Have you ever wondered why so many people jump into cryptocurrency right after a strong Bitcoin run? It is not always about complex trading strategies or deep technical analysis. Sometimes it is simply because they saw the numbers and started believing the next year could look similar. A recent Federal Reserve Bank of Cleveland working paper digs into exactly that idea and the results feel both obvious and surprisingly precise once you sit with them.
What the Cleveland Fed Working Paper Actually Revealed
Researchers looked at large recurring surveys that regularly pulled between 15,000 and 25,000 responses. The participants came from a well-known consumer panel, so the sample felt closer to ordinary American households than to hardcore traders. In the second quarter of 2025 they ran a clean randomized information experiment. Some people saw Bitcoin’s trailing twelve-month return. Others received different pieces of market or inflation data. The goal was simple: measure whether that specific information changed what people said they wanted to own and, later, what they actually bought.
The answer turned out to be yes. Households shown Bitcoin’s previous-year performance raised their desired crypto allocation by roughly two percentage points. That may sound modest until you realize the control group averaged only about 4.3 percent. In relative terms it was a 47 percent jump. Even more interesting, the same people became about 2.5 percentage points more likely to report an actual cryptocurrency purchase in a later survey wave. Before the experiment roughly 11 percent of participants already held some crypto. The treatment therefore lifted the unconditional probability of buying by around 23 percent.
I have always suspected that past performance acts as a quiet marketing campaign for Bitcoin. This study puts numbers on that suspicion. It also shows the effect is not uniform. People who previously said they avoided crypto mainly because of limited knowledge responded most strongly. Those who already viewed it as a poor investment barely moved. That distinction matters.
The Wide Gap Between Owners and Non-Owners
One of the clearest findings sits in the expectations data. Back in the third quarter of 2021, crypto owners who offered a forecast expected an average return of 22 percent over the following year. Non-owners expected only 7 percent. Uncertainty was massive on both sides. Nearly 87 percent of non-owners simply answered “don’t know.” Among owners the share was still 54 percent. By 2025 both groups had become more cautious, yet the gap remained. Owners expected 13.8 percent while non-owners sat at 4.7 percent.
Expected returns also showed a tighter statistical link to ownership than standard personal characteristics such as age, income, gender or wealth. Each extra percentage point of expected return was associated with a 0.8-percentage-point higher probability of holding crypto. The researchers carefully note this is an association, not pure proof that optimistic beliefs alone created every existing position. Still, the randomized experiment supplies cleaner causal evidence about how specific historical-return information can shift later decisions.
Positive returns attract new participants, which raises the price further.
That short sentence from the paper captures a possible feedback loop. Rising prices lift expectations, new buyers enter, and the extra demand can push prices higher still. The authors present it as a potential bubble mechanism rather than a prediction that every rally will self-reinforce forever. In my view the caution is warranted. Markets have a long history of turning feedback into overshooting and then correcting sharply.
How the Information Experiment Worked in Practice
Participants were randomly assigned different pieces of information. One group saw Bitcoin’s 14.3 percent previous-year return. Another saw a simple price chart. Control groups received data on the S&P 500, GameStop, or an inflation forecast. Both Bitcoin treatments raised desired crypto allocations. The money appeared to come partly from cash, checking and savings accounts. Interestingly, desired stock allocations also rose a bit, suggesting the information encouraged broader appetite for risky assets rather than a pure substitution into crypto alone.
Later survey waves confirmed the shift was not just talk. Actual reported purchases increased by about 2.5 percentage points. The researchers pooled the two Bitcoin treatment arms to gain statistical power and reported a p-value of 0.017. Relatively few people changed ownership status between waves, which is why the pooling step mattered. Still, the directional result held up.
Perhaps the most practical takeaway is who responded. The strongest reaction came from non-owners who had earlier cited limited knowledge as their main reason for staying out. Once they received concrete return information, many of them reconsidered. People who already believed crypto was a poor investment showed almost no response. Information works best on the undecided, not the convinced skeptics.
Crypto Gains and Everyday Household Spending
The paper also examined whether Bitcoin price moves influence spending. The estimates suggest that doubling Bitcoin’s price made a household holding its entire financial portfolio in crypto 1.4 percentage points more likely to buy a durable good. With durable purchases occurring among roughly 20 percent of surveyed households, that represented a 7 percent relative increase. The effect was strongest for items such as computers and refrigerators and weaker for cars and homes. Nondurable spending barely moved.
The authors interpret this pattern as households treating crypto gains more like lottery winnings than a permanent rise in wealth. That framing feels familiar. Many of us have watched friends celebrate paper profits with a new gadget while leaving the bigger financial picture unchanged. Of course the interpretation remains the researchers’ reading of the statistics. It does not claim every crypto investor behaves the same way under every market condition.
Why Expected Returns Matter More Than Demographics
Traditional explanations for crypto ownership often lean on age, income or education. Younger people are more open. Higher earners have more risk capital. The Cleveland Fed data push back against that simple story. Once expected returns enter the model, the demographic variables lose much of their explanatory power. Beliefs about future performance simply dominate.
I find this reassuring in one sense and concerning in another. It is reassuring because it means people can update their views when new information arrives. It is concerning because historical returns are not guarantees. A strong trailing year can create overly optimistic forecasts that later disappoint. The study does not claim investors are irrational. It simply shows that recent performance is a powerful input into their mental models.
A Possible Price Feedback Loop in Plain Language
Imagine Bitcoin climbs 40 percent over twelve months. Media coverage increases. Friends mention their gains. Survey respondents who previously said “I don’t know enough” suddenly see a concrete number and raise their desired allocation. Some of them actually buy. The extra demand can support further price strength, which then becomes the next year’s trailing return for a new group of observers. The loop is not automatic or endless, but the mechanism is visible in the data.
This pattern aligns with broader observations that investor attention often follows market performance. Search interest and retail inflows tend to rise after strong periods and fade during prolonged declines. The working paper supplies experimental evidence that the attention-to-action step is real for a measurable slice of households.
Limitations Worth Keeping in Mind
The paper is a working paper. It circulates for discussion and may not receive the same formal review as official Federal Reserve publications. The findings represent the authors’ research and do not establish a position of the Cleveland Fed or the broader Federal Reserve System. Sample sizes are large, yet ownership changes between waves remain relatively rare, which is why statistical pooling was needed. Survey responses can also contain noise. People may report intentions more readily than they later execute trades.
Still, the combination of observational data on expectations and a randomized information treatment gives the results more weight than pure correlational studies. The fact that the strongest response came from the knowledge-constrained group adds a useful nuance. Information does not move everyone equally.
What This Means for Everyday Investors
If you already hold crypto, the study may simply confirm what you sensed: recent performance shapes how others view the asset and can bring new buyers into the market. If you have stayed on the sidelines because the technology felt opaque, the findings suggest that concrete return information can shift that hesitation for many people. The key is recognizing that past returns are one data point among many. They do not eliminate risk or guarantee future outcomes.
In my experience the healthiest approach is to treat any single year’s performance as interesting but incomplete. Portfolio decisions work better when they rest on a longer view of volatility, liquidity needs and personal risk tolerance. The Cleveland Fed research does not contradict that advice. It simply shows how powerful a recent strong run can be in the moment.
- Strong trailing Bitcoin returns raised desired crypto allocations by about two percentage points
- Actual subsequent purchases rose by roughly 2.5 percentage points among treated households
- Owners consistently expected higher future returns than non-owners
- Knowledge gaps, more than demographics, explained much of the ownership difference
- Crypto gains appeared to influence durable-goods spending more than everyday consumption
Broader Context Around Crypto Use in the United States
Separate Federal Reserve survey work has already shown that most Americans who hold cryptocurrency treat it primarily as an investment rather than a payment tool. Activity reached about 10 percent of U.S. adults in 2025 while everyday payment use stayed limited. The new working paper fits neatly into that picture. People enter the market largely because they expect the price to rise, not because they need a new way to buy coffee.
That investment motive makes the return-information channel especially relevant. When prices climb, the investment case becomes more visible to the next wave of potential participants. When prices fall for an extended period, the same channel can work in reverse as fewer people update their expectations upward.
Putting the Numbers in Perspective
A two-percentage-point rise in desired allocation may sound small if you think in absolute terms. Relative to a 4.3 percent baseline it is substantial. A 2.5-percentage-point rise in the probability of buying is likewise meaningful when starting from an 11 percent ownership rate. These shifts matter at the margin. Markets are often driven by marginal buyers and sellers. Small percentage-point changes across tens of millions of households can add up to noticeable flows.
The researchers also found that the Bitcoin information treatments raised desired stock allocations slightly. That spillover suggests the information may have increased general appetite for risk rather than simply shifting money from cash into crypto. Whether that broader risk appetite persists or fades is an open question the paper does not fully settle.
Personal Reflections on the Feedback Mechanism
I have watched several Bitcoin cycles now. Each time a strong year arrives, the conversation changes. Friends who previously dismissed the asset start asking questions. News coverage softens. Search volumes rise. The Cleveland Fed experiment essentially recreates a controlled version of that process and measures the behavioral response. The measured effect is real. It is also limited. Not everyone moves, and those who already hold negative views rarely change them on the basis of one year’s return.
Perhaps the most useful insight is the distinction between knowledge-constrained non-owners and convinced skeptics. Educational efforts and clear performance data may bring the first group into the market. The second group requires a different kind of evidence or may never participate. Policy makers, educators and market participants who want broader adoption would do well to remember that distinction.
Looking Ahead Without Overclaiming
The working paper does not forecast the next Bitcoin move. It does not claim that every rally will become self-reinforcing. It simply documents that historical return information can shift both stated preferences and subsequent behavior for a measurable share of households. That finding is useful for anyone trying to understand retail participation dynamics.
In the end the study reinforces a simple truth many long-time observers already sensed: past performance is a powerful attention magnet. When Bitcoin delivers strong trailing returns, more people notice, more people update their expectations, and some of them act. Whether those new participants stay for the long term or leave after the next downturn remains an open question. The data, however, make the entry channel clearer than before.
Understanding that channel does not require believing in permanent upward trends. It only requires recognizing that human attention and belief updating respond to recent results. The Cleveland Fed researchers have given us careful evidence of how that process works in the specific case of cryptocurrency. For investors, the practical lesson is to stay aware of the influence recent returns can exert on both others and ourselves. Decisions grounded in a longer horizon and clear personal risk parameters will usually serve better than those driven purely by the latest twelve-month number.
The research also reminds us that crypto remains primarily an investment vehicle for most U.S. holders. Payment use stays limited. That reality keeps the focus on expected returns as the central driver of ownership. As long as that remains true, information about past performance will continue to shape who enters the market and when. The feedback loop described in the paper is therefore worth watching in future cycles, even if it never becomes mechanical or endless.
Finally, the durable-goods spending result offers a small window into how crypto wealth is experienced. Gains that feel like lottery wins tend to fund discretionary upgrades rather than everyday consumption. That behavioral detail may matter for understanding the real-economy footprint of large crypto price moves. It also humanizes the numbers. Behind the percentage points sit ordinary households deciding whether a new computer or refrigerator feels affordable after a strong market year.
Taken together, the findings paint a coherent picture. Expected returns dominate ownership decisions. Concrete information about Bitcoin’s recent performance can raise both desired allocations and actual purchases. The effect is strongest among those who previously felt under-informed. And some of the resulting wealth gains appear to spill into selective durable spending. None of these results is revolutionary on its own. Collectively they give a clearer empirical foundation to patterns many market participants have observed for years.
As with any working paper, further research will refine or challenge the conclusions. Larger samples, different market environments, and longer follow-up periods could all add nuance. For now the evidence is strong enough to take seriously. Past Bitcoin returns do more than decorate price charts. They shape beliefs, open doors for new participants, and leave measurable traces in household portfolio choices and even spending patterns. That is worth remembering the next time a strong year appears in the rear-view mirror.
