Documentos Departamento de Economía Universidad del Norte.
Resumen
The literature has continuously examined the relationship between market structure and innovation. In particular, Joseph Schumpeter's Mark I and Mark II innovation patterns inquired on what type of rm drives innovation based on competitive structures. Schumpeter marks, together with innovation systems theory have been widely used to classify productive sectors across the world. This article aims to deploy an exercise of this nature for Colombia in two steps. First, exploit innovationrelated metrics, like stability of innovation, technological opportunities and market concentration, through an unsupervised learning method. Second, take advantage of geographical data to map and compare the sectors of each group. Statistical analysis of the cluster reveals that each group's features align with Mark I and II archetypes. Moreover, geographical information shows patterns of agglomeration similar to those previously explored in the literature. Policy implications are discussed afterwards, concluding that the need of heterogeneous policy designs that tackle specic regional features is paramount for the country, and a potential Colombian policy instrument to channel these ndings is suggested.
