SERC affiliate Christian Helmers has been doing some interesting research on science parks. His findings suggest that the composition of parks can matter a lot and that, from the evidence available, specialisation of science parks might be important for success.
This contrasts with the innovation process in cities, where the evidence suggests that diversity is good for ideas generation. Some time ago, urban economists Gilles Duranton and Diego Puga developed this insight in their American Economic Review Paper on Nursery Cities. They argued that it was good to be in cities to help develop initial ideas but that production then took place outside cities where land was cheap. Data from France seemed to support this theory.
But Christian's research highlights a continued gap in our knowledge. If specialisation is important in science parks, why is diversity more important in cities? When we still face uncertainties about such crucial aspects of the innovation process it is not surprising that some of us argue that policy makers lack the capacity to develop complex local innovation policies that could somehow 'fix' the externalities present in the ideas generation process.