The correlation between Bitcoin and physical gold (XAU/USD)

The correlation between Bitcoin and physical gold (XAU/USD) remains weak and unstable over the long term, making them an excellent diversification pair.
Long-term (since 2015/3-5 years): average correlation ≈ +0.10 to +0.13.
Over one year: typically between -0.17 and +0.32.
Short-term (30 days): currently negative (recent readings between -0.31 and -0.46, with extremes at -0.88 earlier in 2026). 90 days: near zero or slightly negative (≈ -0.08). These levels confirm that BTC and gold react to different macro drivers, despite the “digital gold” narrative.

https://www.steelldy-indices.com

Recent correlation data (public sources 2026) shows:
30 days: –0.31 to –0.46 (AhaSignals, Newhedge, late July/early August);
90 days: –0.08 (AhaSignals);
1 year: –0.17 to +0.12/+0.32 (Mudrex, Bitbo, Portfolio Lab);
3–5 years/long term: +0.05 to +0.13 (J.P. Morgan LTCMA 2026, NYDIG, average since 2015 ≈ 0.10).
The 30-day correlation turned negative in 2026, hitting a low of –0.88 in March, the lowest since 2022. This remains non-structural, shifting quickly from negative to positive depending on liquidity, real rates, and risk appetite regimes.

Gold’s price is primarily driven by real interest rates, central bank purchases, geopolitical uncertainty, and inflation, acting as a classic safe-haven asset. Bitcoin is more sensitive to global liquidity, risk appetite (Nasdaq), ETF flows, the halving cycle, and retail/institutional sentiment, functioning as a liquidity barometer rather than a pure real rates hedge. Although both assets share some common factors like the dollar and liquidity, their sensitivities diverge sufficiently to maintain a low average correlation.

The current Bitcoin-to-gold ratio is approximately 15.4–15.8 ounces of gold per Bitcoin, which is low compared to the three-year average of about 21 ounces. Gold has significantly outperformed Bitcoin in 2025–2026, with positive returns while Bitcoin experienced deep negative returns over 12 months. Historically, very low ratio levels have preceded Bitcoin catch-up phases, but this is not a reliable short-term timing signal.

Sharpe Ratios

_{We treat the –0.88 print in March 2026 as a short-term regime-dependent episode, not a structural regime shift. Key points from our framework:Long-term average remains low and stable
Since 2015 the average correlation sits around +0.10 to +0.13 (J.P. Morgan LTCMA, NYDIG).
This is the number that matters for portfolio construction.
Short-term correlations are highly unstable by nature
30-day and 90-day readings regularly swing between –0.5 and +0.5 (and occasionally more extreme).
The March –0.88 was the lowest since 2022, but previous extreme negative prints (including in 2022) reversed relatively quickly once liquidity and risk-appetite regimes changed.
Different drivers, not a broken relationship
Gold remains primarily a real-rates / central-bank / geopolitical asset. Bitcoin behaves more like a high-beta liquidity and risk-appetite asset. When these drivers temporarily diverge strongly (as they did in early 2026), the correlation can go deeply negative. That does not invalidate the long-term low average.
At STEELLDY we therefore:Monitor the 30-day and 90-day correlations as tactical signals
Use the long-term average (~0.10–0.13) for strategic asset allocation and diversification analysis
Do not interpret temporary extremes as a permanent rejection of the “complementary stores of value” thesis
In short: the –0.88 was real and informative about that specific regime, but it is noise relative to the multi-year relationship. The diversification benefit between BTC and physical gold remains intact on the horizon that matters for portfolio construction.}