An ill-considered outcome of early morning horse-trading

John Cox says AV for Westminster elections would have devastating consequences in Wales

Next May’s referendum on a new voting system for Westminster elections will package three proposals into a single yes or no decision: fewer MPs, more equal electorates, and AV, the Alternative Vote described in the panel (see below). Although elements of this package have some merit, taken as a whole it will diminish democracy and have devastating consequences in Wales.

This article is from the current issue of the IWA’s journal Agenda, issued three times a year. To receive Agenda and get unlimited access to the IWA’s online archive, click here.

Whatever its merits, AV is not proportional – as was shown in Australia in August 2010, when both the leading parties won far more seats in the House of Representatives than justified by their share of the vote (see Table, below).

On the same day, the elections to the Australian Senate elections, which used a version of STV, did give the three leading parties a reasonably fair share of seats. If the UK government wanted to introduce a fairer voting system, it would have proposed STV, not AV.

Moreover, unlike STV, AV fails to reduce voter alienation and actually increases the power of party elites and the hegemony of the leading parties. Despite these well attested defects of AV, a few supporters of STV (notably the Electoral Reform Society) propose to campaign for a Yes vote in the referendum, apparently in the vain hope that AV will be a first step towards genuine reform.

To the contrary, however, AV will entrench the present two party dominance and will put a stop on further change. Certainly this is what has happened in Australia. The politicians elected by AV to the House of Representatives since 1919 naturally prefer it to the more proportional STV system used for the Senate elections.

All forecasts agree that AV will exclude minority parties with less than 20 per cent of the popular vote, and prevent newcomers gaining a foothold at Westminster. So the in-built unfairness of the current system would remain.

The precise outcomes of each constituency election under AV is hard to predict as each is, ultimately, determined by second, third and even lower preference votes. Although a right-wing voter may cast right-wing preference votes and a left-wing voter left-wing preference, many voters are as much influenced by style as by substance. In practice, attitudes of mind also influence voter preferences:

  • Those whose first choice is Conservative or LibDem might well give their preference vote to the other. In this situation Labour could be reduced to a rump at Westminster even it won the most first preference votes. This would be bad for democracy.
  • Those opposed to an incumbent may give their preference votes to ‘anyone but…’ to keep a leading candidate out. This attitude favours the ‘least-disliked’ candidates over the ‘most-liked” – and is equally bad for democracy.

Although pundits often assume that preference votes follow a similar pattern to first choices, there is no hard evidence to support this assumption. Consequently,  it is very difficult to predict outcomes.

The package is to include unspecified but systematic changes to constituency boundaries to ensure that all MPs elected by similar numbers of voters. While there is much to be said for parity, there are better and simpler ways of achieving this end.

In Wales the current 40 constituencies will be reduced to 30 with only Cardiff South and Penarth likely to survive unaltered. Everywhere else, the historic community and administrative boundaries will have to be redrawn. Whereas the Boundary Commission normally may take a decade to consult, put forward changes and consult again before finalising its proposals, the new boundaries will have to be ready in time for the next Westminster election expected in 2015.

If voter to MP parity is really vital, we could use the five existing Welsh electoral regions as 5-7 member STV constituencies. The merit of multi-member constituencies with STV is that it ensures all major parties in Wales will be represented in proportion to their votes. In contrast, using AV for new 75,000+ constituencies, it is possible that neither Plaid Cymru nor the LibDems will win a single seat.

The Alternative Vote system

Rather than simply marking one solitary ‘X’ on the ballot paper, the voter ranks the candidates on offer by preference. Voters thus puts a ‘1’ by their first-preference candidate, and can continue, if they wish, to put a ‘2’ by their second-preference, and so on, In some AV elections, such as most Australian elections, electors are required to rank all candidates.

If a candidate receives a majority of first-preference votes (more people put them as number one than all the rest combined), then they are elected.

If no candidate gains a majority on first preferences, then the second-preference votes of the candidate who finished last on the first count are redistributed. This process is repeated until someone gets over 50 per cent.

Australian August 2010 election results

House of Representatives

(elected by AV)


(elected by STV)


% vote

% seats

% vote

% seats

Liberal/National parties





Australian Labor





Australian Greens










It could be argued that these regions would be rather large. The option that I prefer would be to group the current parliamentary constituencies into ten 3-member STV constituencies. This very simple change would deliver the desired voter to MP parity without the hassle of hasty boundary changes.

Whilst there is merit in equalising voter to MP ratios and reducing the number of MPs, it could be achieved without destroying historic community boundaries. Sadly, STV is not an option for the referendum. Whether we vote Yes or No we still will have the unfairness of a winner-takes-all system: We are not to be allowed to voice an opinion in favour of a genuinely fair proportional system.

The proposed legislation ‘freezes’ all existing constituency boundaries for the elections to the National Assembly. Unless the timetables are altered this means that the May 2015 elections could descend into farce. May 2015 is when the Westminster and the National Assembly elections next coincide. It could mean that Welsh voters will have the challenge of grappling with three distinct voting systems and constituencies on the same day.

In this scenario, Welsh voters will have to choose 30 MPs by AV from new hastily created constituencies. Meanwhile, on the same day, they will have to vote for 40 AMs by first-past-the-post from the old constituencies, and simultaneously, 20 regional AMs from party lists in the obsolete regions created over a decade ago for Euro elections.

But pare a thought for Scotland though. Their local elections employ STV  so they could be using four different election systems on the same day.

Mention of Scotland reminds me that Wales has relatively fewer AMs in our National Assembly than Scotland has MSPs. Parity would have given us 15 more than the 60 currently elected. If, as the government claims, Wales needs to have 10 fewer MPs to achieve parity with the rest of Britain, at the very least they might let us have 10 additional AMs – 15 extra were recommended by the Richard Commission.

AV will ensure that every MP has to win at least half the votes. In May 2010 only 12 Welsh MPs polled 50 per cent of the votes, while 14 obtained less than 40 per cent. At first glance it may seem that this will enhance MP’s legitimacy. However, even a cursory glance at Australia’s post-election traumas shows that election by AV is not the same as winning the support of half the voters. The most common gripe is that many who were elected secured very few first choice votes but were promoted by the preference votes of minor parties, some with racist or other dubious policies.

Suppose, in the UK context, an AV winner depended on the preference votes of those whose first choice was, say, the BNP. Would the winner be comfortable about this? Would the loser, if he or she might otherwise have been the winner, accept the legitimacy of such a result?

Far from bestowing legitimacy on those elected, AV creates even more post-election angst than does first-past-the-post. This is an inevitable consequence of  the elimination of other political parties, even those with 20-30 per cent of the first choice votes, to bestow a false legitimacy to the eventual ‘winner’.

As became evident after the election of Ed Milliband as leader of the Labour Party by AV – and whiuch occurs after every Australian election – the transparency of AV leads to endless post-election recriminations if the leading candidate in the first round is defeated by the transference of votes from the least popular candidates.

Currently five political parties represent Wales in Westminster and Europe. All should be represented in proportion to their voting support. The use of AV in significantly larger constituencies will reduce the limited plurality that currently does exist (even with first-past-the-post) and could re-establish a two-party hegemony.

The proposition in the AV referendum is an ill-considered outcome of early morning horse-trading to establish the UK coalition government. It is not the product of an in depth discussion and debate, merely an excuse for Liberal Democrats to support a minority Conservative government. As we are not to be allowed the option of voting for a fairer voting system, the best way to keep genuine electoral reform alive will be to bury the AV distraction with a resounding No vote .

John Cox is an engineer, a former Torfaen independent councillor, and Vice President of CND. He submitted persuasive evidence to the Richard Commission that advocated STV elections to the National Assembly.

Comments are closed.

Also within Politics and Policy