Polity Atlas
Theory

Varieties of capitalism

Hall and Soskice's (2001) firm-centred theory that capitalist economies cluster into liberal market economies, coordinated by markets and contracts, and coordinated market economies, coordinated by strategic bargaining among firms, unions and banks, with complementary institutions in each.

Compare
Source attached · not independently reviewedLast verified 2026-10-01 · revision 10 · 1 reviewersRevision historyЧитать на русском

Idea

Capitalist economies do not converge on one model. Firms must solve coordination problems with workers, investors, suppliers and other firms, and national institutions shape how they do it. In liberal market economies (LMEs), such as the United States, the United Kingdom and Australia, coordination runs mainly through competitive markets, formal contracts and hierarchies. In coordinated market economies (CMEs), such as Germany, Austria, Sweden and Japan, it runs through non-market relationships: industry-wide wage bargaining, employer associations, long-term bank finance, vocational training run jointly by firms and unions, and codetermination.

Mechanisms

  • Corporate finance: dispersed equity markets and takeovers (LME) or patient bank and block-holder capital (CME).
  • Industrial relations: decentralised bargaining (LME) or sectoral bargaining and works councils (CME).
  • Skills: general education and mobile labour (LME) or firm- and industry-specific vocational training (CME).
  • Inter-firm relations: arm's-length competition (LME) or standard-setting and technology transfer through associations (CME).
  • Complementarity: each institution works better when the others match it.

Theory

Hall and Soskice argue that institutional complementarities give each type a "comparative institutional advantage": LMEs in radical innovation and fast-moving sectors such as software and finance, CMEs in incremental innovation in sectors such as machine tools and automobiles. Because institutions reinforce each other, the theory predicts path dependence: countries should not converge on the liberal model under globalisation. Hall and Gingerich (2009) built indices of coordination for OECD countries and reported that countries cluster along one dimension, and that growth was higher where institutions in different spheres were consistently liberal or consistently coordinated.

Real examples

  • Germany is the standard CME: sectoral collective bargaining, works councils and board-level codetermination, a dual vocational training system.
  • The United States and the United Kingdom are standard LMEs.
  • France, Italy, Spain and Portugal fit neither type well and are sometimes called "mixed" or state-influenced market economies.

Critique

Hancké, Rhodes and Thatcher's Beyond Varieties of Capitalism (2007) argue that the dichotomy neglects the state, conflict and change, and squeezes many economies into two boxes. Critics also note that German coordination eroded after the 1990s, with falling collective-bargaining coverage, which the static model did not anticipate. The theory is about rich democracies and says little about state capitalism or developing economies.

Results

The LME/CME distinction is one of the most used frameworks in comparative political economy and fits the contrasts between the United States and Germany well. Evidence on complementarities and on predicted specialisation patterns is mixed and depends on measurement.

Claims

EstablishedCLM-1149 · checked 2026-10-30

Hall and Soskice (2001) distinguish liberal market economies, where firms coordinate mainly through markets and formal contracts, from coordinated market economies, where they rely on non-market institutions such as sectoral bargaining, employer associations and patient capital.[1]

DisputedCLM-1150 · checked 2026-10-30

Hall and Gingerich (2009) report that OECD economies cluster along a single dimension from liberal to coordinated and that growth was higher where institutions across spheres were consistently of one type, a test of complementarity that critics consider sensitive to measurement.[1][2]

Assess it in the Governance Explorer →

Revision history

  1. 2026-10-30rev 10 — Translations import
  2. 2026-10-30rev 9 — Translations import
  3. 2026-10-30rev 8 — Translations import
  4. 2026-10-30rev 7 — Translations import
  5. 2026-10-30rev 6 — Translations import
  6. 2026-10-30rev 5 — Translations import
  7. 2026-10-30rev 4 — Translations import
  8. 2026-10-30rev 3 — Translations import
  9. 2026-10-30rev 2 — Translations import
  10. 2026-10-30rev 1 — Seed import