Important if underwhelming, Erzberger v TUI AG (2017) C-566/15 has held national codetermination laws are not an ‘impediment to free movement of workers’ in the European Union. Codetermination laws, from the Oxford University Act 1854 to Germany’s Codetermination Act 1976, now benefit people in the majority of EU and OECD countries (map below). Quite rightly, the Court of Justice’s Grand Chamber resisted what has been an absurd challenge to democratic life. But it missed the opportunity (seen in AG Saugmandsgaard Øe’s Opinion [103]) to acknowledge that votes at work are at the heart of the social state principle, both in Germany’s constitution (Grundgesetz 1949 §20(1)) and the Treaty on European Union articles 2-4. In EU law worker votes in corporate governance are not only ‘overriding requirements’ and public policy. They precede all free movement rights. Votes at work are the quintessence of democracy

This blog (1) summarises the Erzberger litigation, (2) explains the constitutional significance of the ‘codetermination bargains’ that developed in Germany, and (3) concludes on what courts can do.

1. Erzberger v TUI AG 

Erzberger, a shareholder, claimed the travel company TUI AG violated EU law by following German law’s minimum requirement for workers to vote for its supervisory board. The Codetermination Act 1976 (Mitbestimmungsgesetz 1976) §1 is usually thought only to cover workers in Germany, but not those working in other countries. For example, the Board Representation (Private sector employees) Act 1987 might cover TUI workers in Sweden, the Code de Commerce article L225-27-1 workers in France, and so on. Of course, TUI AG could voluntarily allow workers abroad to vote. Shareholders like Erzberger could have organised a resolution to achieve this (Aktiengesetz 1965 §122) and, like in a Societas Europaea, TUI could have bargained with the recognised union for a new codetermination plan. Or Erzberger might have attempted, as people in democracies do, to lobby for an extension to codetermination law. Instead, Erzberger went to court. He argued the whole Codetermination Act 1976 violated EU law. He lost. 

Specifically, Erzberger argued the ‘supervisory board should have been composed only of members designated by the company’s shareholders.’ (AG Opinion, [15], what Erzberger said publicly differs.) Supposedly, because workers might lose their votes if they moved countries within TUI (e.g. from Germany to Belgium) their free choice of movement was hindered. Erzberger said this violated TFEU article 18, on non-discrimination, and article 45 on free movement. Of course Konrad Erzberger, whose blog is called ‘Life is good’, cares about as much for workers’ voice, as the CJEU cared for his article 18 argument: this was irrelevant, because article 45 is the more specific rule (at [25] and [32]). 

The CJEU held article 45 was not engaged at all. Workers who always worked abroad would have no factor linking them to ‘any of the situations governed by EU law’ (at [28]). Workers who moved from Germany to Belgium, etc, might lose the right to vote. But this is the consequence of any right (like, say, social security) being limited to territory. Territorial limits are ‘objective and non-discriminatory’ ([38]) precisely because people can move freely. Taken to its absurd conclusion, Erzberger’s argument might have scrapped every member state’s labour law, or indeed corporate law, because moving or selling shares in one country and buying in another could be ‘deterred’ by any difference in rules. 

The Grand Chamber’s reasoning can only be faulted in two respects. First, it said worker representation is ‘a field which, to date, has not been harmonised or even coordinated at Union level’ (at [37]). This goes a bit too far: for example, the Credit Institutions Directive 2013/36/EU article 95(2) requires employee representation on bank board remuneration committees, if national law requires employees in a management body. Like France’s new 2013 law, this exemplifies an authoritarian economy giving way to democracy in the 21st century. A majority of EU countries now have codetermination laws:

Second, the Grand Chamber missed the opportunity to say that codetermination is ‘an essential element’ of ‘the German social order’ (AG Opinion, [103]). AG Saugmandsgaard Øe did, so acknowledging that codetermination is a value expressed in article 2 (‘solidarity and equality’) and article 3 (the ‘social market economy’ and ‘social justice’). It precedes all free movement law. The Advocate General only hesitated to frame this as a part of Germany’s ‘national identity’ under TEU article 4(2), but perhaps he should have. In any event, votes at work are definitely part of the ‘fundamental structures, political and constitutional’ in Germany. The next part shows why. 

2. The codetermination bargains

What I have called ‘the codetermination bargains’ precede and justified the modern German state: once in 1919, again by 1949. Two business theorists, Jensen and Meckling, argued in 1979: ‘Without fiat, codetermination would be virtually nonexistent’. But this view (that codetermination could only be forced by law) is an alternative historical fact. Codetermination came from voluntary collective bargaining. Laws codified the social and constitutional consensus. 


Over the 19th century, codetermination experiments spanned the struggle against Prussian militarism. In the Revolutions of 1848, the short-lived democratic Paulskirche Parlament drafted a new worker council law, with binding rights over management issues. The aristocracy swiftly crushed the movement. But one member, Carl Degenkolb, put the ideas into practice at his Saxon factory till 1862. Otto von Bismarck repressed all labour and democratic voice as he was in power, to 1890. Then, Kaiser Wilhelm II piloted a limited worker consultation right, and other Acts followed for Prussian mines in 1892 and 1905, and Bavarian mines in 1902. During WW1, the Auxiliary Service Act 1916 forced labour on all German citizens, but the Social Democrats secured work council rights, beyond consultation, to bring employers to arbitration. The historian Hans Jürgen Teuteberg, called this the ‘end of the unilateral right to manage’.


The real beginning of codetermination emerged on the collapse of the Imperial military. Trade unions met the employer’s federation to conclude a post-war settlement, with ‘common resolution of all economic and social questions in German industry’. Like the African National Congress and the Apartheid regime, in this brief moment the state was stripped, and the real social forces bargained openly. The Stinnes-Legien Accord (15 November 1918) became the Weimar Constitution article 165. Written by Hugo Sinzheimer, workers and employers ‘on an equal footing’ would cooperate ‘in the entire field of the economic development of the forces of production.’ It was implemented by the Betriebsrätegesetz 1920 (elected work councils with binding rights) and Aufsichtsratsgesetz 1922 (two worker representatives on company supervisory boards). Unions bargained for codetermination. Laws codified the social consensus.


The tragedy of Weimar democracy, and its fascist destruction, was also a judicial assault on labour. The great European labour lawyer, Otto Kahn-Freund, explained in 1931 how as a Berlin judge, he watched the Empire Labour Court dismantle the rights of work councils. It forced trade unions to pledge allegiance to their corporate employers, not their members. Fascism was already an unconscious culture. On 4 January 1933, Hitler met the conservative leader at a bankers’ house, to join in government. On 20 February, Hitler got a million Deutschmarks by promising the Krupp household’s guests to destroy democracy, for capitalism to prevail. Labour voice at work and the trade unions were crushed first. In their place was the nationalised Nazi union, the Deutsche Arbeitsfront. Like a mandatory business school, its motto was ‘the Leader is always right.’ The same ethic prevailed in company law: banks formally usurped shareholder voting rights, because the ultimate investors are ‘irresponsible’. Directors became irremovable without court consent. Fascist corporate law was a nexus of contracts between financial capital, corporate boards and a business-bent state. As Alfried Krupp put it in his Nuremberg trial, ‘Politics is not our thing.’ And for a few years, the top 1% did indeed profit from the destruction of organised labour.


Post-WW2 reconstruction undid the cartel economy and enabled free collective bargaining. This was how codetermination revived. By the Control Council Law No. 22 of 1946, unions bargained with employers to re-establish work councils. For board seats, unions bargained with coal and steel owners before any law. There is a myth among some corporate lawyers that British occupiers imposed codetermination on Germany, perhaps to weaken its industry. The evidence shows the reverse is true. The Allied Military Government delayed German states’ legislative action, so unions remade the ‘codetermination bargains’. Like after WW1, with business in ruins financially and morally, labour’s bargaining power was less unequal. As the Grundgesetz 1949 was drafted, the principle of voice at work was cross-political party consensus. Consolidating the traditions of Paulskirche and the essence of Weimar’s article 165, the Federal Republic was declared a ‘democratic and social’ state, bound by the rule of law (GG art 20). The principle of codetermination is in the core of the German constitution, and has been for close to a century.

3. Conclusion: what courts can do

Today codetermination is as inviolable to Germany as the NHS is to Britain. Voter turnouts are consistently high: 70-80% in work councils. The corporate lawyers who still oppose it are a shrinking minority, because the evidence is becoming too persuasive. Leximetric data, like at the Centre for Business Research, is showing that productive, sustainable, long-term corporate governance needs votes at work. The systems that encourage unjust enrichment of boardrooms, asset managers, or banks are outperformed by models where stakeholders have votes. In fact, the German model is far from the best. But now the debate is more than economics. From the AfD to Le Pen, to Putin, Brexiteers and Trump, authoritarians in the economy again are threatening democrats in politics. So, the economy must become democratic. True freedom needs votes at work, as much as votes in Parliament. 

What can courts do? Certainly the judiciary is at its worst when it sees itself as limiting democratic society, rather than using rights to empower every person’s voice, to realise their full potential. Erzberger was not a decisive end to the poorly justified free movement cases, like Commission v Netherlands (2006) C 282/04 that struck down public participation in systemically important enterprises, or ITF v Viking Line ABP (2007) C-438/05 that placed unprincipled limits on labour and human rights. (cf Cartesio Oktató és Szolgáltató bt (2008) C-210/06, fortunately rejecting AG Maduro’s anti-social opinions that time.) There is still more to do, to ensure free movement rules are compatible with democracy and social justice, not a Hayekian nightmare. Workers move faster than companies establish. But capital moves faster still. The only things that outrun them all are ideas and collective will. More than ever there is a need to reverse the causes of Eurozone austerity and the Brexit disgrace. A Court of Justice committed to protecting Europe’s integrity takes democratic governance seriously, in politics, corporations and the workplace. 

Summary of Germany’s codetermination system:

  • With over 5 staff, elected work councils get binding rights on 13 issues, including hours, breaks, health and safety, pensions and social affairs (Work Constitution Act 1972 §§1 and 87). 
  • With over 20 staff, work councils may defer redundancy dismissals to arbitration on a ‘social plan’ (WCA 1972 §111-113). 
  • With over 500 staff, workers can elect one third of a company’s supervisory board (Aufsichtsrat), which in turn elects the executive (Vorstand) with day to day power to run the company (One Third Participation Act 2003 §1 ff). 
  • With over 2000 staff, workers elect half the supervisory board, but the chair with a casting vote is appointed by shareholders. (Around 60% of shareholder votes are cast by banks, under Aktiengesetz 1965 §135, but statistics were shut down from 2005: chart below, contrast Switzerland since 2013). Just one executive is a worker or union representative (Codetermination Act 1976 §§1, 7, 27-29, 33)). 
  • With over 8000 staff, workers by default delegate their trade union their votes, but can opt out (CA 1976 §§9 and 18). 
  • In mining companies over 1000 staff, today only a few companies, employees elect half the supervisory board. Shareholder/bank representatives have no casting votes (Mining Codetermination Act 1951 §1 ff). 

This blog post draws on E McGaughey, ‘The Codetermination Bargains: The History of German Corporate and Labour Law’ available at SSRN.

Dr Ewan McGaughey is a lecturer in private law at King’s College, London and a Research Associate at the Centre for Business Research, University of Cambridge.