When the Tax Is on the Bottle, the Bottle Stops Travelling

Provence sells 42 per cent of its rosé outside France. The United States alone takes 127,934 hectolitres — 30.59 per cent of the export book and 11.68 per cent of everything the three appellations make. A tariff written on the wrong object can move a meaningful share of that trade from a bottle to a tank without a single consumer noticing, and the region has already run the experiment once.

What happened in 2020

When the United States imposed a 25 per cent tariff on French still wine in 2019, the measure was drafted on containers of two litres or less. Bulk wine fell outside it. The result was immediate and mechanical.

In January 2020, French bulk wine shipments to the United States rose 1,094 per cent year on year, from 1,000 to 12,200 hectolitres. February held the gain at 6,000 hectolitres, up 80 per cent. Provence accounted for more than 75 per cent of French bulk wine export volumes in that period; the Rhône for about 20 per cent.

The regional council’s own account is the part worth keeping. Over January and February 2020, 51,000 hectolitres of Provence AOP went to the United States — a figure described as stable on the prior year — and for the first time 10,000 hectolitres of it went as bulk. That is 19.6 per cent of the shipment, converted from bottle to tank inside one season, to escape a tax that was never aimed at wine quality.

The producers who did it were explicit that this was a bottling decision, not a product decision. One domaine described shipping in bulk and contracting a East Coast winegrower as a service bottler, while keeping control of the dry goods — labels, closures, cartons — ordered in France and sent over. The wine was the same wine. What crossed the Atlantic changed.

Why the statistic lies when this happens

Here is the part that matters to anyone reading a trade series rather than a customs return.

The EUR 5.90 average bottle price Provence earns on export is a bottled number. It is the free-on-board value of a package: the wine, the glass, the closure, the label, the carton, the pallet, and the labour of filling it. Ship the same hectolitres as bulk and every one of those downstream components stops being invoiced by a French producer. The recorded export value falls, the recorded volume does not, and the recorded price per litre drops — even though nothing whatever has happened in the vineyard.

An ad valorem tariff has the same property in reverse. Fifteen per cent of a bottled FOB price is a larger absolute charge than fifteen per cent of a bulk one, so moving the filling abroad shrinks the tax base as well as the tax bill. Both effects push in the same direction, and both are invisible in a headline number that says a region’s exports “fell”.

The 2026 exposure, in numbers

A 15 per cent measure on European wine and spirits has been in force since 1 July 2026, with a sunset on 31 December 2029. I have not been able to establish from the published record whether its product scope runs to bulk as well as bottled, so I am not going to assert that the 2020 mechanism is or is not available this time. What can be said is that the incentive structure is unchanged, that Provence now has a documented route for exploiting it, and that the window — three and a half years — is a short enough horizon for a distributor to plan a switch inside.

The precedent also tells you who moves first. In 2020 the largest, most brand-conscious houses were the ones who shipped bulk and kept the dry goods, because they could control quality at the far end of the chain. The small producer with no US importer of scale simply paid the tax. An instrument aimed at a country lands hardest on the smallest sellers in it.

What actually leaves when the bottling leaves

The wine is the part that stays. Everything downstream of the tank — the bottle, the label, the fill, the pack, the pallet — is the part that goes, and at a EUR 5.90 FOB average those components are a substantial fraction of what a buyer is paying for. Whatever the dry goods actually cost, they are a far larger share of a EUR 5.90 bottle than of a EUR 24.00 one, which is why the bulk-shifting trade only ever appears at the bottom of the price ladder. Nobody bottles a EUR 60 wine in New Jersey to save duty.

For Provence specifically, the loss is not only a bottling line. It is the last stage at which a French producer holds the article that carries the mark — and the region is currently developing a shared recognition mark across its three appellations, for launch in 2026, precisely because being legible on a shelf is its only remaining defence against a generic category. A mark printed on a label applied in New Jersey is a mark whose application someone else controls.

A secondary signal worth watching

The 2025 quality monitoring gives a rough map of where the region’s wine physically ends up. Of 211 samples drawn at the point of sale, 34 per cent came from hypermarkets, 23 per cent from supermarkets, 27 per cent from independent merchants, 11 per cent from other outlets and 5 per cent from discount stores. Twenty of the samples were bag-in-box.

Bag-in-box is bulk in miniature: wine shipped in a flexible container and packed at the destination. It is already 9.5 per cent of the sampled pool, which means the practice of packing downstream of the winery is not foreign to this category at all. Moving a bag-in-box operation to a bottling operation across an ocean is the same decision with a bigger container.

What would falsify this

If the 15 per cent measure is drafted to cover bulk as well as bottled wine, the arbitrage is closed and this mechanism cannot repeat. That is checkable against the implementing text, and it is the first thing to check.

If no bulk-bound Provence shipment to the United States appears in the twelve months after 1 July 2026, the incentive was not binding this time — most likely because the tariff is at 15 rather than 25 per cent, or because US importers absorbed it rather than restructuring.

If unit values in the export series fall while volumes hold, the switch has happened and the series is measuring packaging rather than wine. That is the signal to look for in the data before believing any story about a decline in demand.

Rosé is a category that grows 4.80 per cent a year in value on 0.20 per cent volume growth. When the entire forecast is price, the physical form in which the wine crosses a border stops being logistics and becomes the forecast.