CMStatistics 2018: Start Registration
View Submission - CMStatistics
B1504
Title: Business perceptions of corruption in Europe: A multilevel analysis Authors:  Nikolai Witulski - Instituto Universitario de Lisboa ISCTE (Portugal) [presenting]
Jose Dias - Instituto Universitario de Lisboa ISCTE (Portugal)
Abstract: A two-level latent-variable model is applied to Eurobarometer data regarding employees corruption perception in the 28 countries of the European Union (EU28). For the first time the explicit impact of company and country characteristics on the perception of business corruption is studied in EU28. Both characteristics are added as first and second levels of the hierarchy, respectively. In particular, the first-order latent factor structure covers four dimensions of employees corruption perception: general perception, current practices, widespread \& influence, and punishment. The goodness of fit of all models, including the two conditional models that result from adding company characteristics and country-level indicators, can be confirmed (SRMR, CFI, TLI, and RMSEA). Company characteristics (e.g., number of employees, sector, and turnover) are statistically significant and explain the corruption perception of employees in the EU28. However, when the country-level indicators are added (e.g., Hofstedes cultural dimensions, free press, civil liberties, women representation in parliament, and the three pillars of sustainability), company characteristics become less important (only three variables remain statistically significant: sector, number of employees, and turnover). Five dimensions of Hofstede framework, two pillars of sustainable development, representation of women in parliament, free press, and civil liberties are statistically significant country-level indicators.