Bhide, innovation and basic research

I highly recommend this podcast interview with Amar Bhide at Columbia Business School. Bhide studies entrepreneurship and has a very good understanding of how business innovation actually works. (This is quite rare for an academic -- even those in economics departments or business schools.)

The whole podcast is worthwhile, but particularly the last 25 minutes (starting at about minute 40), in which Bhide discusses Schumpeterian vs Hayekian models of innovation, Knightian uncertainty, securitization and the financial crisis.

I would fault Bhide in one area: the catchy point of his recent book is that high level scientific breakthroughs (e.g., Archimedes' Principle of buoyancy) rapidly become public goods, whereas the nitty gritty skills and know-how necessary to create a useful product (e.g., shipbuilding) are more readily localized and the real source of competitive advantage. He focuses on small optimizations of existing technology to solve everyday problems; it's true that this is the more common type of innovation. But he neglects how whole new industries are sometimes created by real technological breakthroughs.

To take a particular example, look at the Fairchild Semiconductor guys. They weren't just average engineers -- they were trained in advanced labs and hired by the guy who actually invented the transistor (Shockley). It's no coincidence that the country that paid for the basic R&D went on to invent and capture the semiconductor industry. Of course, it was also necessary to have venture capital (actually, they sort of invented it), big capital markets and proximity to customers. But Bhide sounds like he wants to leave out the R&D aspect -- that can be done anywhere, he says. Yes, it can be done anywhere, but the commercialization is more likely to happen nearby, as long as the other necessary factors are also available. I could tell a similar story about biotech or internet companies in the bay area.

Basic research tends to be underfunded primarily because the inventor seldom captures the lion's share of future returns. This makes it less appealing to private investors compared to small optimizations (applications of existing technology) which lead more directly to products and profits. Therefore, funding for basic research must come from the public sector. Societies typically spend too little rather than too much, as future returns are diffuse and have no singular advocate.