Institutional features that make the model work
›Peer review grant councils
›Mission agencies with procurement power
›Long horizon research institutes
›University research plus tech transfer
›Complementary private applied r and d
Supporting cases
Post-Vannevar-Bush US federal research funding (NIH, NSF, DARPA, DOE national labs) produced the foundational science underlying semiconductors, the internet, GPS, mRNA vaccines, GLP-1 drugs, CRISPR. Mazzucato documents the state origin of most core technologies in a modern iPhone.
Max Planck institutes for basic research plus Fraunhofer institutes for applied research form a durable public-private bridge supporting German industrial strength.
CERN as multi-decade publicly-financed basic physics infrastructure produced the Web as a by-product and trained generations of researchers who dispersed into industry.
Decades of NIH-funded mRNA research (Katalin Karikó and others) enabled the 2020 rapid-response vaccines. Private development was fast only because the foundational science existed. Counterfactual of no NIH funding is an order of magnitude slower response.
Disconfirming cases
Soviet science produced strong outcomes in narrow directed fields (physics, mathematics, space) but failed in distributed applied innovation (computing, consumer tech, biotech). Illustrates that state funding requires peer-review and decentralised allocation, not top-down direction.
State-directed commercial R&D in picked firms (Concorde, various European "national champions" in semiconductors, post-war British computing) frequently underperformed. The case for state R&D is upstream of commercialisation, not in picking commercial winners.
The counterfactual claim that pure private pharmaceutical R&D would have produced the same drug pipeline is inconsistent with the evidence: most novel molecular entities trace to publicly-funded upstream research; private pharma's comparative advantage is in development, trials, and distribution, not target discovery.
What this condition is NOT
- An argument for state-directed commercial R&D or national-champion industrial policy
- A claim that every publicly-funded research programme is productive — many are not
- A claim that private R&D is wasteful — private applied R&D is complementary and large
- An endorsement of political direction of research topics — peer review and scientific autonomy matter
- A claim that state research funding in weak-institutional settings produces the same returns
Policy implications
Public funding should concentrate on basic research, long-horizon applied research, and general-purpose-technology challenges where private appropriability is weakest. Grant councils should remain peer-reviewed and insulated from political direction. Mission agencies can set directional challenges but should buy outcomes rather than pick firms. Tech-transfer and strong IP enforcement downstream let private firms capture the commercialisable slice. Countries underinvesting in upstream research (UK R&D intensity below OECD median; post-2010 US federal research in relative decline) face predictable long-run productivity drag.
Framework position
Foundational R&D is one of the strongest conditions favouring intervention because the market failure (non-appropriable spillovers) is quantitatively large, empirically confirmed (social returns to R&D estimated at multiples of private returns in cross-country panels), and has a well-functioning institutional solution (peer-reviewed grant councils plus mission agencies plus long-horizon institutes). The framework endorses sustained public funding of upstream research with private-sector complementarity downstream, while rejecting state-directed commercial R&D as a distinct and weaker claim. The institutional detail — peer review, scientific autonomy, multi-decade horizons, complementary rather than substitutive positioning relative to private R&D — is what distinguishes successful public research ecosystems from failures.