(You may also be interested in this post for an overview of how First-Past-the-Post distorts elections results and makes it more difficult for smaller parties to win seats.)
Despite the release of studies (such as this one by the IPPR) that have concluded that AV will not help fringe parties or extremist parties win seats in the UK House of Commons, the No to AV campaign argues that extremists such as the British National Party would flourish under an Alternative Vote (AV) system.
For example, Coalition cabinet member (and Conservative Party Chair) Baroness Warsi said recently that, under AV, BNP votes would be counted several times – thus “back(ing) a system which rewards extremism and gives oxygen to extremist groups”. The IPPR report linked to above clearly states:
Minor parties might increase their share of the vote but they will still struggle to win seats. Their influence will grow by virtue of having the major parties seek their supporters’ second preference votes. However, extremist parties like the BNP will be penalised by AV and their recycled votes will not influence election outcomes. (p. 3)
it is clear that the BNP would find it almost impossible to win a seat under AV because they would fail to win the backing of at least half the electorate. FPTP proves that they would fail to win a majority of first preferences: their highest vote share in 2010 was 14.6 per cent in Barking. Nor would they pick up many second preference votes: the BES mock AV ballot has them winning just 3.2 per cent. (p. 22)
First of all, I think it’s important to distinguish between fringe and extremist parties. Not all minor parties are extremist parties. A fringe party might have radical ideas, such as legalizing all drugs or abolishing all taxes, but it will pursue these goals in a peaceful, democratic way. Extremist parties are those that advocate the violation of democratic rights, promote violence and disorder or have racist ideologies.
In the 2010 UK general election, the BNP ran candidates in 338* of the UK’s 650 constituencies, received a total of 1.9% of the vote (just over half a million votes) and failed to elect a single MP. Their best result was in the constituency of Barking, where party leader Nick Griffin won 14.6% of the vote, finishing third. The seat was won by Labour, with 54% of the vote.
Would this result have been different under AV? The No2AV side’s arguments to the affirmative seem to hang on three assumptions: 1) some BNP supporters don’t vote for the party under FPTP because they feel it would be a wasted vote, but might vote for the party (as their first choice) under AV; 2) the BNP would pick up vote transfers and would manage to move ahead of other candidates, maybe even win; and 3) the mainstream parties would start deliberately courting the BNP vote in order to gain their second preferences.
There is little evidence to support any of these assumptions. The BNP’s overall vote did increase marginally in the 2010 election (over 2005), but there is little evidence that there exists millions of closet BNP supporters not voting for the party because they think it would be a wasted vote. The BNP has limited appeal to most voters. Because of this, very few non-BNP voters would rank the BNP as their second choice (or third, or fourth, etc.), as the British Election Study 2010 found (see page 8 of that report). In that mock AV election survey, only 3.2% of voters ranked the BNP as their second choice. Most voters probably wouldn’t rank the BNP at all, or would rank them last. Meanwhile, voters who indicated their first choice was the BNP ranked the following parties as their second preferences (in order of support): UKIP 45%, Conservatives 29%, Labour 10%, Greens 9%, Liberal Democrats 7%. But even if the BNP are the second choice of some voters, odds are that the BNP candidate would be eliminated before those transfers came into play.
In constituencies where the BNP finished mid-pack (for example, if there were 9-10 candidates contesting and the BNP ended up 5th or 6th), it is extremely unlikely that they would pick up sufficient vote transfers from the candidates who finished below them and were eliminated to move up enough to avoid being eliminated themselves, never mind actually moving into contention. In some constituencies, they might well be the first candidate eliminated.
As for the argument that the major parties would try to curry the favour of BNP voters under AV, this also doesn’t make much sense. There is little to gain for any mainstream party trying to appeal to BNP supporters. Indeed, the leadership of all three of the major parties in the UK routinely denounce the BNP; it would be difficult to imagine any of them actively courting BNP supporters. Also, as stated above, the other parties BNP supporters would have a choice to rank are all fairly mainstream, or at least, not extremist parties. BNP supporters, if they bothered to rank other candidates on their ballots, would be tranferring their votes to the three major parties and other minor parties such as the Greens and UKIP. They would be doing so voluntarily, not because they believed the Greens (or any other party) would suddenly adopt racist policies in exchange for a smattering of second preference votes from BNP supporters.
One thing is certain, as the IPPR report makes clear: the BNP could win a seat under FPTP (and indeed, they do just that at the local council level – due mainly to much lower voter turnout) and under PR (European Parliament elections). Indeed, low voter turnout definitely favours smaller parties, and BNP leader Nick Griffin himself blamed the higher turnout in the 2010 election for his party’s poor showing. But does that mean they’d never win a seat under AV?
While some on the Yes side might like to say that this would be the case, it isn’t. Of course an extremist party could win a seat under AV. For an example of an extremist party finding some success under AV, one only has to look at the case of Pauline Hanson’s One Nation Party in Australia. You can read about that in detail on Antony Green’s blog (scroll down – it’s after the part about the Greens, you can also take a look at this post).
What is important to note in the example of the One Nation Party is that there exists the means under AV for the mainstream parties to block the BNP vote. This option does not exist under FPTP. Essentially, if an extremist party is treated just like any other party, they might be successful under AV (as they would under FPTP), but if the party is actively treated as an unacceptable pariah party, they can be shut out, or at least have their success reduced under AV:
So looking at One Nation only as a minor party, it has done better under AV rules than FPTP. However, it has been squeezed out of Parliament when the major parties have chosen to defeat the party. When treated as an unacceptable pariah party, One Nation would have done better under FPTP rules. (emphasis added)
To apply that to the UK under AV, if the BNP was looking particularly strong in certain parts of the county, the main parties could agree to swap preferences in order to ensure that the BNP candidate would be shut out. For example, if a BNP candidate looked like he or she might finish first after the first count in a certain constituency (but short of 50% of the vote), with the Liberal Democrat candidate most likely to finish second, Labour and the Conservatives could agree to encourage their supporters to rank the Lib Dem candidate second in order to ensure that the BNP candidate did not win the seat. Such maneuvering would not be possible under FPTP, where the BNP candidate would win the seat.
In short, a minor party that can appeal to mainstream voters does have a chance of winning seats under AV, as it does under FPTP (i.e. the Green Party), but a party whose appeal is extremely limited to begin with, and which can be treated as an unacceptable choice by the main parties can be shut out, or at least be much less successful, under AV than would be the case under FPTP.
*I’ve seen 325, 338 and 339 as the total number of candidates fielded by the BNP in the 2010 general election. In one instance, different numbers appeared on the same page, one in a table, the other in the text below the table. I’ve opted to go with 338, but one candidate more or less doesn’t greatly alter the main point.