Quantifying Family Business Owners' Attitudes towards Succession: Delving Deeper by Random Forests

Abstract

Family businesses are the lifeblood of the economic growth of the nations. However, a large gap exists about the application of machine-learning algorithms such as Random Forests (RF) to the quantification of patterns, drivers, and interactions in the succession process of family businesses. The primary aim and novelty of this study lie in the quantification of variable importance based on machine-learning algorithms, and the differences among the characteristics of family businesses, family employees, and family business owners (FBOs) for multivariate responses. For this reason, a field study was carried out in family businesses in Sivas and Ardahan provinces. The questionnaire form created by the researchers was used in this study. In this research, RF classification model was applied. RF classification models of 17 response variables were constructed as a function of 32 predictors. Implications for Central European audience: Impacts of characteristics of FBOs, family businesses, and family employees on FBOs' willingness to transfer to successors and preferences about the successor's qualities were modelled. High-dimensional data were collected from 53 family business owners (FBOs) in two cities for a total of 49 variables. As a result, the domain of the FBOs' characteristics was found to have a more profound impact on both FBO's willingness to transfer to a successor and what successor's qualities were most valued than did the domains of the family business and employee characteristics.