US spot Bitcoin ETF flows are often treated as a real-time read on market direction, but new research from DWF Labs argues they are better understood as a reaction to price moves rather than a reliable signal of what comes next. In a report released Thursday, the firm found that ETF investors were more likely to withdraw funds when Bitcoin experienced sharp drawdowns—and add during similarly sized rallies.

DWF’s analysis examined 694 trading days dating back to January 2024, when spot Bitcoin ETFs began operating in the US. Over that period, investors pulled money from the funds during seven of the eight instances when Bitcoin fell by at least 15%. By contrast, every one of the nine rallies of that magnitude coincided with net inflows into the ETFs.

Key takeaways

DWF Labs found ETF investors tended to react to large moves: outflows followed 15%+ sell-offs in 7 of 8 cases, while 15%+ rallies saw inflows in all 9 cases.
ETF flows show confirmation of what already happened more than they predict direction over the next week.
The relationship between daily ETF flows and Bitcoin’s performance was moderate for the prior five trading sessions (correlation 0.54) but nearly absent for the following five sessions (correlation 0.03).
DWF estimated roughly a fifth of ETF-related activity may have been hedged using CME Bitcoin futures, though the firm said it cannot prove the same investors were on both sides.
Bitcoin’s large sell-offs have become less frequent since ETFs launched, while still producing declines of similar size to earlier periods.

Why ETF flows struggle as a “forecast”

DWF’s central message is that ETF flows largely validate market moves rather than drive them. “ETF flows confirm moves; they don’t predict them,” the firm wrote in its report. “The evidence points mostly to investors reacting to price, not driving it.”

That distinction matters for anyone using ETF data as a directional trading tool. According to DWF, inflows and outflows provide little guidance on whether Bitcoin will rise or fall over the subsequent five trading sessions—precisely the window many investors look to when interpreting flow headlines.

To quantify that lack of predictive power, DWF reported that each day’s ETF flows had a moderate relationship with Bitcoin’s price changes over the five trading sessions before that day, with a correlation of 0.54. But for the five sessions that followed, the correlation between ETF flows and Bitcoin’s moves dropped to 0.03—close to what you would expect if the two were effectively unrelated.

In practical terms, the findings challenge a common assumption: that consistent flow patterns can be read like a leading indicator. Instead, DWF’s research points to a market where investors adjust exposure after price action has already occurred.

Outflows cluster during the biggest drawdowns

DWF also looked at when withdrawals were largest. The heaviest outflows occurred during the two most recent sell-offs in its dataset, with investors withdrawing a net $6.9 billion between October 2025 and February 2026. A second wave followed during the May–June decline, when net withdrawals totaled $8.2 billion.

While the numbers highlight that large drawdowns tend to coincide with ETF selling pressure, the research does not clarify who is behind those moves. DWF said it could not separate whether the outflows were driven by institutions or retail investors.

“We don’t have a clear split on the exact buyers/sellers of the ETFs,” a DWF Labs spokesperson told Cointelegraph.

That limitation is important for market interpretation. ETF flows are aggregate data, and the motivations behind buying and selling can differ across investor types—particularly during periods of volatility when hedging strategies become more common.

Hedging may explain part of ETF activity

Another nuance in DWF’s analysis is that some ETF-related activity could reflect hedged positioning rather than a straightforward bet on Bitcoin’s spot price. The firm noted that a portion of ETF activity may be linked to investors who buy fund shares while simultaneously taking positions in Bitcoin futures on the CME—potentially aiming to capture the price difference between the two exposures.

As DWF put it, the research estimated that roughly a fifth of flows may have been hedged in this manner. However, the firm said its data cannot confirm that the same investors held both the ETF exposure and the futures positions.

This distinction reinforces why flows can be difficult to interpret. If some investors are using ETFs as part of a broader hedging strategy, net flows may reflect portfolio management mechanics instead of directional conviction. Even where flows do correlate with previous price changes, that does not necessarily mean they can be used to forecast the next move.

DWF’s hedging discussion also aligns with a separate finding from CoinShares published in June. That analysis, according to Cointelegraph coverage, reported that professional investors reduced their Bitcoin ETF exposure by 17% in the first quarter of 2026, with hedge funds and brokerages accounting for roughly 96% of the reduction while banks increased their holdings. While that work does not replicate DWF’s flow-versus-price methodology, it supports the broader idea that ETF positioning can shift for reasons beyond immediate spot-price prediction.

Sell-offs look different since ETFs launched

DWF’s report suggests that Bitcoin’s path around large drawdowns has changed since the ETFs entered the market. The firm found that Bitcoin has had about three sell-offs of 15% or more per year since ETFs launched, compared with roughly four a year in 2022–2023 and eight a year in 2021.

Declines have also become somewhat more gradual. Even so, the magnitude of losses appears broadly comparable. DWF reported that the median sell-off in the post-ETF period erased 21.3%, versus 19.4% in 2022–2023.

Importantly, DWF’s spokesperson cautioned that the ETF rollout is unlikely to be the sole driver. “The changes aren’t solely due to the ETF,” they said, noting that volatility and the frequency of sell-offs were already declining before the funds began in 2024.

For investors, that framing matters because it reduces the temptation to attribute every shift in market behavior directly to spot ETF flows. Instead, DWF’s findings point toward a more complex environment in which broader volatility trends—and potentially new hedging and positioning tools—interact with ETF trading.

As the market continues to adapt, the key question for traders and long-term allocators is whether future ETF flow data becomes more predictive—or whether DWF’s pattern of “reactive flows” remains the dominant feature. Readers should watch for shifts in the flow-to-price relationship over time, particularly whether the correlation structure changes after large moves, and whether better data emerges to distinguish who is behind ETF inflows and outflows.