Cheddar Rises 8.6% At Global Dairy Trade While Butter Falls: The Milk Price Behind It Does Not Count Cheese

The one thing I keep asking cheese buyers is simple. Which line do you actually price from? Most of them watch the headline dairy index. That index rose 3.0% at the sale on 3 November. Cheddar rose 8.6%.

What the November sale actually printed

Global Dairy Trade held its sale on 3 November. Cheddar closed at 3,970 dollars a tonne, up 8.6% on the previous sale. Mozzarella rose 5.8%. Buyers took 39,806 tonnes across every product on the platform. That matters because this platform is not a spot market. Bidders were buying contracts for delivery between December and May. So the price sets what sellers will commit before the milk is collected. A buyer who reads it as this week's cheese price reads it wrong. It is next season's price, agreed now. That distinction changes who should care today. A trader covering a January shipment has just been repriced. A plant quoting a twelve month retail contract has not been, yet. Both cheeses on the platform moved the same way. Cheddar and mozzarella leave the same milk pool in the same week. When both rise, the cause usually sits in the milk rather than in one lane or one plant.

Every milkfat line went the other way

Fat did the opposite. Butter closed at 4,503 dollars a tonne, down 4.7%. Anhydrous milkfat fell 3.2% to 5,306 dollars a tonne. NZX blamed heavy supply from Europe, the United States, Argentina and New Zealand. The reason is worth stating plainly. Fat is the one dairy stream four large exporters can all lift in the same season. A churn is cheap to run and quick to fill. Spare cream finds a churn within days. Cheese does not work that way. A new cheese vat takes years to build. A cheese room also needs brine, ripening space and skilled people. So a fat surplus arrives inside a quarter. A cheese shortage takes seasons to repair. That difference in speed is why the two lines can part company for months.

Why that widens a cheese margin

Put the two moves side by side. A cheese plant gains twice over on the same sale. That happens because a cheddar vat buys two things and sells two things. It buys fat and protein in the milk it takes in. It sells cheese and whey out the other end. Cheaper fat lowers the cost of one of those inputs. A dearer cheese lifts the value of the main output. So the gap between cost and return widens from both ends at once. A butter and powder plant gets the mirror image of that. Its main output is the line that fell this week. Its co-product, skim powder, held up instead. The same milk, split two ways, now pays very differently. That split is the whole commercial content of this sale. It is also the reason a blended dairy index tells a cheese maker almost nothing.

The milk price behind it does not count cheese

Here is the part the coverage left out. Fonterra lifted its forecast milk price midpoint on 30 October to 9.75 New Zealand dollars a kilo of milk solids. The old midpoint was 9.50 New Zealand dollars. Chief executive Richard Allen named whole milk powder and skim milk powder as the cause. He did not mention cheese once. The reason sits in the rules rather than in the market. Fonterra's milk price manual builds the price from five reference products. Those are whole milk powder, skim milk powder, butter, anhydrous milkfat and buttermilk powder. The manual states that the method excludes returns from non-powder commodities such as cheese and casein. So a cheddar rally cannot reach the farmgate at all. It sits outside the basket by design. New Zealand farmers get paid for powder and fat whatever cheddar does next.

So the gain lands on whoever already has the vats

A price signal only changes behaviour where somebody can act on it. The New Zealand milk price is built from powder and fat, so a cheese rise sends nothing to the farm. It sends nothing to the plant manager either. That happens because the co-op pays its members the same for milk whatever it makes from that milk. The make decision then turns on what each plant can physically run. Capacity decides who captures a cheese rally, rather than the week's cheese price. Northern hemisphere cheddar plants sit in a different position entirely. A Wisconsin or Irish cheddar plant is a fixed cheese asset. The milk has nowhere else to go once it arrives. So that plant takes the whole price move into its own margin. Zenith's cheese forecasts place more than 40% of all new volume to 2030 in two markets. Those are the EU27 and the United States. Both of them already have the cheese rooms built.

The Zenith read on where this price goes

We think the cheese lines hold most of this 8.6% gain into the December contract while fat keeps sliding. The reason is the shape of the two supply bases. Fat output can be raised by four exporters inside one season. Cheese output cannot be raised that fast anywhere. The figure that would change our call is the American cheddar print. United States September cheese production is published on 5 November. A return to growth in cheddar output would push fresh supply into the export pool within a quarter. Flat or falling cheddar output would leave this rally standing. Our opening claim settles here. The headline dairy index is a fat and powder index wearing a dairy name. A cheese buyer who prices from it is pricing somebody else's market. Price from the cheddar line instead. Then check who is actually allowed to make more of it.

One caution belongs with any auction print. A fortnightly sale measures what sellers chose to offer and what buyers chose to take. It does not measure what the world ate. Two sales in a row pointing the same way carry far more weight than one. The sale that follows this one is the real test of the move.

This also raises a wider issue around where cheese demand growth actually sits. That is one of the areas we track across cheese markets in our 2026 research. Our Global Cheese Database 2026 holds production, imports, exports and consumption for 39 cheese types across 48 countries. Every figure is in tonnes. Actuals run to 2025 and forecasts run to 2030.