The count
In the Energy Information Administration's weekly series, US crude stocks excluding the Strategic Petroleum Reserve fell to 404.508 million barrels in the week ended 24 July. A week earlier they stood at 411.675 million barrels, making the decline 7.167 million. The four preceding weeks had shown 411.675, 409.665, 411.357 and 408.359 million barrels. For a month the series had moved inside roughly 408 to 412 million barrels; this week it dropped below that band.[1]
A weekly decline of that size is a change in the material balance rather than a price story: those barrels either went into refineries, were exported, or were not imported in the first place. The stock series does not say which — a stock figure separates neither inflow nor outflow. It is worth remembering too that the weekly figures are estimates and can be revised; a single cargo arriving a day late can move the weekly difference by millions of barrels. One week does not show whether the break holds.[1]
The difference between a total and a place
A national total does not show where the tightness is. This series gives only the figure for the United States as a whole, with no split by region or grade. Yet what sets the prompt price is the tightest physical point rather than the national total: which grade of crude sits in which tank, at the end of which pipeline. If the difference between 404.5 million and 411.7 million barrels is concentrated in one region, it means something different from being spread evenly across the country, and this series does not answer that question.[1]
When I wrote on 28 July, I argued that the fall in the futures price was pricing the probability of talks rather than a delivered barrel. This is exactly the kind of physical evidence I was after: 7.167 million barrels measures a volume that has left storage rather than a probability. Yet the evidence does not confirm the thesis, it only makes it testable — because it does not say why the barrels left, and a supply interruption and strong refinery demand can produce the same number.[1], [2]
The threshold to watch
So one figure is not enough; the next two are what matter. A single week's decline can be erased by a build the following week, and the administration's own caution says the weekly estimates can be revised. Dropping below the band and staying below it are two different events.[1]
The measurable threshold is this: if the decline is not reversed in the next two weekly releases, I expect stocks to stay below 405 million barrels through 31 August. The signal will be read in the same series — if the figures for the weeks ended 31 July and 7 August print below 405 million barrels, the band has truly broken; if either returns above it, this week was the shadow of a cargo schedule.[1]