Original Research
The cost of debt in South Africa: A reassessment of commonly used control variables
Submitted: 16 October 2025 | Published: 23 July 2026
About the author(s)
Anet Boshoff-Knoetze, School of Accountancy, Faculty of Economics and Management Sciences, Stellenbosch University, Stellenbosch, South AfricaGeorge F. Nel, School of Accountancy, Faculty of Economics and Management Sciences, Stellenbosch University, Stellenbosch, South Africa
Pierre D. Erasmus, Department of Business Management, Faculty of Economics and Management Sciences, Stellenbosch University, Stellenbosch, South Africa
Abstract
Orientation: The cost of debt (COD) is incorporated into the weighted average cost of capital, which is used in valuations, capital budgeting and costing applications.
Research purpose: The study investigated factors that are correlated with the COD of listed companies in South Africa.
Motivation for the study: The results of previous studies are divergent as to which factors are reliably correlated with the COD, measured as interest divided by average borrowings.
Research approach/design and method: Potential determinants were identified from existing literature. Panel data from 229 companies listed on the Johannesburg Stock Exchange (JSE) were analysed using regression techniques.
Main findings: Only one of the 18 commonly included control variables for COD, as identified from the existing literature, provided robust support for the hypothesised directional relationship with the COD: A binary variable for loss-making entities. Leverage, return on assets (ROA), asset turnover and listing age were statistically significant in the opposite direction to what was hypothesised, whereas the other 13 potential determinants did not demonstrate a stable, statistically significant relationship with the COD across different model specifications.
Practical/managerial implications: Existing perceptions of the impact of factors, such as leverage and ROA, on the COD might be misguided, which brings the results of previous studies into question.
Contribution/value add: The study raises questions about whether researchers include the correct control variables when performing regression analyses on the COD. The results also indicate that measuring the COD as interest divided by average borrowings warrants further scrutiny.
Keywords
JEL Codes
Sustainable Development Goal
Metrics
Total abstract views: 307Total article views: 491