Why do statistics agencies revise numbers months later, and does it matter to a trader?

Economic numbers are estimates, and agencies publish a first estimate quickly and improve it as more information arrives. This month offered three examples. The September jobs report cut July's payroll change from +21,000 to −10,000 and August's from +162,000 to +133,000 (report). US second-quarter GDP growth was revised from 1.5% to 2.2% in the third estimate (report), a release published alongside BEA's annual update of the national accounts, which also revised monthly income and spending back to January 2021 (PCE report). Eurostat's flash estimate of euro area inflation for August was 3.3% and the confirmed figure 3.2% (report).

The reasons differ. Surveys keep receiving responses after the first publication; some data are replaced by more complete administrative records; and agencies periodically update their methods and base years. The first number is the one markets trade, but the revised number is the one that goes into the history.

We would like members, especially beginners, to discuss:

• Do you look at revisions when you read a release, or only the headline?
• Has a revision ever changed a trade you were in, or your view of the economy?
• Do you go back and compare your trade against the revised data, or does only the first number matter for how the market moved?
• Which releases do you think are most often revised, and how do you keep track?

How to read an economic calendar explains actual, forecast and previous readings, and the "previous" figure is often a revised one. Please keep replies friendly and practical. Posts that promise a direction or sell signals will be removed.