Newsroom
23 August, 2026 / News / AI / Tags: beliefs, paper, households, authors, percentage

A Cleveland Fed working paper shows expectations about crypto returns drive ownership more than demographics, while recent Bitcoin performance data can increase both planned holdings and actual purchases
A Federal Reserve Bank of Cleveland working paper examines why cryptocurrency ownership and price swings differ so markedly from traditional assets. Researchers Michael Weber, Bernardo Candia, Olivier Coibion and Yuriy Gorodnichenko analyzed repeated surveys of as many as 25,000 U.S. households per wave and conducted a randomized information experiment. Their findings center on the role of investor beliefs rather than standard demographic or financial traits.
The study, titled “Do You Even Crypto, Bro? Cryptocurrencies in Household Finance,” finds that expectations about future crypto returns account for more of the variation in who owns digital assets than factors such as age, income or gender. A one-percentage-point rise in an individual’s expected crypto return was associated with a 0.8-percentage-point increase in the probability of ownership. When combined with views on risk, these expectations explained substantially more variation than observable characteristics.
This pattern stands in contrast to stocks, bonds and gold, where demographic and financial variables typically carry greater explanatory power than differences in return forecasts. Crypto ownership appears driven more by what people believe the assets will deliver than by who they are.
Demographic differences remain visible. People under 40 were 13 percentage points more likely to own cryptocurrency than those over 60, even after controls. Men were about 4 percentage points more likely than women to hold crypto, and higher-income and wealthier households showed higher participation rates. These patterns, however, proved secondary to beliefs about returns and risk.
Knowledge of crypto returns remains limited across the population. In a 2021 survey referenced in the paper, 87 percent of non-owners said they did not know what return to expect over the following year. Among owners the share was still 54 percent.
Among those willing to offer a forecast, the divergence was large. Crypto owners expected an average 22 percent return over the next year, compared with 7 percent among non-owners. Owners also tended to view crypto as less risky than non-owners did. The researchers link this dispersion in beliefs to the potential for self-reinforcing price movements.
Rising prices can strengthen optimistic expectations, drawing in additional buyers whose purchases then support further gains. The authors note that this dynamic does not require fundamentals alone to drive prices; disagreement and the process of updating beliefs based on past outcomes can themselves become sources of volatility.
The paper’s randomized experiment, conducted in 2025, tested the effect of providing information about Bitcoin’s previous 12-month return. Households were randomly assigned to receive data on Bitcoin alongside other topics such as stocks, GameStop or inflation.
Participants shown Bitcoin’s recent performance increased their desired crypto portfolio allocation by roughly 2 percentage points, about a 47 percent rise relative to the 4.3 percent desired allocation in the control group. Actual subsequent crypto purchases rose by about 2.5 percentage points.
The response was concentrated among people who said they did not own crypto because they lacked sufficient information. Those who already viewed crypto as a poor investment generally did not change their behavior after receiving the data.
The researchers also examined how crypto price changes affect consumption. When Bitcoin’s price doubled, a household whose entire financial portfolio was held in crypto became 1.4 percentage points more likely to purchase a durable good. That represented roughly a 7 percent increase relative to the unconditional probability of such a purchase. The effect did not extend to ordinary everyday spending.
The authors interpret this pattern as evidence that crypto gains are treated more like gambling income or lottery winnings than as a permanent rise in wealth. The short-lived nature of the spending response suggests crypto’s influence on household behavior may be episodic rather than lasting.
The combination of fragmented beliefs, limited common information and sensitivity to recent performance points to a structural source of volatility. Because many households lack clear expectations and can be swayed by data on past returns, price increases themselves can expand the pool of buyers.
The findings indicate that future retail participation may depend in part on how information about recent price moves reaches potential investors and how those investors update their expectations. The paper frames cryptocurrency as an asset whose ownership patterns and price dynamics are shaped unusually strongly by belief formation and learning rather than by conventional demographic or risk-preference factors alone.









