What Do Trade Unions Do? Bargaining, Voice and a Long Decline

The functions extend well beyond the strike that dominates the public image:

• •Collective bargaining over pay, hours, holidays, pensions, redundancy terms and health and safety, producing an agreement that governs everyone covered
• •Representing individual members in disciplinary hearings, grievances and dismissal cases, which is the daily work of most local officials and rarely reported
• •Enforcing existing law, since a right to a rest break or to safe equipment means nothing if nobody is willing to raise it, and unionised workplaces have measurably better compliance
• •Providing a voice mechanism, in the economists' phrase, so that problems are raised and addressed rather than being answered by employees leaving, which reduces turnover
• •Political and legislative campaigning, which produced much of the basic architecture of employment law including limits on working hours and safety regulation
• •Training, legal services, insurance and, historically, mutual insurance against sickness and unemployment before states provided it

The empirical literature is large and its findings are reasonably consistent. There is a union wage premium, typically estimated in the range of ten to twenty percent in the United States and somewhat lower in countries with wider coverage, meaning a comparable worker earns that much more in a unionised job. Unions compress the wage distribution, both by raising the bottom more than the top and by imposing transparent pay scales that reduce arbitrary variation, and several studies attribute a substantial share of the rise in wage inequality since 1980 to declining union coverage. Non-union employers in unionised industries raise pay to avoid organisation, which is a spillover benefit to non-members. Effects on productivity are contested, with the classic argument by Richard Freeman and James Medoff that the voice function reduces turnover and raises productivity enough to offset the wage cost, and other work finding negative effects where restrictive practices persist. Effects on employment are similarly debated, with the standard prediction that a higher wage reduces hiring qualified by evidence that employers hold wage-setting power in many labour markets, in which case union bargaining can raise both wages and employment.

The institutional arrangement matters more than the membership rate. In the Nordic countries and in much of continental Europe, agreements are negotiated at sector level between employer associations and unions and are then extended by law or custom to the whole sector, so that coverage far exceeds membership: France has union membership under ten percent and collective agreement coverage above ninety. In Britain and the United States, bargaining is at enterprise level, so coverage tracks membership closely and has fallen with it. Germany adds co-determination, under which employees elect representatives to company supervisory boards and works councils handle workplace-level consultation separately from pay bargaining. The practical consequence is that a country can have low membership and high coverage, or the reverse, and comparing membership figures across countries without accounting for this produces nonsense.

Membership in most rich countries peaked between the 1950s and the late 1970s and has fallen substantially since, from around half the British workforce to roughly a fifth, and from a third in the United States to about a tenth. Several causes operate together. Employment shifted from manufacturing, which was concentrated, stable and easy to organise, to services, which are dispersed, high-turnover and much harder. Legal changes restricted organising and industrial action in several countries, notably Britain in the 1980s and various American states through right-to-work provisions removing the requirement to contribute to a union that represents you. Employer opposition became professionalised, with an industry of consultants specialising in avoiding unionisation. Globalisation weakened the threat of industrial action where production could be moved. And the composition of work changed toward contracting, agency employment and platform work, in which the legal employer is unclear and the traditional model does not fit.

Two developments run against the trend. Public approval of unions in several countries is at its highest level in decades even as membership stays low, which is a gap that organisers are actively trying to close. And organising has spread into sectors that were previously untouched, including technology firms, coffee chains, video game studios, museums and graduate teaching, frequently driven by younger workers and using publicity and social media rather than traditional structures. Against that, the legal environment in several countries makes recognition slow and contestable, and the structural problems are unchanged. The largest open question concerns platform work, where the classification of a worker as an employee or an independent contractor determines whether any of the apparatus applies at all, and where courts and legislatures across Europe and North America have reached opposite conclusions on very similar facts.

Unions exist because an individual negotiating alone faces an employer who can replace them more easily than they can replace their income, and they bargain collectively, represent members in disputes, enforce existing law and provide a voice that reduces turnover. They raise member pay by something in the range of ten to twenty percent and compress the wage distribution, and declining coverage explains part of rising inequality. Sector-level bargaining means coverage can far exceed membership, and the long decline traces to sectoral shift, legal restriction, professional opposition and changing employment forms.