A machine learning approach to analyze regulatory developments in Latin America

In Latin American countries, Regulatory Agencies were created in a context of privatization and liberalization of different economic, and utilities sectors. Given this institutional change, most of the countries in the region stopped providing public services directly to give way to private provision supervised by independent regulatory authorities Initially, the creation of an independent agencies was associated with positive results in sectors such as telecommunications and energy. In particular, the introduction of an independent regulatory agency was shown to be correlated with increases in efficiency and coverage. Despite the initial positive results, recent literature has failed to address a fundamental aspect for the understanding of Latin American regulatory policy: the notion of development. Recent literature has highlighted that, since the end of the 1990s, regulatory agencies, became more and more involved in social regulation and that since then they regulation has have a great role as policy tool. However, how the specific relation between role of regulatory agencies and public policy has not been explicitly addressed by the literature related to regulatory governance in Latin America. This paper, then, seeks to explore how regulatory agencies have evolved in the specific topics and how they have responded to demands beyond efficiency and competition. For that we created a new database using machine learning techniques that captures the different topics addressed by regulatory agencies in Colombia in the last two decades. We aim to answer the following question how has been the thematic evolution in the regulatory policy of regulatory agencies in Latin America? And does that evolution respond to development and redistribution pressures?