Patent Boxes, Tax credits, or Both?

We analyze the relative desirability of R&D tax credits (TCs) and patent boxes (PBs) as instruments for stimulating R&D, in a setting which allows for several organizational forms within the R&D sector and where there are several market failures, all of which imply under-investment in R&D. There are two key features of the model. The first is that it is a closed economy, so the (international) profit shifting role for the PB is absent. The second is that there may be an unobservable input to R&D (e.g. managerial effort) that cannot be subsidized by a TC. The government can choose a TC, a PB, and also the main rate of CIT. We find that (i) when the unobservable input is absent, a PB should never be used, but a TC may be, if market failure is severe enough; (ii) when the unobservable input is present, the optimal policy depends on the need for tax revenue, as measured by the MCPF. When this is low, the TC should never be used, but a PB may be, but as need rises, both instruments should be used, and when it is very high,