Study on Employees’ Turnover in a Human Capital Network

Abstract

This study investigates employees' turnover problems in a human capital network. We take 2015 ICM problem C as an example to form a network in which employees are nodes, and links are relationships of employees in an organization. We model the probability of job hopping as a sigmoid function of a linear combination of satisfaction and relationships with those hopped (logistic regression). By using job hopping probability, we put forward a resignation decision model and design a basic structure of employees' turnover process, setting forth human capital movement mechanisms. Good employee mobility mechanisms are important indicators of the level of talent management for organizations. Market conditions, competitors, the organizations themselves and the organization talented people are changing constantly, thus employee mobility management becomes a deciding factor in the company's future development. According to international human resource authority's demonstration, reasonable flow ability rate of talents should be controlled between 5% and 10%. For internationally acclaimed companies, this rate is no more than 15%. However, more and more companies have high flow ability rate of talents. When the flow has a poor quality, that is, who the company wants to keep frequently outflows, the flow of talent becomes a talent loss (1). In this study we investigate employees' turnover problems and build a resignation decision model to estimate and forecast a company's employees churn rate in order to help business managers make decisions. Our paper has three parts. First, we discuss properties of network structure in a network in which employees are nodes, and links are relationships of employees in an organization. Second, we consider combine properties to characterize job hopping probabilities through logistic regression. Third, we build a resignation decision model considering both employees' satisfaction and their ties, which is helpful to improve employee turnover management.