Viewpoint
Viewpoint
"The alternative vote is the worst of all possible plans. It is the stupidist, the least scientific and the most unreal," warned former British Prime Minister, Winston Churchill. "An element of blind chance and accident will enter far more largely into our electoral decisions than ever before, and respect for parliament and parliamentary processes will decline," he anticipated during a House of Commons debate in 1931.
In the midst of the Pacific Ocean, Churchill's fears have been strikingly confirmed. Fiji introduced the alternative vote system as part of a bold experiment in electoral engineering back in 1997. It was aimed at bringing together politicians representing the country¹s 52% indigenous Fijian population and 44% Indo-Fijian population, and providing electoral incentives for moderation and compromise. Top American and Australian political scientists extolled the virtues of the system as a tool for promoting the formation of multi-ethnic governments.
- ADVERTISEMENT -
But the first election under the new system in May 1999, gave the largely Indo-Fijian backed Labour Party an absolute majority in parliament, despite negligible ethnic Fijian support. It lasted only a year in office before being overthrown by a group of indigenous Fijian extremists led by George Speight. Now, a second election in August 2001, has brought to power a hard-line ethnic Fijian government with scarcely any Indo-Fijian support.
Even last year's coup leader, Speight, managed to attract sufficient preference votes to get elected from his prison cell.
How the system works
The alternative vote system requires voters to rank candidates in order of
preference. If no candidate gets 50% of first preference votes, then the
lowest polling candidate is eliminated from the contest and his or her
voters¹ second preferences are re-distributed to the remaining candidates.
This process of elimination of lowest polling candidates, and redistribution
of their votes, continues until one candidate emerges with an overall
majority.
Back in the mid-1990s, Fiji¹s Constitutional Review Commissioners considered that this system would encourage voting across ethnic lines. Although the country has a history of race-based voting under a preferential system, voters might be willing to give a second or third preference vote to candidates from the other ethnic groups, particularly if these candidates adopted more moderate political platforms.
Political parties, the commissioners further reasoned, would thereby acquire incentives to strike deals with each other to acquire such preference votes.
"Only moderate parties with conciliatory politics will agree to trade preferences," they stated in their final report. "The system, therefore, encourages the emergence of such parties. For this reason, we favour the adoption of alternative voting system." (The Fiji Islands: Towards a United Future, 1996).
So enthusiastic were the commissioners about providing parties with incentives to strike such deals that they proposed an exceptionally complex split format ballot paper. Voters could either rank candidates in order of preference (voting below-the-line), or they could simply tick next to their favoured party symbol (voting above-the-line). Ballot papers ticked above-the-line were deemed to endorse an order of preferences expressed by political parties or candidates. This provision gave party officials extraordinary power over the redistribution of votes.
Finally, the commissioners anticipated that pre-election deals over preference votes would lead to the formation of robust and enduring multi-ethnic coalitions, unlike those precarious alliances often forged between hostile parties in the aftermath of an election.
Why the System Didn't Work
The alternative vote system was intended to favour moderates, and penalise
ethnic extremists. But is this a sensible strategy? In a country with a
history of coups, devising electoral laws effectively aimed at driving
dissent underground, is a dangerous approach.
Even if one accepts this objective, the alternative vote system did not work as intended, either in 1999 or in 2001. First, it was not only moderate parties with conciliatory policies that struck deals with each other over preferences. Everyone did so, even the most extremist parties.
Second, compelling voters (and parties) to rank preferences lessens the likelihood that these reflect any genuine ideological affinity. It boosts the potential for relatively arbitrary preference rankings and alliances between strange bedfellows.
Third, in both elections, voters and parties did not simply rank candidates from the top downwards, but also from the bottom upwards. They put their most hated party last, and second most hated immediately above, and so on. In 1999, the Opposition parties ganged up against the former governing party, Sitiveni Rabuka's Soqosoqo ni Vakavulewa ni Taukei (SVT) and its coalition allies. This year, it was the Labour Party¹s turn to be placed last on party preference lists. In both elections, key marginal constituency contests were decided in favour of the party that was immediately above these bottom-placed parties, even though this could involve victory for a party drawing on 6th or 7th preferences.
Comparing 2001 and 1999
After the 1999 election, many commentators acknowledged defects with the
operation of Fiji¹s new voting system. But they suggested that the country
needed time to adjust. Perhaps, 1999 was a freak result, they argued. "You
can't judge the system after just one election."
Yet the new 2001 results are, in several respects, just as problematic as those in 1999, and, in some respect, even worse.
-In 1999, 90.1% of registered voters cast their ballots. This year, the turnout fell to 78.6%, well below the level witnessed in other countries where voting is compulsory. 100,000 people may now be prosecuted for not voting.
-In 1999, Fiji witnessed its highest ever level of invalid or informal voting (8.7%), as voters struggled with the complexity of an over-elaborate system. In 2001, invalid votes rose to 12% - one of the highest figures witnessed anywhere in the world.
-In 1999, 92% of voters filled out ballot papers above-the-line, by simply ticking next to a party symbol. Defenders of alternative voting argued that, over time, voters would adjust to the new voting system and learn to rank preferences themselves. Yet, for this year's election, provisional estimates by the Commonwealth Observers suggested that 94% of voters completed their ballot papers above-the-line.
-In 1999, the election result was highly disproportionate. Party shares in the overall vote were very different from their share in the total seats won. The accompanying table uses Gallagher¹s well-known index as a means to calculate the overall votes/seats disproportionality for all elections since 1966. This shows that both elections under the alternative voting system were the worst, in this respect, ever witnessed in Fiji.
Defeat of the Moderates
Most importantly, did the alternative voting system achieve its stated
objective of favouring the more moderate political parties? Certainly, the
local press took up this theme, with headlines such as Moderates Look Good
on Preferences in the run up to the polls.
Before the election, a moderates forum emerged, combining parties that drew support from Fiji's different ethnic groups - the New Labour Unity Party (NLUP), National Federation Party (NFP), Soqosoqo ni Vakavulewa ni Taukei (SVT), Fijian Association Party (FAP) and United Generals Party (UGP). These parties made deals with each other to share preferences, and extolled the merits of a more conciliatory style of politics.
Yet the moderates were obliterated at the polls. In total, they secured only four seats. And of these, two were in the tiny general voters constituencies (where Europeans, part-Europeans and others vote separately).
Fiji's parliament has 71 members - 23 are elected by ethnic Fijians alone, 19 by Indo-Fijians, three by the general voters and one by people from the island of Rotuma. In addition, there are 25 open or common roll constituencies, where citizens of all races vote together. Each voter has two votes; one in their own reserved ethnic constituency, and the other in one of the open constituencies.
All of the 19 Indian reserved seats fell to the Fiji Labour Party, while all of the 23 Fijian seats fell to new Prime Minister Laisenia Qarase's Soqosoqo ni Duavata ni Lewenivanua (SDL) or to the Matanitu Vanua (MV).
As in 1999, the outcome of the election was therefore decided in the 25 open constituencies. The accompanying graph shows the combination of 1st preference and transferred preference votes for the winning candidates in each of these 25 open constituencies. Constituencies are ranked from left to right in accordance with their ratios of indigenous Fijians to Indo-Fijians. On the right-hand side are the mainly Indo-Fijian cane farming constituencies in the north and west of Fiji's two largest islands. Most of these were obtained by the Fiji Labour Party (shown in black). On the left-hand side are those constituencies where indigenous Fijians form the majority. These were largely won by the SDL (shown in white).
Towards the centre of the graph are the more ethnically-mixed constituencies, where the proportions of Indo-Fijians and ethnic Fijians are roughly similar. As in 1999, these were constituencies where the transfers of preference votes exerted a powerful influence. But these vote transfers only led to victory for moderate parties in two constituencies; Suva and Nadi. Otherwise, it was the transfer of votes from the moderate parties that gave victory to the more extremist parties.
In 1999, the Labour Party was the major recipient of these transferred preference votes, and it won most of the seats shown in the centre of the graph. This year, the SDL was the major recipient. It won nine seats where victory was dependent upon the transfer of preference votes from one or other of the moderate parties (three from the NFP, three from the NLUP, two from the SVT, and even one from the FLP).
In 1999, leaders of the three small Fijian parties whose preference votes gave the FLP its absolute majority did join the People's Coalition government. But this year, the moderate parties felt unable to join the new SDL/MV coalition government, and the new government refused to work with the Labour Party. One New Labour Unity Party MP who joined the government is currently threatened with expulsion for so doing.
This exposes the fallacy inherent in the idea that party exchanges of preference votes would necessarily underpin the adoption of more moderate policies, or that this would lead to the formation of robust multi-ethnic coalitions.
Fiji has witnessed a failed experiment in electoral engineering. The constitution-makers juggled with the electoral rules, with the aim of artificially advancing the position of moderate parties. But it didn¹t work as intended, and the objective itself was misguided.
The country would have been far better served by a simpler, and more straightforward, proportional representation system.
(Dr Jonathan Fraenkel, is a lecturer in Economic History at the University of South Pacific.)
