This paper describes the identification, specification, estimation, and evaluation of econometric models of knowledge productivity and the returns-to-research. General issues related to these models are discussed and placed in context of the literature. The path from R&D investment to economic benefit is complex, convoluted, generally unknown and possibly misrepresented. The complexity arises from the intricate spatial spillover relationships and the very long time periods involved, which complicate any econometric analysis. The relevant R&D investment data are typically unavailable, incomplete and poorly measured (or approximated). The appropriate calculation of the financial benefit is not entirely clear. Some parsimonious suggestions for future research are presented.