IESET.
Conditions › Conditions favoring markets

Innovation and technology diffusion

Private-sector-led commercialisation of new technologies, operating within ecosystems that combine publicly-funded basic research, risk-tolerant private capital, flexible labour markets, and effective IP regimes, has empirically outperformed state-directed alternatives at producing and diffusing frontier technology. The full claim requires acknowledging that basic research is largely publicly funded and that purely private systems without that foundation also underperform.

confidence: medium highConditions favoring marketsentry added 2026-04-29innovation_and_technology_diffusion

Institutional features that make the model work

›Risk tolerant private capital
Venture capital provides staged financing with high tolerance for failure, exit via acquisition or IPO, and concentrated returns from successful bets. Lerner's work on innovation finance documents VC-backed firms' disproportionate share of patents and productivity contribution.
›Entry exit and labour mobility
Frontier-technology development requires rapid firm entry, firm death, and engineer mobility across firms. US state non-compete enforcement variation (California's ban is widely cited) is associated with higher regional innovation (Gilson 1999, Marx et al. 2009).
›Publicly funded basic research foundation
NIH, NSF, DARPA, and equivalent public funders supply most of the basic-research input that downstream private commercialisation exploits. Mazzucato's critique is correct that the US system is not a purely private innovation system; it is a public-basic / private-applied hybrid.
›Ip regime that balances appropriability and diffusion
Patent systems create some appropriability for private R&D; excessively strong or broad patents retard follow-on innovation. Bayh-Dole (1980) enabled US university commercialisation of federally-funded research and is associated with large biotech cluster formation.
›Deep capital markets and acquisition ecology
Public equity markets and a large-firm acquirer ecology enable startup exit and scale-up financing. European innovation clusters have historically been capital- constrained on late-stage financing relative to the US.
›University industry proximity and cluster effects
Geographic clusters (Stanford/Silicon Valley, MIT/Route 128, Oxbridge/Cambridge biotech) concentrate tacit knowledge, labour-market matching, and VC/founder density that generate persistent local advantage.

Supporting cases

silicon_valley_persistence_as_global_tech_frontier

A narrow geographic cluster has sustained decades of frontier-technology leadership across semiconductors, PCs, internet, smartphones, cloud, and AI. The institutional combination has not been successfully replicated despite sustained attempts.

  • Saxenian (1994). Regional Advantage.
  • Kerr & Robert-Nicoud (2020). Tech Clusters. JEP.
eu_tech_scale_up_gap

The European Union, despite comparable R&D spending and higher-education quality, has not produced tech unicorns at a rate comparable to the US. Explanations emphasise fragmented single market, shallower late-stage VC, bankruptcy stigma, and non-compete enforcement.

  • Draghi (2024). The Future of European Competitiveness. European Commission.
us_pharma_biotech_cluster_formation_post_bayh_dole

The Boston/Cambridge biotech cluster emerged largely after Bayh-Dole (1980) allowed universities to patent federally-funded research; university spinouts and VC-backed biotech became a primary source of new drug modalities. Combines public basic science (NIH) with private commercialisation.

israel_tech_ecosystem_small_country_exemplar

Israeli tech sector developed through a combination of defence-sector basic technology, military-service networks, US-Israel binational R&D, and an active VC industry. A small country demonstrating the institutional combination can work at small scale.

  • Senor & Singer (2009). Start-up Nation.

Failed replications

european_national_champion_tech_programs

Successive French and European attempts at state-led national-champion tech programs (Plan Calcul, Minitel scaling, various cloud-sovereignty initiatives) have not produced globally competitive firms at the rate that a market-led ecosystem with equivalent resources has in the US.

soviet_consumer_electronics_and_computing

The Soviet Union had strong basic-physics research capacity but consistently underperformed on commercialising civilian technology; by the 1980s the gap in PCs, consumer electronics, and semiconductors was an order of magnitude.

Disconfirming cases

china_state_directed_industrial_policy_frontier_advances_2015_present

China has made genuine frontier advances in EVs, batteries, solar PV, high-speed rail, and selected AI applications under a heavily state-directed model combining subsidies, preferential credit, and strategic coordination. This is the strongest contemporary counterexample to a pure market-led innovation story and should not be dismissed. Honest interpretation: the state-directed model appears to work in scaling and diffusing technologies with relatively well-understood engineering, less clearly in frontier-creating domains.

  • Breznitz (2021). Innovation in Real Places.
japanese_government_research_coordination_1970s_1980s

MITI-coordinated research consortia in semiconductors (VLSI project) and other domains produced genuine catch-up progress in the 1970s-80s. The model worked as catch-up industrial policy but was less successful at generating frontier-leading technologies in the 1990s- 2000s.

What this condition is NOT

  • A claim that private-sector R&D alone, without public basic research, produces optimal innovation
  • A denial that industrial policy can work in catch-up and diffusion phases, as East Asian experience shows
  • A claim that IP rights should be maximalised — excessive patent breadth retards follow-on innovation
  • A claim that market failures in innovation (spillovers, public-good character of knowledge) do not justify R&D subsidies
  • A claim that antitrust should be weak in tech markets

Policy implications

Policy should sustain public funding of basic research at generous levels, maintain a permissive environment for entry/exit and labour mobility (including non-compete restrictions), keep capital markets deep and bankruptcy procedure efficient, and use narrow, conditional industrial policy in specific coordination-failure domains. The hybrid public-basic / private-applied model is the empirically strongest arrangement; neither pole alone suffices.

Framework position

Conditional on publicly-funded basic research, deep and risk-tolerant private capital, labour mobility, and IP institutions that balance appropriability and diffusion, private-sector-led commercialisation of innovation is the empirically dominant arrangement for frontier-technology production. The framework accepts that state-directed industrial policy has a real track record in catch-up and diffusion contexts (East Asia, contemporary China in selected sectors) and does not universalise a pure market-led claim.