Ghostbusters: it’s time to deal with Vanuatu’s phantom voters

By Anna Naupa and Nick Howlett

Part 1 of a 3-part series on Vanuatu’s Electoral Integrity

Read Part 2: ‘Is diversity of political representation possible in Vanuatu?’ >
Read Part 3: ‘Is dual citizenship a threat to Vanuatu? No, but unregulated political financing is’ >
When Vanuatu goes to the polls on 22 January, 2016, an estimated 2,000 ni-Vanuatu youth, and several hundred recently naturalized adult citizens, will be denied a vote due to the impossibly short deadline for voter registration. Not only is this a denial of the democratic right to vote, it highlights the inaccuracy of Vanuatu’s electoral roll.

In Vanuatu’s 2012 General Election, an estimated 55,000 dead people were included on the electoral roll as eligible voters, about 140% more voters than official census data says there should be. This article suggests a simple and cost-effective way of ensuring the integrity of Vanuatu’s electoral roll so these ‘phantom voters’ don’t come back to haunt us.

As in any democratic country, Vanuatu’s electoral roll changes as people reach voting age, migrate, or die. The Vanuatu Electoral Office (VEO) faces a number of unique challenges that make updating the electoral roll difficult: a dispersed island geography, high transport costs, a lack of permanent administrative staff, and inadequate voter registration accountability systems. On top of this are local allegations of political interference and electoral fraud.

There is no easy way to overcome these difficulties, but there are some quick statistical tests that can help pinpoint polling districts. Recent research has shown that analyzing the rate of voter turnout in each voting district, and preferably at each polling station, is not only accurate at predicting voting irregularities, it is also cost-effective. The number of ‘phantom voters’ can be determined simply by comparing voter turnout rates with the electoral roll and calculating the statistical likelihood of the different turnout rates.

So, how does this method work? First, we need to separate the dead from the living using the official census data to give us a more accurate figure of the total number of registered voters. There were officially 192,632 registered voters at the time of the 2012 general election. But the census data from the Vanuatu National Statistics Office (VNSO) shows that this is highly unlikely: in 2012, almost 50% of Vanuatu’s 250,000 population was under the age of 15, and therefore below the voting age of 18. Census-based estimates of the 2012 voting age population suggest the electoral roll contains around 55,000 additional voters (see Table 1).

Table 1. 2012 Predicted 18-21 year old voters

Age Total Male Female
Total Population (Usually Resident) 233,561 118,826 114,735
15 years and under 96,122 50,155 45,967
16-17 years 9,425 4,674 4,751
18 years and over 128,014 63,997 64,017
Predicted Voting Age Population in 2012 137,439

(compared to 192,632 on the electoral roll)

68,671 68,768

Note: authors’ calculations based on 2009 National Census Results (Usually Resident Population by Age, Table 2.6), assuming no deaths during this time. Even allowing for a 5% margin of statistical error, the upper limit of the population at voting age in 2012 is 144,311; still 48,321 less than those on the electoral roll.

We should also note that this figure is likely to be even higher, because voting is not compulsory in Vanuatu, and a proportion of the population choose not to register to vote for or other reasons.

Having identified the estimated number of ‘phantom’ voters on the electoral roll, the next step is to work out the true size of Vanuatu’s voting population. This is where the voter turnout rate method comes in handy: the official count of voter turnout in the 2012 general election was 118,256 voters.

This is much closer to the estimate of voter population we got above from using the official census data, and is a much more accurate estimate of the total voting population. The voter turnout figure is even lower than the census-based estimate because voting is not compulsory in Vanuatu, and a proportion of the population may choose not to register to vote for religious or other reasons.

But it doesn’t stop there: we can use voter turnout rates to analyse election results at the local level. This is where the voter turnout method is at its most powerful.

Closer analysis of voter turnout at the district level during the 2012 General Election indicates that voter turnout was unusually low in several districts, Santo, Malo/Aore, Luganville, Ambae, Efate and Port Vila.

These districts showed low rates of voter turnout compared to the overall national voter turnout rate in the 2012 Election, as well as the voter turnout rates in other electorates elsewhere in the country.

It should be noted that low voter turnout in these districts doesn’t necessarily indicate electoral fraud, however; there are a number of possible reasons for this, for example high population mobility in urban centres, poor weather conditions or local events. But the anomalies in the electoral roll in these districts do indicate that reviewing the electoral roll in these districts should be a top priority.

As we have shown above, using census data to get an accurate number of the total voting population is the first step in ridding Vanuatu of its phantom voters. The voter turnout rate method can then be used to further refine this figure. Most importantly, this method can be used at a local level to highlight the districts with the most irregularities so that they can be reviewed as a matter of priority.

The VEO needs to do some ghostbusting as soon as possible and deal with Vanuatu’s phantom voters once and for all.

Anna Naupa conducted research into Vanuatu’s electoral integrity while studying at the Harvard Kennedy School of Government, Cambridge, Massachusetts, in 2013-14. Her research paper can be read here.

Nick Howlett is a Vanuatu communications professional currently completing an MA at Griffith University, Australia.

Opinions expressed in this post are solely those of the authors and do not express the views or opinions of any organization or imply endorsement of any political party or position.

5 Comments on “Ghostbusters: it’s time to deal with Vanuatu’s phantom voters”

  1. Charles Kick says:

    Good to see some reporting on this surreality. The same situation presented in Solomon at the last election. But attracted no research.


  2. Rick Balkonan says:

    A fitting and timely article which highlights an important method to deal with ‘voting irregularities’.

    The issue may run deeper than this, and I am sure the next series of articles may begin to identify other underlying issues as well, e.g. it has been common knowledge that an individual in Santo can vote twice, using two separate cards. Once in a rural-based polling station and the second time in Luganville (or vice versa), for example.

    There are also allegations that corrupt officials within the electoral office, who are also on the payroll of certain politicians contribute to this “perpetual electoral fraud”, especially in superficially increasing the number of votes of certain former MP’s, one or two of whom are currently in prison. Also that a certain female staff who recently got transferred back to the Electoral Office after she was transferred to the Education Office may be the worst culprit.

    These are issues which international observers may not immediately pick-up on, and may only be identified though a comprehensive audit of electoral office processes and procedures and Staff.


    • Anna Naupa says:

      Thanks for the insights Rick – another area that warrants close scrutiny is the application of the proxy voting system. While data does exist somewhere, it is not a requirement to publicly report on the number of proxy voters and in which constituencies and polling stations – the main data publicly tracked is that of votes cast and void votes, but additional context as to where there are unusual volumes of proxy voting would also help to iron out some electoral integrity concerns. Perhaps something election observers could request proxy information later this month.


  3. Sue Farran says:

    Very interesting. So is this a fraud being deliberately carried out (and if so by whom?) or is it a systemic failure over a number of years to remove dead or migrated voters?


    • Anna Naupa says:

      Thanks Sue. My preliminary research pointed to systemic failure over a number of years (basically if we’re comparing incomplete data with incomplete data detecting fraud is even more challenging) – at least we have pinpointed a few ‘odd polling stations’ to assist the VEO in focusing its scarce resources to systematically clean up the electoral roll.