Nauru

This microstate in the Pacific, with an area of only 21 sq-km (8.1 square miles) and a population of only 12,000, is (also) special because it is currently the only one to use a Borda-like positional voting for legislative elections. The modified Borda method was also used in Slovenia until recently, to elect the Hungarian and Italian minority representatives.

Unlike Australia, however, Nauru does not use instant runoff voting or STV. The so-called ‘Dowdall system’ is similar to Borda’s positional voting method, but with different scoring (Plus, Nauru uses a variant, where multiple winners are always chosen in each constituency). Dowdall scoring converts rankings into points not by equal distances, but according to the harmonic series.

Constituencies of Nauru

For example, if there are 5 candidates, then according to Borda scoring, the candidate ranked first gets 5 points, the second gets 4, etc. — but we could also say that the points range from 4 to 0, or from +2 to -2. As long as the distances are equal, these are all equivalent. In the Nauruan system, first place is worth 1 point, second place is worth 1/2, third place is worth 1/3, etc. — This method gives higher weights to the relatively higher positions (top-heavy), as opposed to the unbiased scoring of Borda. It can be said that it is somewhere “halfway” between FPTP/SNTV (where only first preferences are worth 1 point, everything else is zero) and the Borda count.

The Borda system is known in voting theorist circles for being easily manipulable (“a system only for honest men” according to its inventor) and for its significant potential for tactical voting. Nevertheless, it can still be called a kind of “compromise-seeking” method. Its flaws also stem from the way in which compromise is “forced”. Especially when all candidates are required to be ranked (as in Nauru), voters do well to put their strongest opponents last — even if there is a candidate they would like to see elected even less. In addition, it is worth running decoy candidates (clones), whose presence can increase the effectiveness of tactical voting. (And if it is not mandatory to rank all candidates — as was the case in Slovenia — then it is often not worth it; and thus the search for compromise will not be as successful.)

In Nauru, due to compulsory ranking, the phenomena of both tactical voting and strategic nomination can be observed. Although the system is theoretically less exposed to the tactics observed under the Borda rules, the incentives are very similar. Even if the voters themselves do not want to undermine the compromise-seeking nature of the system, the candidates do want to achieve the best possible result for themselves. Therefore (similarly to what parties and candidates do in Australia), the candidates hand out “How to vote” cards to explain, well, how to vote — taking into account what tactics would be favorable to them. However, Nauruan politics is apparently — at least on the surface — centered much more on individual candidates, not so much on formal party lines, as we see in much larger countries. It may well be that the system itself is partly responsible for the country’s relatively non-partisan politics to this day, but it may still primarily be due to social and historical reasons (plus the small size of the country and legal reasons?).

Nauru’s modified Borda system is also interesting because it is only used in multi-member districts. Most districts are two-member, meaning the race is primarily for the second seat. For this reason, and because of how points are scored, some similarities can be found with the old Chilean ‘binomial’ system. Although the system is not proportional in any sense (unlike, for example, the Australian STV), it is not winner-take-all all either: if there are a few similar candidates, even with the most optimal tactical voting, they cannot guarantee that they will win all the seats with a mere plurality.

Most electoral experts would probably consider the Dowdall system too arbitrary and theoretically problematic. Still, as a sort of compromise between the SNTV (note: in mostly 2-member districts, SNTV may be even more “loser-friendly” — biased towards the second largest faction — than D’Hondt…) and Borda systems, it may have some virtues. One of its goals might be to avoid a straight-up bloc voting system, where the minority has no serious chance, and the plurality faction can easily win all the seats in the district. But at the same time, the system also makes it more difficult for the minority to have too much influence. Plurality bloc voting is, unfortunately, often too tempting a choice compared to the other simple multi-winner system, the capricious SNTV. So while in Nauru, the system might be more winner-take-all than SNTV, it’s still more nuanced than simple plurality bloc voting.

In addition, it preserves the possibility of ranking, and thus the option for voters to provide their preferences in detail, and — for better or worse — it uses these rankings by looking at the entire preference profile of voters at once. (Again, it is worth comparing it with the Australian proportional STV, where further rankings are not considered until the top one is “in play”). This is how the aforementioned (Borda-like) “compromise-seeking” may emerge, but this may also be perceived as a relative advantage in terms of simplicity. For many, interpreting rankings as points is easier and more intuitive than the world of quotas, fractional surplus votes and multiple elimination rounds.

Results in Ubenide constituency in 2025 (Nauru Electoral Commission)

The image above shows the district with the most complex results of all the constituencies. Although this may seem complicated at first because of the 4 seats and 12 candidates, a large part of this might just be a lack of familiarity. In fact, what we see just that all the rankings (positions on each ballot paper) for each candidate were compiled into a table, then they were multiplied by the Dowdall points, and the results were then summed up. It can be seen that the person elected to the 4th seat would have only finished 6th according to first preferences, so the ranked ballots did seem to have a significant direct effect in this case. It is also visible that the person elected to replace him had a strikingly high rejection rate (last preferences), and he came 6th in the final result.

Term length

In Nauru, similarly to Australia and New Zealand, the parliamentary term is 3 years. A referendum was held alongside this election to amend the constitution to extend the mandate of parliaments to four years. Although parliament voted to do so, the 44% ‘yes’ vote in the referendum fell far short of the required two-thirds majority, and the amendment failed. The president of Nauru, also the head of government, is elected by parliament.

Notes

In case you are interested in more details about the real-world use of the Borda and Dowdall methods in Nauru and Slovenia, I recommend Fraenkel & Grofman’s article on the subject:

Page updated

Google Sites

Report abuse