Poor countries invest far less in research and development (R&D) as a share of their GDP than rich countries. Even middle income countries often invest well under 0.5%, compared to 3% and above in advanced countries.
This fact poses a profound development mystery, and at the surface, suggests huge missed opportunities. Estimates of the social rates of return to R&D - often above 40% - in advanced countries are so high, as to justify levels of investment in developing countries that are multiples greater than those actually found. The case appears to be particularly strong for poor countries, where R&D is essential to the "absorptive'' or "national learning'' capacity that is needed to exploit technological advance originating from rich countries.
Evidence from the OECD (Griffith et al), suggests that the returns to R&D, in fact, rise steeply with distance from the technological frontier, reflecting the greater gains from catch-up afforded to ‘follower’ countries. Extending these estimates to middle income countries, the implied returns are truly large, and suggest an even greater effort in R&D is needed in developing countries than found in the advanced. To re-frame Lucas' famous observation about growth, confronted with the rates of return found in the literature, it would be hard for governments (or the World Bank) to think of anything else.
A paper  written by my colleagues and I argues that policy makers, and even the World Bank may in fact, not be so irrational. We confirm Griffith et al’s findings that the gains from “Schumpeterian catch up” rise with distance from the frontier until a country reach the income level of present day Mexico, Hungary or Malaysia. However, we find that the gains begin to fall and potentially go negative for the poorest countries, following an inverted U. That is, despite the fact that gains from technology transfer would seem greater for poorer countries, this is somehow not reflected in the data on returns to R&D.
This is because the further countries are from the frontier, the more they lack the complementary factors necessary for R&D spending to effectively contribute to growth - in particular, the people needed to generate and employ knowledge. These deficiencies represent a countervailing force to the catch up effect that ultimately makes other investments more worthwhile.
What does this imply for policy?
First, very narrow measures of policy recommendations to increased R&D spending based on simple comparisons of investment levels of advanced country benchmarks are seriously incomplete and potentially counterproductive. Without knowing the status of the other factors of production, we can’t know if a country has deficient - or even excessive - R&D investment.
That said, these findings in no way diminish the potentially enormous gains from technological catch up. Rather, they suggest that the policy package to take advantage of gains is potentially more complex and spans many dimensions.
On the supply side of knowledge, higher quality education, incentives for universities and think tanks to produce research, and research that is of relevance to private sector are all central for an innovative society.
However, the agenda goes beyond this to the demand for innovation by firms. Spurring demand requires raising the quality of firm management and entrepreneurs, explicit incentives to innovate, clear trade and competition rules, deeper capital markets, and stronger contract enforcement.
The challenge of working on so many fronts cannot be underestimated. Often, even the narrow systems of government support to innovation are fragmented and disarticulated, making them ineffective. The reforms to the national innovation system more broadly considered necessary to take advantage of gains from technological catch up will require a longer term and more comprehensive view.