Voting reform 3: Single Transferrable Vote (STV)

Ok so it took a while, but here comes the third and final part of the Hippo voting reform series.


If you missed parts one and two look here for a discussion on FPTP, it's faults and the reason that we should never use the term 'winner' in an election.
Once you are done with that, have a look here for a gander at the rules regarding an AV election, the system that we may get a referendum on as part of the Voting Reform Bill.


Those following the goings-on in parliament will have seen that the Liberal Democrats tried to pass an amendment to the Voting Reform Bill, to get STV included on the referendum form. This motion got voted down, so you may wonder why I'm bothering to mention it at all.


Well in my opinion STV is the best way of selecting representatives for a given constituency, indeed instead of a single 'victorious' representative being chosen, a number of representatives are selected based on the spread of political opinion in that constituency. Of course whether or not this is an improvement on the current state of affairs is a matter of opinion. Before we look into the positives and negatives let's first observe how it works.


In each constituency there will be a greater number of parliamentary seats up for grabs. This will mean that each constituency is larger, so rather than a single election determining the representative of apporximately 70,000 people, an STV election would simutaneously produce four representatives for a ward containing 280,000 members.

The ballot paper for an election in this system will look the same as one in an AV election (see image) where each voter, instead of putting a single 'x' in a single box, expresses their order of preference for the candidates. A '1' will indicate the candidate that the voter most prefers, followed by a '2' for the second most favoured all the way down to the last candidate that the voter can stomach.

So far, so familiar. But of course this time we need to select (for the sake of our argument) four representatives, how to do that with the information gathered from voters?

In the current case where we are selecting four representatives, we look for the four candidates who recieve one vote more than 20%* once we have taken into account all the preferences. That is any candidate who recieves 28001 votes in our 280,000 constituent ward gains a seat in parliament (assuming a 100% turnout in the kind of ward described above).

The first thing that we do is look at the first choice votes cast. If any candidate gets the required number of votes, they are elected. If more than one candidate gets the required number of votes, all of those candidates are elected. As it takes a set number of votes to pass the threshold, second choice votes are transeferred proportionally to other candiates. How this is done is a little hard to explain so please bear with me for the next passage.

Leaving our example for a second, imagine that a candidate that requires 10,000 votes to be elected actually gains 20,000 votes. this means that the candidate has an excess of 10,000 votes (and is good for my on-the-fly arithmetic). Now for each of the 20,000 votes cast the second choice is considered and counted up. Once totalled, half of those votes are then transferred to the respective candidate. We only transfer half the votes, because half of the votes needed to be used to elect the first choice, but we counted all the papers because we need to discover the second choice votes of all voters in question.

When the second choice votes are considered for all of the candidates that have reached the required number of votes** we then check to see if any of the other candidates have now reached that mark. If not, the lowest ranked candidate is eliminated, and their votes are completely redistributed to the next preference candidate on that paper.

If a Candidate qualifies at this point, then any surplus is again redistributed in the way outlined above and the process is repeated until all places are filled. If this was a little hard to follow, check out the STV page on wikipedia which offers a simple scenario involving treats.

Hopefully now with a (sketchy maybe) grasp of how STV works, we can look at the pros and cons of introducing such a system. Writing on the AV post from this series an anonymous poster said.

"Each party can put up as many candidates as they like without fear of splitting the vote. And independents who broadly support a party (and party rebels) can also choose to stand without needing to worry about taking votes away from a more viable candidate and letting a less preferable opponent in."

This is of course true of AV systems, but with only one seat available in the house where is the incentive to field more than one candidate? At the end of the day political parties care only that the candidate from their party gets voted in. Not that the constituents are happy. If they feel that the mood in a constituency is such that their party will get the votes (safe-seats)  there is no need to field other potential representatives if you know there will be someone voting with your whip once the votes are in!

With STV there is an incentive to field at least enough candidates to fill all the available seats. That then gives the voters a choice to vote for a party but to exclude a candidate that they do not approve of. Just think of the chance for labour voters to oust Hazel Blears or Jacqui Smith!

There is another regularly raised issue with STV, is that with larger, multi-represented constsiuencies, there will no longer be an obvious link between MP's and their constituency. While this may be undoubtedly true if taken flat, the information on the ballot papers could be used to assign candidates to constituencies once the election is complete. Alternatively, maybe having more than one representative would be a good thing anyway. If voters had more than one person to field an issue to, they would be more likely to vote again for the one that responded most positively, even if it meant using one of their preferences for a candidate that is not part of the party they would choose for government. The best performing MP's in this system would be returned to sit again, while the weakest would only manage to serve a single term. What we'd end up with is basically market forces aiding in the survival and ousting of MP's!

That's about it. This was quite tough to explain properly, so if you can suggest any clarifications I'd love to hear from you via the box below. Also, if you agree disagree or have anything to add, join in below. I look forward to hearing from you.

*The reason for this choice of number is actually a logical follow on from what constitutes a working majority (ie 50% of votes +1) In a typical (FPTP) election, gaining one more than 50% of the votes guarantees that no other candidate can have more votes. In this case, only 4 candidates can possibly have more than a 20% share of the vote so we use 20% of the votes cast + 1vote. This system for working out how many votes are required to be elected is known as the 'Droop quota'

**Of course if more than one candidate has already been selected, we may need to look at the third choice or lower preference candidate on any paper. In all cases we just look for the highest useable preference.