Study · Original data

How much do CPI, Fed decisions and jobs reports move markets?

By Swiftedge · Published 6 Nov 2026

Fed decisions moved Bitcoin 2.9× as far as a quiet hour and gold 3.9×. On Bitcoin, CPI came in at 2.0× and the jobs report at 2.1×. Rate decisions outside the US moved Bitcoin about as much as a quiet hour.

Impact by release and market

First hour after each release as a multiple of a quiet hour. 1.0× is no more than a normal hour.

* Early number: fewer than six releases or four-hour data so far. It firms up with every release. Numbers from hourly or four-hour candles cover a wider window than one hour; the candles behind each number are in the CSV.

The detail on Bitcoin and gold

How the hour plays out on Bitcoin

"First 5 minutes held" is how often the direction after 5 minutes was still the direction at the end of the hour. The sample is too small to read a direction from it.

Next dates: CPI, FOMC, jobs report, or the full economic calendar.

How it was measured

For every scheduled release we take the range from high to low in the hour after the release, as a percentage of the price at the release. We compare it with up to twelve control days for that same release: the same weekday and local time, from six weeks before to six weeks after, on dates without a major release. A release needs at least three usable control days. Impact is the median release range divided by the median of the control days, so 2.0× means the hour moved twice as far as a quiet one.

The 95% interval is a bootstrap of the releases with 4,000 draws. It shows how far the multiple could move with a different sample of releases. Each number needs at least four releases.

When several US releases fall on the same date, only the first counts, in this order: CPI, the jobs report, PCE and GDP, PPI, retail sales. PCE and GDP share one row because they now come out on the same day. A central bank decision is left out when a US release or Fed decision falls in the same hour.

Markets with at least six releases use 1-minute candles. Where the minute history is still short, the number comes from whole hourly or four-hour candles that touch the hour after the release. These cover a wider window than one hour, which dilutes the effect, so they are not exact first-hour numbers. An early number has a star: fewer than six releases, or four-hour candles.

All numbers are recorded on perpetual futures. Bitcoin is measured from January 2025, and the other markets start later. The US 500 and US 100 are 24/7 markets of their own, not the CME futures.

Limits of this study

• Small samples: the shortest histories have only 4 releases, so those numbers can still move a lot.
• Under two years of data cover a single stretch of interest rate policy. Another period could give different numbers.
• The 95% intervals of the top releases overlap, so their order is not certain.
• We measure the size of the move, not why it happened, and other news can fall in the same hour.
• Control days only exclude major US releases and the largest central banks, so smaller data from other countries can fall on them.
• We do not use forecasts, so the numbers do not separate a surprise from a release that matched expectations.

Cite this

Common questions

Keep reading