The Monthly Review
Signals rarely shout. Noise usually does. A read on the month beneath the headlines: the conversation we have been having internally.
One shock, two outcomes
The same barrel of oil was background noise in one market and a tax in another.
Brent crude rose 14.41% in September as the Iran conflict and disruption around the Strait of Hormuz held through the month. Gold fell 6.31% despite it, and the dollar index rose 2.03%.
Gold falling into a geopolitical shock is the tell. Real yields rose as four of the five central banks we track tightened, and that cost of carry outweighed the safe-haven bid.
The dollar did the rest. A 2.03% rise in the index raises the local-currency cost of every imported barrel, so the shock reached some economies through the currency before it reached them through the oil price.
Month-end close to month-end close. A rise in USDZAR means a weaker rand. Source: Bloomberg, as at 30 September 2026.
Seven of the nine indices we follow fell, and the MSCI ACWI gave back 1.1%. What separated the two that rose from the one that fell hardest was not the shock itself but where it landed.
A table that split in two
Only the NASDAQ at 1.9% and the Nikkei 225 at 1.2% finished the month higher. Seven and a half points separated best from worst, against five and a half in August.
The year reads the same way. The Nikkei leads at 30.7% and the NASDAQ at 16.1%, while September pushed the JSE All Share back under water for 2026 at −3.3%.
Total returns in local currency, which removes currency-translation effects but is not directly comparable with older figures reported in US dollars. Source: Bloomberg, as at 30 September 2026.
The NASDAQ’s month was not evenly earned. The gains concentrated into three windows: enterprise AI guidance in the first week, a semiconductor rally on 21 September, and Nvidia’s buyback in the closing sessions.
Whether that is durable is contested. Nvidia’s forward multiple compressed to near its cheapest in a decade, which one camp reads as a derating and the other as an entry point.
Leadership changed hands at both ends
The JSE All Share led all nine indices in August at 4.6% and finished last in September at −5.8%. That is a swing of 10.4 points in a single month, the largest of the nine by a factor of nearly three.
Total returns in local currency, prior month-end close to month-end close, ordered by the September return. Source: Bloomberg, as at 30 September 2026.
The top of the table moved less than it looks. The NASDAQ went from second to first without rising faster, and the Hang Seng was second from bottom in both months.
Our reading, rather than a finding in the data, is that this was not a style rotation. It was one commodity shock sorting markets by whether they import the barrel or earn from it, and South Africa does both in the wrong proportions.
The same reading applies at the top. The NASDAQ led on an AI cycle that is not paying the same way everywhere, and Chinese hardware stocks closed their worst quarter on record over the same three months.
Prepared beats predicted.
Undone by what we import
The JSE All Share returned −5.79%, the weakest of the nine, and is now −3.3% for the calendar year and 4.5% over twelve months. The oil price did most of it, and it did it through the current account rather than through the index.
The second-quarter current account swung to a deficit of 2.6% of GDP, the widest in seven years, as import values rose. Foreign investors took R9.0 billion out of the portfolio account over the same quarter, reversing an identical inflow in the first.
The rand weakened 1.91% to 16.4226, its worst month since March. It was at 15.9207 on 4 September and 16.4764 on 24 September, a range of 3.5% travelled almost entirely in one direction.
The Reserve Bank raised the repo rate 25 basis points to 7.25% on 23 September, which 19 of 22 surveyed economists had expected. August CPI printed at 4.4% on the same day, and the ten-year yield reached 9.09% on 28 September, its highest since April.
We are publishing no sector figures this month. The only sector data available to us are unweighted member averages that do not reconcile to the index return, and one of them is a single stock wearing a sector label.
Two points are contested and we would not settle either. Goldman Sachs argued on 1 September that the market is not pricing South Africa’s path back to investment grade, which Morningstar’s published decision to stay underweight directly contradicts.
For a client holding only South African equities, September cost 5.8% in rand with no offset. The nine-index table is the argument for holding the other eight, and it reads better after a month like this than before one.
The second is the rate path. Bank of America is alone in forecasting a further hike in November to 7.5%, while the Reserve Bank’s own model points to a rate that holds here through the rest of 2026.
Signals beneath the noise
Signals rarely shout. Noise usually does.
Credit
US five-year CDS widened from 66.7 to 77.9 and European CDS from 53.9 to 69.4. Both now sit at their widest since the spring, the US since April and Europe since March.
The European move of 15.5 points is the sixth-largest monthly widening in a series that starts in April 2021. Credit noticed something in September that equity volatility largely did not.
Credit and equity volatility disagreeing is worth sitting with. Credit prices default risk while volatility prices dispersion, and a shock that raises financing costs shows up in the first before the second.
Volatility
Thirty-day volatility rose in five of the nine and fell in four, a quieter picture than the credit market’s. The Nikkei 225 remains the most volatile on both horizons at 20.7 and 27.3.
The JSE fell furthest, from 19.0 to 12.7, which is the opposite of what a 5.8% drawdown usually produces. August’s gold-driven range, not September’s decline, was the unusual reading.
Hover any point to see its August position and 30-day move. Source: Bloomberg, end-August and end-September 2026.
Rates and policy
Four of the five banks we track raised rates inside a fortnight, and the Bank of England held. It is the first month in this series in which every tracked bank met and only one stood still.
Rates before and after each bank’s September meeting; the Federal Reserve figure is the upper bound of its target range. Each next meeting date confirmed against that bank’s own calendar. Source: SARB, Federal Reserve, Bank of England, ECB, Bank of Japan.
The Federal Reserve’s 25 basis point increase on 16 September was its first in more than three years and was unanimous. Sixteen of the eighteen officials expect at least one more before the year ends.
The Bank of Japan’s hike to 1.25% carried two dissents and arrived under open pressure from the US Treasury. The Bank of England held at 3.75% on a 6–3 split and warned that the energy shock may yet force its hand.
How far the European Central Bank goes from here is contested. Markets were pricing four more quarter-point increases over twelve months while the Council declined to pre-commit to any, and we would not present either as the settled view.
The questions we are sitting with into October
Three decisions in three days
The Federal Reserve on 28 October, the European Central Bank on 29 October and the Bank of Japan on 30 October. The Reserve Bank follows on 19 November.
Whether credit is early or wrong
Spreads widened sharply on both sides of the Atlantic in a month when equity volatility mostly did not. One of those two markets is mispricing the oil shock.
The rand and the current account
A 2.6% deficit is financed by the portfolio flows that left in the second quarter. If Brent holds above $100, the third-quarter number is the one to watch.
Hormuz, still unresolved
Reports of a US-Iran channel eased prices briefly on 25 September. A separate risk sits alongside it: the US Energy Secretary warned on 24 September of possible curbs on diesel exports.
Whether the AI trade is one trade
The NASDAQ led the month while Chinese hardware stocks closed their worst quarter on record. The same theme is not paying the same way in every market.
Curiosity is a surprisingly effective risk management tool.