Kumo Research · Bitcoin
Bitcoin on Fed, inflation and jobs-report days: how much the hour after the news really moves (2017–2026)
Crypto traders now watch three US releases closely: the Fed's interest-rate decision, the CPI inflation report and the monthly jobs report. We took every one since late 2017, 284 in all, dated from the official Federal Reserve and Bureau of Labor Statistics calendars, and compared Bitcoin's moves in the two hours after each release with the same two hours on normal days. The answer changed completely over time. Until 2020 these releases made no difference to Bitcoin. Since 2021 they clearly do, and the Fed decision matters most. What they do not do is point in a direction.
By Alex Wu · Published · 6 min read
Key findings
- Since 2021, Bitcoin's range in the two hours after a Fed decision has been a median 2.8× the range of the same two hours on a normal day. That was true in 94% of the 47 meetings, and in 70% it was more than double.
- Before 2021 it was the opposite: the Fed hour was a median 0.9× a normal hour, so slightly calmer. CPI (1.0×) and jobs reports (1.1×) were also normal hours. Crypto began reacting to US data in 2021–2022, when inflation and rate hikes took over all markets.
- 2022 was the peak. The Fed hour was 4.2× a normal hour and the CPI hour 3.1×. Since then the reaction has settled at a lower level, but it has not gone away: 2.9× for the Fed in 2026.
- Direction is a coin flip. Since 2021, Bitcoin was higher two hours after the Fed decision 47% of the time and higher a day later 49% of the time. The releases make the price move more, not move one way.
- Ethereum shows the same pattern: across all years its Fed-hour range is 2.0× normal, CPI 1.4× and jobs report 1.2×. Most of the extra movement happens in the release hour itself. By the third hour after the news the market is back to normal.
Why these three releases?
When people say "the market is waiting for the Fed", they usually mean one of three scheduled US events. The Fed decision (FOMC) sets interest rates eight times a year at 2:00 pm New York time. The CPI report is the monthly inflation number. The jobs report (non-farm payrolls) shows how many jobs the US added. Both monthly reports come out at 8:30 am New York time.
Stock and bond traders have watched these for decades. Crypto was supposed to be different: it trades 24/7 and was once sold as having nothing to do with central banks. We wanted to know if that is still true, and if it ever was.
Until 2020, crypto ignored the Fed
Bitcoin's release-hour range vs a normal hour, by year
Median of (range in the 2 hours after the release ÷ range of the same 2 hours on nearby normal days). 1 = no difference
- Fed decision
- CPI
- Jobs report
Source: Binance BTCUSDT hourly candles; Federal Reserve and BLS calendars · Chart: Kumo Research, kumocast.io
From 2018 to 2020, the hours after these releases looked like any other hours for Bitcoin. The Fed hour was even a little calmer than normal (a median 0.9×), and CPI (1.0×) and jobs reports (1.1×) were within normal noise. In those years Bitcoin's big moves came from crypto news (exchange hacks, the 2020 halving, the March 2020 crash), not from US data.
That changed in 2021 and especially in 2022. Inflation in the US went above 8%, the Fed raised rates faster than at any time since the 1980s, and every market started trading on each new number. In 2022 the Fed hour was 4.2× a normal hour for Bitcoin and the CPI hour 3.1×. One CPI morning (11 May 2022) moved Bitcoin almost 10% inside two hours.
Since 2021, the Fed hour is the big one
| Release | Period | Releases | Range vs normal | Bigger than normal | More than double | Up after 2 hours | Up after 24 hours |
|---|---|---|---|---|---|---|---|
| Fed decision (FOMC) | 2017–2020 | 25 | 0.9× | 48% | 16% | 56% | 52% |
| Fed decision (FOMC) | 2021–2026 | 47 | 2.8× | 94% | 70% | 47% | 49% |
| Inflation report (CPI) | 2017–2020 | 38 | 1.0× | 50% | 8% | 61% | 53% |
| Inflation report (CPI) | 2021–2026 | 68 | 1.8× | 76% | 41% | 54% | 56% |
| Jobs report (NFP) | 2017–2020 | 38 | 1.1× | 55% | 13% | 68% | 63% |
| Jobs report (NFP) | 2021–2026 | 68 | 1.5× | 78% | 15% | 51% | 43% |
Range = highest high ÷ lowest low − 1 over the two hourly candles starting at the release hour (8:00 am New York for CPI and jobs reports, released 8:30; 2:00 pm for the Fed). Normal = median of the same two hours on the same weekday in the 8 weeks before and after, skipping other release days.
Since 2021, the two hours after a Fed decision have been a median 2.8× as wide as the same two hours on normal days. That was true in 94% of the 47 meetings, and in 70% of them the range was more than double. In plain numbers: a typical Fed afternoon moves Bitcoin about 2.8% from high to low within two hours, against about 1.2% on an ordinary afternoon.
CPI comes second (1.8× since 2021) and the jobs report third (1.5×). The reaction peaked in 2022 and has settled since, but it has not gone away. In 2026 the Fed hour still averaged 2.9× normal.
Most of it happens in the first hour
Hour by hour: when the extra movement happens
Bitcoin's 1-hour range vs the same hour on normal days, from 3 hours before to 6 hours after the release (all years)
- Fed decision
- CPI
- Jobs report
Source: Binance BTCUSDT hourly candles · Chart: Kumo Research, kumocast.io
The extra movement is short. In the hour of the Fed decision, Bitcoin's range is 2.4× normal. The next hour, when the Fed chair's press conference starts at 2:30 pm, is still 1.5×. By the third hour after the news, things are close to normal again. For CPI and the jobs report the burst is smaller and even shorter: 1.4× and 1.5× in the release hour, then back near 1.
There is no sign of a big build-up before the news. The hours before the Fed decision are slightly calmer than usual (0.8× three hours before), which fits traders waiting on the sidelines.
Bigger moves, but no direction
Here is the part that matters most for anyone tempted to "trade the Fed". Since 2021, Bitcoin was higher two hours after the decision 47% of the time, and higher 24 hours later 49% of the time. For CPI the numbers are 54% and 56%, and for the jobs report 51% and 43%. All are close to a coin flip.
So these releases make the price move more, not move one way. For someone with a leveraged position, that is a risk to plan for: wider swings make stop-losses and liquidation levels easier to hit, in either direction.
The 8 biggest Fed-decision hours for Bitcoin
High-to-low range in the two hours after the 2:00 pm decision, %
- 2022-06-1511.1%
- 2023-03-228.3%
- 2022-07-276.9%
- 2022-09-216.9%
- 2021-01-276.2%
- 2022-01-266.1%
- 2017-12-135.2%
- 2021-07-284.9%
Source: Binance BTCUSDT hourly candles · Chart: Kumo Research, kumocast.io
The biggest Fed afternoons were all in the 2022–2023 rate-hike period. The largest was 15 June 2022, the Fed's first 0.75-point hike since 1994. Bitcoin swung 11% from high to low in two hours. Another was 22 March 2023, in the middle of the US banking crisis, when Bitcoin dropped 6.5% within two hours.
Ethereum does the same
We repeated everything with Ethereum. Over all years its Fed-hour range is 2.0× normal, CPI 1.4× and jobs report 1.2×, almost the same as Bitcoin's 2.1×, 1.4× and 1.4×. It shows the same shift: ignored before 2021, a clear reaction since. This is a whole-market effect, not a Bitcoin quirk.
What this means
- Expect wider swings, not a direction. On Fed days, Bitcoin's two-hour range has been about 2–3 times normal since 2021. Nothing in the data says which way.
- Leverage is riskier around 2:00 pm and 8:30 am New York time on these days. A liquidation price that is safe on a quiet afternoon can be hit in a release hour.
- The effect is recent. Studies or rules of thumb from before 2021 do not apply. Crypto now trades like other markets when US data comes out.
The dates of the next Fed decision, CPI and jobs report are on the Kumo Engine: ask "when is the next Fed meeting". You can check your own position's liquidation level there too.
Method, data and limits
Data: Binance BTC/USDT and ETH/USDT hourly candles from the public Binance data archive (August 2017 – October 2026).
Release dates: every scheduled FOMC decision (2:00 pm New York time) from the Federal Reserve's meeting calendars, and every CPI and Employment Situation release (8:30 am) from the Bureau of Labor Statistics release schedules, November 2017 – October 2026. The two emergency Fed cuts in March 2020 are left out because they were not 2:00 pm decisions. 72 Fed decisions, 106 CPI and 106 jobs reports in total.
Measure: the high-to-low range of the two hourly candles that start at the release hour, divided by the median range of the same two hours on the same weekday in the 8 weeks before and after (skipping other release days). This compares each release with nearby ordinary days, so a calm year and a wild year are judged fairly. Medians are used so that one crazy day cannot drive the result.
Limits: we only look at timing, not at whether the number beat forecasts. Some big days also had other news (the March 2023 Fed meeting came during the US banking crisis). Hourly candles cannot show moves inside the first minutes. Nothing here predicts a direction.
Download the data (CSV, CC BY 4.0)
Sources
How to cite
Kumo Research (2026). "Bitcoin on Fed, inflation and jobs-report days: how much the hour after the news really moves (2017–2026)". Kumocast. https://www.kumocast.io/research/bitcoin-fed-cpi-jobs-report-days
This study describes past market data. It is not investment advice and past behaviour does not guarantee future results.
Live data for these markets
More from Kumo Research
- Funding ratesCrypto funding rates 2020–2026: longs used to pay over 35% a year, now they pay almost nothing6 min read →
- Funding ratesHyperliquid vs Binance funding rates: longs pay about twice as much on-chain (2023–2026)5 min read →
- BitcoinBitcoin halving cycles compared: price, peaks and crashes after all four halvings (2012–2026)6 min read →
About the researcher
Alex Wu
Crypto, DeFi, perps and FX · driven by data
Alex designed the Kumocast model behind these signals, sets its rules, reviews its accuracy record and updates the methodology whenever the model changes. Signals are refreshed every 30 minutes with live market data.