NextFin News - South Africa's 3% inflation target is robust enough to absorb a range of economic shocks, the central bank said in a new study, estimating a 78% probability that prices stay inside its 2%-4% tolerance band as long as no major shock arrives. The finding lands at an awkward moment: headline inflation printed at 4.3% in July, the Monetary Policy Committee split 4-2 last month with two members pushing for a rate hike, and the next policy decision on 23 September coincides with the August inflation print. The study's comfort is conditional, and the condition is being stress-tested in real time.
The Target Versus the Cycle
The South African Reserve Bank's economic note, published late Tuesday by researchers Theo Janse van Rensburg, Jeffrey Rakgalakane and Rudi Steinbach, argues that inflation expectations have become better anchored and more forward-looking, giving the 3% target the resilience to withstand shocks. In the absence of major disturbances, the authors put the odds of inflation remaining within the 2%-4% band at 78%.
That framework is new. In November 2025, Finance Minister Enoch Godongwana cut the inflation target to 3% from the 3%-6% range that had been in place since 2000, the first adjustment in 25 years, with a tolerance band of one percentage point on either side. The change was implemented over two years, giving businesses, unions and households time to adjust. When the target was flagged in May 2025, the rand surged to a two-month high against the dollar and government bond yields fell, a clear market reward for the credibility move.
The backdrop had been unusually favorable. Headline inflation averaged 3.2% in 2025, a 21-year low, down from 4.4% in 2024, extending the disinflation that began after the 2022 peak. In February 2026, headline and core inflation both sat at exactly 3.0%, aligned with the new target. Governor Lesetja Kganyago told reporters in January that the bank was on course to meet its new inflation target even in 2026, projecting a 3.6% average for this year and leaving room for two more 25 basis-point rate cuts.
Then the cycle turned. Consumer prices accelerated to 5.0% in June 2026, the highest reading since June 2024, before cooling to 4.3% in July, below the 4.5% economists had expected. Fuel inflation, which had run at 34.3% in June, slowed to 20.6% as crude pulled back from its Middle East conflict highs. Food and municipal tariffs also moderated.
Monetary policy has moved the other way. After cutting rates from a 15-year high of 8.25% in September 2024, the bank reversed course in late May 2026 with a 25 basis-point hike to 7.00%. On 23 July the Monetary Policy Committee held at 7%, but only 4-2: two of six members voted for another quarter-point increase. The next meeting is 23 September, the same day Statistics South Africa releases the August inflation figure.
The study's 78% probability is therefore not a forecast of where inflation is today. It is a statement about the structure of expectations under the new regime, and about how much shock the system can absorb before the target slips out of reach.
What the 78% Number Actually Measures
The headline figure is easy to misread. A 78% chance of staying inside the band is not a guarantee; it is a model-based probability conditional on the absence of major shocks. That qualifier does a lot of work.
The transmission mechanism the authors rely on is expectations anchoring. When firms, unions and financial analysts form price and wage expectations around the central bank's target rather than around last year's inflation, shocks pass through more weakly. A temporary oil spike raises the price level once but does not become embedded in second-round wage demands. That is the definition of a well-anchored regime, and it is what makes a point target credible rather than decorative.
The bank's own research, however, shows the anchoring is uneven. A working paper published in April by Anis Foresto, Monique Reid, Rakgalakane and Steinbach found that financial analysts and trade unions had quickly increased their focus on the new target, while firms continued to rely more heavily on prior beliefs and historical inflation. Drawing on the 2017 shift from a 6% to a 4.5% target, the paper showed that adjustment speeds vary across groups and macroeconomic conditions.
That matters because firms set prices and, ultimately, wages. If the fastest-adjusting agents are analysts and unions while the slowest are the price-setters, the second-round channel the 78% probability depends on is the weakest link in the chain. The study's robustness claim is therefore strongest for financial markets and weakest for the real economy, precisely the opposite of what a central bank would prefer.
Cyclical Relief, Structural Anchor
The disinflation South Africa has enjoyed since 2022 is, on the evidence, more cyclical than structural. Three features point that way.
First, the drivers have been reversible. The 2022 peak came from the global energy and food shock; the 2025-26 low came from the partial reversal of that shock. Fuel inflation of 20.6% year-on-year in July is still elevated, and the July moderation came from crude prices pulling back, not from a structural change in South Africa's energy market. The country remains a net crude importer exposed to the same geopolitical risks that pushed prices higher.
Second, administered prices are moderating because of a policy cycle, not a regime change. Electricity tariffs rose 8.1% in 2026, down from 10.4% in 2025; water tariffs rose 10.2% versus 12.1%. Municipal tariff increases are implemented each July, and the 2026 increases were softer across most categories. That is helpful, but tariff paths are set politically and can reverse.
Third, the mean-reversion pattern is visible in the data. Inflation fell from 4.4% in 2024 to a 3.2% average in 2025, touched 3.0% in February 2026, then jumped to 5.0% in June before easing to 4.3%. That is not a smooth structural descent to 3%; it is a cyclical oscillation around a moving target.
The structural change is the target itself, not the inflation outcome. Lowering the target to 3% is a regime shift in the policy framework: it changes the nominal anchor, the communication strategy, and the loss function the MPC optimizes. That part is durable, because a target change is not reversed by a single hot print. But the credibility of the new anchor is earned by outcomes, not declared by framework. Here the record is mixed: the bank hit 3.0% in February, overshot badly by June, and now sits at 4.3% with a committee split on whether to tighten further.
So the correct reading is that the framework is structurally tighter, but the inflation path is cyclical and mean-reverting, which means the 78% probability is most vulnerable exactly when the cycle turns against it.
The Second-Order Risk: When Robustness Becomes a Trap
The first-order reading of the study is comforting: the target can take a hit. The second-order question is what the bank does when it does.
A credible low-inflation regime should buy the central bank room to look through temporary shocks without losing credibility. That is the flexibility the inflation-targeting framework promises. But the July 4-2 vote reveals the tension. Two members were willing to hike into a slowing economy because inflation was above target; four held back. If inflation re-accelerates toward or above the 5% June print, the split could flip and a 25 basis-point hike would follow. That is the normal transmission mechanism working.
The trap is subtler. If the bank believes its own robustness finding too literally, if it interprets a 78% in-band probability as permission to wait, it risks being late. The cost of being late is not just a higher inflation print; it is a de-anchoring of the very expectations the study says are now well-anchored. The April working paper showed that prior beliefs dominate current expectations for most survey respondents. Once those priors reset higher, they are slow to come back down. Firms, the slowest-adjusting group, would be the last to believe the bank is serious again.
There is also a market-pricing channel. The May 2025 rally in the rand and bonds priced in the expectation of lower inflation and, eventually, lower rates. If the bank delivers inflation above 4% for an extended period while holding rates at 7%, the real policy rate compresses, the credibility premium unwinds, and the currency and bond market reaction can reverse. The robustness study, ironically, makes that reversal more likely if it is read as a reason to delay.
Kganyago has already framed the trade-off explicitly. In March, after the Middle East conflict pushed energy prices higher, he said: "Higher energy prices will raise inflation in the near term. We expect headline will soon accelerate to around 4%." He laid out two alternative scenarios, both more adverse than the baseline: "In the first scenario we are back to target during 2027. In the second scenario this only happens in 2028."
"In the first scenario we are back to target during 2027. In the second scenario this only happens in 2028."
That is the honest version of the 78% number: even under the bank's own alternatives, a return to target is measured in years, not months, and the adverse scenario pushes it to 2028. A target that is robust enough to survive a miss is also a target that tolerates being missed for a long time.
What Decides the Path
The immediate test is mechanical and near. On 23 September the MPC meets with the August inflation print in hand. If August holds around the July 4.3% or lower, the 4-2 majority for holding is likely to hold, and the bank keeps the repo rate at 7% while inflation grinds back toward target. If August surprises to the upside, back toward 5%, the two dissenters gain ground and a hike moves into play. Either way, the study's 78% probability is not a policy input; the data are.
Who benefits and who is exposed depends on which leg of the story wins. If the structural credibility leg holds, with expectations anchored and inflation drifting back to 3% by late 2027 as the bank's benign scenario projects, then South African assets keep their credibility premium: a firmer rand, lower real yields, and the rate-cut path Kganyago flagged in January remains viable. Borrowers benefit from the eventual easing; savers and bondholders benefit from stable real returns.
If the cyclical leg dominates, with another energy shock, a tariff reversal, or wage settlements that embed the higher prints, then the exposed parties are holders of rand assets and anyone betting on the 2026 rate cuts. The 7% repo rate would prove insufficiently restrictive, the split committee would be forced to tighten into weaker growth, and the robustness finding would be remembered as the analysis that justified waiting too long.
Three signals decide which path South Africa takes. First, the August and September inflation prints: two consecutive readings at or below 4% would confirm the July cooldown was more than a fuel-price blip; a print back above 4.5% would put the upper half of the band firmly in play. Second, the 23 September MPC vote: a unanimous hold signals confidence; a 4-2 hold with hawkish language, or a hike, signals the bank sees the cycle turning against it. Third, administered prices and wages: the July municipal tariff moderation was helpful, but the 2027 tariff cycle and any public-sector wage settlements will show whether the low-inflation norm has reached the price-setters the April paper found slowest to adjust.
Base case: inflation grinds back toward the low-4% area through late 2026 and the MPC holds at 7% until the data confirm a sustained descent, reaching the 3% target around 2027 as the bank's benign scenario projects. Upside case: energy prices stabilize, administered-price increases stay soft, and expectations converge faster than the April paper's firm-adjustment lag suggests, bringing the target forward into 2026 and reopening the rate-cut path. Downside case: a renewed commodity shock or a wage-price spiral pushes inflation back above 5%, forcing a hike and delaying the target to 2028 or beyond.
The study's real contribution is not the 78% number. It is the admission that robustness is conditional, on no major shock, on anchored expectations, on a committee willing to act before the anchor slips. South Africa's 3% target is structurally sounder than the 3%-6% range it replaced. But a target that can withstand shocks is not the same as a target that prevents them, and the difference is measured in the next two inflation prints.
Explore more exclusive insights at nextfin.ai.
