If Welsh Labour wants a two-member-constituency voting system, this is the one they should adopt

In the recent row over possible changes to the voting system used for elections to the Welsh Assembly, one of the alternatives proposed by the Labour Party was a system of two-member-constituency First Past the Post (see the Devolution Matters blog for an overview of the row). In other words, to expand the number of Assembly Members (AMs) to 80 from the present total of 60 (made up of 40 constituency AMs and 20 top-up regional AMs under the proportional AMS voting system), Labour was proposing having two AMs per constituency and using FPTP to elect them.

Presumably, the model of FPTP they had in mind was that voters would get two votes each, thereby ensuring that where Labour was the most popular party, it would be guaranteed to win both seats even if it were not the choice of a majority of voters. Labour is not known for its enthusiastic backing for fair voting systems, after all. FPTP wouldn’t be so bad if people had only one vote, so that the Labour vote would be split between both candidates, giving other parties more of a chance, especially if they fielded only a single candidate in constituencies where they knew they had no hope of winning both seats.

However, a fairer, more rational and more proportional electoral system for two-member constituencies would be the following, which I’m calling ‘TMPR2’: Two-Member Proportional Represenation (version two). This is a simpler and more practical version of the TMPR system I have previously discussed. TMPR2 works as follows:

  • There are two representatives (AMs, MPs, etc.) per constituency
  • Each voter has two votes. Voters are not obliged to use both votes: they can vote for just one candidate if they wish
  • The individual candidate obtaining the most votes automatically wins one of the seats
  • The individual winner may be either the representative of a party or an independent
  • In addition, if any independent candidate wins the second-highest total of individual votes, that independent candidate is elected
  • However, assuming the second-highest total of votes is not won by an independent, the winner of the second seat is decided on the basis of the share of the vote won by each party:
    • If any party wins over 50% of all votes (that is, the number of actual votes cast, which is higher than the number of voters, as people can vote for two candidates), then both of their candidates are elected (unless one of the candidates obtaining the highest or second-highest total of votes is an independent, in which case the party obtaining over 50% of the vote wins only one seat)
    • In the instance where one of the seats is in fact won by an independent, the party candidate elected is the one that has obtained more votes than the running mate from their own party
    • If, however, no party wins more than 50% of the vote, then the two parties obtaining the highest shares of the vote win one seat each (except in the case where one or more independent candidate are elected, whereby only the top-ranked party or no party respectively wins a seat)
    • In the case that two parties win one seat each, the successful candidates are those who obtained more individual votes than the running mates from their own parties

Advantages of TMPR

  • This is a reasonably proportional system
  • It encourages trans-party voting: voters could and would vote for candidates from different parties. This would equalise the parties’ share of the vote, with the established parties’ share coming down and the smaller parties’ share rising. For instance, quite a lot of right-of-centre voters, if the system were applied in England, would vote for one Conservative and one UKIP candidate; whereas many left-of-centre voters would vote for a Green candidate alongside a Labour or Lib Dem candidate. This means that the vote share parties need to win in order to be elected could be considerably lower than under FPTP. In fact, there is no lower percentage limit on eligibility for a seat. And TMPR2 encourages this pluralism by allowing voters to divide their loyalty between more than one party
  • It incorporates some of the best features of established, familiar voting systems:
    • Like FPTP, the candidate obtaining the largest number of individual votes automatically wins a seat
    • Like AV, if any party wins over 50% of the vote, it takes the whole constituency (i.e. both seats), unless an independent candidate has won either the highest or second-highest individual vote
    • It’s a crude form of PR, similar to STV in the sense that a party, as opposed to an individual candidate, needs to win more than a ‘quota’ of 50% of the vote to win both seats
  • It encourages voting for individuals – and hence, for independents – alongside parties: as voters have two votes each, they will be freer to choose candidates on their individual merits alongside their membership of a particular political party. There would be more of an incentive for independent candidates to run, such as high-profile, respected local figures taking a stand on important issues for the local community
  • It’s easy to understand and operate: there are no complicated voting or counting mechanisms involved, and the result is a clear and direct expression of voters’ preferences. There are no unexpected consequences and fewer tactical-voting constraints for voters. Voters would know that the way they voted would have a direct impact on the result: each of their two votes increases the chances of that individual candidate or party; and if voters are torn between the party / candidate they genuinely prefer and the party they feel they need to vote for in order to ensure that another party does not win (tactical voting), they can hedge their bets and vote both ways.

Disadvantage of TMPR2

TMPR is probably not as proportional as the existing system – AMS – used for elections to the Welsh Assembly and Scottish Parliament. In fact, in an Electoral Reform Society analysis of the 2011 Welsh Assembly election had it been conducted using AMS with 30 constituency seats and 30 regional top-up seats (instead of the present 40/20 ratio) compared with an 80-seat Assembly elected using STV, AMS emerges as the more proportional system. It would be interesting to see the outcome if they ran the same analysis on TMPR2.

However, pure proportionality is not everything; and TMPR does preserve the close links between individual AMs / MPs and relatively small constituencies. By comparison, AMS gives more power to the parties, as top-up AMs / MPs are predominantly elected because of their party affiliation rather than their individual merit or on the basis of local issues. In addition, TMPR is much simpler to understand and operate than either STV or AMS.

Real-world prospects for TMPR2

In reality, TMPR2 has very little chance of ever being implemented, at least not for the Welsh Assembly. As the ‘inventor’ of TMPR2, I don’t exactly have a lot of influence. But as the possibility of two-seat constituencies was being mooted, it seemed timely to bring forward TMPR2 as another alternative: as a possible compromise between FPTP and proportionality. The Labour Party wouldn’t like it, because it’s too fair and proportional. The experts at the Electoral Reform Society probably wouldn’t like it because it’s not proportional enough. But maybe the people would like it if they were offered the choice, precisely because it is fairer than FPTP but less complex and fussy than STV and AMS, with a more transparent link between how people vote in each constituency and the winners.

Anyway, I’m just throwing it out there to see if there are any takers.

English parliament


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: