Fed Study: Crypto Investors Follow Beliefs, Respond to Returns

A new working paper from the Federal Reserve Bank of Cleveland argues that the main reason cryptocurrencies don’t behave like other financial assets may have less to do with demographics and more to do with beliefs. According to the authors, Americans who own crypto—and those who plan to buy—often hold sharply different expectations about what digital assets will return, and those expectations help explain who participates in the market in the first place.

The paper also presents experimental evidence suggesting that information about Bitcoin’s recent performance can meaningfully change what households say they want to hold, and can translate into higher actual purchases. If those findings are broadly applicable, they offer a mechanism for why crypto can stay volatile and why rallies can pull in new buyers in a reinforcing loop.

Key takeaways

• Beliefs about future crypto returns explain participation better than standard demographics, according to a Cleveland Fed study using large household survey waves.
• Most non-owners report they don’t know what crypto returns to expect, while crypto owners forecast substantially higher returns.
• In an information experiment, showing households Bitcoin’s past 12-month return increased desired allocation and subsequent purchases.
• The study frames crypto volatility as partly driven by disagreement and learning, not only fundamentals.

Return expectations are the clearest predictor of ownership

Researchers Michael Weber, Bernardo Candia, Olivier Coibion, and Yuriy Gorodnichenko analyze repeated surveys of up to 25,000 U.S. households per wave. Their central finding is straightforward: expectations about crypto returns account for more of the differences in who owns cryptocurrency than a wide range of demographic characteristics.

In the paper, the authors highlight that the typical person outside the crypto market struggles to form a return forecast at all. In a 2021 Federal Reserve survey referenced by the authors, 87% of people who did not own crypto said they didn’t know what return to expect over the following year. Among crypto owners, the share who didn’t know was still high at 54%. The gap in knowledge is large, but the gap in forecasts that do exist is even larger.

For people willing to estimate returns, crypto ownership aligns with materially higher expectations. The study reports that crypto owners expected an average 22% return over the next year, versus 7% among non-owners. Owners also tended to judge crypto as less risky than non-owners did.

Most importantly for understanding market behavior, the authors show that these expectations are unusually powerful statistically. They find that a one-percentage-point increase in an individual’s expected crypto return is associated with a 0.8-percentage-point increase in the probability of owning cryptocurrency. When return expectations and perceived risk are considered together, they explain far more variation in ownership than observable characteristics such as age, income, and gender.

The researchers argue this makes crypto an outlier relative to other asset classes. For stocks, bonds, and gold, standard demographic and financial factors typically have more explanatory power than differences in expected returns. In crypto, the direction of explanatory power appears to flip.

Crypto participation still looks demographic—but beliefs dominate

Even with expectations taking center stage, the paper does not claim ownership is entirely divorced from who people are. The demographic pattern remains distinct after controlling for other variables. People under 40 are reported as 13 percentage points more likely to own cryptocurrency than those over 60.

Gender and household wealth also show up in the data. The study finds that men are about 4 percentage points more likely than women to own crypto, and that higher-income and wealthier households participate at higher rates. But the authors’ comparison is that these effects—distinct as they may be—are secondary to the role of beliefs about returns and risk.

This matters because it reframes a common debate. Instead of treating crypto’s unusual ownership pattern as mainly a story about who is “more risk tolerant,” the paper pushes investors to focus on what market participants think crypto will do—and how those beliefs differ from one another.

Information about recent Bitcoin returns can change behavior

The paper’s most actionable evidence comes from a randomized information experiment described by the authors. In 2025, households were randomly assigned to receive information that included Bitcoin’s previous 12-month return, alongside other choices that related to stocks and topics such as GameStop or inflation. The paper reports that participants shown Bitcoin’s recent performance increased both what they wanted to hold and what they went on to buy.

Specifically, the treatment increased desired crypto allocation by about 2 percentage points—roughly a 47% increase relative to the 4.3% desired allocation among the control group. The authors also report an increase in subsequent crypto purchases of about 2.5 percentage points.

The paper describes the result as induced demand: “providing information about recent Bitcoin returns induces some households to start buying cryptocurrency.” The effect is not uniform, however. It is concentrated among individuals who said they did not own crypto because they lacked sufficient information. Those who already believed crypto was a bad investment did not meaningfully respond to the information treatment.

In practical terms, the study suggests that crypto participation can be changed by what people are told about what happened recently—not just by long-term narratives or individual risk preferences. For traders and market observers, that implies that retail demand may be unusually sensitive to framing, past performance information, and perceived signal quality during momentum periods.

Why volatility may persist: learning, disagreement, and “past gains”

Beyond ownership and demand, the paper connects its survey and experimental findings to a broader market question: why crypto markets can produce persistent volatility. The authors describe a mechanism often discussed in behavioral finance—past positive returns can attract new participants, whose purchases can push prices higher and potentially draw still more buyers.

They present the logic explicitly, writing that “positive returns attract new participants, which raises the price further.” The authors further argue that this dynamic may be particularly relevant because crypto remains poorly understood by a large share of the population, and beliefs about future returns are therefore likely to be fragmented.

The study also examines whether crypto gains show up in household spending in a way consistent with lasting wealth effects. It reports that when a household’s entire financial portfolio was in crypto, a doubling of Bitcoin’s price increased the probability of buying a durable good by 1.4 percentage points—about a 7% increase relative to the unconditional likelihood of such a purchase. However, the effect did not persist into everyday spending.

That pattern leads the authors to a sharp interpretation: crypto gains appear to be treated more like gambling income or lottery winnings than as a stable increase in wealth. If the market consistently attracts new entrants based on recent performance, the resulting buying-and-repricing cycle could reinforce the very uncertainty and disagreement that make volatility more likely.

What to watch next

Investors will likely want to monitor whether crypto demand remains highly responsive to messages about recent performance—and how quickly beliefs converge or diverge after price moves. The Cleveland Fed paper’s central warning is that if disagreement and learning continue to shape participation, volatility may remain one of crypto’s defining features for the foreseeable future.