R&D in developing countries: what should governments do?

I consider the implications of recent research for R&D policy in developing countries. Typical new growth models, which assume free entry and no strategic behaviour by R&D producers, are less appropriate for policy guidance than strategic oligopoly models. But the latter have ambiguous implications for targeted R&D subsidies, and caution against the anti-competitive effects of research joint ventures. A better policy is to raise the economy-wide level of research expertise. This avoids the need for governments to pick winners, is less prone to capture, and dilutes the strategic disincentive to undertake R&D with unappropriable spillovers.