Are Perceptions of Corruption Matching Reality? Theory and Evidence from Microdata
Corrado GermanaCorrado LuisaDe Michele GiuseppeSalustri Francesco
CEIS Research Paper
Some criticism has been raised on the actual capability of corruption perception-based indices to gauge the essence of concepts they aim to measure. One can argue that perceptions about corruption are not matching reality and could be the reection of distorted truth. Based on this evidence we provide a theoretical ground for the corruption decision-making process (objective corruption) and the corruption perception-making process (subjective corruption) which accounts for the role of media attention. From the theoretical model we are able to derive testable implications for the empirical analysis, i.e. whether socio and cultural norms can explain the gap between the two measures of corruption across Europe. We employ a generalised setting of the structural equation models to build latent indices of objective and subjective corruption from our microdata exploiting the information on various economic, geographic and socio-demographic factors that can a¤ect the perception and the experience of corruption practices. The resulting indices allow us to de ne country rankings for both types of corruption and draw a geopolitical picture of the phenomenon across Europe. We also show that countries where the quality of media is higher are associated with lower di¤erences between perceived and real corruption.
Keywords: Perceived and experienced corruption, Latent variables, Latent multi- dimensional index, Multiple indicators multiple causes models, Generalized SEM MIMIC.
JEL codes: C43, C83, D73, H11, R50.
Date: Thursday 14 December 2017
Revision Date: Thursday 14 December 2017