A Federal Reserve Bank of Cleveland crypto investor belief study argues that divergent expectations about returns, not demographics, are the primary driver of cryptocurrency ownership, and that showing households data on Bitcoin’s past performance can move both their desired allocations and their actual purchases.
The paper, catalogued as Cleveland Fed Working Paper 26-16 and also circulated as NBER Working Paper 31284, is titled ‘Do You Even Crypto, Bro? Cryptocurrencies in Household Finance.’ Its authors are Michael Weber, Bernardo Candia, Olivier Coibion and Yuriy Gorodnichenko. The paper has accumulated 66 to 67 citations as of mid-2026, per the Cleveland Fed’s publications page.
Crypto Investor Belief Study: What the Survey Found
The researchers drew on repeated surveys covering up to 25,000 US households per wave. In their 2021 survey, 87% of non-owners said they had no idea what return to expect from crypto over the following year. Among owners, the figure was still 54%.
Among those willing to forecast, the gap was blunt: crypto owners expected an average 22% return over the following year, against just 7% for non-owners. Owners also rated crypto as less risky than non-owners did, a divergence that runs in the opposite direction from how most asset classes work.
Expected returns turned out to be unusually elastic in explaining ownership. 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 owning cryptocurrency. Return expectations and perceived risk together explained more variation in ownership than age, income, and gender combined.
For stocks, bonds, and gold, that relationship broadly runs the other way: demographic and financial characteristics carry more explanatory weight than belief heterogeneity. Crypto inverts it.
The demographic skew is still present. People under 40 were 13 percentage points more likely to own crypto than those over 60, controlling for other characteristics. Men were about 4 percentage points more likely than women. Higher-income households were also over-represented. The NBER version of the paper adds that crypto holders tended to be more libertarian in political orientation relative to non-holders, a dimension absent from most retail-ownership analyses.
Information Treatment: How Past Returns Move Allocation
The paper’s randomised experiment, run in 2025, is where the crypto investor belief study produces its most market-relevant result. Households were randomly assigned information on one of four topics: Bitcoin’s recent performance, stocks, GameStop, or inflation.
Participants shown Bitcoin’s previous 12-month return increased their desired crypto portfolio allocation by roughly 2 percentage points. The control group’s baseline desired allocation was 4.3%, so that 2-point shift represents approximately a 47% relative increase. Actual subsequent crypto purchases rose by about 2.5 percentage points.
The authors frame it directly: ‘providing information about recent Bitcoin returns induces some households to start buying cryptocurrency.’ The effect was concentrated among those who cited insufficient information as their reason for not owning crypto. Households that already viewed crypto unfavourably were largely unresponsive to the treatment.
The mechanism the paper identifies is straightforward and uncomfortable for market structure. ‘Positive returns attract new participants, which raises the price further,’ the authors write. Past price drives information exposure drives belief formation drives allocation shifts drives price. The loop is self-reinforcing.
Co-author Bernardo Candia, a Research Economist at the Cleveland Fed with a PhD from UC Berkeley, is part of a team whose prior work on household inflation expectations feeds directly into this framework: belief heterogeneity, not fundamental disagreement, moves household financial behaviour.
The paper also finds that a doubling in BTC’s price made a household whose entire portfolio was in crypto 1.4 percentage points more likely to buy a durable good, equivalent to roughly a 7% increase relative to the unconditional probability of such a purchase. The effect did not carry through to ordinary spending, leading the researchers to compare crypto gains to gambling income or lottery winnings rather than a permanent wealth increase.
The conclusion from Yuriy Gorodnichenko and co-authors is direct: ‘The absence of common information and beliefs about crypto across investors suggests that price volatility will continue to be one of the most defining characteristics of this new asset for the foreseeable future.’
For anyone watching retail positioning, the implication is concrete. The next wave of inflows may be less a function of macro conditions or on-chain fundamentals, and more a function of how widely Bitcoin’s most recent return gets distributed to households who have never bought it before. Narrative distribution is the product; price is the input.
