To test the efficacy of subsidised university-industry research collaboration programs on firm performance, we model data on 5000 grant applications from 2002 until 2014. The data includes firms that were successful (i.e. funded) and those that were not successful (i.e. not funded), as well as ex-ante information about the quality of the proposed project. If we compare firms whose application was rated within 10 per cent of the cut-off mark for success, the long-run effect of the subsidy (1-7 years after application) was to raise firm sales by 11.4 per cent.
Paul H Jensen’s web page at Melbourne University