A recent stream in the program evaluation literature has focussed on the estimation of causal effects in the presence of continuous treatments. Dose-response functions based on propensity-score methodologies can be employed, under the unconfoundedness assumption, to perform this analysis. An interesting area of application is that of r&d subsidisation programmes, where little is known about what is the right size of subsidies or of the underlying private investments to be targeted. Focussing on a regional small-business r&d programme implemented in Italy, we estimate a flexible dose-response function and find a roughly inverse U-shaped relation between subsidy and future r&d investment.
Keywords: R&d subsidies; dose-response functions; generalized propensity score.
Jel Code: C21, L53, O38