GDP is published in several versions, and each has a different market impact.
For short-term traders, the most useful part of a GDP release is often the composition. Growth driven by inventories or government spending is viewed differently from growth driven by consumption and investment, and the reaction can fade when the details are weaker than the headline.
Advance (first) estimate: the biggest surprise potential.
Second and third estimates: usually smaller revisions and smaller reactions.
Components: consumption, investment, inventories, government spending and net exports.
Why GDP often surprises less than expected:
Much of the information is already known from monthly data.
GDP nowcasts (e.g. Atlanta Fed GDPNow) track the estimate in real time.
Do you trade GDP releases, or treat them as background?
I mostly treat it as background, with one exception: the US advance estimate when it disagrees with the nowcasts. If GDPNow and consensus are a full point apart going in, the market tends to be priced somewhere between, so a print matching consensus can still move the dollar.
The line I read straight after the headline is final sales to private domestic purchasers. It strips out inventories, government and trade, and a headline beat built on inventories alone has faded on me within the first hour more than once.
The price index in the same release is the other trap. Quarterly core PCE comes out with the advance estimate, and on some days it moved yields more than growth did, because it shifts expectations for the monthly PCE figure due later.
For the UK the quarterly figure barely matters to me now. Monthly GDP at 07:00 already tells most of the story, so the quarterly print usually just confirms it with a small reaction.
Second and third estimates I do not trade at all. Has anyone found a revision big enough to matter outside of a recession call?
