“They have fiddled the figures!” exclaimed former UK prime minister Rishi Sunak in response to Rachel Reeves’s maiden budget as chancellor of the exchequer last year. His accusation related to how public debt is defined and measured.
Indeed, how debt is measured plays a role in how fiscal policy is managed. The “fiddle” that so outraged Sunak increased the UK government’s scope to borrow without changing the debt target or limit set out in the UK’s fiscal rules.
Fiscal sustainability analysis often starts with an estimate of debt, usually expressed as a ratio of GDP. This ratio is also used in determining a country’s sovereign credit risk, which directly affects the cost of government borrowing.
Depending on how public debt is measured, SA’s debt for 2024/25 ranged at 69.2%-129% of GDP. This may seem at odds with the gross and net loan debt ratios of 76.9% and 73.8% as presented by the National Treasury in the national budget this week. However, there are different measurement approaches particularly in relation to debt coverage. It depends on which spheres of government and which public institutions are included in the debt calculation, as well as which debt instruments are considered.
The Treasury uses a relatively narrow definition for public debt in the budget. It essentially covers only the national government, for which scholars use the term the “budgetary central government”.
This excludes extrabudgetary units, such as the SA Revenue Service, and social security funds such as the Road Accident Fund (RAF). From the third quarter of 2024, these entities’ debt amounted to just under R300bn. It also excludes the debt of provincial and local governments and their extrabudgetary units. Their combined debt totalled a further R346bn.
The budget figures also don’t include the debt issued by non-financial state-owned companies (SOCs), such as Eskom, Transnet and Denel. From the third quarter of 2024 their debt stood at R927.4bn.
Lastly, the budget numbers also exclude the liabilities and debts of the Reserve Bank and other public financial corporations such as the Development Bank of Southern Africa and the Government Employees Pension Fund. Other public financial corporations’ debt totalled about R2.75-trillion. One could extend this definition to include the Reserve Bank, which would add R1.2-trillion.
Accounting for all of these entities would broaden the debt coverage from the “budgetary central government” to the “consolidated public sector”. This would increase the 2024/25 gross debt ratio from 76.9% of GDP to about 129%.
But that doesn’t mean that the Treasury has “fiddled the figures” since using the consolidated public sector measure is not necessarily the standard approach. Generally, the quality of SA’s debt statistics is not in question. We have not seen a strategic use or change in the definition of debt to artificially meet targets or “fiddle” with debt statistics.
Misleading comparisons
However, it is important to be aware of the different approaches to measuring public debt for two main reasons. First, failing to account for differences in debt coverage often leads to misleading comparisons between countries. Second, these less examined debts can still pose a risk to the government’s balance sheet.
Consider the IMF’s World Economic Outlook database, which is often used to make cross-country comparisons. It put general government gross debt for Finland and SA at 81.3% and 75%, respectively, for 2024, making SA’s public finances appear healthier. However, the debt coverage of these figures differs significantly. In SA’s case it covers only the “budgetary central government”. In Finland the definition is much broader. Finland’s central government debt ratio, which is still more broadly defined than ours, was just 61.2% in 2024.
The risk that consolidated debt poses to the public finances depends on the extent to which the underlying public entities and funds have a claim on the fiscus. Eskom’s recent debt transfer to the national government’s balance sheet is a clear example of this.
One way to at least partially account for these potential risks is to adjust the official debt numbers for the state’s contingent liabilities — those obligations that could result in state expenditure if specific events occur. In SA, the state’s contingent liabilities consist mainly of guarantees to SOCs.
From 2004/05 to 2024/25, these contingent liabilities increased from R15.9bn to R1.1-trillion, with Eskom and claims against the RAF constituting the largest increases. Adjusting the Treasury’s measure of gross debt to include all these contingent liabilities would increase the ratio from 76.9% to 92.5% of GDP in 2024/25.
Arguably, contingent cash balances in the form of the Gold and Foreign Exchange Contingency Reserve Account (GFECRA) should also be included. After all, net loan debt is calculated by deducting the government’s cash balances from gross loan debt and effectively the GFECRA operates as a second cash buffer. In years where the account sustained losses, including the GFECRA in the net debt calculation would increase net debt — and vice versa.
If the GFECRA is included in the calculation of SA’s net debt for 2024/25, it would fall to 69.2% of GDP from 73.8% because we made GFECRA “profits’’ in that year. But if contingent liabilities were then added, it would rise to 84.9%.
The bottom line is that as much as we fixate on the official debt numbers, it is important not to lose sight of the broader, less examined risks that lurk on the public balance sheet.
• Botha is an independent consultant and researcher; Bisseker is an economics writer and researcher at the Bureau for Economic Research.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.