La corrupción y sus causas. Análisis cuantitativo de la corrupción utilizando proxy datasets
The purpose of this paper is to evaluate popular academic theories believed to cause corruption through quantitative dataset proxies. In undertaking the exercise, the author examines various (and often competing) schools of thought on the topic, while showcasing the challenges that burden the objective study of corruption in a global context. The paper obtains a list of sixteen (16) variables extrapolated from academic literature; each (independent) variable is tied to a proxy dataset. The variables are first analysed through univariate statistics, before being subjected to bivariate correlation analysis against the (dependent) variable of corruption (itself tied to a proxy dataset, the Corruption Perception Index). The methodology employed in the analysis involves a standard mixture of statistical techniques—descriptive statistics & charts, logarithmic normalisation, Q-Q plotting, distribution curve overlays, etc.—as well as regression techniques aimed at the analysis of possible associations. The process uncovers data limitations for at least three variables (monitoring institutions, monotheistic religion, and campaign expenditure limits), while also revealing an unexpected (negative) relationship between corruption and national levels of debt. Several variables believed to impact corruption levels are confirmed, showing that rule of law, violence and instability, and national wealth all exert a strong impact on levels of corruption; other variables exhibit smaller-thanexpected associations (e.g. freedom of the press). The paper outlines future research avenues (multicollinearity analysis coupled with a robust stepwise regression model) that would generate valuable insights into global corruption trends that can then be scaled-down to accommodate local idiosyncrasies.
Derechos de autor 2018 Revista de Ciencia Política y Gobierno
Esta obra está bajo licencia internacional Creative Commons Reconocimiento 4.0.