A Multilevel Analysis of Innovation in Developing Countries

Innovation is a multilevel phenomenon. Not only attributes of firms but also the framework conditions within which firms operate matter. Although this has been recognized in the literature for a long time, a quantitative test that explicitly considers this hypothesis has been lacking. Using a large sample of firms from many developing countries, we estimate a multilevel model of innovation which connects micro and macro levels of analysis in an integrated framework. National economic, technological and institutional framework conditions are shown to directly predict the likelihood of firms to innovate. More specifically, general education, taxation of income and democratic political institutions turn out to be highly relevant, while macroeconomic stability is somewhat less relevant. But what tends to be perceived as “good” governance and the extent of public research infrastructure do not seem to make much difference. The latter result is interpreted as indicating that what matters is not how much money governments spend on the research infrastructure but rather how effectively these resources are used to leverage innovation. Nevertheless, the results also draw attention to the limits of the existing models, methods and data.