Skip to main content

Fed Study: Crypto Investors Follow Beliefs, Adjust to Returns



A new working paper from researchers at the Federal Reserve Bank of Cleveland argues that much of crypto’s unusual behavior may come down to how people form beliefs about digital assets—more than standard demographic or financial factors. In their analysis of household survey data and a randomized information experiment, the authors find that expectations about future crypto returns strongly track who owns cryptocurrency, and that learning about recent Bitcoin performance can meaningfully change both planned allocations and actual buying.



The work, titled “Do You Even Crypto, Bro? Cryptocurrencies in Household Finance”, also points to a mechanism that could help explain crypto’s persistent volatility: when prices rise, they may reinforce bullish expectations, drawing in additional buyers and further amplifying movements.



Key takeaways



  • Return expectations explain crypto ownership better than demographics. The paper finds expected returns and perceived risk account for more variation in ownership than age, income, or gender.

  • Many non-owners lack clear beliefs about crypto returns. In a 2021 Fed survey, most non-owners reported they did not know what return to expect over the following year.

  • Information about recent Bitcoin performance changes behavior. In a 2025 randomized trial, showing households Bitcoin’s past 12-month return increased desired crypto allocations and subsequent purchases.

  • Crypto gains may be treated like “gambling income,” not durable wealth. The study links BTC price increases to more durable-good spending, but not lasting increases in consumption.



Beliefs, not just backgrounds, shape who buys crypto


The researchers—Michael Weber, Bernardo Candia, Olivier Coibion, and Yuriy Gorodnichenko—use repeated survey waves covering up to 25,000 U.S. households per wave. Their central finding is that what people think crypto returns will be does a far better job of explaining ownership than observable characteristics.



According to the paper, expected returns are particularly influential: a one-percentage-point increase in an individual’s expected crypto return corresponds to a 0.8-percentage-point rise in the probability of owning cryptocurrency. When expectations about returns are combined with expectations about risk, the explanatory power rises further—surpassing the impact of factors like age, income, and gender.



This makes crypto an outlier compared with traditional asset categories such as stocks, bonds, and gold. For those markets, demographic and financial characteristics tend to matter more for ownership patterns than differences in return expectations. The paper’s implication is that crypto participation is driven less by “who you are” and more by “what you believe crypto will do.”



Large gaps in understanding may widen the volatility loop


The study also highlights how uneven knowledge and beliefs are across the population. In a 2021 Federal Reserve survey referenced by the authors, 87% of respondents who did not own crypto said they did not know what return to expect from it over the following year. Among crypto owners, the figure was still high—54% reported not knowing what return to expect.



For those who were willing to make a forecast, the study describes a substantial divergence. Crypto owners expected an average 22% return over the next year, compared with 7% among non-owners. Owners also tended to perceive crypto as less risky than non-owners did.



The authors connect this belief gap to a potentially self-reinforcing market dynamic. If rising prices strengthen optimistic expectations, those expectations can attract new participants whose buying then supports higher prices. In the paper’s words, “Positive returns attract new participants, which raises the price further.”



Notably, this mechanism does not require the underlying asset fundamentals to be the only driver of price action. Instead, it suggests that disagreement and learning—how investors update beliefs based on past outcomes—can become a major source of volatility.



Experimental evidence: showing Bitcoin’s past performance moves allocations and purchases


The paper’s most actionable part for market participants comes from a randomized information experiment. In 2025, the researchers randomly assigned households to receive information comparing Bitcoin with other topics, including stocks, GameStop, and inflation. Participants who were shown Bitcoin’s previous 12-month return increased their desired crypto portfolio allocation by roughly 2 percentage points.



The authors report this as about a 47% increase relative to a 4.3% desired allocation in the control group. They also find an increase in actual subsequent crypto purchases of about 2.5 percentage points.



Crucially, the effect was concentrated among people who said they did not own crypto because they felt they lacked sufficient information. Those who already believed crypto was a bad investment did not respond in the same way to the informational treatment.



Framed for readers, the experiment implies that retail demand may not be driven solely by price headlines or broad narratives. It may also be driven by what people are prompted to focus on—specifically, whether they are given recent performance data that reshapes expectations about future returns.



What happens to consumption when crypto wealth rises?


Beyond trading behavior, the paper examines whether crypto gains translate into broader spending patterns. The authors report a spillover effect from crypto price changes into household consumption, particularly for durable goods. When BTC doubled in price, a household whose entire financial portfolio was in crypto became 1.4 percentage points more likely to buy a durable good.



They describe this as roughly a 7% increase relative to the unconditional probability of such a purchase. However, the effect did not persist into ordinary spending, and the pattern leads to a sharper interpretation: crypto gains appear to function more like “gambling income” or lottery winnings than as a steady, confidence-building increase in long-term wealth.



That distinction matters because it suggests crypto’s influence may be episodic. Even if price surges provide short-term boosts to certain spending categories, they may not reshape households’ longer-term financial behavior in the same way as more stable forms of wealth.



For investors and builders, the study’s core warning is uncomfortable but practical: if participation is belief-driven and information-sensitive, volatility may remain structurally high. Readers should watch next for how new retail entrants interpret Bitcoin’s recent performance, and whether changes in public messaging or access to return information amplify the feedback loop the paper describes.



https://www.cryptobreaking.com/fed-study-crypto-investors-follow/?utm_source=blogger%20&utm_medium=social_auto&utm_campaign=Fed%20Study:%20Crypto%20Investors%20Follow%20Beliefs,%20Adjust%20to%20Returns%20

Comments

Popular posts from this blog

Top Cryptocurrencies to Watch: BTC, ETH, BNB, XRP, Solana, Dogecoin & More

Market Analysis and Price Predictions for Key Cryptocurrencies Recent market dynamics reveal a cautious sentiment across the cryptocurrency landscape, with Bitcoin struggling to maintain levels above $90,000 and many major altcoins facing downward pressure. Indicators point toward reduced participation from both institutional and retail investors, raising concerns about a potential consolidation phase after notable gains earlier in the year. Bitcoin has fallen below $87,000, reflecting waning demand at higher price points. Institutional fund flows into BTC and ETH ETFs have turned negative, indicating a period of subdued market activity. Active addresses and Binance deposit/withdrawal activities are at annual lows, suggesting market indecision. Most leading altcoins are approaching support levels, with some poised for potential breakdowns. Tickers mentioned: Bitcoin, Ethereum, Binance Coin, XRP, Solana, Dogecoin, Cardano, Bitcoin Cash, Chainlink, Hyperliquid Sentiment: Neutral to Sli...

Mastercard Launches AI Agent Pay System With Ripple and Solana Help

Mastercard has launched Agent Pay for Machines, a payments system built for autonomous software agents. The service allows AI agents to send and receive payments without direct human action. It brings Ripple, Coinbase, and Solana Foundation into Mastercard’s push for automated digital commerce. Ripple Brings XRPL and RLUSD to Mastercard’s Agent Pay System Mastercard introduced Agent Pay for Machines on June 10 as a tool for machine-led payments. The system targets high-volume and low-value transactions across business and consumer use cases. It also supports automated settlement between software agents and connected machines. Ripple will support the system through the XRP Ledger and its RLUSD stablecoin. The company said that settlement will become more important as automated commerce grows. It also sees blockchain rails as useful for fast and rule-based payments. RippleX senior vice president Markus Infanger said XRPL and RLUSD support enterprise-grade agent payments. He said the tool...

Coinbase's x402 launches AI agents app store for payments

Coinbase-backed x402 has unveiled Agentic.market, a dedicated marketplace aimed at increasing the usefulness of AI agents by aggregating thousands of apps and services that agents can access without any API keys. The rollout positions the platform as a central hub for agents to discover, evaluate, and deploy capabilities across a standardized payments layer. Coinbase product lead Nick Prince described Agentic.market in a video posted on X as a storefront for discovering, comparing, and using x402 services. The marketplace is designed to give both humans and their AI agents access to a wide range of tools—from data feeds to consumer apps—without the friction of managing API credentials. A storefront for discovering, comparing, and using x402 services. Thousands of services. Zero API keys. Powered by x402. Prince added that the market offers a web interface for humans to browse and assess services, alongside a programming layer that lets AI agents autonomously search, filter, and integra...