Over the past fortnight the ocean signal strengthened, Ghana's regulator forecast a 16 percent production fall, and visible cocoa stocks drew down from their mid-July high. Every one of those is bullish. Cocoa fell about 21 percent from its 9 July peak anyway. If you want a single demonstration that a commodity price cannot be read off one weather narrative, this fortnight is it.
A scheduling note first. This series is meant to run on Fridays and the last edition was 17 July. Two Fridays went by without one. This edition therefore covers a fortnight rather than a week, and the gap is worth naming rather than papering over, because a weekly series that quietly skips weeks is not a record you can check.
The rules of the series, restated. This is education, not advice. Every price move has multiple drivers and El Niño is one input among many. Where a connection is speculative we say so. Where a market moved the "wrong" way despite El Niño headlines, we report that too, because that is half the lesson.
This fortnight is almost entirely that second category.
The latest ocean observations remained consistent with a strengthening El Niño, while CPC's 9 July probability forecast remained the most recent official assessment. The relative weekly Niño 3.4 index stood at +1.4°C for the week centred 22 July, unchanged from the prior week and up from +1.2°C at the start of the month. Niño 1+2 off the South American coast reached +3.0°C. The 9 July Diagnostic Discussion put an 81 percent probability on a very strong event during October to December and a 97 percent probability that El Niño conditions persist through early spring 2027.
Worth being precise about what that means. CPC has not published a new monthly probability since 9 July, so the 81 percent figure has not been reassessed rather than having been reaffirmed. The next chance to revise it is 13 August. What did move over the fortnight is the observed ocean, not the official forecast.
Two notes on that number. It is the relative index, which subtracts the tropical mean, and NOAA's conventional weekly table shows a warmer figure for the same box and week. We wrote about that gap in One Ocean, Two Numbers, and it matters here only because market commentary quotes both interchangeably. Second, no new weekly value had posted at the time of writing; the next arrives Monday, and the next Diagnostic Discussion is 13 August.
So the signal every market below is trying to price is, if anything, marginally firmer than it was on 17 July. Hold that thought.
Cocoa reached an eight-month high of about $6,455 per tonne on 9 July. That happens to be the same day CPC published the 81 percent figure. By 30 July it had fallen to roughly $5,112, its lowest in nearly three weeks and about 21 percent below the peak, a give-back of some $1,343 per tonne. It remains roughly 40 percent below where it sat a year ago.
Here is the awkward part. Several things that happened over the fortnight were bullish, and it fell anyway:
Read that list back. A 16 percent Ghanaian shortfall, a greater than 10 percent Ivorian shortfall, drawing stocks and a strengthening ocean signal, against softer European grindings and a bank stepping to the sidelines. On a simple weather-narrative reading, cocoa should not have fallen 21 percent. It did.
Edition 1 flagged the counterweight at the time: Barry Callebaut had said the market enters the coming crop year with surplus stocks and that it did not expect a repeat of the recent cocoa crisis. Two weeks on the price has moved toward the processor's view, and the fear trade we described on 17 July has partly unwound. That is worth recording plainly.
Citi's phrasing names the missing ingredient: evidence of El Niño-related crop damage. It is worth being careful about what is now established. There is preliminary evidence of crop stress, including low cherelle counts and official expectations of lower Ghanaian and Ivorian production. What is still missing is final arrival data showing the size of the regional shortfall, and how much of it, if any, can properly be attributed to El Niño.
COCOBOD's own list makes that last point for us. It named El Niño alongside excessive May and June rainfall, the cocoa tree's natural alternating bearing cycle, swollen shoot disease, ageing farms and illegal gold mining. A regulator forecasting a 16 percent fall is real evidence of a smaller crop. It is not evidence that El Niño caused 16 percent, and the regulator did not claim it was.
Coffee went the other way, and the reason is instructive. Trading Economics showed arabica at about 323 cents per pound on 30 July, up 4.2 percent over the month and 9.2 percent over the year. September arabica had jumped 16.2 percent in a single session on 6 July, reportedly the largest one-day rise this century, carrying the contract toward $3.50 per pound for the first time since late January.
The drivers being reported are physical and countable:
Coffee's immediate rally was attributed primarily to falling certified stocks, short covering and rain-delayed Brazilian harvesting. That is not evidence of the drought-driven crop failure commonly invoked in generic El Niño commodity coverage. A wet harvest delay and a dry crop failure are different physical events with different resolutions, and a delay often means the coffee arrives later rather than never.
One caution against overcorrecting. "Rain rather than drought" does not automatically mean "nothing to do with ENSO," because El Niño produces wet anomalies in some regions as readily as dry ones in others. The defensible claim is narrower: the mechanism being reported here is stock depletion and harvest timing, not the crop failure the narrative usually assumes.
Edition 1's counterexample stayed a counterexample. Raw sugar was around 14.43 cents per pound on 30 July, down 3.7 percent on the month and 11.7 percent on the year, having eased from roughly 14.8 cents when we last wrote. Sugar sits on every El Niño commodity list ever published, the forecast strengthened, and the price fell again.
Front-month wheat was around 658 cents on 31 July, up 11.1 percent on the month and 27.3 percent on the year. That is a global complex responding to many things, and pinning it on ENSO would be unsupportable.
On the Australian side, the number that matters has not been updated. ABARES' June crop report forecast wheat production at 26.7 million tonnes for 2026-27 against 36 million the prior season, a 26 percent fall, with national winter crop production down 21 percent to 54.5 million tonnes. That is the same figure Edition 1 carried. ABARES publishes its next crop report in September, so there is genuinely no new Australian data this fortnight, and saying so is more useful than dressing up the old number as news.
Peru's north-central anchovy season set a total allowable catch of 1.9 million tonnes, reported as the lowest in a decade and about 36 percent below the prior first season. Fishing was halted from 12 May because of the share of juveniles in the catch, and as of 3 June accumulated landings stood at 471,111 tonnes, roughly a quarter of the authorised quota.
We have found no verified update to those figures since early June. Given that this is the market where El Niño has already produced a physical effect rather than a forecast, that data gap is the most frustrating item in this edition, and we are flagging it as unresolved rather than filling it with inference.
| Market | Level | Month | Year | Read |
|---|---|---|---|---|
| Cocoa | $5,112/t | +0.4% | -39.9% | Possible weather-premium unwind |
| Coffee | 323c/lb | +4.2% | +9.2% | Physical stocks, not drought |
| Sugar | 14.43c/lb | -3.7% | -11.7% | Counterexample holds |
| Wheat | 658c | +11.1% | +27.3% | Global, not ENSO-attributable |
| Rice | 13.77 | +7.0% | +12.4% | Worth watching, no trigger yet |
| Palm oil | 4,683 | +2.8% | +10.3% | Slow fuse, still slow |
Levels are as of 30 or 31 July. Percentages are the one-month and one-year changes as published alongside them.
In the fortnight since Edition 1 the ocean signal firmed, West African crop stress appeared in official forecasts, visible stocks drew down, and cocoa fell 21 percent. Sugar extended a decline it had already started. The one soft that rallied hard did so on stock draws and a rain-delayed harvest.
The lesson is not that El Niño will not matter. It is that no single weather narrative explains a commodity price. The price action suggests a substantial weather premium may already have been present by 9 July, although the size of that premium cannot be isolated from positioning, demand and broader supply expectations. Once a premium is in the price, more forecast does not necessarily add to it, and even genuinely bearish-to-supply news can arrive into a market that is busy repricing something else.
This is what it can look like when a market stops adding weather premium without new evidence of damage, even as some evidence begins to arrive. The evidence that would settle it is not a probability or a regulator's projection but main crop arrivals from October, which is the same window CPC's 81 percent probability refers to.
Anybody presenting "forecast up, price down" as a contradiction has misunderstood which of the two is a claim about the future. Both are. Only one of them gets marked to market every afternoon.
For the physical side of the same event, see our note on the widening east-west gradient, and for what all this eventually means at the till, what El Niño does and does not do to a grocery bill.