Does South Africa's Repo Rate Lead Inflation — Or Just React to It?

Gwen N Mlondobozi

Jul 31

4 min read

What 24 years of CPI and Repo Rate data reveal about the timing of South African monetary policy

A descriptive macroeconomic analysis prepared for online publication

July 2026

When inflation rises, it's tempting to assume that raising interest rates brings it straight back down. A review of 24 years of monthly data from Stats SA and the SARB shows the relationship is real but not simple — and that timing matters more than intuition suggests.

The headline findings

The overall correlation between CPI inflation and the Repo Rate is moderate and positive, at r = 0.555.
The relationship is strongest at zero lag — the Repo Rate moves with inflation, not consistently ahead of it.
Correlation weakens steadily as the lag increases: r = 0.527 at 1 month, 0.446 at 3 months, 0.289 at 6 months, and turns negative at 12 months (r = -0.059).
The strength of the relationship shifts over time, tightening during active hiking or cutting cycles and weakening — or briefly turning negative — when the Repo Rate holds flat while CPI keeps moving on its own.
Major shocks — the 2008 financial crisis, COVID-19, and fuel-price spikes — moved inflation largely independent of policy.

Two decades of moving together

CPI inflation and the Repo Rate have broadly tracked each other since December 2001. Both climbed into the 2008 Global Financial Crisis, fell together through 2009 and 2010, and rose again during the 2022–23 hiking cycle. A scatter of all 294 months together confirms the same moderate positive relationship at a glance (r = 0.555).

Figure 1. CPI Inflation vs Repo Rate, December 2001–May 2026.

Tracking together over the long run isn't the same as one driving the other — that distinction is what the rest of the analysis tests.

The relationship is strongest right now — not months ahead

Testing the correlation at increasing time lags — comparing today's CPI against the Repo Rate from 0, 1, 2, 3, 6, 9, and 12 months earlier — shows it peaks at lag zero, at r = 0.555, then declines steadily: 0.527 at 1 month, 0.489 at 2 months, 0.446 at 3 months, 0.289 at 6 months, 0.118 at 9 months, turning negative by 12 months (r = -0.059).

Figure 4. Correlation between CPI inflation and the Repo Rate at increasing time lags (Repo Rate leading by L months).

That pattern is more consistent with a Reserve Bank reacting to inflation that has already happened than one setting rates months ahead of an anticipated move.

The relationship isn't constant

A rolling 12-month correlation shows the relationship tightens during active hiking or cutting cycles — visibly so through the 2000s hiking run into 2008, and again through the 2022–23 hiking cycle — and weakens, or dips negative, during stretches where the Repo Rate held flat while CPI kept moving on its own, most noticeably in the mid-2010s and around 2019–2020.

Figure 2. 12-month rolling correlation between CPI inflation and the Repo Rate.

This time-varying pattern is a reminder that a single overall correlation coefficient can hide as much as it reveals.

External shocks move inflation on their own

Four major shock periods stand out against the CPI and Repo Rate series: the 2008–09 Global Financial Crisis, the 2015–16 rand depreciation and drought, the 2020 COVID-19 shock, and the 2022–23 fuel-price spike tied to the war in Ukraine.

Figure 5. CPI vs Repo Rate with major external shock periods marked.

In each case, inflation moved on forces well outside the Reserve Bank's direct control — a reminder that policy is only ever part of the story.

What this means for businesses and investors

Read Repo Rate decisions in the context of the broader macro picture, not as a stand-alone signal.
Track the rolling correlation as an early sign of when the policy–inflation relationship is shifting.
Watch external shocks — commodity prices, exchange-rate swings, global crises — as leading movers of inflation in their own right.
Treat this as a starting point for causality testing rather than a final answer.
Revisit the analysis once the next full hiking or cutting cycle completes, to see whether the pattern holds.

Methodology in brief

The analysis combined monthly CPI headline year-on-year inflation data from Statistics South Africa (Table B2, Jan 1975–May 2026) with daily Repo Rate data from the South African Reserve Bank. Daily rates were converted to monthly averages and merged with CPI on a common monthly index, giving 294 overlapping months from December 2001 to May 2026. Pearson correlation, a 12-month rolling-window correlation, and lag correlation (CPI against the Repo Rate at 0–12 months earlier, using one consistent shift direction throughout) were then calculated and visualised in Python.

Limitations

This is an exploratory, correlation-based analysis. Correlation doesn't imply causation, and CPI inflation is shaped by far more than the Repo Rate alone. No Granger causality testing, VAR modelling, or Core CPI comparison was applied at this stage — all three are natural next steps.

Conclusion

South Africa's Repo Rate and CPI inflation move together over the long run, but the evidence doesn't support the Repo Rate consistently leading inflation. Instead, the SARB appears to set policy mainly in response to inflation that has already happened, while external shocks continue to drive a meaningful share of inflation on their own.