“It is much easier to forecast the inflation rate and it might even be easier, complex as it is, to forecast the exchange rate than to forecast what ratings agencies are going to do.”
So said Reserve Bank governor Letsetja Kganyago at the end of last week’s monetary policy meeting, reacting to S&P Global’s decision to upgrade South Africa’s credit ratings and keep a positive outlook in a double nudge that says policy moves are sticking.
His comments are a protest against secret exercises of ratings agencies’ judgment, and that opacity turns technical assessments into political shocks. Kganyago couldn’t be more blunt about the opacity that pushes yields higher, saying South Africa’s macro models sit on the table for anyone to test, including publishing quarterly projection models so analysts can tinker with assumptions, poke holes and tell the Bank where the monetary policy committee has been gullible.
That said, when it comes to borrowing costs, there’s no substitute for boring fiscal numbers, even in the eyes of ratings agencies. And the Treasury has numbers it can hold up.
Finance minister Enoch Godongwana has reported consecutive primary surpluses for the first time in 15 years. That’s the single clearest lever ratings agencies watch because when a government stops borrowing to pay interest, the automatic debt-accumulation engine slows and agencies must reprice the probability of fiscal stress.
What’s more, S&P once pencilled debt at about 85% of GDP for 2023, but outturns landed nearly 10 percentage points lower in the mid-70s. The latest medium-term numbers now show debt stabilising at about 78% and trending down. That divergence between forecast stress paths and realised outcomes directly undercuts the downside scenarios agencies emphasise.
These, plus a host of other reasons as outlined in a rare co-ordinated push by business leaders and the government to ratings agencies in a recent meeting, are the exact inputs agencies use. Deliver the surpluses, show the debt is falling, make SOE risks transparent and operational, and you, in theory, force the models to update from “what could go wrong” to “what is going right”.
To be clear, when South Africa’s cost of capital jumps, it’s not because a cabal of ratings agency analysts is cackling in a smoke-filled room, plotting the next downgrade. No, the real story is far more mundane. The technical, nonconspiratorial biases baked into the sovereign credit process are the financial equivalent of death by a thousand cuts. A tweak here, a conservative scenario there, a dash of opacity, and suddenly, the cost of borrowing for an entire nation is ratcheted up. The upshot is insidious. Billions are siphoned from development and into the pockets of global investors, all in the name of methodology.
Sovereign credit ratings are the gold standard of risk assessment. The process, as advertised, is a blend of statistical modelling, peer comparisons and committee deliberation, all wrapped in a veneer of objectivity.
Technical biases
The big three — Moody’s, S&P Global and Fitch — each claim to blend econometric models, scorecards and other analyses to crunch macroeconomic and fiscal data with qualitative overlays. But the weighting of these factors is as clear as mud, resulting in a system in which the so-called expert judgement is the ultimate trump card, and where technical choices about which variables to include, how to weight them and when to override the model are anything but neutral.
Credit ratings agencies love a good scorecard, using scientific metrics. Still, after all the number crunching, the final rating is subject to a qualitative overlay, a euphemism for “we’ll adjust the score if we don’t like what the model says”. The overlay culture is where technical biases take root. The lack of transparency, as called out in a politely furious paper by the UN Department of Economic and Social Affairs in how overlays are applied, means the negative narratives about governance, political risk or SOE liabilities can be overemphasised, while positive developments such as reform progress or fiscal consolidation are underplayed.
South Africa remains trapped in a ratings regime that is structurally biased against it.
Stress testing is the agencies’ favourite party trick. It’s a bit like saying, “let’s see how you’d fare if the sky fell in”. But the section and weighing of the stress scenario is anything but scientific. Adverse scenarios are acknowledged as severe but plausible, even as the probability of realising is rarely disclosed and the severity is often exaggerated.
For instance, Fitch forecast in 2020 that South Africa’s debt-to-GDP would hit 95% in 2023 and 100% in 2024. The actual outcome was 70% in 2023 — more than 24 percentage points lower than forecast. This is not an isolated incident. Both the Treasury and IMF have tended to under- or overestimate South Africa’s debt trajectory, but the credit agencies have consistently erred on the side of pessimism, justifying lower ratings and higher risk premiums.
Even when the numbers look good, ratings agencies reserve the right to adjust for implementation risk. This is the technical equivalent of moving the goal posts. South Africa’s reform progress has been acknowledged, but the benefits are discounted by overlays citing uncertainty and execution risk.
The technical machinery of sovereign credit rating is not neutral. Methodological choices have real-world consequences for the cost of capital, fiscal space and development prospects of African countries. South Africa remains trapped in a ratings regime that is structurally biased against it.
The solution, as outlined by South Africa’s G20 Africa Expert Panel, is full disclosure of ratings data, methodologies that make markets worse. It’s time they make models visible and stop letting unexplained judgment become fiscal tax.
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.