An interesting piece from Centre for Cities suggests that by far the biggest economic impact that a university has on its city economy is through its local employment and the spending of its students rather than the interaction that it has with the local business base.
I thought this was interesting because it urges caution about the role universities might play in the development of complex local innovation policies (as is often advocated by academics studying "regional innovation systems"). Those of us sceptical about the capacity to develop such policies tend to advocate spatially neutral innovation policy that focuses on rewarding success as the best means of encouraging innovation in cities.
This doesn't mean that there are no questions to address when it comes to innovation policy and cities. For example, to the extent that innovation policy focuses on formal R&D it may not apply to much of the innovation in cities that occurs in service industries through investment in intangible assets such as design. In addition, many other policies may be more important in negatively affecting innovation in cities. High rents limit physical proximity (which benefits innovation) and constrains entrepreneurs from starting new businesses. Failing urban schools reduce the supply of highly educated workers, as does restrictive immigration policy. Tackling these problems is likely to prove far more important than any local variation in innovation policy.