The 2019 Bougainville referendum offered valuable lessons for future elections, Kerryn Baker and Thiago Oppermann write.
In late 2019, Bougainville held a non-binding referendum on its future political status, with voters asked to choose between greater autonomy within Papua New Guinea (PNG) or independence. A team of researchers from The Australian National University and the University of PNG observed the referendum, including the polling and counting processes.
The Bougainville referendum of 2019 was conducted in a social setting of great community involvement, with robust polling procedures and generally effective administration and implementation. Whilst there were some minor shortcomings in compliance and isolated security concerns, as an exercise in polling in a Melanesian context, it was exemplary. It met international standards of free, fair and credible polling practices.
The Bougainville referendum was administered by the Bougainville Referendum Commission (BRC). In preparing for the referendum, the BRC specifically targeted common concerns in previous polling exercises held in Bougainville. Most notably, the integrity of the electoral roll has been a continuous struggle in both Autonomous Bougainville Government (ABG) and national general elections in Bougainville.
The Bougainville referendum was underpinned by a significant investment in cleaning and updating the roll, with efforts to increase community involvement in the process. While duplicate names were not eliminated entirely, the result was generally considered to be the most accurate electoral roll ever produced in Bougainville and substantially increased public faith in arrangements for the referendum.
In the 2015 ABG elections, observers noted up to 30 per cent of people who wished to vote were being turned away from polling stations due to their names not being found on the roll. In the referendum, this was mostly avoided — first through the investment in cleaning and updating the roll and then through the introduction of a provisional ballot system. Voters could vote at a polling station that was not the one where they had registered by casting a provisional ballot; these ballots then underwent scrutiny at the counting centre and, if accepted as valid, were entered into the count.
Provisional ballots proved to be an effective means of removing disputes from the polling place and ensuring smooth and fast conduct of polling. The majority of provisional votes were accepted into the count, suggesting it is a valuable way of ensuring people who wish to cast a vote are able to do so. The scrutiny of provisional ballots, however, substantially increased the time the counting of votes took, which could be a serious concern in future elections given the already protracted counting processes in elections using the limited preferential voting system in Bougainville and PNG.
The Bougainville referendum also used significantly more polling stations than in past elections. Polling was scheduled to take place in every ward in Bougainville. This required a substantial increase in personnel and posed some logistical challenges — for instance, the police vehicles transporting the ballot boxes to polling places were frequently delayed — but ensured greater access to polling stations, especially for voters with limited mobility.
The smooth conduct of the referendum was a result of the high community engagement, locally led but bolstered by voter awareness and reconciliation initiatives in the lead-up to the election, supported by the BRC, ABG and development partners. Overwhelmingly, Bougainvillean voters were not only committed to respecting the outcome, but also to the integrity of the process.
This community buy-in created a secure environment for polling to take place just as much as, if not more so than, the deployment of security personnel from Bougainville, PNG and neighbouring countries. Voter turnout was relatively high, at 85 per cent, compared to just over 60 per cent in the 2015 ABG elections. This is perhaps to be expected, given the highly anticipated nature of the referendum, yet it also reflects well on the initiatives undertaken by the BRC to enable greater voter access to the polls.
In his speech announcing the result, BRC Chairman Bertie Ahern noted that the informal count was low, suggesting that voters were well informed on how to cast their vote. While this outcome implies that voter education initiatives were successful, it is also indicative of the simplicity of the vote, which did not require preferencing as in ABG or national general elections. From our observations in the counting room, we estimate that a quarter to a half of informal ballots were in fact errors on the part of presiding officers — who had not signed the back of the ballot to verify it — rather than voter error.
In the counting room, the ballots were ‘mixed’ after reconciliation procedures had concluded but before the count began. The mixing of ballots was to ensure secrecy, by removing the ability to trace votes back to the particular ballot box they came from. Our observations suggested that many voters were unaware of this aspect of the count, and more awareness about it before the vote could have increased faith in the secrecy of the ballot. Mixing ballots prior to the count could potentially have an impact in easing community tensions in post-election periods in the future; a similar system was also implemented in the 2019 Solomon Islands general elections.
The referendum’s polling processes set a new benchmark for electoral administration in Bougainville and PNG. They are of a standard unlikely to be reproduced in the course of regular elections, as the referendum enjoyed a minimalist ballot, a near consensus on the option to be selected, and an absence of competitive politics, including money politics. Nevertheless, many challenges in standard elections also affected the referendum: logistical difficulties, uncertain funding, staffing and training, and the task of compiling an accurate roll. The analogy to a regular election is thus limited, but valuable.
The Bougainville referendum can be an example of best practice in terms of electoral roll updating and cleaning. Problems with electoral roll accuracy are pervasive in Melanesia, and the referendum demonstrates that significant investment in properly maintaining the roll can greatly increase public faith in the integrity of the electoral process. Other initiatives — the use of provisional ballots and a greater number of polling stations — while successful in this referendum, need to be considered carefully in light of the particular contexts of other elections, including financial and time constraints.
The referendum benefitted from the high level of engagement and support that Bougainvilleans gave to the electoral process. This reflects community sentiments around independence, but also a broader commitment to the democratic process than has been observed in past elections. This commitment is perhaps the most valuable tool in preparing for future elections in Bougainville, and there may be lessons from the referendum on how to strengthen and promote community engagement with elections elsewhere in the region.
This article is based upon a paper published by the ANU Department of Pacific Affairs (DPA) as part of its ‘In brief’ series. The original paper can be found here.