Some improvements to STV

I went to an event on Monday organised by COSLA (the Confederation of Scottish Local Government) on STV, the voting system used in Scottish local elections. Some of the comments that the presenters and attendees made got me thinking about STV and how it could be improved.

One point that some of the attendees made is that some voters don’t understand STV. (If you’re one of these people, I wrote a short guide to STV some time ago). Some voters just mark an X against their first choice, which is counted as a first preference, but other voters might mark two X’s, one against each other candidates of their favoured party. This cannot be counted, as it is impossible to tell which of the two is the voter’s 1st preference, and which the 2nd preference.

Another phenomenon with STV elections is that if a party runs multiple candidates in a ward, they want each candidate to get roughly the same number of votes. This makes it more likely that they will both be elected than if one gets lots of votes and the other just a few. However, if you’re a voter who favours a party, you have to rank one of them higher than the other; you cannot give them both the same preference.

A third phenomenon is where a voter mentally ranks candidates into three groups: ones she likes, ones she hates, and a middle group who she has no strong opinions about. Under STV, to make her vote count fully, she has to rank all the liked and middle candidates to fully express her preference of the middle ones over the hated ones.Sometimes voters use “donkey voting” — giving them successive numbers from the top of the ballot paper — to achieve this.

I suggest that STV, as used in local government elections, be changed as follows:

(1) Allow equal preferences

So a voter would be able to give her first preference equally to two candidates, perhaps one so f the same party. This would count a 1/2 a vote for each candidate. If one was elected, the surplus would go to the other. If one was eliminated, the the vote would go to the other first-preference candidate.

This reform means that voters would be allowed to just mark an “X” against multiple preferred candidates and the vote could be counted.

(2) Score candidates out of 10, instead of ranking them

At the moment, a voter must scan the whole ballot paper to mentally pick out their favourite candidate, then do the same thing for their next favoured, then continue until they have no more preferences. So it’s not surprising that most voters give up after about 3 preferences, thereby not making full use of their vote.

But if voters allocated a score to each candidate, instead of a preference, it’d be easier for them. Our hypothetical voter could simply start at the top of the ballot paper, and for each candidate in turn, and give them marks out of 10: 10/10 for a candidate they like, 0/10 for a disliked candidate, etc.

Coincidentally, X is the Roman numeral for 10, so it would be appropriate for it to count as a score of 10.

(3) Give a low default score to unranked candidates

Under scored voting, a voter might not have an opinion about some candidates, and may not give them a score. When this happens, they could be given a default score, which would be not the highest score, or the lowest score, but a low score of perhaps 2/10 or 3/10. The rationale for this is a voter probably knows who they like and who they dislike, and so an unscored candidate is one who they don’t strongly like or dislike and should therefore get a middling score. But because we want to elect candidates who people like, not merely ones they’ve never heard of, it should be a score lower than halfway, so that they score lower than a candidate who the voter marks with a middle score of 5/10.

(4) Have top-up seats

In Scottish parliamentary elections there are both constituency seats, and top-up seats allocated to make the final result more proportional. I envisage doing the same with Scottish councils.

At the moment, if a party gets 5% of first preferences in every ward in a council area, they will get no councillors elected. But for the result to be proportional, they would deserve to get 5% of the councillors (in Edinburgh’s case, 3 of the 58 councillors). The result could be made more proportional to be increasing the number of councillors per ward, but you’d have to more than double the current number to allow people to get elected with 5% of the vote, which might prove unwieldy. It would also mean larger wards, which would mean they had less of a local flavour.

There is an easier way to get proportionality. In Scottish parliamentary elections there are both constituency seats, and top-up seats allocated to make the final result more proportional. I envisage doing the same with councils. As well as ward-based seats, there could also be a small number of seats — maybe 10% — which would be top-up seats. These would be allocated according to the number of 1st preferences each party gets, taking into account the number of seats they’ve already won. The actual method of allocation would be d’Hondt, the same used in parliamentary elections.


Leave a Reply

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

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

Google+ photo

You are commenting using your Google+ 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 )


Connecting to %s