Key takeaways
- In the 27 months since January 2015 when US CPI inflation ran at 4% or more, bitcoin beat that month's inflation 10 times. Across those months it lost 55.4%.
- In the 48 months when annual inflation ran below 2%, bitcoin beat CPI in 32 of them, or 67%. It performed best when there was least to hedge.
- From 30 November 2021 to 30 November 2022, bitcoin fell 66.5% measured in sterling, while UK CPI rose 10.7% over the same twelve months.
- Bitcoin regained its November 2021 dollar high on 4 March 2024, 847 days later. After adjusting for US CPI it took 1,096 days, to 8 November 2024.
- Over the full period bitcoin rose 30,524% in sterling against UK CPI's 43.9%. Its monthly returns annualise to 67.5% volatility, which swamps any CPI signal.
The bitcoin inflation hedge worked in every period except the one that tested it
You've heard that bitcoin is a hedge against inflation. The fixed supply, the halving schedule, the phrase "digital gold". You want to know whether the price actually did what the story says it does.
Tested against realised inflation rather than the argument for it, the answer is no, and the failure is concentrated exactly where it matters. Take Coinbase's daily bitcoin price as published by the Federal Reserve Bank of St Louis, from January 2015 to July 2026, and split every month by what US CPI inflation was running at year on year. There are 137 complete months. Bitcoin beat that month's inflation in 75 of them, which is 55%, a coin flip with a slight tilt.
Split those months by regime and the tilt goes somewhere awkward. In the 27 months when annual CPI was 4% or above, bitcoin outpaced inflation 10 times, a 37% hit rate. In the 48 months when inflation was under 2%, it managed 32 out of 48, or 67%. The chart above plots those three hit rates against the whole sample.
That is the shape of the evidence. Bitcoin's record against inflation is strongest when inflation is absent and weakest when it's high. An asset that only outpaces prices while prices are quiet isn't doing the job people buy it for.
The 27 high-inflation months cost 55.4%, not counting the inflation
Hit rates hide magnitudes, so here's the magnitude. Chain the monthly returns across those 27 high-inflation months and bitcoin lost 55.4% in dollars. The same months in the moderate bucket, with CPI between 2% and 4%, produced a cumulative gain of 636.5%. The low-inflation months produced 7,540%.
The 27 months aren't scattered. Twenty-six of them run consecutively from April 2021 to May 2023, the one genuine inflation shock in bitcoin's traded history. The twenty-seventh is May 2026, when the seasonally adjusted CPI-U series showed annual inflation at 4.2%. So this isn't 27 independent experiments. It's one high inflation regime, observed monthly, and bitcoin went through it badly.
US CPI inflation peaked in that episode at 9.0% year on year in June 2022 on the seasonally adjusted CPI-U index. Bitcoin's dollar price had already peaked at $67,510 on 8 November 2021, before inflation topped out, and bottomed at $15,756 on 21 November 2022. That's a 76.7% fall. Over the same twelve months the CPI index rose 7.1%.
In sterling the shock cost 66.5% while UK prices rose 10.7%
UK readers hold a different currency, and the arithmetic changes. Converting the dollar price at the Federal Reserve's daily sterling exchange rate, bitcoin fell 66.5% between 30 November 2021 and 30 November 2022. The pound's weakness that year cushioned some of the dollar fall. The same mechanism runs through equities, where the currency contribution to returns has moved against the local market for a sterling holder more often than with it. The same mechanism runs through equities, where the currency contribution to returns has moved against the local market for a sterling holder more often than with it.
Set that against what UK prices did. The ONS CPI index went from 114.5 in November 2021 to 126.7 in November 2022, a rise of 10.7%. The annual rate peaked at 11.1% in October 2022, the highest monthly reading in the ONS CPI series, which starts in 1989. That is the single largest inflation event a UK investor has faced since bitcoin existed, and bitcoin lost two thirds of its sterling value through it.
The gap between those two numbers is the whole test. If you'd bought bitcoin specifically because you were worried about UK inflation, the year your worry came true is the year the position fell by two thirds.
Getting the purchasing power back took 1,096 days, not 847
Recovery arithmetic is where inflation quietly changes the answer. Bitcoin's dollar price closed back above its 8 November 2021 high of $67,510 on 4 March 2024. That's 847 days to break even in nominal terms.
But US consumer prices rose 12.0% over that stretch and kept going. To buy what $67,510 bought in November 2021, you needed $76,613. Bitcoin first cleared that level on 8 November 2024, which is 1,096 days after the peak, or three years to the day. The nominal recovery arrived roughly eight months before the real one.
This isn't unique to bitcoin. Every asset that falls in an inflationary period faces the same second hurdle. It matters here because the claim being tested is specifically about purchasing power, and the recovery that most people mark on a chart is the nominal one.
What a bitcoin sleeve did to the rest of the portfolio through the shock
The portfolio question isn't whether bitcoin hedged, but what a given sleeve size did to the whole. Take the same UK window, 30 November 2021 to 30 November 2022, with bitcoin down 66.5% in sterling and CPI up 10.7%. Hold the sleeve at a fixed weight and ask what the other assets had to earn for the portfolio to stand still in real terms.
- No bitcoin: the rest of the portfolio had to return 10.7% to hold purchasing power flat.
- 1% sleeve: it cost 0.67 percentage points of portfolio value, and the other 99% had to return 11.4%.
- 3% sleeve: a 2.00 point drag, and the remaining 97% had to return 13.0%.
- 5% sleeve: a 3.33 point drag, and the remaining 95% had to return 14.7%.
- 10% sleeve: a 6.65 point drag, and the remaining 90% had to return 19.2%.
Those are static-weight figures, so they ignore any rebalancing into the fall, which would have changed both the drag and the recovery. What the table shows is how quickly the hurdle moves. At 1% the sleeve is noise. At 10% it turns a hard year into an arithmetic problem the other holdings had almost no chance of solving.
Sizing that sleeve against long-run returns rather than a single shock is a separate question, and the bitcoin allocation evidence at 1% to 5% of a 60/40 answers it on different data.
The strongest version of the other side, stated properly
Anyone holding bitcoin for inflation protection has a serious rebuttal, and it deserves the numbers rather than a paraphrase.
Over the whole period, bitcoin in sterling went from £152 on 30 January 2015 to £46,686 on 31 July 2026. That's a rise of 30,524%. UK CPI rose 43.9% over the same span, from 99.3 to 142.9 on the ONS index. On the only measure that matters over a lifetime, cumulative purchasing power, nothing else in a normal portfolio comes close.
The rebuttal continues: eleven and a half years contains one inflation cycle, and judging a monetary asset on one cycle is judging it on noise. On that view the 2021 to 2023 episode is a single draw from a distribution nobody has enough observations of.
That case is strong on the long horizon and weak on the specific claim. Bitcoin real returns over eleven years and bitcoin as inflation protection are two arguments, not one. "Bitcoin has produced enormous real returns" and "bitcoin protects you when inflation rises" are different statements. The first is supported by the data. The second requires the price to respond to inflation, and it doesn't. The correlation between monthly bitcoin returns and that month's CPI change over 137 months is -0.02, which is nothing. Against the monthly change in the annual inflation rate it's -0.001.
The mechanism people assume isn't the one the data finds
The intuitive story is that a capped supply protects the holder when the money supply expands. The ECB's Fabio Panetta put the structural objection at the BIS Annual Conference in June 2023: "unbacked cryptos do not improve our capacity to hedge against inflation. Indeed, their price developments exhibit an increasing correlation with stock markets." In the same speech he noted that bitcoin's "limited supply", capped at 21 million coins, "does not offer protection against the risk of structural deflation."
Research on high-frequency data points the same way. Mykola Pinchuk's 2023 working paper "Bitcoin Does Not Hedge Inflation" measures how the price moves in the minutes around US inflation releases. His finding: "Price of Bitcoin decreases by 24 bps in response to a 1 standard deviation inflationary surprise." A hedge moves up when the risk arrives. This one moves down. The paper is a preprint and hasn't been through journal review, so it's evidence, not settlement.
The simpler explanation is scale. Bitcoin's monthly returns over these 137 months annualise to 67.5% volatility. UK CPI has run between -0.1% and 11.1% a year in that time. A signal that small can't survive inside noise that large, whatever the underlying mechanism does. The is bitcoin digital gold question runs into the same wall from the correlation side.
It's also worth being clear about who is exposed. The FCA found in research published on 26 November 2024 that 12% of UK adults own crypto, up from 10%, and some fraction of them bought for crypto inflation protection. The financial promotions regime for cryptoassets came into force on 8 October 2023, which is why this piece describes evidence rather than recommending anything.
What this data can't tell you
Four limits, and each of them is real.
The sample is one inflation cycle. Twenty-six of the 27 high-inflation months are consecutive, so the effective sample size for the central finding is closer to one than to 27. A second inflation shock could produce a different answer.
The regime split uses US CPI on a seasonally adjusted basis while the sterling numbers use the ONS index, which isn't seasonally adjusted. The two series aren't like for like, and the buckets would shift by a month here and there on the unadjusted US figures.
The price series is one exchange's. Coinbase's dollar quote, as republished by the St Louis Fed, is a reasonable proxy but not a consolidated tape.
And realised experience isn't the same as the index. The BIS studied crypto app users from August 2015 to December 2022 and reported that "a majority of crypto app users in nearly all economies made losses on their bitcoin holdings." Entry timing did most of that damage, and no backtest of a buy-and-hold price series captures it.
What would change this conclusion
One clean test would settle it. A second inflation episode, with CPI above 4% for a sustained run, driven by something other than a post-pandemic supply shock and arriving without a simultaneous tightening of real interest rates. In that setup the 2021 to 2023 episode stops being the whole sample, and if bitcoin's hit rate in high-inflation months rises from 37% toward its 67% low-inflation rate, the mechanism starts to look real rather than incidental.
May 2026 was the first month since 2023 to enter the high bucket, at 4.2%. By July 2026 the annual rate was back to 3.3% in the US and 2.9% in the UK, so the test hasn't started. If it does, the number to watch isn't bitcoin's price. It's whether the price moves with the CPI release, or with the equity market on the same day. Instruments that hedge inflation by construction, such as index-linked gilts, answer that question mechanically rather than statistically, which is the distinction the whole argument turns on. Tracking a crypto sleeve's real, CPI-adjusted contribution rather than its headline price is the version of the question LedgerTouch was built to answer.