Direct democracy and trade union action

Introduction

Throughout much of the 20th century, trade unions exerted political influence using their strong ties to labour-friendly parties. Over time, however, as centre-left parties increasingly implemented ‘austerity’ cuts and rolled back welfare and labour protections, such party-union alliances increasingly deteriorated. This is very evident in the Italian and French cases (Mattina and Carrieri, 2017; Parson, 2017), but also visible in Germany and other European countries (Allern and Bale, 2017; Spier, 2017). This has forced unions to reconsider their political activities and their broader political mandate. What complementary strategies are open to them to make their voices heard inside the political system? Incidentally, the more unions increase their own capacity for political action, the more they will also be able to regain influence within their sister parties (Erne and Schief, 2017).

Most answers to this question can be categorised along a radical ‘social movement unionism’ versus a moderate ‘social partnership unionism’ axis (Gumbrell-McCormick and Hyman, 2013). Moderate union leaders may try to convince political leaders to talk to them again by pointing out unions’ role as agents able to help governments to manufacture consent. The leadership of the Irish Congress of Trade Unions, for example, managed to convince the governing coalition that it would be easier to handle the contentious policy changes demanded by the Troika in cooperation with the unions. Unions, in turn, made substantial concessions, but were at least able to keep a seat at the policy-making table (Geary, 2016). Unilateral government cutbacks and labour policy changes, however, also led to a substantial increase in strikes and other forms of contentious action, such as factory occupations (Della Porta et al., 2016; Gumbrell-McCormick and Hyman, 2013; Stan et al., 2015; Vogiatzoglou, 2015). Such counter-movements are important, as even the advocates of neo-corporatist compromises recognise that these ultimately depend on a ‘relative equilibrium of class forces’ (Grote and Schmitter, 1999: 58). And yet, social actors also require institutional access-points to a given political system to be effective. But what could these access-points be at a time when centre-left parties – including the French Socialist Party (Amable, 2017) – are prepared to pay the electoral price for the implementation of business-friendly policies? In other words, what access-points to the political system are still open to unions when their sister parties prefer to be ‘responsible’ for austerity and business-friendly labour policies rather than ‘responsive’ to their traditional working-class basis (Mair, 2013)?

If the leadership of a labour party is no longer responsive to working people’s concerns, trade unionists have two options. They can stay within the realm of party politics or try to pursue their political mandate more autonomously. The first option includes attempts by organised labour to reconquer its party. Jeremy Corbyn would hardly have been able to attain and consolidate his Labour Party leadership in the United Kingdom without union support. In other parts of Europe, the shift to the right by the main centre-left party has led to the rise of new left parties, which may or may not have strong union connections. However, labour politics no longer represents the primary focus of either centre-left or new left parties. Furthermore, political parties can always try to appeal to a new social base, including culturally progressive affluent voters, whereas unions have to represent workers’ interests. It follows that the logic of party politics and the logic of worker interest representation do not necessarily converge at all times. Thus, there is a need for autonomous union action, also in the political sphere.

This article reviews trade unions’ use of a complementary channel to political influence, namely direct democratic instruments at local, national and EU levels. First, we will briefly discuss the rising popularity of citizens’ initiatives and referenda, which contradicts Mair’s account of a ‘hollowing out of democracy’ (Mair, 2013) or Crouch’s dystopian vision of ‘post-democracy’ (Crouch, 2004). In fact, the increasing use of popular consultations across Europe seems to be related to an increasing dissatisfaction with party politics. If one analyses popular votes one after another one has to realise that not all votes fall into the same category. Our article therefore distinguishes different types of direct democratic consultation. Not all popular votes give social actors, such as unions, the same opportunities to access the political system. It matters whether a vote is initiated as a plebiscite from above, as a constitutional referendum by a constitutional requirement or as an initiative or referendum from below. In addition, in several countries and the EU direct democratic mechanisms exist that do not lead to a popular vote, such as the European Citizens’ Initiative (ECI) introduced by the Treaty of Lisbon. The ECI allows one million EU citizens to participate directly in the development of EU policies by calling on the European Commission to make a legislative proposal.

The article discusses the usefulness and pitfalls of different types of direct democracy for autonomous union action in the political sphere, drawing on union experiences at the local, national and EU level across Europe. We will show that plebiscites initiated from above do not often enable unions to make their voices heard. By contrast, unions have succeeded in using compulsory constitutional referenda to their advantage. This happened, for example, in the Swiss case, where the union movement was able to use the votes on the bilateral EU-Switzerland agreements as a ‘lever’ to strengthen collective labour laws, by demanding effective ‘flanking measures’ against social dumping in exchange for union support for Switzerland’s association with the EU’s single market (Afonso et al., 2010; Erne and Imboden, 2015; Wyler, 2012). Unions have also played an active role in citizens’ initiatives and referenda triggered from below; for example in Germany at the local and regional level; in Switzerland and Italy at the local and national level; and at the EU level. Success, however, is not guaranteed, as shown by the unequal trajectories of the right2water ECI launched by the European Federation of Public Service Unions (EPSU) by comparison with the fair transport ECI of the European Transport Workers Federation (ETF). This article therefore does not aim to discuss whether direct democracy is good for unions, but rather under what conditions direct democratic instruments enable unions to obtain better access to the political system.

On the rise of direct democracy in post-democratic times

Until the 1980s, direct democratic consultations were very rare. Referenda occurred on a regular basis only in Switzerland and, to a lesser extent, Italy. Since 1990, however, popular votes have been held much more regularly across Europe. According to the Direct Democracy Database (www.c2d.ch), there were between six and nine national referenda per year in the 28 EU Member States between 1990 and 2016. This rise is surprising given the recurrent accounts of a crisis of democracy, due to globalisation, Europeanisation and the corresponding decline in political leaders’ responsiveness towards their electorates. And yet the trend towards post-democracy also seems to have increased the need of political leaders to seek additional direct democratic legitimisation for contentious policy decisions, for example, regarding the EU integration process.

The declining responsiveness of political leaders has also led to increased demands for direct citizen participation from below. This is notable even in Germany, where there has not been a single national referendum since the Second World War. But whereas proposals to introduce direct democratic instruments at the national level failed to get the required two-thirds majority support in the federal parliament, there has been a dramatic increase in direct democracy at the regional and local levels. In several Bundesländer and municipalities social movements, unions and civic groups, such as Mehr Demokratie (‘More democracy’) have fought successfully for an extension of direct democratic instruments. Whereas there was not a single citizen-initiated referendum in a Bundesland throughout the 1950s, there were 116 from 2000 to 2009 and 86 since 2010 (Rehmet and Webern, 2017: 13).

The increased need for additional democratic legitimacy also led to changes in the EU Treaties. In order to keep abreast with the growing transnationalisation of national economies, transnational social movements, such as Eurotopia (1995), called in the mid-1990s for the introduction of direct democratic citizenship rights at EU level, as documented in a volume on transnational democracy sponsored by the Hans Böckler Foundation of the DGB (Erne et al., 1995). A watered-down version of the Eurotopia proposals was taken up by the Austrian and Italian governments during the subsequent Inter-Governmental Conference that led to the Treaty of Amsterdam. European political leaders, however, conceded only the first ‘baby steps’ (Kaufmann, 2012) towards European direct democratic citizenship rights at the Convention on the Future of Europe, which was asked to produce a draft EU Constitution (Papadopoulos, 2005). After the failed referenda on the draft EU Constitution in France and the Netherlands, the Lisbon Treaty nevertheless introduced the first element of transnational direct democracy – the European Citizens’ Initiative (ECI) – into Article 11(4) of the Treaty of the European Union (TEU).

The ECI is a formal request to the European Commission to propose legislation on matters on which the EU has competence to legislate. An ECI has to be backed by at least one million EU citizens. In addition, a minimum number of signatories is required in seven Member States, ranging from 4500 signatures in the four smallest EU Member States to 72,000 signatures in Germany. The rules and procedures governing citizens’ initiatives are set out in an EU Regulation adopted by the European Parliament and the Council in February 2011. Even if the European Commission is neither obliged to propose legislation nor to call a referendum as a result of an initiative – by contrast to a number of national citizens’ initiative laws – the ECI was immediately quite popular, arguably due to its suggested EU policy ‘agenda-setting’ function (Setälä and Schiller, 2012).

By the end of July 2017, 66 ECIs had been initiated, although only three were successful in attracting one million signatures, including the ‘right2water’ ECI of the European Federation of Public Service Unions (EPSU). Eighteen ECIs failed to reach the necessary number of signatures, including in the case of the ‘fair transport’ ECI of the European Transport Workers Federation (ETF). In 19 cases, the European Commission refused to register an ECI for ‘legal’ reasons. In the very contentious case of the ECI against the proposed US-EU free trade agreement (TIPP), however, the General Court of Justice of the European Union found that the Commission’s reasons for doing so were invalid. ¹ Fourteen ECIs were withdrawn and 12 ECIs were still in process as of July 2017. But even if many initiators have learned that organising a successful citizens’ initiative is very demanding, social movements and unions are increasingly making use of direct democratic procedures to make their voices heard, at local, national but also EU levels (Rehmet, 2017). Before being able to assess the usefulness and pitfalls of direct democracy for union action in the political sphere, however, we must distinguish between different types of direct democracy.

Who decides whether a popular vote has to take place? This is the key question for many typologies of direct democracy (Erne, 2002). A vote can be triggered ‘from above’, namely by a constitutional requirement (mandatory referendum) or by an ad hoc decision of the executive (plebiscite). Or it can be called from below, by a number of citizens (citizens’ initiatives and facultative referendum). To the latter, we may add the ECI, which does not force the Commission to make a legislative proposal, but nevertheless exerts considerable agenda-setting pressure, as shown by the Commission’s relatively positive response to the right2water ECI (European Commission, 2017), but also the Commission’s invalid attempt to declare the troublesome ‘Stop TTIP’ ECI inadmissible. The next sections therefore review the different logics behind different direct democratic procedures before assessing their implications for unions and union action.

Direct democracy from above: mandatory referenda and plebiscites

Mandatory referenda are votes that are required by law. This normally applies to changes in the fundamental rules of a polity, such as a complete or partial revision of the constitution. Mandatory referenda are found in 16 European states, but the circumstances that trigger a popular vote vary substantially from country to country. The Swiss and Irish constitutions require a mandatory vote on every change of the constitution. In Italy, the government can pre-empt a popular vote, if a constitutional law is voted by a qualified majority in each of the two chambers of parliament. A popular vote will also take place only if one-fifth of a chamber of parliament or 500,000 citizens request it. This happened in December 2016, when Prime Minister Renzi sought a fundamental revision of the Italian constitution.

The Italian case shows also that votes on constitutional issues may be ‘mandatory’ in political rather than strictly legal terms. The principle that fundamental decisions have to be taken by the people themselves is widely accepted across Europe. It has been followed even in countries that do not allow for mandatory referenda in their constitutions. Referenda were called on proposals to reorganise states’ territories (for example, on devolution in Portugal and in the United Kingdom, as well as on Scottish independence) and on the transfer of power between national and supranational organisations. The latter category includes not only the mandatory Irish referenda on all EU treaties, but also the EU referenda on the Maastricht Treaty in France and Denmark, the Amsterdam Treaty in Denmark, Euro membership in Denmark and Sweden, the EU Constitution in Spain, France, the Netherlands and Luxembourg, and EU membership and Brexit in the United Kingdom. Pre-accession votes on EU membership were also held in Norway, Sweden, Finland and Austria (1994); Malta, Slovenia, Hungary, Lithuania, Estonia, Latvia, Slovakia, Czech Republic and Poland (2003); as well as Croatia (2012). However, it is not always easy to ascertain which decisions are ‘fundamental’. If this question is not decided by constitutional law, there is always a danger that referenda will be called by political leaders for party political gain. This leads us to the second type of direct democracy from above, namely the plebiscite.

Plebiscites are votes called on an ad hoc basis by political rulers. Plebiscites are held on the basis of several motives. First, they can be called to overcome a blockade in the governmental system. Here, the plebiscite does not aim to support political deliberation but to end it by a definite verdict. An example is the (failed) attempt by then French President de Gaulle to weaken the powers of the Senate, against its will, in 1969. A blockade may also be electorally motivated, as in the case of the (failed) Portuguese plebiscite on the legalisation of abortion in 1988: citizens were called on to vote because the governing Socialist Party did not want to take a clear stance on the contentious abortion question.

Second, plebiscites may also be called following the logic of the mandatory referendum, as outlined above. If referenda are not mandatory in strictly legal terms, however, instrumental motivations always play an important role. President Mitterrand officially called the French vote on the Maastricht Treaty for the constitutional motive used to justify mandatory referenda. However, the plebiscitary motivation to strengthen his legitimacy as political leader certainly also mattered. The same is true for David Cameron’s misjudged calling of the Brexit vote. The Italian Prime Minister Renzi also failed to acknowledge that referenda held for plebiscitary motives can backfire.

This short account highlights the need to distinguish plebiscites from other forms of direct democracy. In a plebiscitary democracy, a responsible body can call for a referendum from above, which often happens in situations of crisis in order to safeguard power. Therefore the term ‘plebiscitary’ should not be used to describe votes triggered from below (initiative, facultative referendum). They are, in contrast to the plebiscite, part of a political culture that aims to limit top-down government. After all, the plebiscite is the only type of popular ‘consultation’ that has also been used by dictators, from Napoleon III to Pinochet (Erne, 2002).

Direct democracy from above – trade union experiences

Our discussion of referenda triggered from above seems to suggest that unions have little to gain from increased use of such procedures. After all, popular votes called by political leaders are aimed first and foremost at legitimising their political decisions, which seems to leave little room for autonomous union action. However, a negative referendum outcome can terminate the political career of the leader who called it. The list of political leaders forced to resign after a referendum defeat is impressive. It ranges from General De Gaulle in the 1960s to David Cameron and Matteo Renzi in more recent times. Furthermore, the material issues that are at stake in a particular referendum may also be so important for a country that its political leadership cannot easily risk a referendum defeat. These circumstances give unions a lever that could be used to extract government concessions in exchange for union support in a given referendum campaign.

According to the General Secretary of the Irish Congress of Trade Unions (ICTU), David Begg, the centre-right Irish government, for example, would have requested an Irish opt-out from the EU Charter of Fundamental Rights, like the United Kingdom, if ICTU had not threatened to campaign for a No-vote to the Lisbon Treaty if this occurred. In turn, the inclusion of the Charter in the Lisbon Treaty became one of the most important arguments for a Yes-vote within the Irish union movement, notably during the Lisbon II referendum campaign in 2009 (Béthoux et al., 2018).

Another prominent example of a successful use of this particular form of political exchange of power can be found in Switzerland. Swiss unions strategically used the votes on the bilateral EU-Switzerland agreements as a ‘lever’ (Wyler, 2012) for a substantial strengthening of Switzerland’s collective labour law. As a result, Swiss collective bargaining coverage in the private sector rose substantially from 40 per cent in 2001 to 49 per cent in 2013 (Visser, 2015). The ‘flanking measures’ that Swiss unions obtained in exchange for their support for the bilateral EU-Switzerland agreements on the free movement of people also substantially strengthened the enforcement mechanisms for the equal pay by nationality principle for local and migrant workers (Erne and Imboden, 2015). This strengthening of union power is somewhat counterintuitive, given the opposite trends in almost all other countries that are part of the EU’s single market. The obligation to get all EU-Switzerland agreements approved by referendum, combined with the existence of strong anti-European right-wing parties, however, forced organised capital and its political allies to accommodate organised labour’s demands (Afonso et al., 2010). Without the ‘flanking measures’ against social dumping, it would indeed not have been possible to secure the popular Yes-votes for Switzerland’s integration into the single market and the free movement of all EU workers in Switzerland.

The capacity to secure political exchanges in a direct democracy, however, also depends on enabling associational properties. They very much mirror the ideal-typical cluster of conditions identified by scholars of neo-corporatism (Schmitter and Grote, 1997: 4). Unions’ capacity to secure favourable political exchanges in a direct democracy also very much depend on organised labour’s capacity to act in unison. But whereas neo-corporatist political exchanges also require ‘some degree of heteronomy with regard to the selection of [union] leaders and the articulation of [union] demands’ (Schmitter and Grote, 1997: 4), union leaders cannot afford to be ‘out of step’ with their membership if they want to be effective in a direct democracy (Hyman, 2010). Incidentally, this does not imply giving in to anti-immigrant demands that seek to restrict the free movement of people, as argued by several pundits after the UK Brexit vote. Taking working-class concerns about social dumping seriously would have required the enforcement of the principle of ‘the same pay for the same work in the same place’, which the British TUC initially failed to support in the Executive Committee of the European Trade Union Confederation (Pedrina, 2016). By contrast, even the president of the Swiss employers’ federation (Economiesuisse) argued, when asked about the Brexit vote, that the introduction of social flanking measures in the United Kingdom would ‘no doubt have damped down the mood of [popular] annoyance against the EU’ (Heinz Karrer cited by Pedrina, 2016).

Finally, holding a referendum initiated from above can also provide unions with an opportunity to extract indirect political concessions from the government in cases in which they fail to adopt a common position on the referendum topic. The Italian constitutional reform referendum called by Matteo Renzi in 2016 is an interesting case in point. The Italian union confederations did not adopt the same positions on the proposed reform. Whereas the CISL backed the reform, CGIL argued for a No-vote (Pasquino and Valbruzzi, 2017). None of the unions, however, invested many resources into the Yes or No campaign. Instead, all union confederations used the increased dependence of the government on popular support in the wake of the referendum vote to conclude a new national collective agreement for the Italian public sector. After the Renzi government had antagonised many workers and unions by the unilateral imposition of the so-called Jobs Act, it arguably felt the need to return to a more conciliatory approach to unions in the wake of the referendum vote. It is no coincidence that the new Italian public sector collective bargaining agreement was signed just a few days before the referendum. In sum, unions may use the holding of referenda to extract concessions from governments, namely when the political leaders that initiated them require working-class support. But what use are referenda initiated from below for organised labour, notably if unions are already part of the relevant policy circles, as social partners and social and employment policy specialists?

Direct democracy from below: initiatives and facultative referenda

In this section we look at votes that are held because a quorum of citizens have requested it. Two cases can be distinguished, namely the facultative referendum and the initiative. In the case of the facultative referendum, a bill passed by parliament becomes subject to a popular vote if asked for by a minimum number of citizens set out by law. In the case of an initiative, a certain number of citizens can request a popular vote on a proposition formulated by the initiators of the initiative themselves.

The initiative stands at the beginning of a political process. It fulfils a proactive function in that process as it draws attention to topics neglected by political parties. Hence, initiatives may have a political impact even if they fail at the ballot box, as they force new topics onto the political agenda. For this reason, the European Citizens’ Initiative also merits discussion under this heading, even if ECIs do not lead to popular votes. By contrast, the facultative referendum stands at the end of a political process, as it submits a law voted by parliament to a popular vote. The facultative referendum functions therefore as a mechanism to hold political leaders publicly accountable. Accordingly, it has been interpreted as a conservative tool, even if the popular vetoes exercised by facultative referenda increasingly also concern progressive causes, such as the protection of public services from privatisation in German Länder (Greer, 2008) or the defence of social security systems from neoliberal assault in Slovenia and Switzerland. Overall, facultative referenda oblige political leaders to seek broad compromises that include all relevant social interests, including organised labour.

The first experiences with direct democracy from below – in a modern sense – go back to the democratic supporters of the French Revolution. Condorcet, a leading politician of the Girondins, promoted the introduction of the facultative referendum into the new Constitution approved by the people in 1793. However, the Constitution was never put into practice as the counter-revolutionary wars allowed the Jacobin leadership to prevent its coming into force. After that, citizen-initiated referenda became an important feature in Switzerland and the United States at the regional level, following important radical democratic and progressive popular mobilisations at the end of the 19th century against the corruption of the so-called ‘robber barons’ and liberal political elites that betrayed the democratic ideals of the two sister republics. Today, Italy, Liechtenstein, Lithuania, the Netherlands, Romania, Russia, San Marino, Slovenia, Ukraine, Hungary and several Latin American states also have citizen-initiated referenda. Initiatives at a regional or local level are possible also in Belgium, Finland, Luxembourg, Austria and the Czech Republic.

The concrete forms of these popular rights, however, vary markedly across countries. In Switzerland and in most German Länder, parliament and government play an important role in the direct democratic process. Citizen’s initiatives are debated in parliament, which enables parliaments to propose direct or indirect counterproposals that may also become subject to a popular vote. By contrast, in most US states, direct democratic procedures bypass ‘parliament’, which produces quite antagonistic and uncoordinated political processes. Other major differences that have a substantial impact on direct democratic processes include, for example, admission quorums, signature collection periods, as well as permissible topics of popular initiatives.

Italy, the EU state with the largest experience of direct democracy, has a specific instance of the facultative referendum, the so-called abrogative (‘abolishing’) referendum. Article 75 of the Italian Constitution allows 500,000 citizens (or five regional parliaments) to ask for the total or partial abrogation of a national law that is already in force. This differs from the facultative referendum in Switzerland that provides a vote before a new law comes into force. All existing laws can be made subject to an abrogative referendum. In addition, the Italian abrogative referendum also allows a vote on the deletion of a single article of a given law. In this way, the sense of the law can be altered. It can therefore be difficult to distinguish the abrogative referendum from an initiative in some cases. The changes in Italian electoral law that led to the transition to the Second Republic at the beginning of the 1990s, for example, were triggered by abrogative referenda that eliminated particular articles. Furthermore, Italian referenda are only valid if at least 50 per cent of eligible citizens participate in the vote. This, however, creates a perverse incentive for the opponents of a proposal to boycott both the vote and any public debate about it. In addition, a participation quorum undermines the secrecy of the vote, as only Yes-voters will have an incentive to go to the polling stations (Erne, 2002).

Incidentally, in 1997 the Hungarian parliament replaced the 50 per cent turnout quorum for referenda with a 25 per cent approval quorum, precisely to prevent such boycott campaigns. In 2011, however, the Orbán government used its two-thirds majority in Parliament to reintroduce the 50 per cent participation quorum and to adopt a new referendum and popular initiative act. This dramatically restricted the use of direct democracy as an opposition instrument (Pállinger, 2016). Ironically, however, Orbán’s new Constitution of 2011 also led to the failure of his anti-European plebiscite on migrant quota, which was held on 2 October 2016, but failed to reach the 50 per cent turnout quorum.

In sum, the laws governing citizen-initiated referenda and citizens’ initiatives differ substantially across political systems. Nevertheless, all forms of direct democracy from below play important agenda-setting roles. But what can be said about European unions’ experiences with the different types of direct democracy discussed so far?

Direct democracy from below: trade union experiences

Our assessment of direct democracy from below distinguishes two sets of situations: (i) defensive and (ii) offensive circumstances. The former includes referenda called to veto or annul existing laws. The latter encompasses cases in which unions collected enough signatures to propose a new regulation. Our review of defensive cases suggests that unions often use direct democratic procedures to prevent a worsening of working and living conditions. This is evident in cases where their campaign targeted public sector privatisations or a worsening of pension and other social security entitlements.

Conclusion

The first section of this article showed that the holding of referenda – even when initiated from above – can represent an opportunity for unions to extract concessions from governments. After all, political leaders still require workers’ support in order to win certain votes, as demonstrated, for example, by recent EU integration referenda across Europe. In some cases, unions were therefore able to conclude political exchanges with policy-makers that mirror some aspects of the corporatist compromises of the past, namely their involvement in social and political negotiations that go far beyond the confines of shop-floor employment relations. Switzerland provides the most prominent examples of using the EU integration issue as a lever to advance labour rights in exchange for union support in referendum debates. There are, however, also major differences between the political exchanges analysed by neo-corporatist industrial relations scholars and the political exchanges in a direct democracy. Whereas neo-corporatist agreements require ‘responsible’ union leaders who are willing to sign agreements that may in fact be very unpopular among rank-and-file union members, proposals offered to labour in a direct democracy must convince individual union members. As a result, the trade-off between the competing logics of ‘influence’ and ‘membership’ that neo-corporatist scholars invoke to discuss the dilemmas that union leaders are facing (Dølvik, 1997; Schmitter and Streeck, 1999) does not apply to political exchanges in a direct democracy. Instead, the direct democratic influence of trade unions is positively related to their capacity to mobilise their members and large sections of the public.

In addition, our review of referenda initiated from below shows that unions often use such procedures to prevent a worsening of working and living conditions. This is particularly evident in campaigns against public sector privatisations or a worsening of pension and other social security entitlements. Success in these cases, however, very much depended on prevailing circumstances. Successful union campaigns did not focus on sectional interests nor did they present themselves as a left-wing political operation. Instead, the unions waged campaigns that emphasised the negative social impact for all citizens. As a result, their referenda campaigns were supported also by right-wing voters who also opposed the proposed cutbacks and privatisations (Afonso et al., 2010; Bieler, 2017: 309). Comparing the role of unions in the Italian and Berlin water referenda campaigns and the corresponding right2water ECI, on one side, and the labour rights campaigns of CGIL and the fair transport ECI of the ETF, on the other, is instructive. In the water campaigns, unions were very active in the background but kept a relatively low profile in public. In contrast, the Italian and ETF labour rights campaigns were very much perceived as union campaigns. As we have seen, however, this strong labour focus was not without risks, as both the CGIL and the ETF campaigns risked being represented as special interest campaigns. Therefore, CGIL framed its campaign as a campaign for general workers’ rights. By contrast, the fair transport ECI failed not only because the ETF arguably underestimated the organisational requirements, but also because it failed to inspire both union activists and larger sections of society. But precisely because successful direct democratic campaigns require both organisational strength and a capacity to mobilise people, unions may want to make increasing use of direct democratic instruments: not least because both alliances with centre-left parties and social partnership arrangements deliver fewer results than ever. Direct democracy is no panacea. There are occasions for political action, as there are occasions for industrial action. At times, however, the reluctance to choose between political and industrial action only benefited those that did not want to act at all (Gobin and Erne, 2017).