Showing US households bitcoin’s 14.3% return over the previous 12 months made them roughly 23% more likely to report owning cryptocurrency in a later survey, according to a Federal Reserve Bank of Cleveland working paper.
The finding comes from a randomised information experiment embedded in a 2025 household survey, offering rare causal evidence for a dynamic long assumed in speculative markets: that past returns pull in new investors.
Researchers Michael Weber, Bernardo Candia, Olivier Coibion and Yuriy Gorodnichenko randomly divided participants into a control group and six treatment groups shown information about bitcoin, the S&P 500, GameStop or the Federal Reserve’s inflation forecast. One bitcoin group was told the asset’s trailing 12-month return. Another was shown a chart of its price.
Both raised subsequent purchases. The return figure increased the probability of reporting crypto ownership in a later wave by 2.41 percentage points; the price chart by 2.48. With roughly 11% of participants holding crypto beforehand, that represents a relative increase of about 23%.
The paper, titled “Do You Even Crypto, Bro? Cryptocurrencies in Household Finance,” was published on July 14, 2026.
What the experiment measured
The design matters, because it separates correlation from cause in a way survey data alone cannot.
The study drew on quarterly surveys of US households in the Nielsen Homescan Panel running since 2018, with individual waves collecting 15,000 to 25,000 responses. The randomised experiment ran during the second quarter of 2025. The ownership analysis covered 5,352 respondents across the second through fourth quarters and controlled for whether participants already held crypto before receiving the information.
Beyond actual purchases, the treatment shifted stated intentions. Households shown bitcoin’s trailing return raised their desired crypto allocation by roughly two percentage points, a 47% increase relative to the control group’s average target of 4.3%.
That increase came partly at the expense of cash, checking and savings accounts. Participants also raised their desired stock allocations, suggesting the information encouraged broader appetite for risk rather than a narrow rotation into crypto.
Who responded
The effect was not evenly distributed, and the pattern is the most instructive part of the result.
Responses were concentrated among people who said they did not own crypto because they lacked sufficient information. Those who had already concluded crypto was a bad investment did not meaningfully respond to the treatment.
That fits a broader finding in the paper: beliefs about future returns explain more about who owns crypto than demographics do. Crypto holders expect annual returns near 22%, against roughly 7% among non-holders. In a 2021 survey wave, 87% of households without crypto said they did not know what return to expect at all.
Demographics still register. Age was the strongest single predictor, with those under 40 about 13 percentage points more likely to own bitcoin than those over 60. Holders skew young, male, higher-income and politically independent or libertarian. But expectations explained more of the variation than that entire set of characteristics combined.
US household crypto ownership rose from under 2% in 2018 to roughly 11-12% by 2025.
The feedback loop
The authors frame the result as a mechanism that can contribute to asset price bubbles, and they state it plainly.
“High rates of return for a new speculative asset lead new investors to expect similarly high rates in the future and help draw in a growing number of new investors seeking similarly high returns,” they write. Elsewhere they describe the effect as induced demand: providing information about recent bitcoin returns “induces some households to start buying cryptocurrency.”
The loop runs in stages. Rising prices increase paper wealth for existing holders and simultaneously change expectations among potential buyers. Those revised expectations shift portfolio preferences, and in some cases produce actual purchases. New purchases support prices, which regenerates the input.
A separate finding in the paper suggests holders treat crypto gains differently from other wealth. Bitcoin price increases raised spending on durable goods in proportion to holders’ crypto wealth share, but did not produce lasting increases in consumption. The authors compare this to how households spend temporary windfalls such as gambling proceeds, which tend to fund one-time purchases rather than permanently raising spending. Roughly 20% of holders reported that crypto makes up at least half their finances.
The limits
Several caveats deserve weight, and the authors flag them.
The most important is that relatively few respondents actually changed ownership status between survey waves. A 2.5 percentage point shift is statistically meaningful, with the pooled bitcoin treatments producing a p-value of 0.017, but it rests on a modest number of real behavioural changes.
The experiment also does not determine whether bitcoin is overvalued, nor does it prove that new retail buyers are responsible for sustaining any particular rally. It establishes only that information about past gains can change investment behaviour.
The treatment itself was a single piece of information delivered in a survey. Real-world exposure to bitcoin’s performance arrives continuously through media, social platforms and price alerts, which may produce larger or smaller effects than a controlled prompt.
Why the timing is awkward
The paper lands during precisely the conditions it describes.
Bitcoin traded near $77,400 on Monday after gaining 23% over seven days, its best August performance since 2017. XRP rose 46% over the same stretch, ether 28%, Solana 24%. The Crypto Fear and Greed Index has swung from 27 at the start of the month into Greed.
If the mechanism holds, that price action is currently generating exactly the input the experiment tested: a large, visible, widely reported trailing return, arriving in front of a population where roughly 88% of households own no crypto and most non-owners have no formed expectation of returns.
The study does not say whether that is good or bad. It says it is measurable. For anyone assessing whether the current rally has durable demand behind it or is drawing in buyers on the strength of the rally itself, the paper offers a rare piece of evidence that both can be true at once.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. Cryptocurrency investments carry significant risk. Always conduct your own research before making any investment decisions.
