Will Wilkinson recently had a post on the creation of new knowledge by location. Very interesting (even with the caveat about interpreting data mentioned in the comments).
I think this is one of the most important questions of political economy we have: what institutions are successful at creating an environment conducive to the creation of new knowledge? As a corollary question, to what extent does this knowledge need to have property rights protection (i.e., patents) to induce people to bring it into being?
One interesting insight from Will’s post is that at the nation-state level, the US generates a largely disproportionate share of new scientific knowledge, and
American institutions confer a fantastically huge positive externality, in terms of knowledge, to the rest of the world. Science is a root cause of economic growth. New knowledge enables new technologies, which enable increases in the productivity of capital, which enable growth. And good institutions are the root cause of science. If the U.S. produces most of the world’s knowledge and Asia produces most of the world’s technology, then the institutions that underpin epistemic and technical advance are chiefly responsible for growth in states that have different institutions, but which are able to import knowledge. Which is why it is nonsense to compare, say, American and French GDP growth, as if those growth rates were a function of American and French institutions in isolation from one another. Because institutions are not isolated. The interesting question is: what would French GDP growth have looked like if the U.S. had produced, say, only 10% of its actual scientific output? If the Japanese had made only a 10% of their technological advances? My sense is that French growth would not have looked good. (NB: I have picked the French because they are very good in science and tech, but even so, the point is, others are much better.)