Supply and demand reports: EIA, IEA, USDA and how to read them

What you will learn

• Which reports matter for oil, gas and agriculture and when they usually appear
• How to read the EIA weekly petroleum and gas storage reports beyond the headline
• How a WASDE balance sheet is built and how to compute ending stocks and stocks-to-use
• How to judge a report against expectations, seasonal norms and one-off effects

Explanation

The Macro course showed that markets react to the surprise against the consensus, not to the number itself. This lesson applies that idea to commodity reports.

The report calendar

ET means US Eastern Time, normally six hours behind German time (the gap changes briefly when the two countries switch clocks on different dates). Holidays or a US government shutdown can delay reports, so check the official calendars.

Reading the WPSR

• Start with the commercial crude stock change, but also read Cushing stocks, gasoline and distillate stocks and refinery utilisation.
• Product supplied is the EIA's measure of fuels leaving the primary system, a proxy for demand. Use the four-week average because single weeks are noisy.
• Net imports and exports swing widely with tanker timing. One delayed export week can turn a draw into a build.
• SPR changes are reported separately from commercial stocks.
• Compare with the consensus, the API figure and the five-year average for the same week, because stocks have a seasonal pattern.

Monthly oil balances

IEA, EIA and OPEC estimate supply minus demand, which implies a stock build or draw. Their forecasts differ and are revised often; the direction of revisions can tell you more than the level.

The WASDE balance sheet

Total supply = beginning stocks + production + imports

Total use = domestic use + exports

Ending stocks = total supply − total use

Then stocks-to-use = ending stocks ÷ total use (lesson 5). The market compares each number with the average analyst estimate published before the report.

Commitments of Traders

The COT report splits open interest into groups such as producers, swap dealers and managed money (funds). Very large fund positions show crowding and the risk of sharp moves when funds exit. It gives context, not timing.

Worked example

Hypothetical example, part 1: a WASDE corn balance (million tonnes). Beginning stocks 30, production 380, imports 2, domestic use 310, exports 70.

1. Total supply: 30 + 380 + 2 = 412.
2. Total use: 310 + 70 = 380.
3. Ending stocks: 412 − 380 = 32; stocks-to-use: 32 ÷ 380 = 8.4%.
4. The new report cuts production to 372 with use unchanged: ending stocks 24, stocks-to-use 24 ÷ 380 = 6.3%.
5. Analysts expected ending stocks of 28, so the report is 4 million tonnes tighter than expected.

Part 2: an EIA week. Consensus expected a crude draw of 1.5 million barrels; API showed −0.5 million; the EIA reports a build of +2.0 million. Surprise: +2.0 − (−1.5) = +3.5 million barrels. But exports fell by 0.6 million barrels per day versus the prior week: 0.6 × 7 = 4.2 million barrels stayed in the country. The build is explained by export timing, and four-week product supplied is steady, so the report is weak evidence of falling demand.

Excluded: price reaction, revisions to previous weeks and data errors.

Common mistake

Practice activity

Hypothetical data (million tonnes).

1. Wheat: beginning stocks 15, production 120, imports 5, domestic use 95, exports 30. Calculate total supply, total use, ending stocks and stocks-to-use.
2. The next report raises exports by 4 with everything else unchanged. Calculate the new ending stocks and stocks-to-use. Analysts expected 13. Is the report tighter or looser than expected?
3. The EIA reports a gas injection of 60 billion cubic feet (bcf). Consensus was 72 bcf; the five-year average for the week is 75 bcf. Interpret.

Key takeaways

• Know the calendar: API Tuesday, EIA oil Wednesday, EIA gas Thursday, WASDE monthly.
• Judge each figure against consensus, the seasonal norm and one-off effects such as exports.
• Ending stocks = supply − use; stocks-to-use summarises the balance.
• COT data show crowding and context, not timing.

Next step

In the next lesson you see how futures curves turn into the return an ETC investor actually earns, including roll yield and collateral return.