Uncategorized

Electoral Systems in Wales, 4: STV

[Note: Amended as of 14.39, 13/11/13, to include direct link to the Sunderland Commission Report; thanks to Penarth a’r Byd for tracking this down!]

The Single Transferable Vote (STV) electoral system is the only one discussed in this series of Blog Posts which is not currently used in any major Welsh elections. But official Commissions of Inquiry have recommended its use for both National Assembly and local government elections in Wales. It is therefore worth some consideration.

The Richard Commission, which reported in March 2004, examined electoral arrangements for the National Assembly. Unlike in Scotland, there had not been widespread and inclusive public debate on this issue prior to devolution. After considering a wide menu of possibilities, the Richard Commission surprised many observers by recommending wholesale reform to the electoral system. It advocated STV elections for an 80-member Assembly. More AMs were argued to be necessary for the Assembly to wield effectively the additional powers recommended by Richard. (Although, as some have pointed out, no very clear argument was made as to why the number should be 80). STV was seen as the best electoral system in such circumstances: maintaining similar levels of proportionality to AMS with 60 members, and eliminating tensions around the relative status of constituency and list AMs.

The Richard report was vague regarding how the multi-member constituencies necessary for STV might best be created. Proposals that have subsequently been floated include pairing the 40 Westminster constituencies, with each of the 20 pairs returning four AMs. Another possible approach would be to use Wales’ 22 local authority boundaries: the number of AMs per local authority would then vary by size, and the very largest authorities (most obviously Cardiff) might be divided into more than one multi-member district. Allocating seats via the usual (Sainte-Laguë) formula used in such contexts, this would produce the following number per authority in an 80-Seat National Assembly:

Number of AMs

Local Authorities

9

Cardiff

6

Swansea, Rhondda Cynon Taf

5

Carmarthenshire, Caerphilly

4

Flintshire, Newport, Neath Port Talbot, Bridgend, Wrexham

3

Powys, Vale of Glamorgan, Pembrokeshire, Gwynedd, Conwy

2

Denbighshire, Monmouthshire, Torfaen, Ceredigion, Ynys Môn, Blaenau Gwent, Merthyr Tydfil

Calculating the electoral implications of introducing STV is inherently difficult – it is quite unlike working out the consequences of an extra 10 or 20 list seats under AMS, where one can safely assume that voting patterns would be largely unchanged and simply deduce who would have won any extra list seats in a particular election. STV asks voters to cast their ballot in a fundamentally different way – as a series of preferences, rather than a single categorical choice. STV also requires subtle strategic decision-making from parties: as Irish parties could testify, difficult decisions have to be made concerning the number of candidates to run per constituency, and whether or not to clearly prioritise one as the candidate to receive first preference votes.

The 2011 Welsh Election Study did, however, gather some relevant information by asking respondents how they would have voted in the Assembly election if they had been asked to rank parties in order of preference. Labour does well here: not only gaining many first preference votes, but also a lot of second preferences from supporters of Plaid Cymru, the Liberal Democrats and the Greens. Plaid also do reasonably well, being the most popular second preference for Labour supporters. By contrast, few supporters of other parties choose the Conservatives as a second preference.

Based on this information, and the actual 2011 election results, the Electoral Reform Society and I tried to estimate what the 2011 National Assembly election might have produced under STV. For an 80-seat Assembly with four AMs each elected from 20 pairs of constituencies this yielded the following estimated ‘result’:

Labour: 40

Conservatives: 20

Plaid Cymru: 15

Lib-Dems: 5

A comparison of this estimate with that in my previous post, for an 80-seat Assembly elected under AMS in 2011, reinforces the point that Labour would tend to do well from STV. Were the Assembly ever to expand to 80 AMs, it would be strongly in Labour’s interest to have this elected via STV rather than a version of AMS in which constituency and list seats were split evenly.

The introduction of STV for local elections in Wales was recommended by the Sunderland Commission Report in 2002. This recommendation has not, thus far, been implemented: local elections continue to be conducted under a combination of single- and multi-member district plurality. STV was, though, introduced for local elections in Scotland in 2007. We can therefore readily compare the May 2012 local elections held in the two nations under the different electoral systems. Such a comparison shows that the results were more disproportional in Wales than Scotland. Across the whole of Scotland, the local council elections produced a Gallagher Index score of 4.48; compared to 6.09 across the whole of Wales (with, for the purposes of these calculations, Independent and Others treated as a single bloc in both nations). There were no examples in Scotland, unlike in Wales, of parties winning substantial majorities in council chambers on a minority share of the vote. Elections were also less competitive in Wales, where 8.0% of seats saw candidates elected unopposed (plus one seat where no candidates at all were nominated); in Scotland there were no uncontested seats.

In my view, the introduction of STV to local government elections in Scotland is one of the most unambiguously positive developments to have followed from devolution. Local government elections have been rendered far more competitive, and the composition of local government itself more closely tied to popular wishes. Voters have also been given greater choice: they can choose between candidates as well as parties, and offer gradations of preference rather than being forced to make an all-or-nothing categorical choice. There are no good reasons not to introduce STV for local government elections in Wales.

The argument made for STV for National Assembly elections by the Richard Commission also remains strong. STV would retain a broadly proportional system for electing the Assembly while eliminating any further arguments about the status of constituency and list members (and the potential for disputes about ‘dual candidacy’). Wales is the sort of small country in which STV has already been found to work fairly well (the other well-known examples being the Republic of Ireland, Northern Ireland and Malta). Wales’ dominant party might also be well-advised to consider supporting the introduction of STV for more pragmatic reasons: if the Assembly ever is to expand in size, electing it via STV will likely produce outcomes more favourable to Labour than the obvious alternative route of increasing the number of regional list AMs.

Comments

  • Tal Michael

    The main argument against STV is that large multi-member constituencies would reduce the link between voters and their elected member. The danger is that we would end up with lots of “regional members” whereas people feel a closer connection with their constituency members.

    My preferred solution for an improved electoral system for Wales would be two-seat STV, which would ensure the Assembly was representative while maintaining a close link with constituents. It would make sense to require all parties to field one man and a woman, thereby ensuring a much better gender balance in the Assembly (and removing the need for all-women shortlists, the only mechanism any party has used to date which has had any impact on improving gender balance).

    My predictions for the outcome of this are fairly similar to your predictions for four-seat STV – see http://talmichael.wordpress.com/2013/11/12/a-better-system-for-conducting-elections-to-the-welsh-assembly/

    • Roger Scully

      Thanks for this Tal, and for the kind words on your Blog.

      Your suggestion for a two-seat form of STV is interesting. There is, of course, a trade-off: a smaller number of members per seat does help retain more local representation; a larger number helps promote greater overall proportionality. (It’s not impossible to get a proportional result with something like two-seat STV, but it is less likely).

      One other factor that you mentioned on your Blog was the patterns of transfers. This does matter, and I may look in more detail in a future post at this. But the basic picture, I think, is clear. Under nearly all plausible political scenarios, the Conservatives would be the major party in Wales that would do worst out of STV, as they would be the party least likely to receive large numbers of vote transfers from supporters of the other parties (except for some from UKIP). Under most scenarios, Labour also does well out of STV: it gets lots of first preferences, and also plenty of transfers from those giving their first preferences to Plaid, as well as a fair few from the Greens, and the Lib-Dems.

  • Penarth a'r Byd

    I asked for the Sunderland report under FoI and was sent a copy, which I am now sending to you (on email). I think the Welsh Government has deliberately tried to suppress the Sunderland Commission report – why else would there be no link for it even though it exists in electronic form? And why else (as you will note from the email) did it take WG 8 months – following a reminder email – to send me the documents?
    It would be great if you could publish it on your site so it has a publicly accessible URL (I haven’t yet got round to uploading).

  • Rebecca Williams

    “There are no good reasons not to introduce STV for local government elections in Wales” – hear hear!!

Leave a Reply

Your email address will not be published. Required fields are marked *