This paper decomposes productivity growth across highly developed OECD countries in the period 1972-2000 into components attributable to domestic R&D output, a catch-up effect, and international technology diffusion via imports of hi-tech products. Two alternative specifications of 'productivity growth' are considered: (i) increments of the conventional residual measure of total factor productivity (computed under the assumption that the production function takes both physical and human capital as inputs), and (ii) the Malmquist productivity index, computed non-parametrically from a DEA-based growth accounting exercise. The methodology consists in using panel data techniques to regress productivity growth against our constructed measures of domestic R&D output, technological catch-up and technology diffusion. Domestic R&D output is derived using the Schumpeterian specification taken from fully endogenous R&D-based growth models while the catch-up and technology diffusion terms are based upon accumulation of the human-capital augmented and weighted sum of 'technology flows', proxied by appropriate measures of technological distance between the source and destination country in each pair. Our results indicate that all of these components have been important for productivity growth in OECD countries in the considered period. The estimated relative magnitudes vary significantly across countries though. On average, productivity growth has been fueled mainly by technology diffusion and technological catch-up, whereas domestic innovations account on average for only 11% of productivity growth.