EIA crude oil inventories and the weekly petroleum report

What the report measures 

The weekly petroleum status report is a survey of operating companies conducted by the statistical arm of the United States Department of Energy. Respondents report the volumes they hold, produce, refine, import and export, and the agency compiles them into a balance for the week ending on the preceding Friday. Reporting is mandatory for the surveyed companies, which distinguishes this from a voluntary industry survey.

The headline that reaches a calendar is the change in commercial crude oil stocks. Commercial is doing work in that phrase: the strategic petroleum reserve, held by the government for emergency purposes, is reported separately and is not part of the commercial figure, so a government sale or purchase moves the two series in opposite directions.

The report is followed as closely as it is because physical oil is one of the few globally traded commodities for which a weekly, mandatory, government-collected inventory measurement exists at all. Comparable data for other consuming regions is monthly, is compiled with a longer lag, or covers only part of the storage base, so a United States weekly figure carries analytical weight out of proportion to the country's share of world consumption. That is a fact about data availability rather than about the importance of one market.

The lines beneath the headline 

• Refinery utilisation. The share of refining capacity actually running. Refineries consume crude and produce products, so a change in utilisation moves crude stocks in one direction and product stocks in the other, which is why a crude build during a maintenance period carries a different meaning from a crude build during peak running.
• Gasoline and distillate stocks. Distillate covers diesel and heating oil, and its seasonal pattern differs from gasoline's. Products are where consumption actually shows up, so several analysts treat the product lines as the more informative half of the report.
• Domestic crude production, published weekly as a modelled estimate rather than as a survey response, and later reconciled against a monthly survey that is more accurate and much slower.
• Net imports and exports. The United States both imports and exports crude in large volume, and the balance between them swings with freight economics and with the price relationship between the domestic and international benchmarks.
• Implied demand, sometimes called product supplied. This is not a measurement of consumption. It is a residual: the volume that left the primary distribution system, calculated from the other lines, and it therefore inherits the error in all of them.
• Stocks at the Cushing, Oklahoma storage hub, reported separately from the national total.

Why one storage hub is reported separately 

Cushing is the delivery point specified in the contract that defines the main North American crude benchmark. A futures contract that settles by physical delivery is a claim on barrels at a named location, so the storage capacity available at that location is part of the contract's economics rather than a piece of local trivia.

When stocks at the hub approach the limits of usable capacity, the cost of taking delivery rises sharply, and when they approach operational minimums the ability to deliver at all comes into question. The most extreme episode in the history of the contract, in which the expiring month settled below zero, was in substantial part a story about storage capacity at this one location, and it is the standard illustration that a futures price is a price for a specific grade at a specific place at a specific time rather than a price for oil in general.

Cadence, and the survey that precedes it 

The official report is published weekly, in the late morning Eastern Time on Wednesday, covering the week ending the previous Friday. When a public holiday falls in the week, publication shifts by a day, which is a recurring source of confusion on a calendar that usually shows a fixed weekday.

The evening before, an industry association publishes its own estimate of the same stock changes from a voluntary survey of its members. It is a different sample, it is not mandatory, and it is released to subscribers rather than to the general public, though the figures circulate immediately. The two surveys frequently disagree, sometimes on the direction of the change, and the industry figure is not a preliminary version of the official one. Treating it as a preview is a common error rather than a methodology.

The agency also publishes a monthly petroleum supply report which is more complete and more accurate than the weekly, and the weekly estimates are eventually superseded by it. As with every timely statistic, the frequency is bought with precision.

How the release is conventionally read 

The prevailing convention reads the crude change together with refinery utilisation and the product stocks, on the reasoning that crude drawn into refineries reappears as products and that a change in one line without a corresponding change in another indicates a flow that has not been accounted for. A second convention compares stocks against a multi-year seasonal average for the same week rather than against the previous week, since the series has a strong regular within-year pattern. A third reads the hub stocks against usable capacity as described above.

A fourth convention connects inventories to the shape of the futures curve. When stocks are abundant, the market structure that makes storing oil profitable, with later dated contracts trading above nearer ones, is the condition under which storage is financed. When stocks are tight, the relationship can invert. The link between physical inventory and curve shape is an economic relationship rather than an identity, and it is mediated by storage costs, financing rates and the availability of tankage.