Predicting voluntary turnover in employees using demographic characteristics: A South African case study

Abstract

Purpose: Employee turnover presents arguably the biggest threat to business sustainability and is a dynamic challenge faced by businesses globally. In South Africa, organisations compete to attract and retain skilled employees in an environment characterised by a burgeoning skills deficit. Turnover risk management is becoming an important strategy to ensure organisational stability and promote the effective retention of employees. The purpose of this research was to contribute to the practice of turnover risk management by proposing an approach and constructing a model to predict employee turnover based on demographic characteristics readily available in a human resource information system.

Design: An exploratory research design was employed. Secondary quantitative data were extracted from an existing human resources database and analysed. Data obtained for 2592 employees in a general insurance company based in South Africa and Namibia formed the basis for the analysis. Logistic regression analysis was employed to predict employee turnover using various demographic characteristics available within the database. A likelihood ratio test was used to build a predictive model and the Akaike information criterion and Schwarz criterion were used to test how much value each variable added to the model and if its inclusion was warranted. The model was tested by conducting statistical tests of the significance of the coefficients. Deviance and Pearson goodness-of-fit statistics as well as the R-square test of significance were used. The overall goodness-of-fit of the model was also tested using the Hosmer and Lemeshow goodness-of-fit test.

Findings: The current findings provide partial support for a predictive model explaining employee turnover. The model tested 14 demographic variables and the following five variables were found to have statistically significant predictive value: age, years of service, cost centre, performance score and the interaction between number of dependants and years of service. It is proposed that these five demographic variables be used as a model to help identify employees at risk of turnover or termed as flight risks.

Practical implications: Gaining an understanding of the factors that influence employee voluntary turnover can be instrumental in sustaining workforce stability. The proposed model could help human resources professionals identify employees at risk of turnover using data that are readily available to them. This will further enable the use of targeted interventions to prevent turnover before it happens. Decreased levels of turnover will result in cost saving, enhanced talent management and greater competitive advantage.