Knowledge Transfer in R&D Outsourcing (and Linux-Vs-Windows)

Why did Microsoft not hire all those smart programmers who ended up developing Linux through the internet? Because, we answer, the value of the information about its operating system that Microsoft should have transferred to any of them to render her productive would have been too high compared to her expected individual contribution, so that after writing a contract with Microsoft the typical developer would have run away to sell the acquired knowledge on the market. On the other hand, knowledge transfer in R&D outsourcing is not always so critical, and for example in the pharmaceutical and chemical industries research contracts are extensively used, usually in the context of a long term relationship between firm and innovator. We analyze this kind of repeated interaction, and find that when the knowledge-transfer problem is not blocking, the firm should transfer to the innovator as much information as it is compatible with the latter's incentive constraints.