Government and Politics


The 2011 Welsh Referendum

Professor Roger Scully

Democracy means government by the people. This is normally understood in practice as representative democracy: the people voting in elections to choose representatives to occupy key positions of power and take the major decisions. But increasingly around the world, elections are being supplemented by a more direct form of democracy, the referendum, where the people themselves decide. One such referendum was held in Wales on 3rd March 2011.


A lot is known about democratic elections; rather less about the direct democracy of referendums. Here we will examine the Welsh referendum, as an example of what happens when the politicians let the people decide.

The 2011 Welsh Referendum: the Background

The referendum on 3rd March 2011 was the third Wales-wide referendum to be held. However, as Table 1 shows, it was one of a rather larger number of major referendums to have occurred in the UK. As with Wales’ two previous referendums, it concerned devolution. In 1979, the people of Wales had overwhelmingly rejected the idea of having a National Assembly and partial self-government within the UK. By 1997, however, opinion in Wales had moved a long way. This time the people voted narrowly in favour of devolution.


Table 1: Major Referendums in the UK


Territory Year Issue Turnout %
Northern Ireland 1973 Northern Ireland remaining in UK 58.7
UK 1975 European Community membership 64.5
Scotland 1979 Devolution 63.8
Wales 1979 Devolution 58.8
Scotland 1997 Devolution 60.4
Wales 1997 Devolution 50.1
Northern Ireland 1998 Good Friday Agreement 81.0
London 1998 Elected Mayor & GLA 34.1
North East England 2004 Devolution 47.7
Wales 2011 Further devolution 35.6
UK 2011 Voting reform 42.0


The 2011 poll was rather different in nature to the earlier ones. It wasn’t about setting up a National Assembly – that had already happened – or about whether the Assembly should continue to exist. Rather, the 2011 vote concerned what powers the Assembly should have. After the 1997 referendum, the Assembly had been given responsibility over twenty areas of government activity (including much of education, economic development, health and the Welsh language). But the power to make new laws in these areas remained with the Westminster parliament. Welsh devolution was thus very different from that in Scotland: there, the Scottish Parliament was given both the responsibility for policy and law-making powers. The 2011 referendum asked voters whether the National Assembly should be given full law-making powers in the twenty policy areas it was responsible for. Thus, the referendum was not about whether the breadth of the Assembly’s responsibilities should be extended, but about the depth or strength of its powers in areas where it already had responsibility. A Yes vote would make the National Assembly more powerful; a No vote would leave the Assembly with the rather limited powers it already possessed.


The Referendum Campaign

A notable feature of the 2011 Welsh referendum was the imbalance between the two campaigns. The Yes campaign (Yes For Wales) won support from all four parties in the National Assembly (although the Conservative party as a whole remained formally neutral); a Yes vote was also supported by many public and civil society organisations. The Yes campaign successfully established a campaign organisation across most of Wales. Those supporting Yes tended to argue that the warnings of devolution’s opponents in 1997 had not come true: devolution had not divided Wales or alienated it from the rest of Britain. But the Assembly’s ability to tackle Wales’ problems, they claimed, was limited; a Yes victory would give the Assembly the powers to deliver effective policies, and mean that there were no excuses for those in power in Cardiff failing to deliver.


The No campaign (True Wales) won support from almost no senior political figures or major organisations. Its major campaigners tended, therefore, to be less politically experienced than those on the Yes side, and proved less effective at building a functioning campaign organisation across Wales. No campaigners were often people who had opposed devolution in 1997, and many clearly still rejected it. Abolishing devolution was not an option in the referendum – a No victory would leave in place what many No campaigners had opposed in 1997. But a No vote would prevent Welsh devolution going further. True Wales often argued that the Assembly had delivered very little with the powers it already had – so why give it more powers? They also suggested that strengthening devolution would push Wales further along a ‘slippery slope’ leading to the eventual break-up of the UK.


Campaigners on both sides found it difficult to explain to people the importance of the referendum. No clear issue of principle was at stake; the question being asked was more subtle and technical. Given this, and given also the weakness of the No campaign, it is little surprise that the campaign struggled to engage with most voters. Fewer than 1 in 10 of those surveyed by the Welsh Referendum Study (WRS) recalled being contacted by either campaign or anyone else about the referendum, while only just over 1 in 4 felt they had been given ‘enough information to make an informed choice’. But of the two sides, the Yes campaign was clearly the more effective. When asked which side they thought had run the best referendum campaign, almost eighteen times as many WRS respondents chose Yes as chose No.


The Result

The 1997 referendum had shown Wales to be split down the middle on devolution. Half the electorate didn’t vote; the rest divided almost evenly, with 50.3 percent voting Yes and 49.7 percent No. Eleven of Wales’ Local Authority areas supported devolution; eleven opposed it. The 2011 result was very different. Turnout was even lower: indeed, at only 35.6 percent, it was the second lowest yet in a major UK referendum (see Table 1). This partly reflects a general decline in voting participation in recent years; but it also reflects the lacklustre nature of the referendum campaign, and the difficulties many voters had in perceiving anything important being at stake.


Those who did vote delivered a clear Yes victory. The Yes side won by almost two-to-one across Wales as a whole, and in 21 of 22 Local Authority areas. (Only Monmouth, very narrowly, voted No). And, as Table 2 shows, the Yes vote increased from 1997 most in areas which had rejected previously devolution. This was consistent with findings from public opinion research, which had been showing for some time that opposition to devolution had fallen most among those social groups, and in those parts of Wales, which had been most strongly opposed to it in 1997.


Table 2: % Yes vote by Local Authority, 1997 & 2011


Local Authority 1997% Yes 2011 % Yes % Change
Flintshire 38.1 62.1 24.0
Denbighshire 40.8 61.8 21.0
Wrexham 45.3 64.1 18.8
Conwy 41.0 59.7 18.7
Newport 37.4 54.8 17.4
Monmouth 32.1 49.4 17.3
Cardiff 44.4 61.4 17.0
Vale of Glamorgan 36.7 52.5 15.8
Ynys Mon 50.9 64.8 13.9
Bridgend 54.4 68.1 13.7
Blaenau Gwent 56.1 68.9 12.8
Torfaen 50.1 62.8 12.7
Pembrokeshire 42.8 55.0 12.2
Rhondda Cynon Taff 58.5 70.7 12.2
Gwynedd 64.1 76.0 11.9
Swansea 52.0 63.2 11.2
Merthyr Tydfil 58.2 68.9 10.7
Caerphilly 54.7 64.3 9.6
Powys 42.7 51.6 8.9
Ceredigion 59.1 66.2 7.1
Neath/Port Talbot 66.6 73.0 6.4
Carmarthen 65.3 70.8 5.5
WALES 50.3 63.5 13.2


(Note: areas which had voted No in 1997 in italics)

Yes or No?


So why did the Yes side win the 2011 Welsh referendum?


How people vote in elections is well understood to reflect, in most instances, a mixture of long-term influences – such as deep-rooted identities that people may have with a social group or a party – and shorter-term factors – such as current evaluations of the performance of parties in government, or their assessments of party leaders. Referendum voting is less well understood. Referendums are still much less frequent than elections in most democracies, and the nature of the choice before voters is very different. Referendums present far fewer options (Yes, No or not voting) than most elections do, with people voting not for a person or party but a specific outcome.


Previous studies of referendums have identified several factors that often seem to shape how voters decide.


Party Cues: Parties are not ‘on the ballot’ in referendums. However, the positions that parties take in referendums often seem to influence how people vote. If voters lack much knowledge about the issue being voted on, they often seem willing to be guided by what major parties and political leaders do. Popular parties and leaders can attract support to causes they endorse. But equally, an unpopular party or politician may prompt voters to oppose something: Nick Clegg’s support for the Yes campaign in the May 2011 UK Alternative Vote referendum was believed by Yes campaigners to have lost them many votes.


For parties to cue voters in this way, however, requires that the voters actually know what the parties’ positions are. In 2011, the rather lacklustre campaign left many voters confused about at least some of the parties. As Table 3 shows, Plaid Cymru were the only party whose stance was understood by a clear majority. The cues offered to voters by the other parties were much more limited.


Table 3: Party Cues in the Referendum (%)
Which way do you think [PARTY] recommended people should vote in the referendum?


  Conservative Labour LibDems Plaid
Yes 23 52 32 70
No 18 3 6 1
Did Not make a recommendation 15 7 12 4
Don’t Know 44 39 49 25


Performance Evaluations: Much analysis of elections points to the importance of performance and policy delivery, arguing that judgements on the overall competence of the rival parties are the most important factor shaping election voting. Performance considerations have sometimes been found to be important in shaping referendum voting, and could certainly be expected to shape voting in the 2011 Welsh referendum. Voters were being asked to give the National Assembly enhanced powers; their willingness to do this might well be affected by how well they thought it had done with the powers it already possessed. But giving more powers to politicians in Cardiff meant, to some extent, taking those powers away from Westminster. So judgements about the recent performance of the UK government might also be relevant.


As Figure 1 shows, those with overall positive evaluations of the performance of the Assembly government were much more likely to vote Yes in the referendum than those with negative evaluations. But it is also clear that evaluations of the UK government were relevant. If you had a negative view of the UK government’s performance, you were more likely to vote to transfer greater powers from London to Cardiff.


Figure 1: % Yes Vote & Government Performance Evaluations


Figure 1: % Yes Vote & Government Performance Evaluations


Influence of the Campaign: Studies of referendums around the world have often shown big changes in voting intentions during the campaign period leading up to the vote. As the voters are made more familiar with the issues at stake in a referendum, they often change their initial views. This particularly tends to happen in places (such as some states in the USA) where campaigns are able to spend substantial amounts of money prior to the referendum. In the UK, referendum spending is tightly controlled. And, as we saw earlier, the lacklustre campaign in Wales in 2011 struggled to connect to most voters. So it is no surprise that little changed during the campaign period (beyond the number of undecided voters slowly falling). The Yes campaign had a clear lead in all opinion polls published during 2010 and early 2011. And as we see from Figure 2, nothing happened during the final four weeks of campaigning to change that.


Figure 2: Referendum Voting Intention (Rolling 3-day average) %


Figure 2: Referendum Voting Intention (Rolling 3-day average) %


Identity Politics: As mentioned earlier, the 1997 devolution referendum had shown Wales split down the middle on devolution. A major factor behind those divisions was national identity. Those with a more Welsh identity were much more likely to have supported devolution than those with a British identity. Although subsequent research had shown rapidly declining opposition to devolution among British identifiers, it was far from clear that such people would be willing to go as far as supporting even stronger devolution for Wales. And as Figure 3 shows, some differences did remain, with more strongly Welsh identifiers being clearly more likely to vote Yes.


Figure 3: % Yes Vote by National Identity


Figure 3: % Yes Vote by National Identity


Two other differences which were seen in the 1997 referendum persisted in 2011. In 1997, men had been several percentage points more likely to vote Yes than women; this was still the case in 2011. And as Figure 4 shows, older voters remain much the most reluctant to support greater devolution – although even among them, a narrow majority of those voting chose to vote Yes.


Figure 4: % Yes Vote by Age Group


Figure 4: % Yes Vote by Age Group


Views on the Issue: Last, but certainly not least, how people vote in a referendum can depend on what they actually think about the issue on the ballot. Referendums sometimes concern an issue on which many people have clear and deeply-held views, although often this is not the case. The 2011 Welsh referendum, as we have discussed, concerned a rather difficult-to-grasp question. But this question linked directly to a much more basic issue: whether Wales should have some political autonomy within the UK and, specific to this referendum, how far-reaching that autonomy should be. Opinion surveys had, for several years, been suggesting significant public support for strengthening devolution, very much in line with the eventual referendum result. And WRS evidence indicates that over 95 percent of voters saying that they favoured independence for Wales, or thought the Assembly should have more powers, voted Yes in the referendum; well under 20 percent of those who believed that devolution should be abolished, or the Assembly’s powers reduced, voted No. In short, how people voted in the referendum was very closely linked to their broader views about how Wales should be governed.


The direct democracy of referendums remains less frequently used than the representative democracy of elections, in the UK and in most of the democratic world. But the use of referendums is certainly increasing, with the Welsh referendum of March 2011 being but one instance of a global trend. The Welsh referendum offered voters a question that was difficult for many to engage with, and a campaign that failed to engage or interest many of them. Unsurprisingly, turnout at the referendum was low. But those who voted delivered a clear mandate for giving the National Assembly greater powers. This confirmed the Assembly as an important part of government and politics in Wales – it is an institution whose influence over the lives of people in Wales is only likely to grow in the next few years.


The Welsh Referendum Study interviewed a 3029 person representative sample of the Welsh electorate prior to the referendum, and then re-interviewed 2569 of them immediately after the referendum. The study was conducted with the support of the Economic and Social Research Council (grant RES-000-22-4496); interviews were conducted via the internet by the survey agency YouGov.