Prediction and Ranking of Corporate Diversity in European and American Firms

Currently, corporate social responsibility and environmental/social/governance topics are gaining more relevance in business and finance. Attention to corporate diversity in boards and the workforce is included in this trend. Although most studies focus on executive boards and objective scores, the perception of diversity by employees and its rankability are not fully understood or researched. In this paper, we analyze corporate diversity rankings from the perspective of predictive and prescriptive analytics. Inside predictive analytics, the perceived diversity of a sample of 350 European diversity leader companies is predicted by using three different feature sets (raw financial data, ratios and objective diversity variables) and three machine learning algorithms (K Nearest Neighbors, Logistic Regression, Decision Tree). The best performing algorithm is the Decision Tree, and all three feature sets outperform one random dummy algorithm; the best performing set is the financial ratios set. Inside prescriptive analytics, several rankings involving American companies are intersected and compared in three exercises (studying diversity categorization, ethnic origin and comparing diversity with other unrelated metrics). From these, global rankings were built to search for the best possible agreement among the rankings. These results with both predictive and prescriptive analytics encourage managers to strategize and include diversity in management, as well as employ new technologies in their decision-making processes.