Fixed investment D2
Is GFCF as a share of GDP rising or falling, how is it split between private and public, and is the trend accelerating or decelerating?
Is fixed investment rising or falling? D2
Rising = bullish, falling = bearish — your own rule, cited rather than editorialised. The split is the one you asked for: private, public corporations, general government.
Table view — Gross fixed capital formation
| Date | Gross fixed capital formation |
|---|---|
| 30 Jun 2026 | 13.70% |
| 31 Mar 2026 | 13.20% |
| 31 Dec 2025 | 13.53% |
| 30 Sept 2025 | 13.75% |
| 30 Jun 2025 | 14.02% |
| 31 Mar 2025 | 14.25% |
| 31 Dec 2024 | 14.19% |
| 30 Sept 2024 | 14.57% |
| 30 Jun 2024 | 14.71% |
| 31 Mar 2024 | 14.63% |
| 31 Dec 2023 | 14.80% |
| 30 Sept 2023 | 14.96% |
| 30 Jun 2023 | 15.32% |
| 31 Mar 2023 | 14.66% |
| 31 Dec 2022 | 14.62% |
| 30 Sept 2022 | 14.27% |
| 30 Jun 2022 | 14.01% |
| 31 Mar 2022 | 13.81% |
| 31 Dec 2021 | 13.37% |
| 30 Sept 2021 | 13.14% |
| 30 Jun 2021 | 12.85% |
▲Bullish 13.7% of GDP in Q2 2026, unchanged since the September Quarterly Bulletin of 29 September: +0.50 points on Q1's 13.2% and the first rise after four consecutive falls (14.25, 14.02, 13.75, 13.53, 13.2). All three legs rose: private business 9.13 → 9.49, public corporations 1.61 → 1.63, general government 2.45 → 2.58. A year earlier the ratio stood at 14.02%. Bullish on the rule you gave us: fixed investment as a share of GDP rising is bullish, and it rose. Most of the rise came from private business, which is what your thesis most needs to see. One quarter does not reverse a year: the ratio is still below Q2 2025, so this is a turn to watch, and the December Bulletin will show whether it holds.
Is the trend accelerating or decelerating? D2
This is the one chart in the system whose colour says good or bad, because the brief defines the sign as the signal in so many words. Two triggers, both named: steepening — the decline is getting worse — and persistence — a full year of decline, at any rate. Today steepening is not firing and persistence is not firing. A move smaller than 0.05pp is inside the rounding of a revision and cannot start a run.
Table view — Quarter-on-quarter change
| Date | Quarter-on-quarter change |
|---|---|
| 30 Jun 2026 | 0.50% |
| 31 Mar 2026 | -0.33% |
| 31 Dec 2025 | -0.22% |
| 30 Sept 2025 | -0.27% |
| 30 Jun 2025 | -0.23% |
| 31 Mar 2025 | 0.06% |
| 31 Dec 2024 | -0.38% |
| 30 Sept 2024 | -0.14% |
| 30 Jun 2024 | 0.08% |
| 31 Mar 2024 | -0.17% |
| 31 Dec 2023 | -0.16% |
| 30 Sept 2023 | -0.36% |
| 30 Jun 2023 | 0.66% |
| 31 Mar 2023 | 0.04% |
| 31 Dec 2022 | 0.35% |
| 30 Sept 2022 | 0.26% |
| 30 Jun 2022 | 0.20% |
| 31 Mar 2022 | 0.44% |
| 31 Dec 2021 | 0.23% |
| 30 Sept 2021 | 0.29% |
| 30 Jun 2021 | -0.22% |
–Neutral +0.50 at Q2 2026, quoted from scripts/derived.mjs: the last four quarterly deltas are -0.27, -0.22, -0.33 and +0.50. Persistence false, steepening false, signal none. The early warning you asked for has cleared. It fired on persistence, four negative quarters in a row, and the September Bulletin broke the run with the only positive delta in the last five quarters. It is neutral and not bullish because one positive quarter is a turn, not a trend. If the December Bulletin shows a second positive delta, the acceleration will be real.
The amount invested, against last year D2
Clarified 2026-08-31: "a quantum is being invested… the amount being invested in relation to the previous year, in line with chart D4c." So this is D4c’s chart on D2’s data — the rand actually invested, cumulative through the calendar year, this year’s curve against last year’s — and the gap between the curves is the answer. It is drawn twice because rand and constant-2015 rand are different units, and this system puts one unit on one axis: nominal to match tax collection, and real because a nominal year-on-year gap is partly just inflation.
At the same point last year the pace reads +1.3% in nominal rand and +0.8% in real terms — the difference between the two is inflation. Quarterly national accounts, so the curves step four times a year where tax collection steps twelve, and the latest quarter arrives one to two quarters after it ends. Not scoped by the horizon control — a calendar year is the window.
Table view — Invested, cumulative from 1 January
| Date | 2026 | 2025 |
|---|---|---|
| 31 Dec 2000 | — | R1,061bn |
| 30 Sept 2000 | — | R789bn |
| 30 Jun 2000 | R525bn | R518bn |
| 31 Mar 2000 | R262bn | R261bn |
Table view — The same, in constant 2015 rand
| Date | 2026 | 2025 |
|---|---|---|
| 31 Dec 2000 | — | R647bn |
| 30 Sept 2000 | — | R480bn |
| 30 Jun 2000 | R318bn | R316bn |
| 31 Mar 2000 | R160bn | R160bn |
–Neutral R263.13bn invested in Q2 2026, +2.52% on Q2 2025's R256.67bn; in constant 2015 prices R157.58bn, +1.52% on a year earlier. Q1 was restated from R261.26bn to R261.76bn. The first half of 2026 totals R524.89bn against R517.91bn in the first half of 2025, +1.35% (real +0.75%). Neutral. Measured the way you asked, against the same point of the previous year, the rand invested is growing again in real terms, which it had stopped doing in Q1. The nominal pace is still below inflation of 4.4%, and the real gain is small, so this is a halt to the contraction and not yet a boom.
The running total, and the years compared D2
The brief asks for fixed investment "as a percentage of GDP, and if possible as a running total". A running total of a ratio is not a quantity — four percentages of GDP cannot be added together and mean anything — so this shows the four-quarter rolling average of the ratio, a reading you confirmed on 30 August, and beside it the annual comparison you asked for at the same time: each complete calendar year’s average, so one year stands against another directly. On 31 August you added the other half of the ask — the running total as an amount, “a quantum is being invested” — which is the section above, in rand.
Table view — Four-quarter rolling average
| Date | Rolling 4q average |
|---|---|
| 30 Jun 2026 | 13.54% |
| 31 Mar 2026 | 13.63% |
| 31 Dec 2025 | 13.89% |
| 30 Sept 2025 | 14.05% |
| 30 Jun 2025 | 14.26% |
| 31 Mar 2025 | 14.43% |
| 31 Dec 2024 | 14.53% |
| 30 Sept 2024 | 14.68% |
| 30 Jun 2024 | 14.78% |
| 31 Mar 2024 | 14.93% |
| 31 Dec 2023 | 14.94% |
| 30 Sept 2023 | 14.89% |
| 30 Jun 2023 | 14.72% |
| 31 Mar 2023 | 14.39% |
| 31 Dec 2022 | 14.18% |
| 30 Sept 2022 | 13.87% |
| 30 Jun 2022 | 13.58% |
| 31 Mar 2022 | 13.29% |
| 31 Dec 2021 | 13.11% |
| 30 Sept 2021 | 13.16% |
| 30 Jun 2021 | 13.28% |
Table view — Year against year
| Date | Calendar-year average |
|---|---|
| 31 Dec 2025 | 13.89% |
| 31 Dec 2024 | 14.53% |
| 31 Dec 2023 | 14.93% |
| 31 Dec 2022 | 14.18% |
| 31 Dec 2021 | 13.11% |
| 31 Dec 2020 | 13.79% |
Not called in this cycle — the chart and the rail carry the numbers; only the judgment is absent.
Where these numbers come from
| Series | Source | Cadence | Latest |
|---|---|---|---|
| Fixed Investment (GFCF) as % of GDP | SARB Quarterly Bulletin xlsx (KBP6 National Accounts) | quarterly · quantitative | 2026-06-30 · current |
| Fixed investment in rand | SARB Quarterly Bulletin xlsx (KBP6 National Accounts) | quarterly · quantitative | 2026-06-30 · current |
| Investment Acceleration (early warning) | Computed at build time (src/utils/) | quarterly · derived | 2026-06-30 · current |
