Well chaps and chapesses, I’m back to base after a very enjoyable break in foreign climes. I won’t bore you with interminable holiday snaps and the like, however. Suffice to say that, although it was good to get away from it all for a bit, it’s also nice to be back to Blighty. I’ve got quite a few things to catch up with on at home, at work, and on the blog, and I’ll try to return to fairly frequent postings now that I’m home.

I thought I’d start with one of the big events that happened while I was away. Not the Royal Wedding, which I successfully avoided completely although I only narrowly escaped seeing some of it on an outdoor  big screen (which I had assumed would be showing some form of sporting contest). Not the killing of Osama Bin Laden, either. Nor even the AV referendum, which went the way I expected. No, I think the first thing I should comment on is the result of the elections to the National  Assembly for Wales which I followed, as best I could, by Twitter and on the net via my Blackberry while I was away. I wasn’t helped by the fact that North Wales decided not even to start counting votes until the morning after Polling Day, thus holding up the final results by half a day. Perhaps that’s because the count was done in Llandudnno, where people generally go to bed about 10pm?

This was the first Senedd election I have had the opportunity to vote in, even though I had to do it postally. For the Welsh Assembly elections, each voter gets two votes. One is cast just as in a General Election, i.e. by picking one candidate for one’s own consituency – in my case Cardiff West. This is a safe Labour seat, previously held by former First Minister Rhodri Morgan, and it was no surprise to see the Labour candidate romp home with an increased majority. Of the 60 members of the Welsh Assembly, 40 are elected directly through constituency votes like this.

Incidentally, one of the other candidates in this constituency was Neil McEvoy, standing for Plaid Cymru. McEvoy is currently deputy leader of Cardiff City Council (which is run by a coalition of LibDems and Plaid Cymru councillors) and is an enthusiastic champion of the building of a major road into Bute Park for use by heavy lorries. It might have been better for the people of Cardiff – especially those who appreciate its wonderful green spaces – had Councillor McEvoy been elevated to the Senedd, because that would prevent him doing further damage on behalf of the Council. In the end, though, he trailed in third place in the Welsh Assembly poll so will presumably remain on the Council.

Meanwhile, back at the polls. Voters in the Welsh Assembly elections get another regional vote in addition to their constituency vote, which they can cast for a  closed party list. There are 5 regions in Wales, each of which elects 4 members to the Assembly taking its full complement to 60. The so-called Additional Member system uses the d’Hondt divisor formula to allocate regional AMs in accordance with the following algorithm:

  1. Party list votes are totalled from each of the constituencies making up the region.
  2. These totals are then divided by the number of constituency seats each party has won – plus one.
  3. The party with the highest resulting total elects one Additional Member.
  4. That party’s divisor is then increased by one (because of its victory)
  5.  Step 2 is repeated with the updated number of seats.; again, the highest resulting total wins a seat.
  6. The process is then repeated until all Additional Members are elected.

The aim of the system is to compensate parties which pile up lots votes in constituencies but fail to win many seats there. Under the d’Hondt system, they are much more likely to gain additional regional members. Conversely, parties which do well in constituency elections will do less well in the top-up seats. The idea is that the final outcome is much more proportional than it would be based on constituency votes alone. It’s not perfect, of course. Welsh Labour won precisely 50% of the seats in the Senedd, but with considerably less than 50% of the popular vote.

This system probably sounds quite complicated – especially considering the difficulty many people seem to have had understanding the Alternative Vote, which is much simpler! – but it is actually fairly straightforward to operate. It does, however produce a few unexpected consequences.

In the 2011 Welsh Assembly elections the constituency vote held up very well for the Conservative Party. This was probably helped by a relatively low turnout of just over 40%, because ensuring the core Tory voters turned out for the poll was probably all the campaign strategists needed to do. However, the unexpected success of the Conservatives in the constituency vote led to one notable casualty when the additional members were calculated. The Conservative leader in the Welsh Assembly, a regional member, Nick Bourne, found himself a victim of the party’s own success: he lost his seat, and the Tories now need a new leader.

In summary, Welsh Labour did pretty well, returning 30 out of the total of 60 Assembly Members, up 4 on the last election. The Conservatives, somewhat surprisingly, were up 2 on 14. It was a bad night for Plaid Cymru, who lost four members to end on 11. The Liberal Democrats did poorly in the constituency vote, losing all but one of their seats, but picked up 4 regional members courtesy of d’Hondt. No other parties won any seats.

What happens next? Labour could try to form a minority administration on their own, but it seems more likely that they will try to find a coalition partner. The previous administration involved a combination of Labour and Plaid Cymru, but the latter did so badly in these elections that they may decide that they don’t want to play anymore. That would make the LibDems favourites, although they might be considered a bit toxic after their poor showing elsewhere in the UK. We’ll just have to wait and see what emerges from the discussions (which have presumably already started). I’ll be following it all with particular interest because, amongst other things, there might be important implications for Higher Education in Wales if Labour go it alone or the LibDems replace Plaid in the governing coalition.

There were, of course, elections going on last week throughout the United Kingdom. I haven’t got time to comment on all the results, but fortunately I found this interesting and informative summary of the situation Nationwide


19 Responses to “Returning”

  1. Claire Says:

    I agree with the Tories trying to get core voters to turn up, the leaflet we had said something along the lines of: ‘Need a lift to the polling station? Email our North Wales team’!
    I live about 200m from where I vote, I was quite tempted to take them up on it.

  2. Tom Shanks Says:

    I think Welsh politics lost something when Rhodri Morgan retired. On Start the Week from 21/12/09 he speculates why Tony Blair didnt initially back him as leader of the Welsh Labour Party. The relevant bit is from 05.20-06.50 mins in…. talks about an out-of-body experience with Rhodri’s mother-in-law!

    BTW I had a great time at the Llandudno NAM – great plenaries plus excellent student and postdoc talks at the parallels. The Californian weather didnt do any harm either! Havent met anyone who didnt enjoy it!

    • telescoper Says:

      Excellent fellow, Rhodri Morgan. He came to the launch party for Herschel and Planck at Cardiff. Not only was he charismatic and enthusiastic about the projects, he also remembered the names of people he met the last time he visited, ten years previously…

    • telescoper Says:

      I also enjoyed NAM. I think the RAS should go to the seaside every year!

    • Bryn Jones Says:

      Ah, yes, the pivotal role of Lionel Blair in Welsh politics of the 1990s. A classic story.

  3. Tom Shanks Says:

    Er… and of course excellent talks from faculty staff at the parallels too!

  4. Bryn Jones Says:

    An odd characteristic of the electoral system of the National Assembly for Wales is that it has the mechanisms of an additional member system of proportional representation, but it does not produce a result that is closely proportional to parties’ support in all cases.

    Thurday saw Labour receiving 37% of the party list vote and get 50% of the seats, the Conservatives 23% of the vote and 23% of the seats. Plaid Cymru 18% of the vote and 18% of the seats, and the Liberal Democrats 8% of the vote and 8% of the seats. Smaller parties, such as the Greens and UKIP, got the remaining 15% of the party list vote but no seats. So the representation of the Conservatives, Plaid Cymru and the Liberal Democrats very closely matched their party list vote, but the Labour Party received appreciably more seats than its party list share, at the expense of the minor parties. (The proportion of the votes in the first-past-the-post constituencies was Labour 42%, Conservatives 25%, Plaid Cymru 19%, Lib Dem 11%).

    The reason for this is that there are too few regional top-up seats (20 in number) to correct the first-past-the-post elections in the 40 constituencies to produce a proportional result. This was the result of a difficult political compromise in the mid-1990s between more pluralist figures in the Labour Party who advocated proportional representation and other Labour figures who wanted a first-past-the-post system only. The need to adopt a somewhat more pluralist system became essential when Labour declared it would put its Scottish Parliament proposals to a referendum, and a referendum became necessary in Wales too.

    In contrast, the Scottish Parliament has 129 members, of whom 73 are elected in first past-the-post constituencies and 56 from party lists in 8 electoral regions. Although this has produced near-proportionality at times in the past, the SNP on Thursday received 53% of the seats on 44% of the regional party-list vote.

    • telescoper Says:


      The problem is really that the “other” parties were so fragmented that they didn’t pick up any additional seats through the regional vote. The Greens, for example, were fairly confident of a seat through the South Wales Central List, but only managed 5.1% of the vote and failed. The upshot of this is that the additional members did fine by the Conservatives, LibDems and Plaid Cymru but didn’t help any of the smaller parties, their votes effectively giving Labour extra seats.


    • Bryn Jones Says:

      Yes, that’s right, the votes of the smaller parties were fragmented. I am therefore untroubled that the system did not reward the minor parties with seats (although I do feel pity for the Greens in particular). It is odd that only the largest party (Labour) benefited from the exclusion of the minor parties, although it does boost the concept of `strong government’ somewhat.

      There are some legitimate arguments that electoral systems do not reward very small parties (although there are some arguments to the contrary as well, particularly democratic representation and fairness).

  5. Thanks for explaining how the Additional Member system operates. I had been wondering how an AM could possibly lose his seat by doing “too well” in an election – a seemingly bizarre anomaly they made no attempt to explain on the BBC.

    It’s amazing how politicians can be reasonably creative in dreaming up such an algorithm… if only because they are desperately trying to avoid actually implementing full-blown PR.
    (Although I do note it was a mathematician who devised the method.)

    • Bryn Jones Says:

      The case of Nick Bourne, leader of the Welsh Conservatives, is peculiar, and illustrates the strange coupling between results in individual constituencies and in the regional party lists. He lost his seat in the Mid and West Wales region because the Conservatives won three constituency seats and therefore the party did not pick up any seats from the regional list.

      The Conservatives took Montgomeryshire from the Liberal Democrats, and therefore did not get a regional seat unlike in previous elections. This may have been partly due to the strength of the Conservative campaign in Montgomeryshire, but was also due to the weakness of the Liberal Democrats there. The previous Liberal Democrat assembly member was involved in a court case following a drunken night out, and its former representative in the House of Commons was Lembit Opik, famous for his media profile and personal life. The irony here is that a loss of Liberal Democrat support following one Lib Dem’s criminal charge and another’s involvement with a Cheeky Girl contributed to the leader of the Welsh Conservatives losing his assembly seat. That is really odd.

      This would not have happened with a system such as the single transferable vote, as advocated by the Richard Commission of 2004.

    • telescoper Says:

      Just to clarify the point, the Tories did well in the consituency elections in Nick Bourne’s region which meant that they accumulated enough seats that way to give them a fair representation on the Senedd. They therefore lost the regional seat they got last time. It may seem strange, but it does make sense when you think about it.

      The really strange thing was that the Conservative leader was a regional AM and apparently a smart strategist. He might have seen this coming!

      The problem with any truly proportional system is that it removes the link between the representative and the constituency. A PR system would either have to have very large constituencies with many members, or treat the whole electorate as if it were homogeneously distributed. Even a small country like Wales has enormous regional variations in attitudes and voting intent, which is reflected in the fact that there are twice as many constituency members as regional ones. My feeling (as a foreigner living in Wales) is that the consituencies as they are have more meaning than the regions so I think the 2:1 balance is not unreasonable.

      One could suggest increasing the number of AMs to 80 by doubling the number of regional seats, which would give smaller parties a better chance of representation and would also give more room to correct the lack of proportionality in the constituency vote. But does Wales really need an Assembly of 80?

    • Bryn Jones Says:

      Yes, I think Wales needs an assembly with 80 members, or even more. 80 members were advocated by the Richard Commission of 2004.

      A legislature with executive powers must have sufficient people of calibre to form the cabinet of a government and to scrutinise that government. The governing party/parties in the National Assembly for Wales needs/need in excess of half the assembly members for stability, requiring at present at least 31 of the 60 members. The governing parties must find about 14 talented people from their members to provide ministers and deputy ministers. Increasing the total size of the National Assembly to 80 or 100 would provide a greater pool of talent to fill those ministerial posts.

      Having a larger Assembly would also increase the collective experience of the members. It would increase the chances of having people within it with experience of a diverse range of policy fields, such as business, industry, schools, health, universities and agriculture. This would improve the scrutiny of the Welsh Assembly Government.

      The Scottish Parliament has 129 members and the Legislative Assembly of Northern Ireland has 108 members.

    • telescoper Says:

      Here’s an interesting exercise for you:

      Assume each region has 8 regional seats rather than 4. Assume the votes cast in the regional ballot were the same as in the last election. Use D’Hondt’s method to calculate the outcome for an 80 member Senedd.

    • telescoper Says:

      I just did this for South Wales Central, in which Labour won all 8 constituencies. According to my quick calculation (which may be wrong), the regional members elected assuming the same share of the vote as happened in the 2011 election would have been: 4 Conservative, 2 Plaid Cymru, 1 Lib Dem and 1 Green.

    • Bryn Jones Says:

      Right. I tried last night to repeat the d’Hont mechanism calculation for the regional party lists for Thursday’s election to the National Assembly for Wales. I tried to use the process as specified in Part 1, Sections 8 and 9, of the Government of Wales Act (2006).

      (Actually this is not trivial: the wording of the Act is rather abstruse. For example, it states:

      (a) there is to be added together the number of electoral region votes given for the party in the Assembly constituencies included in the Assembly electoral region, and

      “(b) the number arrived at under paragraph (a) is then to be divided by the aggregate of one and the number of candidates of the party returned as Assembly constituency members for any of those Assembly constituencies.”

      It’s all rather like that.)

      Anyway, my calculations give the same result as the declared results, which is unexpected but is reassuring: the returning officers performed the calculations correctly.

      I then went on to consider the hypothetical case of there being an Assembly of 80 members with 40 first-past-the-post constituencies (as at present) and 40 regional party list members (rather than 20). This extends Peter’s calculation for South Wales Central to all five regions of Wales. Each region has 8 party list seats (rather than the 4 we have in reality).

      The results are quite interesting. Yes, I agree with Peter’s calculation, and I also find that a Green would have been elected in South Wales Central, although it is only the last allocation of regional seats that elects a Green: the Greens only just get a place even with 8 regional list members. No Greens are elected in any other regions. The only other minor party that gets seats is UKIP, which succeeds in North Wales (only at the 8th allocation) and South Wales East (the 7th position).

      With 40 regional list seats, I calculate that the hypothetical Assembly would have the following regional list members: 8 Labour, 14 Conservative, 10 Plaid Cymru, 5 Liberal Democrat, 2 UKIP, 1 Green. The total composition of the Assembly would be:

      36 Labour seats (45% of total)

      20 Conservative (25%)

      15 Plaid Cymru (19%)

      6 Liberal Democrat (7%)

      2 UKIP (3%)

      1 Green (1%)

      The numbers of Labour, Conservative and Plaid Cymru members in this hypothetical Assembly closely matches their level of support in the regional ballots, but there is some under-representation of the Liberal Democrats. Such a large number of regional list seats would have let the United Kingdom Independence Party into the Assembly, which would have been interesting given their very odd policy towards the Assembly: the party wants to abolish Assembly members and replace them with members of the House of Commons who would be sent over to Cardiff from London for one week in four!

      The Richard Commission of 2004 advocated an Assembly of 80 members elected through the single transferable vote in multi-member constituencies (electing 3 to 6 members each).

    • telescoper Says:

      Interesting. So we’d still be looking at a Labour-LibDem coalition as the most likely administration…

    • Bryn Jones Says:

      A Labour-Liberal-Democrat coalition would be one possibility with 40 regional members, or a One Wales mark 2 (Labour and Plaid Cymru as there has been for the past 4 years). It would be less likely that a Labour-only administration could survive than as present. So, in practice, things would be quite similar to how they are now.

      A “rainbow coalition” of the Conservatives, Plaid Cymru and the Liberal Democrats would be difficult in an Assembly with 40 regional seats (even if Plaid Cymru could stomach working with the Conservatives).

      Of course, Alex Salmond managed to run a minority administration in Scotland for 4 years. But then Alun Michael tried in Wales in 1999 but failed and was thrown out of office.

  6. telescoper Says:

    Another thing worth mentioning is that Wales (like Scotland) actually has four significant parties (Labour, Conservative, LibDems and Plaid Cymru). The ballot papers for many constituences (mine included) only had four names on it, whereas the regional list was very long indeed. I didn’t vote for the same party in the regional vote as I did in the constituency poll. It would be interesting to correlate voting patterns in the two polls, but that’s impossible because the ballot papers are counted separately.

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

%d bloggers like this: