Main Predictor(s) of Good Corporate Governance in National Government Departments: A Decision-tree Classification and Prediction Analyses

Purpose: Good public sector corporate governance leads to good management, stewardship of public wealth, public engagement, and ultimately, better outcomes for citizens. As South Africa has one of the worst Gini coefficients in the world, its public sector should effectively address the challenges of inequality, poverty, and unemployment. Yet, almost 30% of national and provincial government departments obtain unfavourable audit outcomes. Methodology: Supporting the drive for clean audit outcomes that equate to good corporate governance, the research objective was to determine the predictors of national government departments’ corporate governance success. The study used CHAID (chi-squared automatic interaction detection) analyses, a decision-tree technique based on information reported over a 13-year period. Findings: The CHAID analyses indicated that national government departments’ corporate governance success is primarily explained by three factors, namely: a quality internal audit function (main predictor), the number of fraud incidents, and the number of internal control weaknesses. Implication: Even though all the variables in this study receive attention in literature and regulation, the problem of unfavourable audit outcomes persists. The question is where to focus on resolving the problem. Regulators may take note of where to focus their efforts in strengthening elements of good corporate governance.