Referendums

This section highlights trends about the sequence of UK referendums, topics, and voter turnout across our period of study at the UK and devolved level. As presented in Table 1, referendums in the UK started, in earnest, in the 1970s.Footnote ² These cases were primarily instigated to avoid party splits (Qvortrup Reference Qvortrup2015). After four referendums in that decade there was a break under Conservative governments with none until New Labour in 1997, with various referendums central to their manifesto. More recently the Conservative party has become more accustomed to referendums, initially, as a compromise with the Liberal Democrats to consult on electoral reform in 2011, followed by the Scottish Independence referendum in 2014 and the Brexit referendum in 2016, the latter case being a return to the motivation behind the 1970s referendums, ie, to address party splits. Overall, there has been limited use of referendums – with only 3 UK-wide votes and 10 within the regions – especially compared with countries such as Ireland and Switzerland.

Table 1. National and regional referendums in the UK

Source: Electoral Commission.

There is greater use of referendums at the local level. Again, starting in the 1970s, the Local Government Act 1972 enabled citizens to initiate referendums in parish councils (the most local level of government in the UK at the time). Building on this through the Local Government Act 2003, local authorities in England and Wales were given powers to have nonbinding polls on issues such as service provision and council tax (Qvortrup Reference Qvortrup2015).

Three issues have dominated the UK’s referendum agenda, with European Union (EU) membership dominating the UK cases and national independence and devolution being the primary focus of the regional ones. Gordon (Reference Gordon2020) highlights the importance of focusing on occasions when referendums in the UK have not been held and details the EU and devolution legislation as passed without a public ballot.

Voter turnout is pivotal to the legitimacy of a referendum. Table 1 details the voter turnout at each. It is apparent that this varies depending on context, such as the salience of the issue to the electorate and the predictability of the outcome. Laycock (Reference Laycock2013) demonstrates how voting and campaigns in UK referendums differs to voting in elections. Firstly, party cues and identification are weaker influences on the vote in referendums. Party cues are less consistent in referendums than elections. Voters are also less likely to use a referendum vote to punish the government than in an election. Ultimately, campaigns matter more in referendums than elections in terms of influencing how people vote, especially if the campaigns are intense. Issue preference of voters is the key determinator, without being definitive. Despite their influence, UK referendum campaigns are generally thought to be low quality and not deliberative, include misleading and inaccurate information, and ultimately fail to enable voters to make informed choices, and the Electoral Commission does not have powers to intervene (Qvortrup Reference Qvortrup2015; Organ Reference Organ2019). The campaigns in the 2016 Brexit referendum are seen by some to be particularly divisive (Organ Reference Organ2019) and have led to continued polarisation among the public and politicians (Henderson, Jeffery, Wincott et al. Reference Henderson, Jeffery, Wincott and Jones2017). This highly controversial referendum may have dampened some enthusiasm or ‘accommodation’ for referendums in the UK precisely because it has shown they are not always benign instruments (Gordon Reference Gordon2020), but it is also hard to envision significant legislation on fundamental constitutional issues being passed in the UK without one because of these precedents.

To further understand the nature of referendums in the UK, we can employ institutional forms to analyse them. Legally referendums can be preregulated, where constitutional rules require them, or ad hoc, where they are discretional (Setälä Reference Setälä2006). The British referendums have all been ad hoc without ‘clear criteria which defined when a referendum would (or would not) be held’ (Gordon Reference Gordon2020: 214). This is starting to change via legislation to preregulate their use. For example, the European Union Act 2011 (now repealed because of Brexit) required changes to the powers of the EU to be subject to a referendum, and acts have been passed which require any change to the status of Northern Ireland, Scotland, or Wales to be put to a referendum (Gordon Reference Gordon2020).

A further distinction is between ‘decision-promoting referendums’, where the same political actor instigates and sets the agenda for the referendum, and ‘decision-controlling referendums’, where these roles are divided (Setälä Reference Setälä2006). In the UK they are the former and are controlled by the government, reducing their potential to deepen democracy, as ‘when governments control the referendum, they will tend to use it only when they expect to win’ (Lijphart Reference Lijphart1984: 203). There is therefore a case for introducing citizens’ initiatives to enable the UK public to influence the political agenda and make the use of referendums less top down.

Referendums can also be advisory or make binding decisions (Setälä Reference Setälä2006). In the UK, referendums are advisory. The exception was the referendum on electoral reform, as the Parliamentary Voting System and Constituencies Act 2011 made it a ‘binding referendum’.Footnote ³ Politically, though, it is very difficult for parliaments to ignore referendum results, even in the UK (Coulter, McKee, Pannell et al. Reference Coulter, McKee, Pannell and Sargeant2024).

Clearly, there is scope to go further in passing referendum legislation to embed and regularise this democratic innovation in the UK political system to make their use less ad hoc (Gordon Reference Gordon2020). This could include legislation on when to have referendums; how the campaigns are conducted to instil deliberation (Organ Reference Organ2019), including regulations on campaign funding; how the results are implemented; and whether there are rules on turnout and the size of the majority required. Organ (Reference Organ2019) also advocates for the use of mandated citizens’ assemblies (CA) in the campaigns to help inform voters. It is to a consideration of mini-publics to which we now turn.

Mini-publics

Mini-publics emerged in the UK in the 1990s, but the historical trajectory across space and time also varies according to type. The most prevalent types of mini-publics have been deliberative polls (DP), citizens’ juries (CJ), and CA. We consider the trajectory of each type in turn.

In some respects, the UK has been a global leader in the use of mini-publics, most notably the DP format (which have larger and more statistically representative participant samples but also tend to be shorter in duration and culminate in an aggregation of participants’ opinion). The world’s first DP was in the UK in 1994 on crime where proof of concept was achieved (Davidson and Elstub Reference Davidson and Elstub2014). To date, the UK is level with Japan as the country to have the second most DPs (8), behind only the USA, who are way out front (60). The most recent addressed immigration and regulation post-Brexit (Curtis, Davies, Fishkin et al. Reference Curtis, Davies, Fishkin, Ford and Siu2021), making it the only DP in the UK since 2010 despite their continuing dispersion around the globe.Footnote ⁴

There are several causes of this fate. First, the DP model is copyrighted, placing barriers to replication and institutionalisation. Second, UK DPs were only ever loosely connected to existing political and administrative institutions, and this meant they failed to achieve any real impact, which has affected their status. Third, the first few DPs were partly broadcast on national television, demonstrating the interest in the format. The fact that this level of media coverage was not extended to the subsequent cases indicates that this interest has waned (Davidson and Elstub Reference Davidson and Elstub2014). Fourth, as we will see below, the DP model has been surpassed by other mini-public forms in the UK.

CJs (which have a fewer number of participants than DPs but often last longer and culminate in agreed upon policy proposals) started in the UK at a similar time in 1996 and became established ‘in a remarkably short space of time’ (Delap Reference Delap2001: 39). Considerably cheaper and easier to replicate than DPs, there have been hundreds of cases; they are especially run by local authorities and public agencies on an array of policy issues, with a concentration in health initially (Delap Reference Delap2001: 39) and climate more recently. Furthermore, the New Labour government used CJs to inform their own policy,Footnote ⁵ and Gordon Brown argued they should be used more by local government (Davidson and Elstub Reference Davidson and Elstub2014).

CJs in the UK have varied greatly in policy influence, as they have nearly all been ad hoc, which promotes cherry-picking of recommendations (Elstub and Khoban Reference Elstub and Khoban2023). There are some exceptions. There has been a permanent CJ since 2002 offering proposals on moral and ethical issues in the National Institute for Clinical Health Excellence (NICE), a nongovernmental agency that produces clinical guidelines on new interventions (Rawlins Reference Rawlins2015) as part of the governance-driven democracy (Warren Reference Warren2009) response to depoliticisation. Having run several ad hoc CJs since 2019, the Scottish parliament is seeking to embed these processes within its committee system (Elstub Reference Elstub, Schwarzmantel and Beetham2024).

The CA format is also increasingly popular in the UK, especially on the topic of climate change (King and Wilson Reference King and Wilson2023).Footnote ⁶ A CA tends to have the larger number of participants that characterise DPs and last even longer than CJs. Since the first CA on regional devolution in England in 2016,Footnote ⁷ they have been used often by local governments. At this level, they usually have fewer participants and a shorter duration, so the distinction between CAs and CJs is narrowing, and the format of the CA stretching and diluting. One of the reasons why CAs have been so popular in the UK is the country’s proximity to Ireland, where high-profile CAs occurred on same-sex marriage and abortion between 2012 and 2018.

CAs are also used at the national and regional levels, especially by legislatures. The UK parliament and all the regional parliaments have commissioned, or have run, some form of CA/CJ in recent years. The influence of these CAs has been variable because of their largely ad hoc nature. There has also been a turn towards institutionalisation of CAs to address this. Firstly, in public administration in England, as another consequence of depoliticisation, via NHS Citizen a CJ and CA were used in conjunction with other participatory processes. NHS Citizen did not last more than one iteration though (Dean, Boswell, and Smith Reference Dean, Boswell and Smith2020). More recently, some councils in London (Newham and Camden) are seeking to imbed CAs into their governance structures following some ad hoc cases. The most ambitious plans for institutionalisation of CAs are from the Scottish government (2022), although these have not yet been implemented.

The UK has been a world leader in mini-publics with respect to the number of cases and in some aspects of their design. However, the mechanisms to afford them impact and publicity have been variable and limited. Some attempts at institutionalisation are occurring, but this agenda is in its infancy and is yet to address these weaknesses.

CG

CG is arguably the most populous family of democratic innovations in the UK and possibly worldwide. Much democratic innovation emerges in public administrations and governance networks, rather than traditional political arenas – as per Warren’s (Reference Warren2009) ‘governance-driven democratisation’ thesis. From this perspective, CG aims to compensate for loss of legitimacy in electoral democracy by seeking legitimacy ‘issue by issue, policy by policy, and constituency by constituency’ (Warren Reference Warren2009: 7 – 8). CG thus features great internal diversity (Bussu Reference Bussu, Elstub and Escobar2019; Elstub and Escobar Reference Elstub, Escobar, Elstub and Escobar2019: 27), encompassing public forums and partnerships where citizens assemble with organisational or institutional stakeholders to collaborate in the pursuit of public or common goods.

In the UK, CG can be traced to urban regeneration, community development, and environmental governance in the 1970s, but the turning point was the New Labour era. In the 1990s, the (New) Labour party experimented with pilots in England and Scotland, seeking ‘greater empowerment and involvement by communities in the affairs of local authorities’ while addressing the fragmentation of public services by neoliberal reforms in the 1980s (Lloyd, Coles, Peel et al. Reference Lloyd, Coles, Peel and Houghton2006: 3). CG was thus propelled by ‘Third Way politics’ and the ‘Communitarian Turn’ that emphasised participatory governance, public deliberation, active citizenship, decentralisation, and ‘new forms of partnership’ (Kennett Reference Kennett and Osborne2010: 23).

The 1997 New Labour government made such partnerships central to ‘the “modernisation” and “democratic renewal” of British public services’ (Sullivan and Lowndes Reference Sullivan and Lowndes2004: 51), turbocharging their proliferation across policy areas and levels of governance. Sullivan and Skelcher (Reference Sullivan and Skelcher2002: 24–27) estimated 5,500 partnerships just at the local level and spending £4.3bn annually. CG partnerships may be temporary or ongoing and have a geographic or thematic focus covering issues such as ‘urban regeneration, children’s and young people’s services, community safety, employment, education, environmental sustainability and health care’ (Sullivan and Lowndes Reference Sullivan and Lowndes2004: 52).

The institutionalisation of CG spread across the four UK nations as devolution gained pace (see Sinclair Reference Sinclair2008: 374). England and Wales saw the Local Government Act 2000 prescribing a new power of ‘community wellbeing’ and mandating ‘community strategies’. This generated local strategic partnerships in England and community strategy partnerships in Wales. In Scotland, social inclusion partnerships were succeeded by community planning partnerships in the Local Government in Scotland Act 2003. In Northern Ireland, local strategy partnerships eventually turned into community planning partnerships following the Scottish model. Despite idiosyncrasies, these partnerships converged in aims and form because of New Labour’s leading or co-leading administrations across the UK during this period (Sinclair Reference Sinclair2008: 385).

Subsequent administrations reformed these arrangements. In Scotland, the Community Empowerment Act 2015 sought to strengthen them with some successes (Weakley and Escobar Reference Weakley and Escobar2018). In England, Conservative governments since 2010 (first in coalition with Liberal Democrats) recoiled from centrally mandated partnerships to allow local approaches while watering down participatory mechanisms that connected local and national governance (Bua and Escobar Reference Bua and Escobar2018). In Northern Ireland, community planning partnerships were further developed since the Local Government Act (NI) 2014.

These partnerships are not the only forums in the CG family of democratic innovations, but they were given legislative basis and covered numerous policy areas. Citizens, at least in theory, could participate in governance through ‘new spaces within which citizens and officials meet together to deliberate, make and review policy’ (Barnes, Reference Barnes, Barnes and Prior2009: 33). Assessing the success of these partnerships is difficult. They were introduced as strategies to counter democratic deficits, deal with complex issues, increase problem-solving capacity, foster social capital, improve public services, and restore legitimacy to policy-making. Researchers usually point to mixed and highly contingent results (eg Barnes, Newman, and Sullivan Reference Barnes, Newman and Sullivan2007; Weakley and Escobar Reference Weakley and Escobar2018; Sullivan Reference Sullivan2022). From a democratic innovations perspective, have they contributed to reimagining and deepening the role of citizens? Four tensions have troubled this agenda.

First, mandated CG processes are riddled with power inequalities. In these democratic innovations, citizens and officials are meant to work as ‘partners’, but power-sharing has been limited (Barnes, Newman, and Sullivan Reference Barnes, Newman and Sullivan2007; Sullivan Reference Sullivan2022). Second, CG uncovers tensions between technocratic and democratic imaginaries of governance and their underlying managerial, organisational, and political rationales. Some partners’ motivation is to improve efficiency and effectiveness in policy, whereas others seek to deepen public participation and community empowerment – aims often at odds in practice (eg Sullivan and Lowndes Reference Sullivan and Lowndes2004; Barnes and Prior Reference Barnes and Prior2009). Third, CG is another arena where participatory and representative politics are in tension. Politicians undertake ambiguous roles, as partnerships usually have unclear relations with representative institutions. Finally, the ‘network governance’ logic of these democratic innovations often clashes with the command-and-control culture of politics and public administration in the UK – a highly centralised and elitist system of governance (Cairney, Boswell, Ayres et al. Reference Cairney, Boswell, Ayres, Durose, Elliott, Flinders, Martin and Richardson2024).

These tensions must be placed in the broader context of the new public management reforms carried out by successive governments in the UK since the 1980s, and the concomitant depoliticisation of key areas of public administration (Flinders Reference Flinders2008). Depoliticisation refers to the ‘transfer of functions away from elected politicians’ and the distancing of public governance from traditional democratic control and accountability, which became ‘the dominant model of statecraft’ in the UK by the turn of the century (Flinders and Wood Reference Flinders, Wood, Flinders and Wood2015: 2; Hay Reference Hay2007). The proliferation of CG arrangements can be understood against this backdrop, which opened spaces for democratic innovation – insofar as governance-driven democratisation required new legitimation processes, as without the administrators being elected, there was an absence of principal–agent bonds with the public (Warren Reference Warren2009) – but also imbued them with the tensions outlined above.

In sum, CG has generated myriad interfaces between citizens and authorities across policy arenas and public services, providing new spaces for collaboration and contestation. As Barnes (Reference Barnes, Barnes and Prior2009: 48) argues, the ‘objective of the “empowerment” of citizens is too simplistic but nor is it useful to understand these forums simply as another place in which officials exercise power over citizens’. Although early hopes are tempered, CG continues to pursue ‘a new set of relationships between government, communities, and citizens’ across the UK (Barnes and Prior Reference Barnes and Prior2009: 5).

PB

The UK was an early adopter of PB in Europe (Röcke Reference Röcke2014). The idea travelled from Brazil via activists and community workers through visits in the 1990s and the World Social Forum in Porto Alegre in 2001 (Jackson Reference Jackson and Graham2011: 96; Escobar, Garven, Harkins et al. Reference Escobar, Garven, Harkins, Glazik, Cameron, Stoddart and Dias2018: 314). PB enables citizens to participate in deciding the allocation of public or community budgets, including the development and scrutiny of proposals and monitoring of implementation (Escobar Reference Escobar, Loeffler and Bovaird2020). This makes PB a unique family of democratic innovations insofar as participation is inextricable from direct mobilisation of resources.

The UK history of PB reflects the interplay – collaborative and agonistic – between civil society and public authorities. Civil society organisations (ie community groups, anti-poverty charities, churches) started pilots and advocated PB’s potential to support social justice and democratic renewal. In the mid-2000s the New Labour government linked PB to decentralisation and community empowerment, resulting in the 2008 National Strategy for Participatory Budgeting, alongside legislation for a ‘duty to involve’ (Jackson Reference Jackson and Graham2011: 97). The PB Unit was funded by the Ministry of Community and Local Government and outsourced to Church Action on Poverty. The Home Office funded 20 pilots in 2008, and the UK government declared that all local authorities would feature PB by 2012 (Jackson Reference Jackson and Graham2011: 97–98). This policy context helped build capacity and a community of practice that would become instrumental to develop PB across the four nations (Hall Reference Hall, Dias, Enriquez, Cardita, Julio and Serrano2021: 209).

The two decades of UK PB maps onto ebbs and flows in the four nations. The first decade advanced PB in England and introduced it in Wales, while the second ushered PB in Scotland and later Northern Ireland (Dias, Enriquez, Cardita et al. Reference Dias, Enriquez, Cardita, Julio and Serrano2021). By 2010 there were 65 processes in England, with another 100 local areas planning to start PB (Jackson Reference Jackson and Graham2011: 96). The importance of New Labour’s support became patent when the Conservative–Liberal Democrat coalition took power in 2010 and the PB agenda waned down in England – although some authorities and civil society organisations continued to carry the PB torch (Hall Reference Hall, Dias, Enriquez, Cardita, Julio and Serrano2021: 209). Although Wales had pilots, PB never took off as in England, albeit the Welsh government has reconsidered rekindling this agenda (Williams, St Denny, and Bristow Reference Williams, St Denny and Bristow2017). In Northern Ireland, PB gained momentum since 2017 (Jackson Reference Jackson and Graham2011: 96) via housing associations, community planning partnerships, and the PB Works network – supported by practitioners from England and Scotland (O’Kane Reference O’Kane, Dias, Enriquez, Cardita, Julio and Serrano2021).

PB proliferated in Scotland from 2010 drawing on expertise from England. Civil society organisations, researchers, and reform commissions advocated PB, and the 2014 Independence Referendum provided fertile ground steeped in discourses of community empowerment and democratic renewal (Escobar, Garven, Harkins et al. Reference Escobar, Garven, Harkins, Glazik, Cameron, Stoddart and Dias2018). The Scottish government, led by the Scottish National Party, launched the Community Choices Fund to develop PB with local authorities and community organisations. This was supported by creating PB Scotland and the PB Charter and by policy developments such as the Community Empowerment (Scotland) Act 2015. By 2016, PB was a core commitment in Scotland’s National Action Plan to join the Open Government Partnership. It went from few cases before 2010 to over 200 by 2018 (Escobar, Garven, Harkins et al. Reference Escobar, Garven, Harkins, Glazik, Cameron, Stoddart and Dias2018: 314). Scottish government investment totalled £6.5m, match-funded by local authorities (£5m) and supported by other institutions and community organisations (Escobar and Katz Reference Escobar and Katz2018: 4; Escobar, Garven, Harkins et al. Reference Escobar, Garven, Harkins, Glazik, Cameron, Stoddart and Dias2018: 319). This developed a community of practice of 1,300 PB organisers and enabled participation by 122,000 citizens (National Participatory Budgeting Strategic Group 2021). Since 2017, the Scottish government and the Convention of Scottish Local Authorities agreed to allocate ‘at least 1% of local government budgets’ via PB – an estimated £100m (Escobar and Katz Reference Escobar and Katz2018: 5). A first milestone was reached in 2023, with 1.4% spent (£154m).Footnote ⁸

This trajectory illustrates how democratic innovations can develop apace through the interplay between civil society and public institutions. It also shows the fragility that stems from dependence on the support of a particular administration. UK PB is criticised for lacking radical ambition and the level of investment and impact seen elsewhere (Röcke Reference Röcke2014; Williams, St Denny, and Bristow Reference Williams, St Denny and Bristow2017; Escobar Reference Escobar, Loeffler and Bovaird2020). PB was adapted to the UK context of the early 2000s so that it could be piloted in civil society (Hall Reference Hall, Dias, Enriquez, Cardita, Julio and Serrano2021: 206). This generated the ‘participatory grant-making’ model featuring small funds for community projects – different from models that allocate larger mainstream public budgets (PB Partners 2016a; 2016b; Escobar and Katz Reference Escobar and Katz2018). Path-dependency ensued, and despite exceptions, most PB has been of this nature (DCLG 2011; Röcke Reference Röcke2014), albeit there are efforts to mainstream PB, particularly in Scotland (Escobar Reference Escobar, Loeffler and Bovaird2020). Along the way, PB organisers have faced cultural, political, capacity, and sustainability challenges now exacerbated by the mainstreaming agenda (Escobar and Katz Reference Escobar and Katz2018: 5).

Evaluations have documented positive results that are commensurate with relatively small investments (eg DCLG 2011; O’Hagan, Hill-O’Connor, MacRae et al. Reference O’Hagan, Hill-O’Connor, MacRae and Teedon2019; Harkins, Moore, and Escobar, Reference Harkins, Moore and Escobar2016). PB has improved citizen participation by mobilising resources towards community priorities across policy areas and public services. Nonetheless, this has not amounted to transformative impacts in outcomes for citizens or in democratising the governance of public budgets (Escobar Reference Escobar, Loeffler and Bovaird2020). Lessons from these two decades, sustained through civil society practice and institutional capacity-building, suggest that these democratic innovations have not reached their potential, but have perhaps laid groundwork for PB of greater potency.