Identifying the effects of publicly funded, user-oriented R&D

Identifying causal effects Selection bias is the main challenge when estimating the effect of a government program. Jaffe (2002) suggests a solution that builds on a socalled regression discontinuity design. This requires data on how the granting agency has ranked or evaluated the applicants. In Norway, such data are available through the PROVIS database. Matching these unique data to rich data sources on relevant outcomes, it is possible to implemtent a framework to estimate causal effects of R&D suppo rt on various( socio)economic outcomes. Spillovers in different public R&D support schemes We propose to assess how the extent of spillovers differs between R&D projects supported by different public R&D support schemes, with emphasis on the difference b etween large and small firms, as they tend to self-select into different schemes. Producing a reliable estimate for how externalities differ between various groups of firms using econometric techniques is problematic as the relevant spillover pool and the spillover mechanisms are largely unknown. We suggest instead to use a relatively simple and descriptive approach, where we will utilize unique project information on variables that are thought to be strongly correlated with externalities. Evaluating spi llover effects from user-directed research There is little evidence, in economic terms, of the social return from user-directed research. Such documentation is strongly needed. However, high social returns is not a sufficient criterion to assure correct selection of private projects that warrant public funding. A complicating matter is that high social return in many cases might be equivalent with high private return, implying that public funding, at least partly, displace private resources. This sub-pro ject will extend previous empirical work by MFM on economic returns from user-directed projects. The focus will be on spillovers and external effects that generate high social return.