was the least violent in a year characterized by high levels of ethnic violence. 15The legislative elections had little to do with the pattern of April 25th violence, which occurred only in Ambon city where there were fewer reports of electoral irregularities. Conversely, in the districts of Maluku Tengah, Buru, Maluku Tenggara and Maluku Tenggara Barat where the electoral process was riddled with irregularities and various political parties rejected outcomes, there was no violence. 6 The incidents of conflict related to the elections can be further distinguished by the actors involved. There were four main forms of election-related violence. Inter-party disputes involved clashes between candidates/representatives from different political parties, including party leaders, party cadres and satgas (paramilitary groups). Intra-party conflicts involved disputes between rival candidates and their representatives within the same party. Inter-group/individual disputes among non-state actors were the third type. Lastly, there were disputes between individuals/groups and the state, predominantly the local election commission or KPU. These incidents ranged in intensity from physical injury/property damage to psychological intimidation and other non-violent disputes. Our results, detailed in Table 2 and Figure 2, show that over half of the incidents (53 percent) involved conflicts between parties/individuals and the state, usually over the election process or results. Nearly a fifth of the incidents (18 percent) were inter-group/individual disputes among non-state actors. Conflicts within political parties accounted for 15 percent of the election-related disputes, a slightly higher total than for inter-party conflicts (14 percent). The proportion of conflicts that involved parties/individuals protesting against the state appears to be larger than is the case for other democracies. A study of elections in fourteen countries in 2002 by the International Foundation for Election Systems (IFES) (Fischer 2002), for example, estimated that only 14 percent of election conflict cases related to voters in conflict with the state, claiming unfairness over election processes, compared to over 50 percent in Indonesia. Table 2: Forms of election-related conflict (March 15 ­ May 15) Inter- Parties/individuals Inter-party conflicts Intra-party conflicts group/individual versus the state (No. incidents/% (No. incidents/% conflicts (No. conflicts (No. within province) within province) incidents/% within incidents % within province) province) Bali 21 (20%) 20 (20%) 31 (31%) 29 (29%) East Java 1 (2%) 3 (7%) 2 (5%) 39 (86%) Maluku 7 (14%) 7 (14%) 7 (14%) 28 (57%) NTT 1 (4%) 4 (15%) 1 (4%) 22 (81%) Total 30 (14%) 34 (15%) 41 (18%) 118 (53%) The form of election-related conflict varied sharply between provinces. While the majority of disputes were between individuals/groups and the state, these conflicts were particularly concentrated in East Java and NTT, where they accounted for 86 percent and 81 percent of all incidents respectively. In contrast, such disputes were less dominant in Bali (29 percent of the provincial total), with a larger number of inter-group/individual conflicts among non-state actors being reported there (31 percent). In Bali, there were also relatively high numbers of both inter- and intra-party disputes (21 and 20 percent respectively). In Maluku, conflicts between individuals/groups and the state were also the most prominent (57 percent), but there were a considerable number of incidents of conflicts of the other three types. It is important to note, however, that these findings may be influenced by limitations of the dataset; underreporting in newspapers, especially of intra-party disputes, may have skewed the data. Interviews with respondents in East Java and NTT suggested that in both these provinces more cases occurred than newspaper accounts indicated. These conflicts rarely make the newspapers unless they take violent form. By contrast, in Maluku where readership appetite for news of inter-party and intra-party rivalry was high, the local newspapers tended to be more comprehensive in their coverage of election-related disputes. 7 Figure 2: Type of dispute by province (March 15-May 15) 45 40 st 35 end 30 cinI 25 Violent ctli 20 Non-violent nfoC 15 # 10 5 0 il Ba mit Ja ukul T il Ba mit NT Ja ukul T il Ba mit NT Ja ukul T il Ba mit NT Ja ukul T NT Ma Ma Ma Ma Inter-party Intra-party Inter- Parties/individuals conflicts conflicts group/individual and State conflicts Conflicts Form of conflict by Province 3. The Timing of Conflict As Figure 3 shows, the level of the different forms of conflict varied over time. In general, three distinct spikes are noticeable: first, at the beginning of the campaign period; second, in election week; and third, during the first week of May, the period when the vote counts were formally announced. Inter-party conflict peaked during the last week of campaigning just before polling, when inter-party competition was strongest, yet it almost completely disappeared one week after the poll (April 12th), suggesting that this form of conflict was not significant when the results were announced and that political parties by-and- large accepted the results. Intra-party conflict in contrast rose in the last week of campaigning, peaked two weeks after polling when the results of the informal counts were clearer, and peaked again in the first week of May when the results were formally announced. This reflected the unhappiness of particular candidates within political parties over the results, signifying higher contestation within parties for seats after the results were announced. Conflicts between communities and the state coincided with key events during the electoral cycle: at the start of the campaigning period, around polling day, as well as between late April and early May when the results were announced with protests planned to achieve maximum symbolic effect. This temporal variation in the form of conflict points to the need to have different strategies of conflict prevention at different periods in the electoral process. 4. Urban/Rural Distribution Election-related conflict occurred in both urban and rural areas and varied across provinces. In NTT, most conflict incidents occurred in urban areas, while in Maluku the incidents were rural pre-April 25th. The selection of case studies in East Java was weighted toward rural areas, limiting the ability to assess rural/urban distribution differences. However, while the districts themselves were more rural than urban, the incidents which did take place generally occurred in the more urban district and sub-district capitals. In Bali, conflict occurred in both rural and urban areas, although most cases in rural areas were concentrated around towns in highly populated communities. Most of the conflict that involved the state 8 was concentrated in urban areas, the institutional center of the election bureaucracy and the security services. Figure 3: The timing of election-related conflicts (March 15-May 15) 45 40 35 Inter-party conflicts s nt 30 Intra-party conflicts dei ncItcil 25 Inter-group/individual 20 conflicts Parties/individuals and onfC#15 State Conflicts Total Election related 10 conflicts 5 0 Wk Wk Wk Wk Wk Wk Wk Wk Wk 15th 22nd 29th 5th 12th 19th 26th 3rd 10th March April May Date This finding concurs with other studies of contemporary conflict in Indonesia (e.g., Varshney, Rizal, and Tadjoeddin 2004) that find more violence in urban than in rural areas. This may in part be a product of an urban bias in newspaper reporting (Barron and Sharpe 2005). Yet, it also is a function of the new predominant form of conflict in Indonesia during elections, conflict directed toward institutions of the state, which is concentrated in urban areas. It should also be noted that this pattern does not mean that the causes of such violence have their roots in conditions or tensions in urban areas. Rather, urban centers serve as venues for parties to engage in conflict; often groups from rural areas take their grievances to district capitals, where their protests are more likely to be heard (and covered by local media) and where the appropriate adjudicating institutions of the state are located.16 The geographic distribution of conflict also reflects underlying patterns of preexisting local conflict in the provinces, which illustrates the centrality of the "local" in understating conflict. For example, in Maluku during the period through April 25th, previous inter-village conflicts were rekindled before the election and during the election period. Similarly in East Java, incidents of conflict occurred at the district level in predefined "hot spots" where there were higher frequencies of previous conflicts and violence, and at the sub-district level where a high concentration of bajingan (criminal groups) existed with no counter local leadership. 16See O'Brien (1996) and Wilkinson (2004) for a discussion of this in China and India, respectively. 9 III. Causes of Conflict or Lack of Conflict The patterns of election-related conflict raise questions about the underlying causes of conflict. Why did some areas experience more conflict than others? Why in some places did conflicts take violent form, while in others disputes remained small and non-violent? The likelihood of an area experiencing election- related conflict, and the seriousness of that conflict, was a function of both the underlying factors that caused tensions to arise and escalate and the efficacy of the conflict prevention and management mechanisms that were in place in the locality. This section will consider the former--the factors that determined whether election-related conflict would arise and escalate. The next two sections will examine in more detail the effectiveness of conflict prevention/management mechanisms in the four provinces. While these two dimensions are, of course, interrelated, we treat them separately in order to illustrate better the underlying dynamics of conflict at the local level. Our analysis of the selected provinces identified four main factors that explain why conflict was present (or absent) during the April 2004 elections. First, institutional weaknesses primarily at the local level, in particular vis-à-vis the electoral bureaucracy and security sector, provided a context where tensions could emerge. Where the state was (perceived as being) more effective, in terms of being able to run and manage the elections in a non-corrupt, transparent, accountable, neutral and efficient way, election tensions were less likely. Second, the local history of conflict was also important in shaping the likelihood of election-related tensions, as was the extent to which local actors sought to mobilize based on pre- existing cleavages/tensions. Where pre-existing cleavages were tied to electoral outcomes by local elites, or where the elections exacerbated preexisting tensions, election-related conflicts were more probable. Third, the nature of the local party system also mattered in that the strength of party loyalties, internal party discipline and the level of competition among and within parties helped structure intra- and inter- party dynamics, impacting both the form and level of conflict. In areas where party identity was stronger and competition was high among or within parties, election-related conflict was more probable. Fourth, and finally, the extent to which local populations viewed the elections as `mattering' also helped determine conflict levels. Where elections were perceived to have a significant impact on local power structures, election conflict was more likely. Conversely, where local populations did not view the elections as significant, or where they valued the peaceful conduct of the elections, election-related disputes did not tend to escalate. These explanations are political, historical or institutional in nature and build on a rich understanding of conflict within Indonesia and elsewhere.17 Our analysis draws directly from rich cases studies (see Acciaoli 2001; Aragon 2001; Barron, Clark and Daud 2005; Barron and Madden 2004; Barron, Smith and Woolcock 2004; Clark (ed.) 2004; Cribb 1990; Davidson 2003; Fein 1993; ICG 2002a, 2002b, 2003a; Klinken 2001; Robinson 1995; Smith 2005; Tajima 2004) and an increasing attention to local variation across Indonesia (see Aspinall and Fealy 2003). Recent research on the whole has enriched our understanding of contemporary violence in Indonesia using a variety of frameworks (see Colombijin and Lindblad 2002 and Anwar et al. (eds.) 2005). What distinguishes our study from others is the focus on institutions at the local level during the electoral process18 and the breadth of comparative local data. We show that in Indonesia, examining the effectiveness of the democratizing state as an arbitrator of conflict (highlighted by Wilkinson 2004 for India), rather than seeing the state as an agent of violence, a view which has dominated understanding of violence in Indonesia (see Anderson 2001; Barker 1998; Imparsial 2003; Liem 2002; Lindsey 1998; Malley 2001a, 2002b; Zinoman and Peluso 2002), emerges as central to 17See Snyder (2000) and Tilly (2003) for theoretical discussions of the relationship between violence and elections. 18This varies sharply from other studies of violence during elections in Indonesia, see Emmerson (1999), NDI (2004) and Vermonte and Budiman (2005). For a similar approach looking at the effectiveness of local institutions in the area of education in Indonesia, see Bjork (2003). 10 understanding both the nature of conflict in Indonesia and the dynamics of state-society relations at this point in Indonesia's transition. Other explanatory factors, including local economic and social structures, group inequality and ethnic/religious heterogeneity did not adequately explain variation in election conflict incidence. For example, the availability of natural resources, highlighted by World Bank (2003) and Homer-Dixon (1989) as a key conflict driver, was not a good indicator of levels of conflict. Neither were rich or poor provinces, or districts for that matter, more or less likely to experience conflict, as Galtung (1969), Burton (1990), Easterly (2001) would imply would be the case. Inequality between groups, argued by Tilly (1999) and Stewart (2002) as being a key cause of conflict, was similarly not an important source of election-related violence. Ethnic or religious heterogeneity--emphasized generally by Brown (2002), Gurr (2000), Horowitz (1985), for cross-national investigation by Lande (1999), at the micro level by Miguel and Gugerty (2002), and in the Indonesian context by Bertrand (2004)--was similarly not an accurate predictor of conflict. In fact most of the election-related conflict was not sectarian in nature. This is not to say that ethnic or sectarian factors were irrelevant during elections. Regions like Maluku that experienced sectarian violence from 1999, were conscious of the potential for ethnic violence to flare up and took great pains to ensure that sectarian tensions were well-managed during the 2004 elections. Local executive elections from 1998 through 1999 were affected by the levels of ethnic violence in Indonesia. Klinken (2005) argues that political struggles between status quo groups resisting regime change and emerging elites seeking to secure resources and power in the context of decentralization were shaped by electoral contests. His 2005 research and events in Maluku point to the need to examine the interplay of factors shaping conflict during elections and to move beyond sectarian factors, which have dominated the research, in understanding conflict within Indonesia. Other common explanations did not emerge as critically important in the elections. For example, the presence of preman or bajingan (criminal) groups did not in itself lead to election-related conflict. Rather, only when economic and social factors interacted with the four factors outlined above did they have an impact. Similarly, the level of social capital (argued as being crucial by Vashney 2002) did not emerge as paramount in shaping conflict, although the scope of the interviews limited a full analysis of inter-group relations and in specific regions, Bali and East Java, the presence of social cohesion minimized conflict. While all these factors are important, in the context of election-related conflict, they only gained salience through their interaction with actor-driven political dynamics and with the institutional configurations and historical conditions that help structure the interests and incentives of such actors (Box 1). 1. Institutional Strength/Weakness Of the four causes of election-related conflict, institutional factors were the most central. This reflects the dominance of conflict directed against the state. While the April 2004 elections demonstrated progress towards the acceptance of competitive electoral processes, at the same time institutional weaknesses-- demonstrated in problems in the running and management of the election--created sources of tension (over perceived corruption, procedural misconduct and official bias) and, in some cases, allowed these tensions to escalate. Variations in the strengths or weaknesses of the state institutions that ran and oversaw the elections, and that were mandated to address problems that emerged, helped determine local levels and forms of election-related conflict. 11 BOX 1: OVERVIEW OF THE LOCAL DYNAMICS OF CONFLICT IN THE FOUR PROVINCES In the province that experienced the highest number of incidents, Bali, election-related conflict was largely shaped by the recent history of election-related violence that raised and rekindled political tensions and local competition. The April 5th elections proved to be an opportunity for a number of isolated politically motivated revenge attacks from earlier election- related conflict. Tensions were highest in districts where PDI-P elites, who politically control Bali, refused to accept challengers or where the electoral competition threatened existing power structures. Districts highly dependent on a tourist economy were more sensitive to the negative impact of election violence and, consequently, elites in these areas were more invested in a non-violent election. The prevention mechanisms implemented by the police and local elites were highly effective, not withstanding the high number of incidents, and significantly reduced election-related conflict compared to the 1999 polls. An overwhelming majority of the election-related conflict incidents were non-violent. East Java had a similar pattern of predominantly non-violent election-related conflict, although overall levels were lower, particularly considering its higher population than the other provinces in the study. Existing patterns of tensions among Islamic boarding schools (pesantren), tied to different political parties, underscored a good portion of the conflict. The role of local elites proved equally significant. In areas where local leaders were not dedicated to minimizing conflict, either individually or collectively (particularly where their support bases were roughly equal in size and influence), competition between parties flourished and led to higher levels of conflict and election process irregularities. More often than not, this involved the mobilization of different local crime networks (bajingan), who also had a vested interest in the election outcomes. The most recent catalyst of political competition, the 2003 Bupati (District Head) elections significantly affected local political dynamics, leading to intra-party splits as well as heightened inter-party competition, all of which were exacerbated by vote buying and money politics during the legislative election campaign. These tensions, although high, did not erupt into violence and were managed through strong conflict prevention efforts. Many local elites openly called for non-violence and the decentralization process and security reforms empowered these elites to address local concerns, whilst helping to increase the understanding of local officials. Of the provinces studied, Maluku had the highest level of violence prior to the elections. Religious tensions erupted in 1999, killing thousands, fueled by underlying ethnic tensions stemming from unequal access to jobs and political positions as well as partisan security forces that allowed the infiltration of armed, extremist Muslim groups exacerbating the local conflict. A peace agreement brokered by the central government in 2002 has been accompanied by extensive and widely-supported local peace-building initiatives. During the April elections, the investment in conflict prevention efforts was largely successful, as the election was peaceful and relatively trouble-free in spite of the political reversals experienced throughout the province. Incumbent, status quo parties lost in all districts except in Ambon, while the relatively new Muslim (PKS) and Christian (PDS) parties emerged as political contenders. On April 25th however, the mishandling by security forces of a local anniversary associated with separatist aspirations led to an eruption in violence resulting in 38 deaths through May 15th. While the legislative elections themselves did not lead to violent conflict, locals believe that the start of the presidential contest was a factor. The clear causes, however, are unresolved issues from the 1999-2002 conflict exploited by disgruntled former militia leaders and the poor response of the security sector. NTT stands in sharp contrast to the other provinces. This eastern province had the least number of conflict incidents and those that occurred primarily involved problems with electoral administration, rather than the mobilization of preexisting tensions. NTT lacks the political importance or economic wealth of the other provinces, so there were limited funds and less competition associated with electoral outcomes. Emotional attachment to the election results was low and voter apathy comparatively high. The elections did not exacerbate existing tensions; conflicts over land and across the dominant social cleavage, clans or kinship groups, were not tied to electoral outcomes, and the legislative elections did not challenge the status quo in communities. The conflicts that did arise were related to the electoral process: training of election personnel, lack of witnesses at the polls, and voter disenfranchisement due to shortcomings in voter registration. A strong conflict prevention effort was not present, beyond the national initiatives introduced in policing practices and the campaign process. The lack of conflict in NTT was less the result of elite initiatives, than the broad perception that the outcome of the legislative election would not challenge existing power structures. The extent to which institutional weakness was a source of conflict can be seen by the fact that the predominant form of election-related conflict across the four areas (53 percent), and particularly in East Java (86 percent) and NTT (81 percent), was that of individuals/groups protesting against the state. Complaints centered on voter registration, voter education, training of election officials, lack of neutrality, 12 and the implementation of the elections. Most of these complaints were targeted toward the local or district election commission or the police. In this way, institutional weaknesses directly contributed towards election-related tensions; in some cases these tensions escalated into (violent) conflict. Where the state was (perceived as being) more effective, in terms of being able to run the elections in a non-corrupt, transparent, accountable, neutral, and efficient way, electoral tensions were less likely. Democratic states draw their strength not only from their enforcement ability, the extent to which they hold a monopoly on the means of force, and their ability to unilaterally implement decisions, but their strength is also dependant on the judicious use of state power that the public deems to be legitimate. Legitimacy is in part, a function of the degree to which the citizenry feel they can engage the state, the nature of state response as well as expectations of state accountability. The legislative elections displayed the growing confidence that individuals and groups have in protesting the procedures and outcomes of elections; it stands in stark contrast to previous elections where communities felt far less able to protest apparent corruption and intimidation due to excessive state repression. Thus, in many respects, the fact that these kinds of disputes were relatively common is a sign that Indonesian state institutions are democratizing. The protests also point to the growing confidence of Indonesian civil society, which has strengthened considerably during the democratic reformasi (reform) era. However, state legitimacy is also dependant on communities trusting the state as an arbiter of their complaints. The scale of protest--and of perceived post-election disillusionment--shows the potential for a legitimacy deficit that could affect the likelihood of future elections remaining peaceful and the level of engagement in the electoral process. A failure to address institutional weaknesses, in the election apparatus and beyond, could potentially undermine democracy in Indonesia in the future. Weaknesses in the implementation of the election varied across the four provinces and were associated with three main substantive areas. The first centered on an inability of officials to fully carry out their responsibilities during the election--official incapacity. This was largely a result of a lack of training, funding and logistical support for officials and polling staff. Official incapacity occurred in many forms. One source of contention resulted from a lack of training in voter registration, as well as poor monitoring of officials. In East Java, there were a number of complaints related to election officials not visiting households to register villagers. The result was that people who were dead, underage, or no longer living in the region were registered to vote. Inadequate training and monitoring also meant that some polling station staff were unable to distinguish between votes for candidates and parties, while many did not know how to tabulate the results. Inconsistencies in the method of counting between the village and sub-district level polling stations led to tensions between polling staff and witnesses. Poor preparation by the election bureaucracy moreover resulted in candidates' names not being printed on ballot papers and delivery of papers to the wrong electoral districts. Further, late delivery of reporting and aggregation forms delayed the counting process, increasing tensions in places. In some cases, these inefficiencies were related to problems with budgetary allocations from the central government and poor logistical support from the national KPU or election commission. However, there was also considerable local variation in the performance of the district and provincial level KPU. These problems were most pronounced in NTT and Maluku. East Seram in Maluku Tengah saw rioting by scores of angry voters outside the Sub-district Head's (Camat's) office when the local KPU failed to distribute the ballot papers on time. Voting only took place on April 7th after the provincial KPU took charge of distribution. In addition, polling staff in East Seram received no training from the Maluku Tengah KPU because it claimed to have insufficient funds. In NTT, officials provoked violent reactions from citizens who demanded a fair and efficient accounting of the results. In Sikka district, protests over alleged irregularities in vote counting and the electoral 13 process resulted in a mob of over one hundred people attacking the local KPU with rocks and fishing explosives and death threats for the head of the district KPU branch. The situation was only brought under control by intervention from the police and local party leaders. Nor were the problems confined to Sikka. Two KPUs in nearby districts within NTT were occupied by angry citizens, as was a KPU in western Bali. Voter registration was also poorly carried out in places, notably Maluku, where a lack of coordination between state agencies underscored poor state performance. The district-level KPU relied on figures from the government statistics agency (BPS), which failed to check and update its figures with local neighborhood chiefs and village heads. The result was that more than 20,000 voters, particularly Internally Displaced Persons (IDPs), were not registered to vote. This led to widespread protests and demonstrations because unregistered voters were afraid that they would be stigmatized politically as supporters of the banned Maluku Independence Front (FKM) who had urged citizens to boycott the elections. Such cases highlight the need to improve training, funding and oversight at the local level, especially as the decentralization process continues and the election bureaucracy takes on greater responsibility for managing elections locally. Building local institutional capacity is an essential ingredient for future conflict prevention. A second concern relates to the conduct of state officials during the election. While many officials within the election bureaucracy and security services worked to promote a professional work ethic, official bias created tensions and, in some cases, undermined the standard of the election being free and fair. In East Java, and particularly on the island of Madura, there were instances where the District Heads `encouraged' civil servant staff and village heads to vote for their preferred party. The lack of neutrality by state actors led to increased tensions, especially among excluded groups. Similarly in Maluku Tenggara, the local KPU added ten polling stations to its list in violation of electoral law. According to PDI-P members, the votes from these polling stations were divided between Golkar and the other parties who wanted to oust the incumbent PDI-P from power. In yet another instance, close ties between local officials and local security personnel in Tabanan, Bali prevented a full investigation of a violent campaign attack. Bias of state actors at the local level, whether election officials or security personnel, can potentially undermine the fairness of the electoral process, delegitimizing local democracy. Equally important were claims of misuse of state funds and perceptions of corruption tied to the lack of a transparent procurement process. The poor handling of election-related contracts in a handful of localities, such as those involving the data entry of results in Jembrana, Bali, led to discontent and tensions which ultimately undermined the credibility of the state officials charged with managing and implementing the elections. Provincial and national oversight of local transparency and corruption remains critical. In Jembrana and Maluku Tenggara, the local election commissioner was subsequently replaced, improving the management of subsequent elections. Other areas and provinces did not have the same level of oversight and intervention. Developing and maintaining the professionalism of election personnel remains a challenge in Indonesia's ongoing democratization. A third area of institutional weakness concerns the relationship between officials, political parties and citizens. Voter education by the KPU, political parties and NGOs was lacking in varying degrees across the provinces and at different stages in the electoral process. In parts of Bali, for example, limited voter education affected the registration process, resulting in no informal monitoring by the community to counter anomalies. This lack of monitoring created a space that in places allowed parties to mobilize voter support through the inclusion of "incentives" with registration cards. Voters in all four provinces failed to understand the party list system and the procedures for addressing irregularities during the vote counting on polling day. Significantly, local and national NGOs focusing on voter education and conflict 14 prevention were present to some degree in Bali, East Java and Maluku, but not in NTT. Thus, unsurprisingly, the problem was most pronounced in NTT, and created conditions for misinterpretations and conflict during and after the election. Many of the polling stations in NTT lacked witnesses to observe vote counts, and as such disputes often escalated. This was the result of political parties lacking funds to pay witnesses to remain engaged during the lengthy counting process. Yet this problem existed elsewhere. In Maluku, the provincial KPU was roundly criticized for not socializing the new election law or voting procedures to polling staff or citizens. This was less of a problem in Ambon city due to local NGOs and the press but was more pronounced in more remote parts of the province where voter awareness remained low. In other instances the lack of broad engagement was a by-product of Indonesia's past, specifically the New Order period. While the police force has reformed extensively, some officials were reluctant to cooperate with NGO groups in Bali, which were perceived to be antagonistic to the state. This limited the ability of officials to manage conflict, a theme developed in more detail below. However, perceptions of the local KPU's performance in the socialization of the election laws and voting using the party list system was positive in Bali compared to the other three provinces, with candidates from all parties involved in the socialization process from the outset. There were still problems with voter education in the more rural and isolated villages, particularly among the aged members of the community who were unfamiliar with new procedures and in many cases abstained from voting. The fostering of vibrant and sustainable links between the state and society, and between different institutions of the state, that channel communication and help reduce violent conflict, is an ongoing challenge. The experience in the April 2004 legislative elections emphasizes the importance of strengthening existing institutions. With over half of the conflicts developing as a product of institutional weakness, addressing the capacity, professionalism and societal engagement of officials in the election bureaucracy and security sector is critical. 2. Local Actors and Histories: The Mobilization of Pre-existing Tensions The level and form of election-related conflict in the research sites was also dependent on the extent to which local leaders attempted to mobilize population groups based on existing cleavages. This, in turn, was often a result of the local history of conflict in that area. The case studies show that the likelihood of election-related conflict in the legislative elections was affected by the forms of preexisting tensions in communities, and their relationships to the election, rather than preexisting levels of conflict per se. The dominant form of preexisting conflict proved important because of the way it interacted with the political competition associated with the elections. When the legislative elections acted as a catalyst for preexisting tensions, conflict was higher. In cases where preexisting tensions were perceived as distinct from electoral competition, levels of conflict were considerably lower. In Maluku, for example, with underlying religious tensions persisting from the 1999-2002 conflicts, as well as the presence of religious political parties, there were concerns that preexisting tensions would interact with political competition. However inter-party competition was strongest between parties of the same religious or secularist affiliation; party rivalry was keenest between PDI-P and Golkar, whose supporters are mainly Christian or secularist, the possibility of inter-party competition igniting religious tensions was small. Similarly, intense rivalry between Muslim parties, such as the United Development Party (PPP) and the Prosperous Justice Party (PKS), competing for the urban Muslim vote in Ambon focused on rooting out corruption and did not revisit the inter-religious tensions of the 1999-2002 conflicts. The fact that electoral rivalries did not fit with preexisting patterns of religious conflict in 15 Maluku explains, in part, why violence in this conflict-ridden province did not follow high earlier patterns.19 Similarly, in NTT, the legislative elections were not perceived as significantly affecting conflict over the dominant local problem, disputed land.20 Consequently, election-related conflict was considerably lower. The prominence of revenge killings in East Java illustrated a parallel dynamic; these killings, by and large, were not tied to the elections and thus the level of conflict in East Java was not as high as it would have been if the elections had ignited revenge responses.21 In Bali, however, revenge killings associated with elections were more prevalent. The pattern related to the 1999 contests, particularly in northwestern Bali where conflict escalated between Golkar and PDIP supporters. The stronger link between the elections and preexisting tensions explains in part why the numbers of incidents of conflict in Bali was higher than in other provinces.22 Conflict associated with previous local electoral contests in East Java also extended into the 2004 legislative elections. Forms of political conflict arising from the recent District Head (Bupati) elections in East Java triggered a switch in the allegiances of local elites and pesantren (religious school) networks to different parties for the legislative elections. This heightened tensions between the main parties (PKB, PPP and PBB in Pamekasan, and PKB and PPP in Probolinggo), as well as within the party ranks of PKB in both areas, during the legislative elections. However, in large part because of conscientious actions from local elites, these tensions remained within the political sphere and led to few incidents of violence. While preexisting tensions and local histories are crucial to understanding and addressing "new" conflicts, communities can also break with the past and learn from their history. The legislative elections in Maluku saw citizens vote out incumbents who had not dealt with the 1999-2002 conflict adequately and switch allegiances from mainstream parties fielding candidates from the past, to newer parties with new candidates. Parties or candidates that emphasized division were not popular; voters chose instead parties and candidates who had provided assistance and bridged divisions during the 1999-2002 conflict. Part of the reason the forms of preexisting tensions were important concerns the relationship between preexisting conflicts, control of the state, and party identification. In areas where control of the state was pivotal in shaping the outcome in local conflicts, the elections served to ignite tensions. This occurred, for example, in parts of Bali, where the elections provided an opportunity to consolidate political power within the state for some PDI-P elites at the expense of potential rivals both within and outside of the party. This was not the case in Ambon, Maluku, where the legislative contest was also seen as crucial in determining who wins executive power in the next mayoral contest and tensions were contained. In this case, inter-party rivalry between Golkar and PDI-P was magnified by intense personal rivalry between the incumbent mayor from PDI-P and his challenger from Golkar. During Golkar campaigns, the latter vowed that not only would Golkar win the legislative elections in 2004, but it would go on to wrest the mayoralty from PDI-P in 2006. Yet tensions were not ignited because both contenders were also Christian community leaders committed to peace, so the scope for mobilization both within the Christian and the Muslim community was reduced, underscoring that elite behavior and leadership can significantly shape the interplay between preexisting conflicts, control of the state and party identification in determining conflict. In contrast, in the districts visited in East Java, the legislative elections were seen as a means of regaining access to the political power lost by some parties during the preceding District Head (Bupati) 19On conflict in Maluku and neighboring North Maluku, see Andrianto (2000), ICG (2001b, 2002a); Klinken (2001); Trijono (2001); Wilson (2001, 2005). 20On land conflicts in NTT, see Clark (ed.) (2004), and Prior (2003, 2004). 21On revenge killings in East Java, see Diprose (2005) and Wiyata (2002). 22On political conflict in Bali, see Robinson (1995) and ICG (2003b). 16 elections. Tensions were highest in those geographical areas perceived to be the power bases of the District Heads and their various party allegiances. In all four provinces, dominant local forms of identity were important in shaping the scope and form of conflict. However, elite behavior ultimately affected the timing and level of violence of the conflict during the election period. In the northeastern part of East Java, identity formation is often parochial and most closely associated to affiliations with local Muslim clerics and their pesantren (religious schools). This is particularly the case in rural areas, where voter preferences generally follow that of religious and other associated leadership. While this is beginning to change, the parties must strategically form alliances with the different religious elites to gain access to blocks of voters. However, at the same time, it is often the religious leaders--not the members of the local administration directly--who are eventually held accountable for public discontent with irregularities in government administration. Such strong links between the religious leaders and the identity of the local populace necessarily means that public statements made by religious leaders can easily incite the masses (as occurred in 1999) or can be a form of restraint. The public statements guaranteeing the peaceful implementation of the elections, as well as other efforts by these religious figures, were one reason why the incidence of conflict in East Java was not as high as predicted. The research showed that local elites played varying roles in the extent to which they provoked conflict or, conversely, helped to prevent or manage it. One of the surprising findings of this study concerns the impact of mobilization of party support through paramilitary groups--known as party satgas--or through links to bajingan or preman criminal groups. Although these groups are common in Bali and Ambon, and to a lesser extent in East Java, their presence did not directly contribute to conflict nor was their presence associated with higher levels of intimidation. Rather, it was the management of these groups by elites that proved to be decisive. If elites lacked control or alternatively incited these groups, as was the case in parts of East Java, then bajingan often contributed to election-related tensions. Yet, where elites monitored these organizations carefully, as was the case in parts of Bali and other parts of East Java, or where police intervention fostered non-violent cooperation and dialogue among these groups, then the presence of paramilitary groups did not negatively affect conflict levels within or between parties.23 In Ambon, preman allegiances are tied to individuals rather than parties per se and in the elections preman worked primarily to influence the position of candidates on the party list. Whilst preman have networks that descend to the village level in Ambon and Maluku Tengah, they are regarded as local heroes so intimidation to vote for particular parties/individuals would not be necessary as villagers would follow their example willingly. There were no clashes between preman gangs and youth organizations from rival parties in Ambon during the legislative elections. In fact, in some instances, paramilitary groups worked to maintain peace, clearly demarcating areas of support without clashing with opponents. Paramilitary groups only contributed to election-related violence when leaders mobilized these organizations in disputes or when leaders failed to have effective control over these organizations. An example from Bali reinforces the critical role of elite intervention, or lack thereof. In Tabanan district, the failure of local elites to respond to conflict set a precedent early in the campaign period. A small party challenged the dominance of the PDI-P and as a result its leaders were subjected to intimidation. The most violent incident occurred when a party leader's house was surrounded by "ninjas", believed to be affiliated with PDI-P, who ransacked his residence and set fire to the party leader's car. The police were called and opted not to respond during the incident. Election officials in Panwaslu and the KPUD (the district level KPU) were also informed, but felt the event was a "criminal" action outside of their authority. Local village leaders claimed the action was provoked by outsiders. In essence, all of the actors 23See Meliala (ed.) (2005) on the importance of effective cooperation between the police and political satgas in conflict prevention during the 2004 legislative elections. 17 avoided intervention. There were no witnesses willing to come forward, although tensions persisted beyond the electoral campaign. The incident set a precedent for supporters of alternative parties in the area and for voters who refused to follow the status quo. Tabanan continued to experience high levels of inter-party conflict. Similarly, in East Java, in one sub-district, the bajingan groups were mobilized to purchase votes for one of the major parties. When it became evident that the party concerned was not winning in the vote count, the head of the sub-district electoral committee was threatened and intimidated by the bajingan and eventually run out of town. At the time the research was carried out, he had not yet returned and feared for his life. No efforts were made by local elites in this case to guarantee the safety of the elections official, nor to control the bajingan, demonstrating that elite intervention is essential for the peaceful implementation of the elections. 3. The Nature of the Party System The strength of party loyalties and the level of competition among and within parties directly impacted both the form and level of conflict. Levels of conflict, and inter-party conflict specifically, were higher in areas where the incumbent party faced a viable challenge from another party that threatened the existing power structure. Although inter-party conflict was rare overall, it was most common in Bali. In northwestern Bali, conflict emerged between PDI-P and challengers, especially Golkar, and among Muslim parties, namely PAN and PKB. Inter-party conflict was also high when incumbent elites sought to consolidate their position for the forthcoming Bupati (District Head) elections, as was the case in Tabanan in Bali, or when lingering tensions remained from previous Bupati elections, as was the case in East Java. Moreover, inter-party conflict was higher in areas where party loyalties were stronger. In NTT, for example, party identity was quite weak and, not surprisingly, the level of support and emotional attachment to parties was also correspondingly low. This experience contrasted with that of Bali, where attachment to "Mega" and PDI-P was deeply rooted. In areas like Buleling and Tabanan, party identification, tied to previous conflicts from the 1940s onwards, had evolved over time and was particularly strong. In contrast, while the potential for inter-party conflict was high in Ambon city, Maluku, it was not significant for two reasons. Disillusionment over government mismanagement of the 1999-2002 conflicts and subsequent recovery/reconstruction had weakened former party loyalties to the three largest parties. Moreover, past conflicts had reduced both the public and elite appetite for conflict and rowdiness, with the result that electioneering was more restrained than in the past. While the historical context was important in shaping party relationships, the introduction of the partial open-list system for determining elected candidates in the 2004 elections created a new dynamic that increased the level of intra-party conflict. Unlike the closed-list system of 1999, where party leaders determined candidates' ranking and ultimately who got elected, voters in 2004 could vote for individual candidates as well as parties, potentially giving popular candidates more leverage over party bosses. Unfortunately the new system required candidates to achieve a quota of votes in order to be elected and because the quota was set too high, with almost no candidates reaching the required total, parties reverted to candidates' ranking on party lists determined by party bosses to fill seats.24 In a number of cases, candidates who won the greatest number of votes did not gain a seat from the party leading to conflicts between candidates, between candidates and party bosses, as well as between the supporters' of disappointed candidates and party loyalists. Intra-party infighting was particularly intense where loyalty 24Since the quota ranged from 8.33 percent of valid votes for electoral districts with 12 seats to 33.3 percent for districts with three seats, it was hard enough for many of Indonesia's 24 parties to win a quota of votes, much less individual candidates. 18 was invested in individual candidates rather than parties, as was the case in NTT where clan loyalties around individual candidates predominated or in East Java where candidates were identified with particular pesantren (Islamic schools). The selection process for the party list, coupled with internal party discipline and elite intervention in managing competition, often affected the level of party infighting. When party elites opted for transparent ranking criteria or discussion, rather than responding to posturing by supporters, sometimes tied to paramilitary groups, or financial pay-offs, disputes were more often peacefully resolved. In Maluku, candidates' positions in Golkar were determined by their positions within the party, by the level of support from other Golkar branches in the district, by contribution to party activities, and through letters of support from the public. Unfortunately, endorsement by party managers at the district and provincial level was crucial and this reduced transparency and led to financial pay-offs. Levels of intra-party conflict in Maluku were lower in parties with strong internal party discipline such as Golkar, PDI-P and PKS. This was not the case with the highly factionalized United Development Party (PPP), where candidates staged public protests and attempted to exploit factional differences when they were dissatisfied with their positions on the party list. Tensions over party lists occurred in parts of NTT and Bali, although they were handled differently. In NTT, local clan structures controlled the selection process and shaped conflicts, while in Bali the critical role of leaders with access to financial support proved decisive. Intra-party conflicts over the party list intensified after the outcome of the elections became clear, especially in contexts where parties did not do as well as expected, forcing a reevaluation by parties of who would enter the legislatures. For example, in Probolinggo, East Java, the PKB chair and associates signed an agreement prior to the elections that seats would be allocated based on the number of votes received by each candidate. However, this became a major source of contention following the announcement of the preliminary results, where individuals who did not get the number of votes that they had anticipated, disputed the original agreement. In parts of Bali, candidates bought their way up the list, `encouraging' those higher on the list to drop out. When one candidate refused, he was subjected to intimidation, although ultimately he was able to retain his seat. In NTT and Maluku, candidates discussed the possibility of splitting terms, although this violates election rules. It remains to be seen if these informal agreements will be kept or whether conflicts will re-emerge in the future. The growing importance of individual candidates in the 2004 elections reduced the level of party funds for local campaigns and many candidates had to finance their own campaigns. Given that candidates also had to pay off party bosses to ensure a higher position on the party list, there was less money all round for traditional forms of vote buying that had prevailed in the past such as paying supporters to attend party rallies. Instead candidates and/or parties opted for more selective forms of money politics such as offering development projects or assistance to villages in exchange for votes guaranteed by the Village Head. Overall, the legislative elections did not profoundly affect power dynamics within or between parties in communities. Incumbent parties maintained their control in three of the four provinces studied. At the same time, they provided an opportunity for a small group of new elites to emerge and, since the legislative elections created new offices, diversified political power in communities. Maluku, was an exception, as incumbent elites were thrown out of office and new parties and candidates were chosen. Significantly, many of the newly elected had played important peacekeeping roles during the 1999-2002 conflict. Popular support for peace efforts can be more decisive than inter-party rivalries. 4. Legislative Elections: A Less Tense Contest In general, the devolution of power to local communities has not been accompanied by a rise in tensions associated with legislative contests. Indeed, there was less election-related conflict than was the case in the 1999 elections. However, the levels and nature of public engagement in the 2004 legislative elections varied across the provinces. In NTT, the legislative elections were viewed with high levels of skepticism; 19 in short, people did not feel they were important. In East Java, the district elections, evoked the most tensions, compared to the provincial and national contest, reflecting the local nature of power bases and networks. This was not the case in Bali where candidate selection by party elites outside of Bali reduced the levels of local competition. In contrast, public engagement with the legislative elections in Maluku was high. People felt that a peaceful election could showcase the province's break with its violent past and, consequently, tensions were well-managed. In East Java, local executive elections (e.g., the Bupati elections) were perceived to be much more important and generated a stronger connection between election-related conflict and preexisting local tensions. This was the case because in East Java people tend to strongly identify with the leader of the executive branch of government, rather than any one party or the number of different members of the legislative assemblies. Indeed, reports from the Bupati elections of mid-2005 across Indonesia demonstrate much higher levels of conflict, including the burning of party offices and local KPUs, driven by extensive money politics. In East Sumba district in NTT, the police had to fire warning shots and use tear gas to disperse an angry crowd of five hundred outside the local KPU office who were protesting the outcome of the Bupati elections and accusing both the KPU and the winning pair of electoral fraud. In Poso district, Central Sulawesi, the district KPU office was bombed following a week of demonstrations protesting the election result. While the aggregate number of executive election-related violent incidents is not currently available, the level of violence, organization and numbers of people mobilized appear to be higher than was the case for the legislative contests.25 The lower intensity of conflict associated with the 2004 legislative elections reflected greater voter apathy as well as the reduced appeal of the traditionally strong parties such as PDI-P and Golkar. While the 83 percent voter turnout is high by international standards, it is lower than the 91 percent who turned out to vote in the first post-New Order elections of 1999. Moreover, the share of votes for all five of the largest parties from 1999 went down in the 2004 elections, highlighting dissatisfaction with incumbent parties. The legislative elections provided space for the emergence of non-traditional parties, reducing tensions among long-standing rivals. Indeed, the diffusion of support and power across a broader range of parties appears in the Indonesian experience to have lessoned tensions associated with political competition. The primary beneficiaries of this dissatisfaction with the performance of the largest parties were two relatively new parties: the Prosperous Justice Party (PKS) and Susilo Bambang Yudhoyono's Democrat Party (PD), both campaigning on platforms of clean government and anti-corruption. This pattern was evident in Maluku, for example, where voters' desire for change shaped by government mismanagement of the 1999-2002 conflicts was seized on by new parties such as the Prosperous Justice and Prosperous Peace parties (PKS and PDS respectively) who fielded candidates active in peace building and good governance initiatives. Both parties were also more successful in connecting with voters in new and creative ways using village or neighborhood dialogues and door-to-door canvassing, allowing for greater face-to-face contact and some discussion of issues important to voters. Unsurprisingly, these new parties made the greatest inroads in the elections. IV. Limiting Violence: Successful Conflict Prevention While the causes of conflict discussed above are important in determining levels of conflict, the degree and form of conflict prevention efforts are also key. Tensions associated with electoral contests are inevitable. Conflict prevention involves pre-emptive planning and action to avoid such tensions escalating into outbreaks of conflict. The impressive attention to conflict prevention at the national and local levels accounts in large part for the absence of violence during the legislative elections. A convergence of a broad range of initiatives and actors contributed to the success of conflict prevention efforts. 25See, for example, "Massa Bentrok dengan Polisi di Sumba Timur", Kompas, 5th July 2005 and "Kantor KPU Poso Dilempar Bom", Kompas, 13th July 2005. 20 Law No. 12/2003 on the General Elections set the framework for the peaceful conduct of the elections by setting the parameters of behavior.26 At the same time, the police, KPU and Ministry of Home Affairs provided general policy guidelines including codes of conduct and campaign regulations to be implemented by political parties. The police also instituted a policy to identify potential conflict "hot spots", and to allocate personnel accordingly. Government funding was allocated to ensure the successful implementation of these initiatives. Because of the significance of the 2004 elections for democratization in Indonesia--the elections were the first time citizens could vote directly for the President, as well as for individual candidates in the legislative elections--many NGOs throughout Indonesia formed partnerships and conducted extensive voter education as well as conflict prevention and resolution training to ensure successful elections that were free from violence.27 These measures were jointly implemented by local governments, the police, district KPUs, political parties, and local leaders. In all four provinces, the ultimate success of these conflict prevention measures depended on local knowledge, ownership and implementation, as well as the level and quality of cross-institutional coordination. The communities where these efforts were most successful saw broad cooperation between the state and ongoing community-driven peace-building efforts, particularly in high conflict areas. Overall, no single actor or policy dominated conflict prevention, although in individual provinces or districts specific actors or initiatives took the lead. Four factors collectively stand out as the most decisive in shaping conflict prevention efforts: the regulations governing campaigning; the police operation identifying potential `hotspots'; proactive local leadership; and effective cross-institutional collaboration. Together they account, in large part, for the relative lack of violence during the legislative elections. 1. Campaigning Guidelines and Implementation The campaign period was flagged as being the time when violent conflict was most likely. As such, the KPU required all parties to sign a code of conduct on February 6th, 2004 and introduced a plan to limit potential interaction among political supporters during the campaign period. Key points in the conduct agreement included a rule that party symbols should only be placed in areas designated by local government; acceptance of the campaign restrictions governing open rallies; the assumption of responsibility by parties for the orderliness of supporters during campaigns; and an agreement to refrain from slandering opponents or, in the case of Maluku, raising conflict-triggering issues such as advocating separatism.28 Campaigning was restricted to electoral areas (Daerah Pemilihan) and carefully scheduled. Parties were given specified dates to campaign in individual locations and the schedule was widely 26The law prohibited: (a) insulting a person, religion, ethnicity, race, group, other candidates and/or electoral participants; (b) provoking or leading individuals or community groups into disputes; (c) disturbing public order; (d) threatening to commit violence or advising the use of violence against a person, community group, and/or other electoral participants; (e) destroying and/or removing electoral participant's visual display items; and (f) using government, religious or educational facilities. 27For example, the international NGO Common Ground worked in six provinces (including Bali and East Java) covering 18 districts with 18 local NGOs on voter education and conflict transformation activities involving 9,900 people. Forum Interseksi, a Jakarta-based NGO, worked in four high conflict provinces (including Maluku) with four local NGOs on similar voter and conflict education activities. For further details, see Manoppo et al. (eds.) (2004) and Vermonte and Budiman (eds.) (2005). 28While separatist aspirations are not widely held in Maluku, the Maluku Independence Front (FKM) did express the grievances of many Christians during the 1999-2002 violence who felt that the government was not protecting Christians from attacks by Muslim militias. While the FKM is in general only a "talking shop" that does not pose any real threat, many locals believe that the military and military-backed militias talk up the threat it poses to justify their own existence. In the wake of the April 25 violence, the police arrested FKM supporters for belonging to a banned organization. However, they have not arrested any members of the "red and white" militias involved in attacking and burning local neighborhoods. 21 publicized. These measures ensured that groups of competing political supporters did not congregate in the same area or in high numbers, and limited the number of outsiders campaigning in an individual community. Campaigning was stopped a few days before the election--a regulation generally followed-- limiting potential escalation of tensions in the build-up to election day. These measures had a profound effect. In the 1999 legislative elections, when people voted for political parties and not for individual candidates, there was unlimited mobilization during campaigning. For example in East Java, truckloads of party cadres and supporters from Surabaya traveled all over the province to mobilize support for various parties. Unrestricted numbers and access resulted in conflict, particularly in Bali and East Java. By contrast, in 2004 the contest shifted to individual candidates competing in specific electoral areas with campaigning restricted to candidates' respective electoral areas. The scope for mass mobilization was as a result reduced, lowering the possibility of inter-party and inter- village clashes. Structurally, because parties or candidates often dominated specific electoral areas, fierce competition was limited.29 Campaign guidelines were most successful in conjunction with effective policing, local initiatives and cross-institutional cooperation. In localities where these factors were prevalent, campaign restrictions were extended. In Ambon, where memories of communal violence from 1999 were still fresh, party parades and processions through the main streets were banned on rally days. Instead, neutral meeting and dispersal points were designated collectively by the police, local government, political parties and the KPU. Routes to and from the rally area were chosen to avoid conflict-prone neighborhoods where provocation was likely. More importantly, rallies were limited to two per day--one in the morning and one in the afternoon--with a two-hour interval in between to enable supporters to disperse from an area without encountering supporters from the ensuing rally. As far as possible, with the cooperation of the political parties, rally timetables avoided scheduling two rival parties on the same day. The police worked closely with political parties and local NGOs, giving information on these guidelines before campaigning began, and enforcing them during the event. In contrast, the two-hour interval measure was not adopted in Maluku Tengah, resulting in inter-party clashes when supporters from concurrent rallies crossed paths. 2. Proactive Policing The police were one of the most important actors in preventing conflict. As the institution primarily responsible for internal security, the police were charged with ensuring non-violent elections and adopted a strategy that emphasized conflict prevention and the incorporation of community policing principles, reflecting ongoing security sector reforms.30 As with the effectiveness of campaign guidelines, successful policing was due to the initiative of local officers, supported by regional police leaders. In general, the efforts of the police were effective, although questions remain as to how they would have performed if tensions around the elections had been higher. The main conflict prevention tool employed by the police was the identification of election hotspots. National level guidelines instructed local police to classify electoral areas into three categories: sensitive (rawan), moderate risk (sedang) and normal (biasa). The sensitive category was further divided into areas of high (rawan 1) and middling risk (rawan 2 and 3). In Ambon, for example, all fifty villages were deemed to be `sensitive' but most were in the middling category. In East Java, the `sensitive' areas included the strongholds of crime networks, the incumbent District Head's residential area (relating back to the tensions at the time of his appointment), and areas where more than one party dominated with 29In Ambon, de facto ethnic segregation due to the communal conflict (1999­2002), which resulted in Muslim and Christian neighborhoods, also reduced the possibility of communal clashes during campaigning, because religious parties were active primarily in their respective strongholds. 30On community policing in Indonesia, see Asia Foundation (2005). 22 roughly equal support bases and a history of political violence. Even in areas with a history of lower levels of conflict (such as Sikka district in NTT), a majority of villages were designated as being `at risk'.31 Mapping of potential hot spots was based on local characteristics (such as village size and the presence of more political parties, especially rival parties), crime statistics, and local conflict dynamics (including a past history of election conflict). This mapping provided the basis for the allocation of policemen to polling stations in high risk areas.32 Hansip (civilian defense officers) supported the police, and were also allocated in response to perceived variations in conflict risk. The police were stationed at the polling stations two days before and after polling day. Subsequently, they moved to the stations at the village and the sub-district levels as vote counting got underway and vote tallies were aggregated. In all four provinces, although not in all districts, police mapping of hot spots, and deployment of personnel accordingly, proved to be an effective conflict prevention mechanism. Beyond the identification and deployment of personnel to potential "hot spots", the police initiated and engaged in a series of dialogues with political elites at the local level, from candidates to leaders of political satgas (paramilitary groups).33 Twenty-four political parties came to a consensus with the Indonesian National Police on ensuring the security for the 2004 elections on 29 December 2003 and formed the Political Candidates--National Police Central Headquarters Communication Forum with similar offshoots at the regional, municipal and regency levels. Each political party had three board members assigned to the forum. The forum's command post was situated in the National Police Headquarters and all matters pertaining to election security were directed to the command post. In many areas, Memorandum of Understandings were signed with parties, state institutions, and other influential bodies. In some cases, these dialogues were initiated three months before the campaign began allowing for confidence-building to foster cooperation and shared interest in a non-violent election. The effectiveness of these social capital-generating initiatives varied widely across the four provinces. In some localities, grass-root leaders developed the initiatives and sustained dialogue, while in others local officers took the lead. The quality of policing reflected many of the institutional strengths and weaknesses highlighted above. When police worked together with both election officials and the local actors, the efforts were most effective. Where police officers had broad and early interaction with communities, the initiatives were more successful. In Buleling, Bali, officers began a dialogue in December 2003, and increased the intensity to at least three a week during the campaign including candidates and village leaders as well as youth gang members. In Jembrana, another district in Bali, the sessions were less participatory, less frequent, and less inclusive; here, conflict prevention measures were less successful. In East Java, local leaders were reluctant to involve the police too deeply in addressing conflict, preferring instead to rely on the authority of local elites at the village level and village-based defense mechanisms in the first instance.34 This did not exclude information exchanges or coordination with the police. Another significant departure from the 1999 elections was the new relationship between the police and Panwaslu regulated by Election Law No. 12/2003. The Law states that Panwaslu should draw its members from the police, public prosecutors, academia, public figures and the press. This was done to redress Panwaslu's shortcomings in 1999, in particular the inadequate reporting of electoral crimes that made subsequent investigation difficult. It was held that police membership of Panwaslu would improve the quality of reporting, and familiarity with reports from Panwaslu would also enable the police to 31In Sikka district, 468 out of 684 TPS (local voting areas) were designated as such. 32The number of policemen allocated to "high risk" areas varied by province. For example, in Maluku, high risk areas had two policemen assigned to each polling station and middling risk areas had two policemen assigned to between three and five polling stations; in NTT in the areas at most risk, one policeman covered two polling stations. 33See Meliala (ed.) (2005). 34This was also the case in Poso, Central Sulawesi where community policing was interpreted as the community's own efforts in detecting and resolving escalating conflict without involving the police. For similar findings, see Manoppo et al. (eds.) (2004:25) and Vermonte and Budiman (eds.) (2005:175-176, 198). 23 investigate cases more quickly. This is an important consideration as investigators are given only 30 days to investigate before a report is dropped due to lack of evidence. A comparison of the 1999 and 2004 elections points to a significant improvement in the number of reports that resulted in prosecution. In 1999, over 200 reports were forwarded to the police and only five made it to the courts. For the 2004 elections, 2,392 reports were investigated by the police, 1,289 cases were forward to the public prosecutor and 1,066 cases were settled in court.35 In short, cooperation between the electoral bureaucracy and police strengthened conflict prevention and resolution at the local level. Professionalism at the local level was also significant. In NTT, there was a marked difference between Sikka, where mapping was based on 1999 election-related violence as well as non-election-related conflict and crime statistics dating back five years,36 and neighboring East Flores. In the latter, there was no hot spot mapping. The deputy police chief in East Flores did not appear to fully grasp the mapping concept, reflecting general complacency about potential conflict and a lack of data to form a sound basis for hot spot identification. Not only did the capacity of officers to implement the policy make a difference. Neutrality was also important. In Bali, officers in general performed exceptionally well, yet in the district of Tabanan the clear political loyalties of some policemen limited the scope and efficacy of conflict prevention, resulting in less effective police protection for potential challengers to the incumbent PDI-P candidates. In the districts studied in East Java, police officers worked very closely with both election officials and local elites to anticipate even the smallest conflict triggers, and conflict measures were usually implemented by the three sets of actors in coordination. Interviews indicate that the police performed best at deterrence where their presence and sheer force of numbers reduced the likelihood of physical clashes or attacks. The police were often seen dispersing political supporters congregating on street corners in an effort to reduce tensions. These efforts were supported by the dialogue sessions. Meetings with local leaders in East Java, for example, helped to reduce the number of people mobilized at the grassroots--in some cases, from 300 down to 50. The police were also deemed to have performed well in adapting to changing patterns of conflict, altering tactics of intervention according to the election cycle. The deployment of personnel in Bali as the campaign evolved illustrated continued responsiveness. The greatest success was tied to making greater use of community policing methods in all four provinces by building on local knowledge. The police helped to forge better relations among political rivals' parties in Bali, worked with clan leaders in NTT, and ensured there would be no mobilization along religious lines in Maluku, all in cooperation with local leaders and election officials. The election demonstrated improved and more proactive policing, but it also highlighted certain weaknesses. While police identified election conflict hot spots based on pre-existing tensions or past conflicts, there were few attempts to address the roots of these tensions. The focus was on ameliorating tensions in the short run. When clashes erupted, the police responded by restraining parties and stopping the violence, but they did not follow-up to see if the conflict had been resolved. This highlights a compartmentalized understanding of conflict prevention and management by the police, which emphasizes deterrence, restraint and maintaining order, but which stops short of conflict resolution or transformation where the roots of the problem are addressed. Such tasks were generally left to community leaders and other local actors. It should be noted, however, that there is ambivalence at the local level as to how deeply the police should be involved in conflict resolution and transformation, either because local leaders do not want the police usurping their authority or because the police are not trusted. In both cases, 35See Meliala (ed.) (2005: 23-25). 36Although it should be noted that the quality of crime and conflict statistics was fairly poor. 24 a better understanding of local contexts, and better relations with the community, would enable the police to work with local leaders, in the process increasing trust between communities and the police. This compartmentalized understanding of the policing role carried over into election-specific conflict management. For example, in Maluku Tengah police maintained the peace but failed to intervene when polling officials only meters away from policemen on duty blatantly told voters how to punch their ballot papers in clear violation of the electoral law. When questioned, the police replied that Panwaslu should report the violation first before they could investigate. Given that Panwaslu did not have sufficient personnel to cover all polling stations, there is clearly a need for a better, and more realistic, division of labor between Panwaslu and the police, who did have the manpower. If this compartmentalized mind-set does not change, the public will witness the police standing by when blatant violations occur, with the result that trust and confidence in the police force will be further undermined. The central role of the police in the 2004 elections contrasts sharply with the 1999 elections, where the military played the dominant role. The military showed greater professionalism than in the past, displaying neutrality, and worries about military-police tensions proved to be largely unfounded. In Bali, East Java and NTT, the military maintained neutrality and often cooperated with the police in forwarding information for conflict prevention purposes. This level of cooperation was shaped by the relationship between local military commanders and police officers. Conditions were different in the high conflict area of Maluku. Although the military provided much-needed logistical assistance by transporting election logistics to remote areas so that the legislative elections could go ahead, one military unit disrupted proceedings on polling day pointing to a lapse of judgment at best and the resurfacing of military-police rivalries at worst. 3. Local Leaders and Grassroots Initiatives In all four provinces, local leaders were at the forefront of conflict prevention and grassroots-led initiatives were significant. Religious leaders, village headmen, politicians, government officials, NGOs and everyday citizens all showed leadership in different communities by openly calling for a peaceful election, by discouraging others from using the election period to settle scores, and through facilitating communication among groups in society, often cooperating with security personnel. Many introduced grassroots initiatives based on local knowledge to limit conflict and, especially, violence. The quality and level of local intervention shaped the level, form and timing of election-related conflict. One important grassroots initiative involved early warning systems and intervention. On Ambon island, Maluku, the quick resolution of inter-village clashes has become a way of life after the heavy toll of the 1999-2002 conflicts. Early warning systems operate in all villages and are particularly alert to rumors that circulate as well as the presence of strangers in the village. In the case of rumors, there are established fact-checkers who are well positioned to verify or refute the information.37 Strangers are monitored and if their activities raise suspicion, the local police are contacted. It has become regular practice for Camat (Sub-district Heads) to meet all Village Heads once a month to discuss problems and share information. During campaigning, this was increased to weekly meetings. The ongoing nature of conflict prevention contributed to their success. In Situbondo, East Java, one of the biggest conflict hotspots in the province, a family of influential Kyai (Muslim leaders) were split three ways over the support for and alliances with three parties: the National Awakening Party (PKB), the United Development Party (PPP), and the Star and Crescent Party (PBB). This could have led to tensions and accusations of betrayal amongst previously united followers of the 37Fact-checkers have cross-communal networks that extend throughout Ambon island and beyond so that they can verify rumors or "warnings" about the activities of other villages or communities. 25 Kyai. These leaders, however, prevented this by making a joint public statement guaranteeing peaceful elections and undertaking not to mobilize their followers during the campaign. In Maluku, as a reminder to all parties as well as citizens that peace-building and reconciliation transcended political competition, the first day of campaigning kicked off throughout the province with a symbolic procession on foot of party leaders and local government through the main town centers ending with the joint display of party flags in agreed locations. In parts of Bali, local political elites sat down with party satgas and preman (criminal) groups, to request collegiality between rival organizations during the campaign period, and encouraged these organizations to work toward guaranteeing a safe campaign. From public statements and processions to low-profile meetings, local interventions proved critical. The more effective conflict prevention measures had a ground swell of local support in communities and were often initiated "from below." The breadth of community-led efforts in "hot" zones was impressive in three of the provinces studied: Bali, East Java and Maluku.38 The mechanisms adopted were context- specific, catered to the varied factors shaping the pre-existing conflicts, and were often implemented by local elites. Whether it was early warning systems managed by village leaders in all three provinces, on- going community-driven peace-building efforts and citizens attuned to conflict dynamics in Maluku, or attempts by local religious and political leaders to reduce tensions among political groups in East Java or Bali, conflicts in these hot spots were ameliorated by grassroots intervention. NGO alliances in the three provinces reinforced many of these grassroots initiatives by including conflict mapping and analysis in conflict transformation activities that raised the standards of existing village-based early warning systems, produced more local facilitators, and increased public awareness of the links between political competition and violence.39 NGOs not only conducted voter education sessions but also stressed the structural causes of violent conflict that could only be addressed by effective government and urged citizens to use their vote wisely. In high conflict areas, NGOs were also more likely than local government or district KPUs to target groups vulnerable to manipulation or coercion such as IDPs, first- time voters, women or people with minimal formal education. More often than not, the higher the level of previous violence, the more extensive and developed local conflict prevention efforts tended to be. This was underpinned by populations tired of violence, as was the case in Maluku, or communities opposed to crime committed by groups such as bajingan in East Java. In Maluku, the clear message emanating from the population was that no election was worth fighting over and political candidates harking back to the conflict of 1999 did themselves no favors. One of the hard- line Muslim leaders from the 1999 conflict and head of the local Star and Crescent Party (PBB) advocated segregation of communities along religious lines during his legislative campaign and failed to win votes.40 4. Institutional Coordination and Cooperation Effective conflict prevention also rested on cross-institutional collaboration and relations between state institutions and societal actors. Local leaders, including NGOs, and police officers worked in tandem with the district KPUs, local government officials and political representatives. Early coordination resulted in more successful conflict prevention measures such as the inter-party code of conduct. In Maluku, this 38Significantly, these three provinces also had extensive NGO activities devoted to voter education and conflict management compared to NTT, where there was a more limited NGO presence. 39See Manoppo et al. (eds.) (2004); Vermonte and Budiman (eds.) (2005). 40Of equal significance was the failure of the United Development Party (PPP), the party that had hitherto captured the majority of Muslim votes in Maluku, to maintain its position as the third largest party in Maluku behind Golkar and PDI-P. Instead, Muslim voters, particularly urban ones, were won over by the Prosperous Justice Party (PKS), a Muslim reformasi party formed in 1999. PKS ran on an anti-corruption platform that appealed to voters and downplayed religious issues. In contrast, PPP was deemed to be out of touch with the voters and lacking in new ideas. 26 KPU-sponsored inter-party agreement was socialized only a week before campaigning with mixed levels of success--socialization was not extensive enough in reaching down to the grassroots and good conduct pledges were framed as internal agreements between parties rather than binding pledges with sanctions for violations. This stood in contrast to parts of Bali, where the code of conduct was introduced a month before the campaign, with measures developed even before the KPU in Jakarta signed off on the final agreement. The greater trust and social capital in Bali fostered better coordination and cooperation, and thus had greater impact in reducing conflict. The importance of cooperation with a wide range of actors was evident in all four provinces, whether it was ties between the district KPUs and Kyai in East Java, local leaders and police in Maluku and Bali, or clan (adat) leaders and local government officials in NTT. Election-related violence was less likely in areas where there were higher levels of cross-institutional coordination and cooperation, and NGOs played an important role in organizing focus group discussions, seminars and training workshops that brought together a wide range of relevant institutions and actors. In NTT, there were more tensions as a result of poor cooperation among security agencies and election officials, with the exception of Sikka district, where better cooperation prevented conflict. In contrast, the Bupati's (District Head's) dominance in East Flores in NTT and Tabanan in Bali crowded out other actors, limited effective cooperation, and curbed the impact of joint-initiatives weakening collaborative efforts in conflict prevention. A successful and innovative example of cross-institutional coordination and collaboration occurred in Maluku with the establishment of an elections crisis center in the local police station of each electoral area, manned by party managers from each political party, the local KPU and Panwaslu. The crisis centers were designed to respond quickly to local disputes that could escalate if not nipped in the bud by capitalizing on the influence of local party managers in restraining their supporters and the roles of the KPU and Panwaslu in resolving disputes in accordance with election laws and regulations. They demonstrated the growing use of community policing measures by the local police--i.e., involving a greater range of local institutions and actors in dispute resolution rather than relying on repressive methods. The crisis centers started functioning a week before campaigning and continued until a week after polling day; members were on call 24 hours a day. The role of coordination and collaboration points to the increasing institutional breadth necessary to prevent conflict in local areas and the need for effective cooperation between security personnel and other agencies. The strengthening of a range of local institutions is necessary. The evidence also shows that no one actor can dominate conflict management. Success rests on a variety of ongoing collaborative strategies that draw on the strengths of different institutions and actors. V. Managing and Resolving Conflict: Mixed Success The issue of institutional weaknesses was especially pertinent with regard to the management of conflict--intervention after conflict occurred. While the prevention of conflict was indeed a success story of the legislative elections, the management of tensions and disputes when they broke out was less so. For the most part, potentially violent clashes were resolved peacefully. Yet, non-violent disputes, especially those involving conflict with the state (the majority of conflict incidents), continued to simmer. The experience of the four provinces during the legislative elections points to important lessons for longer-term conflict resolution. Prevention measures and management efforts did not sufficiently address the sources of conflict in communities, with the focus on ameliorating tensions rather than addressing the root causes of conflict. Tensions were only dampened temporarily and remain a reservoir for conflict in the future. 27 For example, in a "hot" district in Bali, Buleling, which had a history of conflict in the 1965 transition and the 1999 election, and where two election-related deaths occurred in October 2003, conflict prevention efforts were extensive, extending from politicians and election officials to village headmen and businessmen in private meetings and public forums. However, tensions between PDI-P and Golkar supporters remain very high in some communities, where revenge cycles had begun years ago and emotions remain strong. The violence in Ambon that erupted on April 25th also showed that if underlying conflicts are allowed to simmer or issues remain unaddressed, then areas of high conflict may fall back into violence even with extensive grassroots peace-building efforts. The violence in Ambon illustrates that unresolved legacies of the 1999-2002 conflict, such as the presence of IDPs, the ready availability of weapons, the presence of disgruntled former militia leaders who see conflict as a way of regaining power and wealth, as well as lingering inter-communal suspicions that have never been laid to rest, can easily lead to conflict reigniting. In large part, this is because the government has refused to reveal the results of the independent fact-finding team commissioned to investigate the violence and has failed to arrest the perpetrators of violence. In such a context, the best grassroots peace-building efforts would be severely taxed. In all four provinces, tensions within parties grew over candidates' positions on the party list after the results were announced. Without more comprehensive interventions addressing inter- and intra-party rivalries, these tensions could easily erupt into significant conflict in the future. 1. Local Intervention Our study reinforces the critical role of local leaders in conflict management during the legislative elections. Local leaders took the initiative to intervene when physical clashes occurred often before the police arrived, and led in peace-making overtures after the police had intervened. The pivotal role that village, religious and grassroots leaders play in communities, and the trust and social capital that is often associated with their positions, highlights the significance of building the capacity of local leaders for long-term conflict management and resolution. In Ambon, for example, a quick response by the mayor prevented major violent conflict from occurring on polling day itself. The local military logistics unit had been embroiled in a protracted land ownership dispute with a local community. Although the court had ruled that no action should be taken by either party pending the outcome, and despite the fact that the town council was about to mediate between both parties on April 7th, the commanding officer of the logistics unit sent a unit of his men to destroy some of the houses while the counting of ballots was underway. Immediately, the local community protested and the soldiers who were carrying guns as well as tools to destroy the houses threatened them and called the protesting residents "separatist loyalists" who would be "dealt with".41 Scores of outsiders were gathered and tensions escalated before the mayor intervened. The mayor called on the district military commander to restrain the logistics unit and cease further dismantling of the houses whilst the municipal police chief dispersed the crowd; the fact that the mayor had excellent working relations with both officers guaranteed a quick response and effective coordination by the municipal police and district-based battalion. He then instructed polling officials to stop the counting process until the crowd had dispersed, saying that he would answer to both the district KPU and Panwaslu for intervening. The crisis was averted and counting resumed later that day. Conflict management typically and regularly occurs at the grassroots where timely intervention is crucial and local knowledge or the authority of community members is decisive, and the elections were no exception. When local youths who were Golkar supporters got into a fight with PDI-P supporters in Maluku Tengah, it was the women of the village who intervened and dragged them away by their ears 41Labeling the protesting residents as "separatist loyalists" could easily have incited bystanders to attack them because separatism is an emotive issue and frequently used to mobilize vigilante groups. This would have diverted attention away from the soldiers whose actions were a clear violation of the court injunction. 28 exclaiming, "You fight today, but tomorrow outsiders will come and destroy our village." It is unlikely that the youths would have responded in such a docile manner to outside intervention. When the car of a Christian passing through a Muslim village in Maluku Tengah was stoned, the Village Head immediately rounded up the culprits and marched them off to the Christian village to apologize. What distinguishes these examples from the prevention efforts discussed earlier was the willingness of individuals to intervene and assume responsibility for their actions. The mayor realized that the police would be reluctant to intervene in a dispute involving the military, and would be slow to act, so he took the initiative, leveraging his excellent relations with both the district police and military command. The similar initiative taken in response to the stoning of the car and the clash between youths shows that effective conflict management involves local ownership and a deeper appreciation of the early need to address non-election-related disputes. Where non-election-related issues are not addressed, there is the potential that underlying cleavages can spillover making election-related conflict more likely. 2. Limited Police Role in Conflict Management In all four areas, the police were effective in responding to violations of the election laws. The police were comparatively less successful in addressing the underlying tensions of communities. The legislative elections point to the profound and positive impact of the ongoing reforms in the security sector. The most important reform involved handing over responsibility for internal security to the police, thereby reducing the power of the military (TNI). Internally, the police has focused on improving its professionalism in its relations with society, through changing force composition so that it better reflects and responds to local conditions. Areas which experienced a greater degree of reform, resulting in larger numbers of more professional personnel on the ground and more in-depth local policing, were more likely to have functioning conflict management mechanisms in place and were less prone to violence. The two provinces that experienced the most extensive security reforms were Bali and Maluku. In Bali, the measures were widespread, resulting in rapid transformation of the composition of the police force and greater professionalism, some of which was the result of measures introduced after the tragic 2002 Bali bombing. The police played a major role in conflict prevention, often leading local efforts. In the district most successful in reducing conflict, Buleling, security reforms were especially important in leading to neutrality on the part of the military, which had not been the case in the 1999 elections. Reforms also enhanced the interaction of the police with community groups, extending from NGOs to local elites. In Maluku, increased local recruitment and adoption of community policing methods, such as working with a broad range of community leaders, have formed the backbone of police reform. However a heavy military presence prevails because of the communal violence that raged from 1999-2002. The professionalism of the police was pivotal throughout the campaign period, polling day and beyond until April 25th when police mishandling of a flag-raising ceremony by a small group supporting independence for Maluku led to renewed violence in Ambon. This incident raises several questions about the depth and extent of police reforms in Maluku. First, the triggering incident boiled down to ineffective crowd control by the police and questions just how ready the police is to deal competently with a mobilized and angry crowd (as opposed to giving in to them or shooting at them). This underscores the point made previously that the police are better at deterrence than conflict management. Second, questions about police neutrality, a legacy from the 1999 conflict, remain unresolved and could stigmatize local recruitment efforts unless compromised personnel are disciplined and removed because police recruited locally are 29 more likely to be suspected of being partisan.42 Last but not least, the continued presence of military personnel may stymie and undermine policing approaches to internal security in Maluku. In the run-up to April 25th, the police chief emphasized that the flag-raisers were not a dangerous separatist movement and favored a persuasive approach. The military chief by comparison painted them as enemies of the state that had to be exterminated and appeared to condone the formation of nationalist ("red and white") militias that contributed to the violence of April 25th. In NTT and East Java the security reforms were not as apparent, although TNI neutrality and community policing measures did have an impact in reducing violence. Yet while police reforms led to effective conflict prevention, the police were less effective at longer-term conflict transformation or resolution. In part, this had to do with the nature of the police security operation, which focused on deterrence and was less concerned with addressing the intensity of tensions. The relatively recent development of community policing and mixed standards of professionalism of individual officers was also a factor, since some officers were not familiar with the underlying tensions in specific communities or were unwilling/unable to intervene. For the most part, the police saw their role in conflict management as limited to keeping tensions from violently erupting and felt that non-violent disputes were beyond their purview. This was reinforced by the compartmentalization of conflict management, which narrowly defined responsibilities for disputes. Officers in Bali, for example, were careful to distinguish conflict according to the actor responsible for management, and at times dismissed issues if they did not pertain to their specific responsibilities. There were exceptions to this overall trend. In Denpasar, Bali, for example, police officers opted to implement and support a workshop that brought together local rival political supporters after a heated non-violent dispute, with a quiet airing of grievances the main focus of the forum. The police mediator role was an important step toward long-term conflict management. It should be noted, however, that not all communities view the police as best suited to long-term conflict transformation or resolution. Where communities are still grappling with the legacy of communal violence, and where questions remain regarding police neutrality, communities prefer to use their own leaders, build bridges with other communities themselves and focus on educating their youth against violence. They feel that the police should focus their energies on law enforcement. In high conflict areas, there was the additional suspicion that security forces benefit if the threat of violent conflict is sustained because of the official and unofficial financial incentives that accrue from being assigned to these areas. Some members of the community suspect that the security forces, including the police, are not interested in conflict resolution and may even be involved in the incidents of violence that justify their presence in larger than usual numbers. In Poso, Central Sulawesi, for example, local NGOs and local leaders worked together to ensure a successful and peaceful election that did not involve the police in their early warning system. In Maluku, concerns about police neutrality reduced the likelihood of the police being called to intervene in certain areas when conflicts occurred. Other assessments of police performance in securing the 2004 elections on a national level highlight similar strengths and weaknesses. At a conference organized by Partnership for Governance that presented findings from the Deputy Police Chief, the KPU, the Committee of Election Observers, NGOs, academics and candidates from the legislative elections, there was broad agreement that the police had 42Many locals believe that the police mishandled the situation because they sympathized with the people involved in the flag-raising ceremony and allowed them and their supporters to march rowdily through town instead of dealing firmly with them. 30 improved considerably on their 1999 performance in handling complaints and legal violations.43 The largest number of complaints and violations were logged during the legislative elections and the number dropped considerably during the first and second round of presidential elections due to political parties realizing that attempts to bribe, dupe or coerce voters to vote a certain way had not been successful. By October 2004, 90 percent of all complaints about elections violations had been investigated. The police were also seen as successful in broadening cooperation with other relevant social actors to reduce conflict, especially political party paramilitary units. However NGOs felt that the police defined their responsibilities too narrowly, waiting for violations to filter through the local Panwaslu and KPU before investigating, thereby limiting their conflict management role. Our findings indicate that successful conflict prevention occurs when police consult and work well with a broad range of actors, ranging from state institutions such as local government and the military to village heads, community, religious and youth leaders, community-based organizations and non-governmental organizations. However, ongoing police efforts need to focus on three broad areas. First, police need to do more than respond to the symptoms of a conflict. They need to work with local leaders to transform underlying conflict dynamics which is the foundation of community policing and successful long-term conflict prevention. Similarly, the police need to redefine their role away from a compartmentalized one focused solely on maintaining order, and must be more pro-active in acting on a broader range of criminal violations. Finally, being effective deterrents through sheer force of numbers begs the question of how effective the police would be if challenged in cases where conflict escalates. Focusing energies on effective crowd management techniques cannot be over-emphasized. 3. Varied Strength of Election Officials Compared with the actors outlined above, election officials were the least effective in managing tensions, as is evidenced by the fact that the highest number of conflicts were directed against the electoral bureaucracy. The dominant form of election conflict involved tensions between citizens and the state, over issues such as dissatisfaction with discrepancies in vote tallies between the district KPU electoral counts and those of the political parties, as well as disagreements between parties and electoral officials over which candidate from the party list should be chosen in the absence of clear winners. In Maluku, tensions centered on inadequate preparation--ballot papers were not delivered on time to outlying polling stations, were delivered to the wrong electoral districts, or left out the names of candidates--as well as the lengthy vote-counting process, a result of inadequately trained polling staff and party witnesses, as well as the absence of reporting and tabulation forms. In such cases, inadequate preparation stoked tensions. On the whole, these conflicts could have been managed better, especially in Maluku and NTT. In contrast to the campaign period where all parties were committed to the successful and peaceful implementation of the elections and pulled together, interests diverged during the counting phase of the elections where selective non-neutral alliances emerged. There were cases where local government, district KPUs and dominant political parties were seen to be colluding to determine outcomes. This appeared to occur in Jembrana in Bali, where the district KPU official was subsequently replaced after the election. In Maluku Tengah, polling staff or civil servants told voters how to vote in some polling booths and marked the ballot papers of some citizens; in other stations, polling staff were threatened by individuals promoting a particular party in clear sight of the police. In these cases, a climate of mutual suspicion and recrimination prevailed and neither local government, the local KPU, Panwaslu or the police were deemed sufficiently neutral or professional to be effective in managing election-related conflict over outcomes. This led to protracted disputes that eventually had to be settled by the Constitutional Court in Jakarta. The 43The conference on October 20th, 2004 at the Hotel Grand Mahakam in Jakarta concluded a series of election- related activities supported by the Partnership for Governance and the National Police. Police participation in these elections-related activities is part of a larger police reform program organized by the Partnership for Governance. 31 Constitutional Court was, for the first time after the General Elections, required to adjudicate disputes arising from the announcement of the final election results. Out of the 300 complaints received, approximately half were rejected or dismissed on the grounds of lack of evidence. Large variations, however, occurred both within and across provinces in terms of responsiveness to problems and complaints during the election process. Conflicts between the community and the state over the electoral process may have had the greatest frequency as a proportion of the conflicts in East Java. Yet the total number of cases (39) across seven districts with a population 5.6 million, was relatively insignificant, with many being small incidents of electoral violations (i.e. fake candidate qualifications) which did not become violent, and which were reported by the district Panwaslu to the police who had jurisdiction over the issue. While the district KPU was relatively responsive in East Java and Bali, with respondents highlighting how quickly problems were solved with the decentralization of the process, the district KPUs in Maluku were deemed slow or unwilling to respond to problems and biased in terms of resolution outcomes. Significantly, there were variations. In Ambon city, where there was greater socialization of the elections due to NGO activity, an alert local press, more monitors and a higher level of awareness amongst the general population compared to other districts. Both the local KPU and party witnesses demonstrated quick thinking and were able to resolve problems on the spot. For example, when the wrong electoral district ballot papers were delivered, the Ambon KPU instructed polling staff to count the votes as party votes if witnesses agreed and the electoral process was not disrupted. In other districts, polling staff and witnesses did not know what to do, and this often led to disruptions and demands for new polls to be conducted. When votes from polling stations were aggregated at the sub-district level, and vote tallies were disputed by political parties, polling staff in all Maluku Tengah sub-districts did not know how to deal with questions and criticisms from party representatives. They failed to distribute reporting, tabulation or complaint forms. Instead they pretended to register complaints but did not take them to higher levels and were often intimidated by witnesses. There was suspicion as well over the neutrality of the KPU, Panwaslu and polling staff in certain districts. In Maluku Tenggara, the local KPU head formerly headed Golkar's Maluku Tenggara branch and the Panwaslu head was the cousin of the current Golkar regional head. The latter was subsequently removed for failing to act on complaints against Golkar during the campaigning period. When an entire village voted for PDI-P, the Maluku Tenggara KPU rejected the result stating it was a statistical impossibility, but PDI-P members detected bias. In addition, complaints to the local KPU that military personnel stationed in villages (Babinsa) were coercing villagers to vote for Golkar were not acted upon. In general, the role of Panwaslu in all provinces was limited in that it did not have the authority under the election laws to resolve problems, only to pass on administrative irregularities to the local KPU and criminal offences to the police. In Sikka in NTT, Panwaslu was left powerless after its request for a recount in some voting areas--a result of widespread accusations of corruption in the election counting process--was rejected by the local KPU. There are no clear provisions in the electoral framework on how to contest KPU decisions. The election laws also provided for a heavy burden of proof of election anomalies, which also impacted on the success of Panwaslu, which generally took charge of collecting evidence. The lack of witnesses willing to come forward, who were under no legal obligation to do so, was a problem, especially in high conflict areas. The interrelationship between the KPU and Panwaslu, and the respective authority of these agencies, was also a problem. KPU's proposal to limit Panwaslu's role and authority was hotly debated at the national level, but appeared to be of less interest in the provinces. KPU at the provincial level in East Java supported an amendment of the Panwaslu mandate, as it would be in accordance with the intention of establishing an election supervising body, and stated that it was Panwaslu which had been formed to 32 oversee the elections, not KPU. In the legislative elections, Panwaslu was criticized by KPU for focusing too much on disputes involving the KPU and neglecting those involving political parties. Panwaslu, on the other hand, expressed concern that KPU was not accountable to anyone. Court cases dealing with election disputes where KPU was accused of manipulating results were mentioned to illustrate the need for an independent election body with a mandate to question the work of KPU. However, the proposal was not pursued in the following presidential elections and the Panwaslu mandate remained the same.44 The problematic dynamic in the electoral bureaucracy remains outstanding and limits the ability of Indonesian citizens to effectively channel conflicts and the government to address complaints. It is important to note, however, that most of the limitations in conflict management by the election bureaucracy were present at the local rather than the provincial level, where there was a greater level of professionalism. In Bali, two KPU district chiefs were replaced due to effective oversight by provincial officials. The provincial KPU in East Java pointed out, however, that their role and responsibilities--and the consequent difficulties they faced--were significantly smaller when compared to that of the district KPUs. The provincial KPU had more of a coordinating and oversight function compared with the district, which was responsible for logistics and implementing the elections. The legislative elections show that citizens were eager to channel conflict non-violently. Protesting citizens largely pursued their grievances through legitimate channels according to set procedures rather than mobilizing supporters and resorting to violence. This shows positive signs of democratic consolidation. Yet it is clear that there is also a need for capacity-building, increased professionalism and broader social engagement on the part of state actors. VI. Reflections: Future Democratic Consolidation Overall, the legislative elections were peaceful. In the four provinces studied there were a limited number of election-related conflicts. Of these, just over 10 percent were violent. This is an impressive achievement. The forms of election-related conflict varied from rivalries between parties, within parties, and among individuals. However, the majority of conflicts were directed against the state. Of the four provinces, Bali had the largest number of conflicts, while Maluku had the largest number of violent incidents. The conflicts occurred in three spikes: at the beginning of the campaign period, during election week, and in the week that the provisional results were announced. In general, most disputes remained small and non-violent. The study shows that the low level of election-related conflict, especially violent conflict, was the product of a combination of, primarily the lack of interaction between pre-existing forms of conflict in localities and the electoral process, as well as a wide range of effective conflict prevention measures. Local elites and grassroots initiatives, especially in high conflict areas, played a pivotal role in minimizing tensions. In general, there were lower levels of emotion attachment to the outcomes of the 2004 legislative elections, reflecting growing public awareness that the elections are not zero-sum games. In high conflict areas such as Maluku, the elections were welcomed as an opportunity for change and a positive demonstration that peace and stability had returned to Maluku after the horrific 1999-2002 conflict. However in the other three provinces, incumbents were by and large returned to office leaving the existing power structures intact.45 Yet while the elections were generally positive, they do highlight institutional weaknesses that could undermine Indonesia's institutionalization of democratic elections and potentially leave room for violence 44Sissener (2004:32). 45On the strategies elites have employed to maintain power in the reformasi era, see Hadiz (2003) and Robison and Hadiz (2004). 33 to escalate in the future. In particular, the study shows the need for increased attention to be directed to three broad areas. First, the cases highlight the need to rethink the way in which we conceptualize conflict and conflict resolution. Conflict in Indonesia is not just an episodic, event-driven occurrence, but is a symptom of broader problems, cleavages and interests. The study found that one of the major factors underpinning whether election-related conflict was likely to arise and/or escalate was the history of conflict in local communities. The history and patterns of conflict in local communities shaped who the groups were, how likely they were to be in tension with each other, and the motivations for leaders to seek to mobilize groups in conflictual ways for political gain. While conflict prevention was by and large successful in reducing violence around the elections, in many cases the measures tended to address symptoms rather than causes; the deterrence-oriented conflict prevention measures did not take sufficient account of broader pre-existing conflicts and clashes within and between communities. The result was that while in general violence did not erupt, unresolved conflicts during the electoral period may potentially feed into existing tensions and lead to future outbreaks of conflict. The most serious incident of violence that occurred during this study, that which took place on April 25th in Ambon, shows that insufficient attention to the roots of conflict can contribute to further cycles of violence. Addressing violent conflict requires not only deeper strategies, but also the realization that election-specific measures must be embedded in longer term, ongoing processes of conflict prevention and management that cannot be merely event-driven. Second, these deeper, ongoing reduction measures can only be undertaken by strengthening institutions to allow for the positive channeling of tensions. One of the major findings of this study was that the performance of institutions in Indonesia (both those focused specifically on managing the elections, and institutions of the state, more broadly) was mixed: capacity, professionalism and constructive societal engagement were lacking in many localities. These institutional shortcomings resulted in high levels of protest and, in some cases, attacks against state institutions. At a macro level, this points to the potential for a crisis of legitimacy in the long run. At the micro level, the research demonstrated weaknesses in specific state institutions that varied across localities. The new electoral bureaucracy successfully administered the legislative elections and should be commended. However, weaknesses were clearly present as evident in the number of complaints about poorly trained KPU and polling officials, corruption and bias. For the most part, the lack of time and resources to prepare adequately for the elections in terms of logistics as well as training were the main problems and require further attention from the national KPU as well as the state budget. Greater oversight of the local KPUs by provincial KPUs would also help address issues of bias and corruption. No amount of oversight, however, can substitute for the transparent recruitment of neutral local KPU and polling officials. KPU officials also need to address complaints in a timely fashion. It is not a coincidence that most of attacks against state institutions were focused on the KPU. The fact that the KPU itself is not accountable to any other authority is a significant structural weakness in the current electoral bureaucracy. The Panwaslu was, for the most part, a powerless, under-resourced monitor that was unable to respond to alleged irregularities and other complaints. Combined, the weaknesses in addressing complaints on the part of the KPU and Panwaslu meant that candidates and parties were forced to take recourse in the expensive and lengthy procedures of the Constitutional Court. In some cases, tensions continued to simmer for some time; in others they persist. In addition to electoral institutional strengthening, the elections law, particularly the partial open-list system for determining elected candidates, needs to be revisited. In the interest of promoting candidates who are beholden to their constituents rather than party leaders, the vote quota should either be lowered to realistic levels or abolished altogether in favor of a district system which would also be easier to understand and implement. 34 The 2004 legislative elections were the first time that the police were solely in charge of internal security and they successfully adopted conflict prevention measures that resulted in peaceful elections. However they performed best at deterrence when their presence by sheer force of numbers reduced the likelihood of clashes. When clashes occurred, police intervention focused on restraining conflicting parties but stopped short of addressing the roots of the conflict. This study emphasizes that successful conflict reduction occurs when police adopt community policing approaches which involve consulting and working with a broad range of actors from local government and the military to community actors and community-based organizations. This underscores the need to deepen and broaden police reform. By adopting a supporting role and remaining neutral, the military has contributed positively to the ongoing security reforms required to consolidate democracy in Indonesia. Finally, the study's most significant finding was the pivotal role local actors and institutions played in preventing and managing conflict. While national agencies provided important policy guidelines, the success of conflict reduction measures depended on local knowledge, ownership and implementation at the regional levels of state institutions, the quality of cross-institutional coordination and cooperation, and the role of other local actors. The importance of local knowledge and ownership underscores the need to recognize and support community-driven peace-building efforts where they are successful and build capacity at this grassroots level where efforts are less developed. 35 References Acciaioli, Greg. 2001. "Ground of Conflict, Idioms of Harmony: Custom, Religion, and Nationalism in Violence Avoidance in Lindu Plain, Central Sulawesi." Indonesia 72 (October): 81-114. Amalinda, Savirani. 2005. "The New Local Strongmen in the Indonesian Decemtralization: The Case of Bantul District, Yogyakarta," Paper presented at the University of Indonesia 4th International Symposium of Journal Anthropologi Indonesia, Depok, 12th-15th July. Anderson, Benedict R. O. G. (ed.). 2001. Violence and the State in Suharto's Indonesia. Ithaca, N.Y.: Cornell University Press. Andrianto, Tuhana Taufiq. 2000. Konflik Maluku. Yogyakarta: Gama Global Media. Anwar, Dewi Fortuna, Helene Bouvier, Glenn Smith, and Roger Tol (eds.) 2005. Violent Internal Conflicts in Asia Pacific: Histories, Political Economies and Policies. Jakarta: KITLV. Aragon, Lorraine V. 2001. "Communal Violence in Poso, Central Sulawesi: Where People East Fish and Fish Eat People." Indonesia 72 (October): 45-74. Ananta, Aris, Evi Nurvidya Arifin and Leo Suryadinata. 2005. Emerging Democracy in Indonesia. Singapore: Institute of Southeast Asian Studies. The Asia Foundation (2001). Survey Report on Citizen's Perceptions of the Indonesian Justice Sector. http://www.asiafoundation.org/pdf/IndoLaw.pdf ------. 2005. Community Policing in Indonesia. http://www.asiafoundation.org/pdf/Indo_Community_Policing_eng.pdf Aspinall, Edward and Greg Fealy (eds.) 2003. Local Power and Politics in Indonesia: Decentralisation and Democratisation. Singapore: Institute of Southeast Asian Studies. Barker, Joshua. 1998. "State of Fear: Controlling the Criminal Contagion in Suharto's New Order." Indonesia 66 (October): 7-42. Barron, Patrick, Samuel Clark, and Muslahuddin Daud. 2005. Conflict and Recovery in Aceh: An Assessment of Conflict Dynamics and Options for Supporting the Peace Process. Jakarta: World Bank. Barron, Patrick, Kai Kaiser, and Menno Pradhan. 2004. "Local Conflict in Indonesia: Measuring Incidence and Identifying Patterns." World Bank Policy Research Working Paper 3384. Washington, D.C.: World Bank. Barron, Patrick and David Madden. 2004. "Violence and Conflict Resolution in `Non-Conflict' Regions: The Case of Lampung, Indonesia." Indonesian Social Development Paper 2. Jakarta: World Bank. 36 Barron, Patrick and Joanne Sharpe. 2005. "Counting Conflicts: Using Newspaper Reports to Understand Violence in Indonesia." Conflict Prevention and Reconstruction Paper 25. Washington, D.C.: World Bank. Barron, Patrick, Claire Q. Smith and Michael Woolcock. 2004. "Understanding Local Level Conflict in Developing Countries: Theory, Evidence and Implications from Indonesia." Conflict Prevention and Reconstruction Working Paper 19. Washington, D.C.: World Bank. Bell, Gary F. 2003. "Indonesia: The New Regional Autonomy Laws, Two Years Later," in Daljit Singh and Chin Kin Wah (eds.), Southeast Asian Affairs 2003. Singapore: ISEAS: 117-131. Bertrand, Jacques. 2004. Nationalism and Ethnic Violence in Indonesia. Cambridge: Cambridge University Press. Bjork, Christopher. 2003. "Local Responses to Decentralization Policy in Indonesia." Comparative Education Review 47(2) (May): 184-216. Brown, Michael. 2002. "Ethnic Conflict and Internal Conflicts: Causes and Implications," in Chester Crocker, Fen Osler Hampson and Pamela Aall (eds.), Turbulent Peace: The Challenges of Managing International Conflict. Washington, D.C.: United States Institute for Peace. Burton, John W. 1990. Conflict: Resolution and Prevention. New York: St Martin's Press. Chanintira na Thalang. 2005. "The Presidential Elections in Indonesia, July and September 2004," Electoral Studies 24(4) (December): 755-761. Clark, Samuel (ed.) 2004. "More Than Ownership: Ten Case Studies of Land and Natural Resource Conflicts from East Java and NTT". Indonesian Social Development Paper 4. Jakarta: World Bank. Colombijn, Freek and J. Thomas Lindblad (eds.) 2002. Roots of Violence in Indonesia: Contemporary Violence in Historical Perspective. London: KITLV Press. Cribb, Robert 1990. "The Indonesian Killings 1965-1966: Studies from Java and Bali." Monash Papers on Southeast Asia 21. Centre of Southeast Asian Studies, Clayton: Monash University. Crouch, Harold. 2000 "Democratization and the Threat of Disintegration," in Daljit Singh (ed.), Southeast Asian Affairs 2000. Singapore: ISEAS:115-133. Davidson, Jamie S. (2003). "The Politics of Violence on an Indonesian Periphery," in South East Asian Research 11(1) (March): 59-90. Diprose, Rachael. 2005. "Conflict Pathways in Indonesia: Conflict, Violence and Development Programs in East Java." World Bank Office Jakarta. Mimeo. The Economist. 2004. "Special Report: Indonesia's Elections". 25th September: 29-32. Easterly, William. 2001. "Can Institutions Resolve Ethnic Conflict?" Economic Development and Cultural Change 49(4): 135-157. 37 Emmerson, Donald K. 1999. "Voting and Violence: Indonesia and East Timor," in Donald K. Emmerson, Indonesia Beyond Suharto: Polity, Economy, Society, Transition. Armonk, NY: M.E. Sharpe. ------. 2004. "A Year of Voting Dangerously?" Journal of Democracy 15(1): 94-108. Fein, Helen. 1993. "Murders in Democratic Kampuchea, 1975-1979 and in Indonesia, 1965 to 1966." Comparative Studies in Society and History 35: 796-823. Fischer, Jeff. 2002. "Election Conflict and Violence: A Strategy for Study and Prevention 2001-2002." IFES White Paper. February. Galtung, Johan. 1969. "Violence, Peace, and Peace Research." Journal of Peace Research 6(3): 167-191. Gurr, Ted. 2000. "Ethnic Warfare on the Wane." Foreign Affairs. 79(3): 52-64. Hadiz, Vedi. 2003. "Reorganizing Political Power in Indonesia: A Reconsideration of so-called `Democratic Transitions'". Pacific Review 16(4) (December): 591-611. Harris, John R. and Michael P. Todaro. 1970. "Migration, Unemployment, and Development: A Two-Sector Analysis." American Economic Review 60(1): 126-142. Homer-Dixon, Thomas. 1989. Environment, Scarcity and Violence. Princeton, N.J.: Princeton University Press. Honna, Jun. 2005. "The Post-Soeharto Local Politics in West, Central and East Java: Power Elites, Concession Hunting and Political Premanism." Paper presented at the University of Indonesia 4th International Symposium of Journal Anthropologi Indonesia, Depok, July 15th. Horowitz, Donald L. 1985. Ethnic Groups in Conflict. Berkeley: University of California Press. Imparsial. 2003. Front Line Indonesia: Murders, Death Threats and Other Forms of Intimidation of Human Rights Defenders, 1998-2002. Jakarta: Front Line Indonesia. International Crisis Group. 2001a. "Indonesia: National Police Reform." ICG Asia Report No. 13. Jakarta & Brussels, 20th February. ------. 2001b. "Indonesia's Maluku Crisis: The Issues." Indonesia Briefing. Jakarta & Brussels. 19th July. ------. 2002a. "The Search for Peace in Maluku." Asia Report No. 31. Jakarta & Brussels. 8th February. ------. 2002b. "Tensions on Flores: Local Symptoms of National Problems." Asia Briefing No. 22. Jakarta & Brussels. 10th October. ------. 2003a. "Indonesia: Managing Decentralisation and Conflict in South Sulawesi." ICG Asia Report No. 60. Jakarta & Brussles. 18th July. ------. 2003b. "The Perils of Private /Security in Indonesia: Guards and Militias on Bali and Lombok." Asia Report No. 67. Jakarta & Brussels. 7th November. 38 Kimura, Ehito. 2005. "The Politics of Pemekaran: Some Preliminary Observations from Northern Sulawesi," Paper presented at the University of Indonesia 4th International Symposium of Journal Anthropologi Indonesia, Depok, 12th-15th July. King, Blair A. 2004. "Empowering the Presidency: Interests and Perceptions in Indonesia's Constitutional Reforms, 1999-2002 (Megawati Soekarnoputri)". Ph.D. Dissertation. Ohio State University. Klinken, Gerry van. 2001. "The Maluku Wars of 1999: Bringing Society Back In." Indonesia 71 (April): 1-26. ------. 2003a. Minorities, Modernity and the Emerging Nation: Christians in Indonesia, a biographical approach. Leiden: KITLV Press. ------. 2003b. "Decentralisation, violence, and democracy: The colonial roots of ethnic conflict in Indonesia", in Stanley Adi Prasetyo, A.E. Priyono, and Olle Tornquist (eds), Indonesia's post- Soeharto democracy movement. Jakarta: Demos: 81-96. ------. 2005 "New actors, new identities; Post-Suharto ethnic violence in Indonesia", in Dewi Fortuna Anwar, Helene Bouvier, Glenn Smith and Roger Tol (eds.) (2005). Violent Internal Conflicts in Asia Pacific: Histories, Political Economies and Policies. Jakarta: KITLV. Kroef, Justus M. van der. 1986. "Petrus: Patterns of Prophylactic murder in Indonesia," Asian Survey 25(1): 745-799. Lande, Carl. 1999. "Ethnic Conflict, Ethnic Accommodation and Nation-Building in Southeast Asia." Studies in Comparative International Development 33(4): 89-117. Liem Soei Liong. 2002. "It's the military, stupid!", in Freek Colomjin and J. Thomas Lindblad (eds.). Roots of Violence in Indonesia: Contemporary Violence in Historical Perspective. London: KITLV Press, 197-225. Lindsey, Timothy. 1998. "The Criminal State: Premanism and the New Indonesia," in Grayson Lloyd and Shannon Smith (eds.). Indonesia Today: Challenges of History. Singapore: Institute of Southeast Asian Studies: 283-297. Malley, Michael S. 2001a. "Class, Region, and Culture: The Sources of Social Conflict in Indonesia," in Nat J. Colleta et al (eds.), Social Cohesion and Conflict Prevention in Asia: Managing Diversity through Development. Washington, D.C.: World Bank: 349-381. ------. 2001b. "Indonesia: Violence and Reform Beyond Jakarta," Southeast Asian Affairs 2001: 159- 176. ------. 2002. "Political centralization and social conflict in Indonesia," in S.A. Giannakos (ed.), Ethnic conflict: Religion, Identity, and Politics. Athens: Ohio University Press: 170-189. ------. 2003. "New Rules, Old Structures and the Limits of Decentralization in Indonesia," in Edward Aspinall and Greg Fealy (eds.). Local Power and Politics in Indonesia: Decentralisation and Democratisation. Singapore: Institute of Southeast Asian Studies: 102-118. Manoppo, Piet G., Mellaz, August, and Muhammad Athoillah (eds.) 2004. Merangkai Perdamaian: Refleksi Atas Partisipasi Perdamaian Pemilu 2004. Jakarta: Common Ground Indonesia. 39 Meliala, Adrianus (ed.) 2005. "Evaluation of the Police Role in the 2004 General Elections". Jakarta: Partnership for Governance Reform in Indonesia: 28-31. Miguel, Edward and Mary Kay Gugerty. 2002. "Ethnic Diversity, Social Sanctions, and Public Goods in Kenya." Mimeo. U. California Berkeley/U. Washington: Berkeley/Seattle (December). Morishita, Akiko. 2005. "Today's Local Politics in Indonesia 1998-2004: Who Controls the Political and Economic Interests in Kalimantan?" Paper presented at the University of Indonesia 4th International Symposium of Journal Anthropologi Indonesia, Depok, 12th-15th July. National Democratic Institute. 2004. "Advancing Democracy in Indonesia: The Second Democratic Legislative Elections since the Transition". Jakarta: NDI. O'Brien, Kevin. 1996. "Rightful Resistance," World Politics 49(1): 31-55. Prior, J.M. 2003. "The Church and Land Disputes: Sobering Thoughts from Flores." Maumere: Candraditya Research Centre for the Study of Religion and Culture. Mimeo. ------. 2004. "Land, Church and State: Forestry Business Packaged in Ecological Concerns in Flores." Inside Indonesia 78 (April - June). Priyambudi Sulistiyanto. 2001. "Whither Aceh?" Third World Quarterly 22(3): 437-452. Rhode, David. 2001. "Indonesia Unraveling?" Foreign Affairs 80(4) (July/August): 110-124. Rieffel, Lex. 2004. "Indonesia's Quiet Revolution." Foreign Affairs 83(5) (July): 98-110. Riefqi Muna. 2004. "Security Reform: Reform of the Security Sector is Fundamental for Achieving a Democratic Indonesia," Inside Indonesia (Jan-March) http://www.insideindonesia.org/edit77/p06-7muna.html Robinson, Geoffrey. 1995. The Dark Side of Paradise: Political Violence in Bali. Ithaca, NY: Cornell University Press. ------. 1998. "Rawan is as rawan does: The origins of disorder in New Order Aceh," Indonesia 66 (October): 127-56. Robison, Richard, and Vedi R. Hadiz. 2004. Reorganising Power in Indonesia: The Politics of Oligarchy in an Age of Markets. London and New York: Routledge Curzon. Saiful Mujani and William R. Liddle. 2004. "Indonesia's approaching elections: politics, Islam, and public opinion." Journal of Democracy 15(1) (January): 109-123. ------. 2005. "Indonesia in 2004: The Rise of Susilo Bambang Yudhoyono." Asian Survey 45(1) (February): 119-126. 40 Sakurai, Masatoshi. 2005. "The Privatizing Party Organization at Local, Considering with the Impacts of Direct Election system for Local Administration Chief: The Case of PDI-P Responding to the Mayor election in Municipality of Semarang," Paper presented at the University of Indonesia 4th International Symposium of Journal Anthropologi Indonesia, Depok, 12th-15th July. Sherlock, Stephen. 2004. "The 2004 Indonesian Elections: How the System Works and What the Parties Stand For." Canberra: Center for Democratic Institutions. Singh, Bilveer. 2003. "The 2004 Presidential Elections in Indonesia: Much ado About Nothing." Contemporary Southeast Asia 25(3) (December): 431-48. Sissener, Tone. 2004. "The Republic of Indonesia: General and Presidential Elections April ­ September 2004." NORDEM Report 12/2004. Oslo: University of Oslo. Smith, Claire Q. 2005. "The Roots of Violence and Prospects for Reconciliation: A Case Study of Ethnic Conflict in Central Kalimantan, Indonesia." Conflict Prevention and Reconstruction Paper 23. Washington, D.C.: World Bank. Snyder, Jack. 2000. From Voting to Violence: Democratization and Nationalist Conflict. New York: W.W. Norton & Co. Stewart, Frances. 2002. "Horizontal Inequalities: A Neglected Dimension of Development." QEH Working Paper 81. Oxford: Queen Elizabeth House, University of Oxford. Suryadinata, Leo, Evi N. Arifin, and Aris Ananta (eds). 2003 Indonesia's Population: Ethnicity and Religion in a Changing Political Landscape. Singapore: Institute of Southeast Asian Studies. Tadjoeddin, Mohammad Zulfan. 2002. "Anatomy of Social Violence in the Context of Transition: The Case of Indonesia', UNSFIR Working Paper 02/01. Jakarta: UNSFIR. Tajima, Yuhki. 2004. "Mobilizing for Violence: the Escalation and Limitation of Identity Conflicts. The Case of Lampung, Indonesia." Indonesian Social Development Paper 3. Jakarta: World Bank. Tan, Paige Johnson. 2002."Anti-party reaction in Indonesia: causes and implications," Contemporary Southeast Asia 24(3) (December): 489-508. Tilly, Charles. 1999. Durable Inequality. Berkeley: University of California Press. ------. 2003. The Politics of Collective Violence. New York: Cambridge University Press. Trijono, Lambang. 2001. Keluar Dari Kemelut Maluku: Reflexi Pengalaman Praktis bekerja untuk Perdamaian Maluku. Yogyakarta: Pustaka Pelajar. Varshney, Ashutosh. 2002. Ethnic Conflict and Civic Life: Hindus and Muslims in India. New Haven: Yale University Press. Varshney, Ashutosh, Rizal Panggabean, and Mohammad Zulfan Tadjoeddin. 2004. "Patterns of Collective Violence in Indonesia (1990-2003)", UNSFIR Working Paper 04/03. Jakarta: UNSFIR. 41 Vermonte, Philips J. and Hikmat Budiman (eds.) 2005. Konflik dan Pemilu: Civic Engagement Dalam Pemilu 2004 (Kasus Empat Daerah Pasca Konflik). Jakarta: Interseksi Foundation. Welsh, Bridget. 2005. "Keroyokan in Indonesia: Lessons." Paper presented at the University of Indonesia 4th International Symposium of Journal Anthropologi Indonesia, Depok, 15th July. Wilkinson, Steven. 2004. Votes and Violence: Electoral Competition and Ethnic Riots in India. Cambridge: Cambridge University Press. Wilson, Chris. 2001. "Internal Conflict in Indonesia: Causes, Symptoms and Sustainable Resolution." Canberra: Australian Foreign Affairs, Defence and Trade Group. 7th August. Wilson, Chris. 2005. "The Ethnic Origins of Religious Conflict in North Maluku Province, Indonesia, 1999-2000," Indonesia 79 (April): 69-91. Wiyata, A. Latief. 2002. Carok: Konflik Kekerasan dan Harga Diri Orang Madura. Yogyakarta: LKiS. World Bank. 2003. Breaking the Conflict Trap: Civil War and Development Policy. New York: Oxford University Press. Zinoman, Peter and Nancy Peluso. 2002. "Rethinking aspects of political violence in twentieth-century Indonesia and East Timor," Asian Survey, 42(4) (July-August): 545-549. Appendix A: Conflict Incidents by Area Table A: Number of Conflict Incidents March 15-May 15, 200446 Number of Number of Number of Non- Number of Province District Election Violent Election election incidents Violent Non- Incidents Incidents election incidents Bali Buleling 6 1 3 3 Tabanan 16 1 1 0 Jembrana 6 1 0 0 Karangasem 14 1 3 2 Klungkung 5 1 3 1 Bangli 5 1 3 1 Gianyar 15 0 9 5 Badung 10 0 4 2 Denpasar 25 3 15 14 Total 102 9 (8%) 41 28 (68%) East Java Pamekasan 9 1 2 0 Probolinggo 4 0 2 0 Sumenap 7 0 10 2 Bangkalan 15 1 16 7 Pasuruan 2 0 5 0 Sampang 8 0 2 0 Total 45 2 (4%) 37 9 (24%) Maluku Ambon 25 5 56 42 Maluku Tengah 14* 2 14 7 Buru 5 1 0 0 Maluku Tenggara 2 1 4 3 Maluku Tenggara 3 1 0 0 Barat Total 49 10 (20%) 74 52 (70%) NTT Sikka 4 0 6 4 East Flores 1 1 4 2 Ende 8 1 6 3 Lembata 4 2 1 1 Manggarai 2 0 4 1 Manggarai Barat 2 0 0 0 Ngada 7 0 6 5 Total 28 4 (14%) 27 16 (59%) Total (All Provinces) 224 25 (11%) 179 105 (57%) 46Source: Newspaper Data Archive (March 15-May 15) As there is no local newspaper in Maluku Tengah, non- violent incidents may be significantly underreported there, by as much as half. Only violent incidents are reported in the Ambon press. Newspaper data was only collected for districts in Flores island within NTT (i.e., not for West Timor or Sumba). Similarly data was only collected for two "clusters" in East Java. Appendix B: Election-related Conflict Incidents by Type Table B: Forms of Election-related Conflict Incidents March 15-May 15, 200447 Inter-party Intra-party Inter- Parties/individua Province District conflicts (Non- conflicts (Non- group/individual ls and State Violent/Violent) Violent/Violent) conflicts (Non- Conflicts (Non- Violent/Violent) Violent/Violent) Bali Buleling 2 / 1 2 / 0 1 / 0 0 / 0 Tabanan 9 / 1 1 / 0 4 / 0 1 / 0 Jembrana 0 / 0 0 / 0 0 / 1 5 / 0 Karangasem 0 / 0 5 / 0 2 / 0 6 / 1 Klungkung 0 / 0 3 / 0 1 / 1 0 / 0 Bangli 1 / 1 1 / 0 1 / 0 1 / 0 Gianyar 3 / 0 1 / 0 8 / 0 3 / 0 Badung 1 / 0 3 / 0 4 / 0 2 / 0 Denpasar 2 / 0 4 / 0 5 / 3 10 / 0 Total 18 / 3 20 / 0 26 / 5 28 / 1 East Java Pamekasan 0 / 0 0 / 0 0 / 0 8 / 1 Probolinggo 0 / 0 1 / 0 0 / 0 3 / 0 Sumenap 0 / 0 0 / 0 0 / 0 7 / 0 Bangkalan 1 / 0 0 / 0 1 / 0 12 / 1 Pasuruan 0 / 0 2 / 0 0 / 0 0 / 0 Sampang 0 / 0 0 / 0 1 / 0 7 / 0 Total 1 / 0 3 / 0 2 / 0 37 / 2 Maluku Ambon 2 / 0 1 / 3 4 / 1 13 / 1 Maluku Tengah 1 / 0 1 / 1 1 / 1 9 / 0 Buru 3 / 0 0 / 0 0 / 0 1 / 1 Maluku 0 / 0 0 / 0 1 / 1 Tenggara Maluku 1 / 0 1 / 0 0 / 0 0 / 1 Tenggara Barat Total 7 / 0 3 / 4 5 / 2 24 / 4 NTT Sikka 0 / 0 3 / 0 0 / 0 1 / 0 East Flores 0 / 1 0 / 0 0 / 0 0 / 0 Ende 0 / 0 0 / 0 0 / 0 7 / 1 Lembata 0 / 0 0 / 1 0 / 0 2 / 1 Manggarai 0 / 0 0 / 0 1 / 0 1 / 0 Manggarai Barat 0 / 0 0 / 0 0 / 0 2 / 0 Ngada 0 / 0 0 / 0 0 / 0 7 / 0 Total 0 / 1 3 / 1 1 / 0 20 / 2 Total (all provinces) 26 / 4 29 / 5 34 / 7 109 / 9 47Source: Newspaper Data Archive (March 15-May 15). CONFLICT PREVENTION AND RECONSTRUCTION UNIT--WORKING PAPER SERIES No. Title Authors Date 16 The Private Sector's Role in the Provision of Infrastructure in Jordan Schwartz August 2004 Post-Conflict Countries: Patterns and Policy Options Shelly Hahn Ian Bannon 17 Within and Beyond Borders: An Independent Review of Post- Swarna Rajagopalan October 2004 Conflict Fund Support to Refugees and the Internally Displaced 18 Drugs and Development in Afghanistan William Byrd December 2004 Christopher Ward 19 Understanding Local Level Conflict Pathways in Developing Patrick Barron December 2004 Countries: Theory, Evidence and Implications from Indonesia Claire Q. Smith Michael Woolcock 20 Addressing Gender in Conflict and Post-Conflict Situations in Sonia Margallo January 2005 the Philippines 21 Community Cohesion in Liberia: A Post-War Rapid Social Paul Richards, et al. January 2005 Assessment 22 International Companies and Post-Conflict Reconstruction: John Bray February 2005 Cross-Sectoral Comparisons 23 The Roots of Violence and Prospects for Reconciliation: A Claire Q. Smith February 2005 Case Study of Ethnic Conflict in Central Kalimantan, Indonesia 24 The Mindanao Conflict in the Philippines: Roots, Costs, and Salvatore Schiavo- February 2005 Potential Peace Dividend Campo Mary Judd 25 Counting Conflicts: Using Newspaper Reports to Understand Patrick Barron May 2005 Violence in Indonesia Joanne Sharpe 26 Young Men and the Construction of Masculinity in Sub- Gary Barker June 2005 Saharan Africa: Implications for HIV/AIDS, Conflict and Christine Ricardo Violence 27 Conflict-Sensitive Development Assistance: The Case of Juana Brachet June 2005 Burundi Howard Wolpe 28 Survey of the German Language Literature on Conflict Barbara Müller October 2005 29 Demand-Driven Approaches to Livelihood Support in Post- Piet Goovaerts October 2005 War Contexts: A Joint ILO-World Bank Study Martin Gasser Aliza Belman Inbal 30 Strategic Environmental Assessments: Capacity Building in Rob Verheem December 2005 Conflict-Affected Countries et al. 31 Consolidating Indonesia's Democracy: Conflict, Institutions Patrick Barron December 2005 and the "Local" in the 2004 Legislative Elections Melina Nathan Bridget Welsh Also published as Social Development Paper No. 91