˘ ˇ ˆ˙ ˝˛ ˚ ˜˜ ˛ ! ˙˘ # $ˇ˘% ˙& *+*, ˆ &-..///˛˘0 ...

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hile R&D activity is relatively dispersed around the world, innovation of new-to-the-world technologies has historically been concentrated in a few countries.)]

4 During the 1970s andthe early 1980s, only Switzerland, a relatively small but technology-intensive economy, achieveda per capita international patenting rate comparable to that of inventors. This variationamong advanced economies in their ability to innovate at the global frontier raises an empiricalpuzzle: if inventors can draw on technological and scientific insights from throughout the world,why does R&D productivity depend on location?This question is important for at least two reasons. First, technological innovation plays acentral role in the process of long-run economic growth. Despite substantial agreement aboutthis proposition, there is a great deal of debate about the underlying drivers of the innovationprocess itself. International variation in R&D productivity presents an opportunity to examinehow different influences contribute to technological innovation and thereby distinguish amongdifferent drivers of productivity growth. Second, understanding international differences in R&Dproductivity informs science, technology, and broader issues of economic policy.

5 Most R&Dproductivity studies have focused on the innovation production function ( , the relationshipbetween inputs for innovation such as R&D expenditures or manpower and visible outputssuch as patenting or new product announcements) within a given public policy environment(Griliches, 1984; 1998). However, estimating the impact on innovation of country-level policydifferences ties more directly to policy by these twin concerns, this paper evaluates the sources of differences amongcountries in terms of their production of visible innovative output. To do so, we introduce anovel framework based on the concept of national innovative capacity. National innovativecapacity is the ability of a country as both a political and economic entity to produce andcommercialize a flow of innovative technology over the long term. Innovative capacity dependson an interrelated set of investments, policies, and resource commitments which underpin theproduction of new-to-the-world technologies.

6 National innovative capacity is not the realizedlevel of innovative output at a single point in time but reflects the more fundamental2determinants of the innovation process. Differences in national innovative capacity reflectvariation in both economic geography ( , the level of spillovers between firms) as well ascross-country differences in innovation policy ( , the level of public support for basic researchor legal protection for intellectual property).We develop the national innovative capacity framework by drawing on three distinctareas of prior research: ideas-driven endogenous growth theory (Romer, 1990), the cluster-basedtheory of national industrial competitive advantage (Porter, 1990), and the literature on nationalinnovation systems (Nelson, 1993). Each of these perspectives seeks to identify factors in anational environment which determines the flow of innovation. Not surprisingly, these theoriesshare common elements; however, each also contributes a distinct perspective.

7 For example, inmodels of ideas-driven growth, the ideas production function depends on two aggregate factorsthat influence the rate of innovation in a national economy: the prior stock of knowledgeaccumulated by that economy and the level of R&D effort devoted towards ideas production (asopposed to final goods production). Porter (1990) emphasizes the microeconomic underpinningsof innovation in country-specific industrial clusters; this relationship depends on subtleinteractions between input supply and local demand conditions, the presence and orientation ofrelated and supporting industries, and the nature of local competitive rivalry. By focusing onindustrial clusters ( , information technology) rather than individual industry segments ( ,printers), this perspective highlights how the rate of innovation depends critically on bothknowledge spillovers and the nature of technological interdependencies across related sectors. Finally, the national innovation systems literature, built on rich descriptive accounts of theorganization of innovation in specific countries, tends to emphasize the role of the overallnational policy environment ( , intellectual property or trade policy), the educational sector, aswell as more idiosyncratic institutions that affect innovation but for which internationalcomparison is difficult ( , the rules of specific funding agencies in individual countries).

8 When considered together, these perspectives suggest that the determinants of nationalinnovative capacity can be divided into several broad areas. First, national innovative capacitydepends on the presence of a strong common innovation infrastructure, or cross-cutting factorswhich contribute broadly to innovativeness throughout the economy. Among other things, the3common innovation infrastructure includes a country s overall science and technology policyenvironment, the mechanisms in place for supporting basic research and higher education, andthe cumulative stock of technological knowledge upon which new ideas are developed andcommercialized. The common innovation infrastructure therefore includes several of the keyelements highlighted by the national innovation systems perspective and ideas-driven growththeory. Second, a country s innovative capacity depends on the more specific innovationenvironments in a country s industrial clusters. As emphasized by Porter (1990), whether firmsinvest and compete on the basis of new-to-the-world innovation depends on the microeconomiclogic inherent in their local competitive environment.

9 Ultimately, it is the microeconomicconditions associated with a nation s clusters which determine whether firms respond totechnological opportunity and innovate at the global frontier. Third, national innovative capacitydepends on the strength of linkages between the common innovation infrastructure and specificclusters. The productivity of a strong national innovation infrastructure is higher when specificmechanisms or institutions, such as a strong domestic university system and funding mechanismsfor new ventures, migrate ideas from the common infrastructure into commercial theoretical framework can be used to evaluate the sources of cross-country R&Dproductivity differences. The framework highlights the potential importance of the compositionof research funding and performance. For example, while public R&D spending contributes tothe common innovation infrastructure, private R&D spending is a more direct reflection of theinnovation environments of a nation s industrial clusters.

10 As well, the framework incorporatesboth the economic and political roles played by national boundaries. More precisely, whereasone stream of prior research focuses on how geography mediates knowledge spillovers anddifferential access to human capital (Porter, 1990; Jaffe, Henderson and Trajtenberg, 1993;Krugman, 1991), a second area of prior work has emphasized how national R&D productivitydifferences may be driven by differential public policies and institutions (Nelson, 1993). Ratherthan concentrating on a single explanation for R&D productivity differences across countries, ouranalysis incorporates a relatively diverse set of potential drivers and then adjudicates theirempirical salience. Finally, the national innovative capacity framework suggests that whileformal ideas-driven growth models offer insight into the broad determinants of innovative4activity, cross-country differences in R&D productivity may be driven by more nuanced factors(such as those related to the composition of funding, public policy, and cluster-specificcircumstances).