Skip to main content
Coronavirus (COVID-19): Latest updates.


Message in a Ballot Box

21 December 2015

This article was written by Aled Morgan Hughes, a PhD student in the Wales Governance Centre at Cardiff University. You can follow Aled on Twitter: @AledMorganH


Forgiveness is on offer to those who might have forgotten that Wales’ next Government won’t be the only thing decided at the polling box on the 5th of May, 2016. It will also be judgement day for Wales’ four Police and Crime Commissioners (PCC).

It’s fair to say that the idea of voting for their PCC hasn’t exactly caught the imagination of the Welsh and English electorate. When the first elections were held, in November 2012, turnout across the four Welsh Police areas was a measly 14.9%. And one polling station in Bettws, Gwent, proudly achieved a 0% turnout! Recent PCC by-elections in England haven’t suggested any growing interest either, with a turnout of 14.8% in South Yorkshire last November – albeit this was an increase from the 10.4% witnessed in the West Midlands in August 2014. Some people remain opposed to the existence of PCCs, while others complain about their cost, the decisions taken by PCCs, or the occasional scandal.

Across Wales in 2012 it was, unsurprisingly, Labour who came out on top. The party was riding high in the Welsh opinion polls at the time, and they won 144,805 of the First Round votes. (As you may recall, PCC elections use the Supplementary Vote system). Across the regions, however, Labour’s fate proved mixed. Former First Minister (sorry, Secretary) Alun Michael won in South Wales. But it was a different story for his son, Tal, in North Wales region, where he lost to independent candidate Winston Roddick. There was a similar story in Gwent, where another independent candidate, Ian Johnston, claimed victory. And in Dyfed-Powys Conservative Christopher Salmon narrowly overcome former Labour AM and Agriculture Minister, Christine Gwyther.

However, PCC elections in 2016 are likely to be very different from the 2012 vintage, for three main reasons.

Turnout: Although not reaching the levels we find in Westminster general elections, turnout for National Assembly elections has averaged a little over 40%. Whatever people think about the idea of PCCs, therefore, we are likely to get a far greater number of people voting in the elections than in November 2012.

Party Competition: Another major difference from 2012 will be the increased level of party competition for the roles. Plaid Cymru refused to put up candidates for the election in 2012, a decision that may have helped the independent candidates. But in 2016 Plaid will field candidates in all the four regions. It is also likely that the Liberal Democrats will field candidates; while representatives of UKIP (who stood unsuccessfully in North Wales last time) may also stand.

Political Context: As well as raising turnout, holding PCC elections on the same as the Assembly poll will also shape the broader political context in which these elections are held. Two of the four races were won in 2012 by independent candidates. Success may be harder for such candidates against a general political context dominated by the parties campaigning for the Assembly election. We don’t yet know much about how voting patterns may differ between the Assembly and PCC elections, particularly when they are held simultaneously.

On this latter point, we can at least look at how people voted in past Assembly elections, and try to map these results onto the four PCC regions – which, of course, differ from the five regions we have for electing party list representatives in the Assembly. In the following tables, I map the results of the constituency vote in the 2007 and 2011 National Assembly election as closely as I can onto the PCC boundaries. (The main complication is the Merthyr Tydfil and Rhymney, the majority of which is in the South Wales PCC region, but some of which is in Gwent. I hope no offence is caused by my including Merthyr Tydfil and Rhymney in South Wales here!)

 National Assembly 2007 Results (Constituency Results by Police Region)

Police Region Labour Conservative Plaid Cymru
North Wales 60,767 53,799 63,492
Dyfed-Powys 39,102 45,855 64,400
South Wales 156,964 80,767 69,272
Gwent 58,092 38,309 21,898


National Assembly 2011 Results (Constituency Results by Police Region)

Police Region Labour Conservative Plaid Cymru
North Wales 73,162 63,465 54,085
Dyfed-Powys 46,619 52,376 51,901
South Wales 199,668 83,377 55,525
Gwent 78,750 37,615 21,196


This exercise indicates that, had people voted in PCC elections in remotely similar ways to how they voted in the National Assembly, Labour would have won both the Gwent and South Wales PCC regions comfortably in 2007, and even more so in 2011. This suggests that Labour must start as strong favourites in both regions in 2016.

The picture looks much less clear-cut in the two other regions. Plaid Cymru must fancy their chances in Dyfed-Powys. Not only did they top the constituency vote there in 2007, and come very close in 2011; this area will include both current strongholds (Ceredigion and Carmarthen East & Dinefwr) and target seats (Llanelli and Carmarthen West and South Pembrokeshire) in the Assembly election; Plaid’s levels of activism within the area should, therefore, be strong. However, the Conservatives came top of the constituency vote in this area in 2011, and with their support levels in Wales generally looking robust, plus with whatever benefits come from PCC incumbency, they will have realistic hopes of holding the region.

North Wales also promises to be an interesting race. In the 2011 Assembly election, Labour came out on top here comfortably, whilst Plaid (who won the most constituency votes here in 2007) slipped to third behind the Conservatives. This time round, the region could well be within the grasp of all three parties. And the incumbent independent candidate will probably also be standing! North Wales could well be the region where second preference votes are most likely to decide the matter.

Thus, when we look forward to next May, most people in Wales are currently focussing on the National Assembly election. It may grab most of the headline. But the Police and Crime Commissioner elections could indeed prove to be an interesting sub-plot!


  1. Christian Schmidt

    With North Wales potentially having 4 viable candidates the issue that the Supplementary Vote system only offers a second preference (compared to unlimited preferences under Alternative Vote) starts playing a role. Voters in the North may end up with their worst (out of the top four) choice because they vote for the wrong middle-choice…

  2. David Willis

    Sorry to go off topic but how did the UK end up with so many different voting systems?

    First past the post for Westminster and English and Welsh local government
    STV in Northern Ireland and Scottish Local government
    Additional member system for the Scottish Parliament
    A slightly different Additional member system for the National Assembly
    The dHondt list system for the European Parliament
    And now the Supplementary vote system for PCC’s
    Have I missed any?

    Do you see a potential problem with two elections being held on the same day, using two different voting methods, one of which involves two different ballot papers?

    • Roger Scully

      I think you’ve got most of them, David…

      Yes, the two different systems could be problematic. Not really a very good idea. Would have been better to combine PCC and next year’s local elections, in my view.

  3. oldnat


    I note that YouGov now includes “Place of Birth” as a weighting factor for Scottish polls – after discovering during the referendum campaign, that those born in the rest of the UK were over-represented in the sample, and that did distort the results a bit.

    I wonder if that has been investigated for use in Welsh polling? It may be that there is no difference in party support on that basis, but it would be useful to know.

    If there is a difference, then it might be worth weighting for that in Wales as well.

  4. J.Jones

    I agree Oldnat. Recently I asked a similar question of YouGov with regard to Welsh speaking ability and place of birth could easily be a significant factor.

    YouGov say that their present weighting gets a sample that, when analysed, actually gives a representative sample for other factors. In the case of the recent question that we asked it’s obvious that ability to speak Welsh would be a major factor and so it proved. The sample actually did over represent Welsh speakers by 4%. This is what YouGov had to say:-

    ” We don’t include Welsh speakers in our sampling and weighting scheme but rather rely on the natural fallout to be broadly similar to the actual figure. From my calculations the weighted percentage in the sample is 19%, which obviously still isn’t the exact census figure but is as close as we would expect when it is not part of our standard sampling frame.”

    So I would guess that the same would be assumed to be true for place of birth. Worth checking though.

    • oldnat

      J Jones

      Thanks. Interesting response from YG.

      After the May polling debacle in England (outwith London), one might have expected the pollsters to be aware that their “standard sampling frame” needs improvement!

      I suspect that client demand is important here. It’s cheaper for companies not to make changes, unless pressured. In Scotland there may simply be more polling than in Wales. It’s certainly clear from the number of folk reporting (on online chat rooms) in Scotland that they have been polled, but no polls are published, that there is a lot of private polling going on here.

      • Roger Scully

        Well, I suspect there probably is some private polling in Scotland. But I suspect that some of the polls being conducted are YouGov ‘test’ polls, prior to them rolling out their revised methodology in the new year.

        On your and JJ’s comments on sampling and weighting, if you don’t mind I’ll forward them to my colleagues at YG for their consideration.

  5. oldnat


    No problem – though if my comments go to Anthony Wells, he’ll be more inclined to snort “Him, again?” 🙂

    Ta for the info on the YG test polls. Anthony did say I’d have to be patient and wait for details of how YG are going to deal with weighting the Scots sample in GB polling.

  6. J.Jones

    By all means Roger. Just as when considering polling for the Scottish referendum it was evident that place of birth was a significant factor, here in Wales place of birth may also be a factor and ability to speak Welsh clearly is. Both these measures have robust census information upon which to base weightings. Why not use them?

Leave a reply

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