Imagine two versions of the same morning. In the first, a fresh U.S. jobs report lands at 8:30 a.m. and traders around the world scramble to reprice everything from stocks to bonds. In the second, the exact same report arrives, but the mood among investors is already sunny. Do markets react to that number in the same way both times, or does the prevailing mood change how the news gets absorbed?
That question sits at the heart of a study of cryptocurrency markets published in the Journal of Behavioral and Experimental Finance. The research offers evidence that crypto prices and trading react sharply and almost instantly to scheduled economic reports, and that the strength of those reactions depends heavily on how optimistic investors are feeling at the time.
Why cryptocurrencies make an unusual test case
Cryptocurrencies behave differently from traditional assets in ways that matter for this kind of study. They trade around the clock, they are decentralized, and a large share of their participants are individual retail investors rather than big institutions. A JPMorgan survey cited in the paper found that only about 11 percent of institutional investors were trading crypto, and most of those who had not started said they were unlikely to.
Retail-heavy markets tend to process information differently. The author points to earlier research suggesting that crypto holders often have high digital literacy but relatively low financial literacy, which pushes them toward mental shortcuts when interpreting new information. Cryptocurrencies also lack the cash flows or earnings that anchor the value of a stock, making their prices more open to speculation and belief-driven swings.
That backdrop left an open question. Studies of whether crypto responds to economic news had produced conflicting results, with some finding strong reactions and others finding almost none. Nhan Huynh of Griffith University in Australia set out to reconcile those findings and to test something that had received little attention: whether investor sentiment shapes how economic news gets translated into prices and trading.
Measuring reactions minute by minute
The study draws on high-frequency trading data sampled at five-minute intervals for the top 100 cryptocurrencies from January 2014 through December 2021. Those coins account for more than 90 percent of total crypto market value. To reduce the risk of only studying survivors, the dataset kept coins that later became inactive or were delisted, letting each contribute data only while it was actively trading.
Huynh tracked two things around each news release: returns (how much prices moved) and turnover (trading volume relative to market value, a gauge of trading intensity). The main window was narrow, running from five minutes before to five minutes after an announcement, which helps isolate the news itself from unrelated market noise.
The economic news came from 32 scheduled U.S. reports, grouped into eight categories including real economic activity, consumption, investment, inflation measures, monetary policy, and forward-looking surveys. For each report, Huynh calculated a “surprise,” meaning the gap between the actual figure and what forecasters had expected. Standardizing these surprises made it possible to compare, say, an inflation reading against a jobs number on the same scale.
To capture mood, the study used cryptocurrency-specific sentiment indices from Refinitiv MarketPsych. These indices apply machine learning and text analysis to more than 2,000 news outlets and hundreds of social media sites, scoring the tone of coverage on a scale from very negative to very positive. Importantly, the sentiment measure reflects investor mood and attention rather than the actual state of the economy, which lets it act as a separate ingredient from the news surprises themselves.
What the analysis revealed
The first finding was that crypto markets do respond, and quickly. Most economic releases produced statistically meaningful moves in both returns and trading within the ten-minute window. Of the 32 announcements, 20 moved returns significantly and 24 moved turnover. The direction made economic sense: better-than-expected GDP growth pushed returns up, while higher-than-expected unemployment pushed them down. On average, a one-standard-deviation surprise moved absolute returns by roughly 9.8 basis points and turnover by 17.9 basis points within ten minutes.
The second finding is the one Huynh treats as central. When sentiment was bullish, the market’s reaction to economic news was consistently weaker. A one-standard-deviation increase in investor sentiment reduced the return response to a surprise by about 16 percent and the turnover response by about 24 percent. In plain terms, the same jobs report tended to move prices and volume less when investors were already feeling optimistic.
Huynh links this pattern to an idea from behavioral research: optimistic investors tend to lean on mental shortcuts and pay less attention to fundamental information, while pessimistic or anxious investors tend to process information more systematically. Under that view, a wave of good feeling can dull the market’s sensitivity to hard economic data.
Testing whether the pattern holds up
Much of the paper is devoted to checking whether these results survive different assumptions. To address the worry that mood and news might be reacting to each other at the same moment, Huynh re-ran the analysis using sentiment measured on the day before an announcement, and again using the average of the three prior days. The dampening effect remained, which the author interprets as sentiment shaping how news is read rather than simply responding to it.
The findings also held when sentiment was measured in other ways, including a broader country-level index, the well-known Baker and Wurgler stock-market sentiment index, and a survey of individual investors. They held across wider event windows of 10 and 30 minutes, across large and small coins, and after controlling for factors such as market volatility, economic policy uncertainty, and search-based investor attention.
A few additional patterns stood out. Bad economic news tended to move the market more than good news of the same size, and sentiment softened the reaction to bad news especially. Reactions were noticeably stronger during the 2020 to 2021 pandemic period, when uncertainty was high, and stronger for large, widely followed coins than for small ones. When Huynh extended the analysis to other economies, crypto reacted meaningfully to news from the European Union and China, but was largely unresponsive to reports from Japan and Germany. In all the responsive cases, sentiment again played a moderating role.
What it might mean for investors
Huynh is careful to frame the practical takeaways modestly. The results do not point to a mechanical trading rule or a promise of profit. Instead, the author suggests that the same economic announcement may carry different weight depending on the mood of the market, which has implications for managing short-term risk and timing portfolio adjustments around scheduled releases. For anyone holding digital assets, the broader message is that a “sentiment-aware” reading of economic news may be more informative than treating every report as equally market-moving.
Several caveats deserve attention. The study focused on large, liquid cryptocurrencies to ensure reliable minute-by-minute data, so the findings may not extend cleanly to smaller or more speculative tokens. The sample was also built from coins that ranked near the top by the end of the period, which the author notes may introduce some survivorship bias despite efforts to limit it. And because this is observational data rather than a controlled experiment, the results describe associations and immediate reactions rather than proof that sentiment causes markets to tune out the news.
