返回Google圖書搜尋

Methodology for the Development of a Tool Focused on Highlighting Companies of Interest Related to Transnational Bribery in Colombia

Daniela Azumendi GóngoraJuan Diego Aristizábal Bernal

出版	Uniandes, 2019*
URL	http://books.google.com.hk/books?id=ov2czQEACAAJ&hl=&source=gbs_api

註釋This paper explores an Operational Research based methodology to identify companies that may have a high risk of committing transnational bribery. The main purpose is to help the Superintendence of Societies, a governmental agency overseen by the Ministry of Commerce, Industry and Tourism of Colombia, to make its investigation management more efficient. Corruption, in this case bribery, cannot be measured by a single variable, so we used different sources to put together four information nuclei: the general characteristics of the company, the risk associated with the countries where it has business abroad, whether the company has risk of financial stress in the short term, and the probability that it has altered its financial statements. Based on this data we built descriptive and predictive models to characterize and evaluate the company's current and future status, aiming to detect potential illegal behaviors. As a result, we provide to the Superintendence a risk index that helps them prioritize the opening of an investigation, allowing them to assign their resources in a more efficient way. The risk index is fed by a Factorial K-Means clustering model to assess the risk associated with having business in a particular country, a Random Forest Classifier to determine whether the company has risk of financial stress in the short term and both Beneish and Modified Jones' models to calculate the probability of financial statement's alteration.