Pivot or persist

• Department of Economics

This event features a discussion with Assistant Professor Christoph Carnehl (Bocconi University).

A resource-constrained researcher seeks to advance knowledge by choosing the ideas to explore and the methods to apply to them. Similar ideas require similar methods, introducing correlation in the researcher's two-dimensional search problem. Our main result proves that the researcher's problem is convex, allowing us to draw parallels to the canonical consumer choice problem. We obtain a rich and tractable framework to study the properties of the optimal research mix that we leverage to generate novel insights in the context of scientific discovery and innovation. We find that ideas are an inferior good, implying that more resources can lead to fewer ideas generated. Applying our results to research teams, we show that task allocation is based on comparative advantages, leading to specialisation. Endogenous team formation relies on absolute advantages as well, generating either maximally diverse teams or no teams at all. When studying the implications of technology adoption, we find that tools supporting idea generation, such as LLMs, have ambiguous effects on the amount of ideas produced.

Co-author: Martino Banchio