Does gold lead Bitcoin?

1 Does gold move before Bitcoin?

Not reliably, in this sample. On weekly data from 4 September 2020 to 23 October 2026, which give 315 pairs of six-week changes (fewer at longer leads), gold’s six-week change was most closely associated with Bitcoin’s six-week change 13 weeks later, at a correlation of 0.11, and a peak at least that high turned up somewhere between 0 and 39 weeks in 85% of the samples in which Bitcoin’s series was shifted in time to break any real link. This is a model estimate as of 28 October 2026, and past results do not indicate future results.

The question starts in the terminal’s Chart Desk. The layout built for this article, Figure 1, puts weekly Bitcoin with PAX Gold, a gold-backed token, laid over it, and below that PAX Gold with the Global Liquidity Model laid over it. Read by eye, the two assets seem to take turns: over the year to 29 December 2023 Bitcoin rose about 153% and PAX Gold about 12%, and over the year to 26 December 2025 PAX Gold rose about 72% while Bitcoin fell about 7%. Turns like these are where the idea of a gold lead comes from, and the sections below test whether they add up to one. A free account can build the same layout and take it further: the interval switched to daily or monthly, the Collateral Base added from the watchlist as a further overlay, or both charts indexed to 100. Past results do not indicate future results.

2 Why would gold lead Bitcoin at all?

Because both are thought to answer to the same two forces, the quantity of money and credit and its price after inflation, and gold might answer first. If it did, gold’s moves would be associated with Bitcoin’s later ones, and a move in gold would say something about Bitcoin before Bitcoin itself moved.

The terminal draws both forces. On its Global Liquidity page, the growth chart in Figure 2 shows the first as the Global Liquidity Model measures it. Twelve-month growth peaked at 28.3% on 26 February 2021 and fell to −5.7% on 11 November 2022. Over the year to 30 December 2022, with the Global Liquidity Model down 3.3%, Bitcoin fell about 64% and PAX Gold about 1%: one contraction, two very different responses. This is a model estimate as of 28 October 2026, the model’s weeks before 31 July 2026 are reconstructed, and past results do not indicate future results.

Note: Growth of the Global Liquidity Model (3-month, annualised; 12-month), weekly from 2019, as a free account sees it: to 7 August 2026. Twelve-month growth peaked at 28.3% on 26 February 2021 and touched −5.7% on 11 November 2022. Weeks before 31 July 2026 are reconstructed. Model estimate as of 28 October 2026; past results do not indicate future results.Source: BlockPhi terminal, Global Liquidity > Global Liquidity growth, captured 28 October 2026.

Free with an account·Open this chart in the terminal (opens in a new tab)

Global Liquidity Model growth at each quarter’s last week, June 2020 to 7 August 2026, the last week a free account saw on the capture day, percent. Weeks before 31 July 2026 are reconstructed. Model estimate as of 28 October 2026; past results do not indicate future results.

Week	3-month, annualised	12-month
26 June 2020	35.24%	10.27%
25 September 2020	24.83%	16.63%
25 December 2020	44.22%	26.05%
26 March 2021	7.89%	27.31%
25 June 2021	12.31%	21.53%
24 September 2021	7.87%	17.18%
31 December 2021	6.28%	8.50%
25 March 2022	−2.49%	5.68%
24 June 2022	−13.01%	−0.86%
30 September 2022	−14.32%	−5.58%
30 December 2022	16.98%	−3.29%
31 March 2023	10.18%	−0.12%
30 June 2023	−4.88%	1.24%
29 September 2023	−5.39%	3.78%
29 December 2023	17.60%	3.92%
29 March 2024	−0.42%	1.32%
28 June 2024	−2.06%	2.06%
27 September 2024	21.54%	8.66%
27 December 2024	−11.20%	1.29%
28 March 2025	14.49%	4.89%
27 June 2025	17.92%	9.87%
26 September 2025	7.29%	6.50%
26 December 2025	−0.19%	9.65%
27 March 2026	11.43%	8.91%
26 June 2026	5.22%	5.86%
7 August 2026	1.13%	6.51%

What moves the model from week to week, from central bank balance sheets to the dollar’s translation of the other currencies, is set out in what moves Global Liquidity.

On the Yield Curve page, the real-yields card splits the 10-year Treasury yield into what an inflation-protected bond pays and the inflation the market expects, and its own reading note says that rising real yields have weighed on gold, growth stocks and Bitcoin. The 10-year real yield went from −1.04% on 31 December 2021 to 1.58% on 30 December 2022, and was 2.64% on 27 October 2026, close to the card’s high since 2019 of 2.95% on 5 October 2026. Yet PAX Gold rose about 72% in the year to 26 December 2025 with the real yield near 2%. In this sample gold rose through conditions the card associates with weakness, which suggests that something other than rates was setting its price. Past results do not indicate future results.

3 Who buys gold, and who buys Bitcoin?

Different buyers at the margin, which is one reason a lead between the two would come and go. World Gold Council data put buying by central banks and other institutions, as the Council groups them, above 1,000 tonnes in each of 2022, 2023 and 2024 (Gold Demand Trends: Full Year 2024, published 5 February 2025) and at 863 tonnes in 2025 (Gold Demand Trends: Q4 and Full Year 2025, published 29 January 2026). Past results do not indicate future results.

The terminal’s Positioning page shows the speculative side of gold, from the CFTC’s weekly Commitments of Traders, in Figure 3. Managed money, the CFTC’s category for commodity pools and other professionally managed funds, held a net long of 166,169 COMEX gold contracts of 100 troy ounces on 31 December 2024 and 126,040 on 30 December 2025: a smaller position at the end of a year in which PAX Gold rose about 72%. The card’s own note reads a falling net long during a rally as a sign that the buying comes from outside futures, from central banks or physical demand. Past results do not indicate future results.

Note: Net position of managed money in COMEX gold futures, contracts of 100 troy ounces, weekly as of each Tuesday. It was 166,169 contracts on 31 December 2024 and 126,040 on 30 December 2025, a smaller net long at the end of a year in which PAX Gold rose about 72%. Past results do not indicate future results.Source: BlockPhi terminal, Positioning > Gold (COMEX): managed money, captured 28 October 2026. Data: CFTC disaggregated Commitments of Traders.

In the terminal for Premium members·Open this chart in the terminal (opens in a new tab)

Managed money’s net position in COMEX gold futures (long less short, futures only), contracts of 100 troy ounces, at the window’s high and low, the last report of each year from 2020 to 2025, and the report of 20 October 2026. Past results do not indicate future results.

Report (Tuesday)	Managed money, net contracts	Note
29 December 2020	108,995	last report of the year
28 December 2021	83,663	last report of the year
27 September 2022	−43,094	low in the window
27 December 2022	50,877	last report of the year
26 December 2023	106,454	last report of the year
24 September 2024	219,029	high in the window
31 December 2024	166,169	last report of the year
30 December 2025	126,040	last report of the year
20 October 2026	104,206	latest report read

One card further down, Figure 4, the same report covers Bitcoin futures on CME. Leveraged funds’ net short reached 23,197 contracts of 5 BTC on 17 December 2024, which the card reads as the cash-and-carry trade: owning spot Bitcoin or exchange-traded funds and selling the future to earn its premium. Asset managers’ net long peaked at 19,802 contracts on 23 July 2024, and on 20 October 2026 the two stood at a net short of 4,942 and a net long of 2,999. Gold’s marginal buyers in this sample include central banks; Bitcoin’s include funds running a basis trade. There is no obvious reason for the two to keep one timetable. Past results do not indicate future results.

Note: Net positions of leveraged funds and asset managers in CME Bitcoin futures, contracts of 5 BTC, weekly as of each Tuesday. Leveraged funds’ net short reached 23,197 contracts on 17 December 2024; asset managers’ net long peaked at 19,802 on 23 July 2024. Past results do not indicate future results.Source: BlockPhi terminal, Positioning > Bitcoin futures (CME), captured 28 October 2026. Data: CFTC Traders in Financial Futures.

Free with an account·Open this chart in the terminal (opens in a new tab)

Net positions in CME Bitcoin futures (long less short, futures only), contracts of 5 BTC, at the last report of each year from 2020 to 2025 and on 20 October 2026. Past results do not indicate future results.

Report (Tuesday)	Leveraged funds, net	Asset managers, net	Note
29 December 2020	−4,787	588	last report of the year
28 December 2021	−5,651	4,695	last report of the year
27 December 2022	−1,922	4,972	last report of the year
26 December 2023	−9,831	11,862	last report of the year
31 December 2024	−20,064	9,770	last report of the year
30 December 2025	−10,911	6,597	last report of the year
20 October 2026	−4,942	2,999	latest report read

4 What do published estimates say?

They disagree on the timing and even on the direction (Table 1). CryptoSlate reported on 14 January 2026 that Bitwise Europe’s head of research, using Granger causality tests, found gold leading Bitcoin by about four to seven months. A liquidity research vendor’s newsletter found on 8 October 2025, in a vector autoregression that also carries its own liquidity measure, that a rise in gold was followed by a fall in Bitcoin some four to six weeks later, and on 28 June 2026 described Bitcoin peaking first, on 6 October 2025, with gold’s peak around early March 2026, while allowing that the timing may be coincidence. A peer-reviewed study published on 12 March 2026, on daily data from 2013 to 2025 with macroeconomic controls, found directional dominance between the two generally weak once many tests were corrected for, with gold’s influence on Bitcoin stronger at long horizons in periods of stress. An earlier BlockPhi version of this analysis, published on 5 May 2025, put the lag at about 100 to 150 days. Each is its author’s estimate, and past results do not indicate future results.

Part of the spread is method: Granger tests, a vector autoregression, peak dating and daily quantile tests ask different questions of the same two prices. Two of the estimates hold other variables beside them, the newsletter’s vector autoregression its own liquidity measure and the peer-reviewed study a set of macroeconomic controls; none tests gold against Bitcoin with the Global Liquidity Model held constant.

5 How does BlockPhi test the lead?

With the lag scan the terminal itself runs on the Global Liquidity Model, so gold, the dollar and Global Liquidity are measured against Bitcoin on one method. Does Global Liquidity lead Bitcoin? and Does the dollar lead Bitcoin? use the same scan.

Gold’s six-week percentage change in each week is set against Bitcoin’s six-week log change k weeks later, for every k from 0 to 39 weeks, on weekly closes to 23 October 2026, and the cross-correlation at each lead forms a profile. The peak is then tested three ways: a 13-week moving-block bootstrap for the interval of the peak lead, a split into the sample’s first 70% and last 30%, and a circular-shift test that slides Bitcoin’s series against gold’s to count how often chance produces a peak as high. A direction test was pre-specified at 18 weeks, the midpoint of the 100 to 150 days in BlockPhi’s May 2025 version. This is a model estimate as of 28 October 2026, and past results do not indicate future results.

The profile in Figure 5 peaks at 13 weeks, at a correlation of 0.11. The bootstrap interval of the peak spans the whole range, 0 to 39 weeks; the last 30% of the sample peaks at 30 weeks, against 11 in the first 70%; and the shift test puts the chance of a peak this high at 85% (Table 2). At the pre-specified 18 weeks, gold’s six-week direction matched Bitcoin’s in 147 of 295 weeks, 49.8%, against 50.5% expected by chance (binomial p 0.62). On data from 18 February 2022 alone the peak moves to zero weeks. The profile also dips below zero between 16 and 28 weeks, reaching −0.19 at 21 weeks; no pre-specified test covered that trough, and because neighbouring weeks share five of their six weeks of change, it is reported here and not read as a finding. This is a model estimate as of 28 October 2026, and past results do not indicate future results.

Note: Correlation of gold’s six-week change with Bitcoin’s six-week change k weeks later, k from 0 to 39, on weekly data from 4 September 2020 to 23 October 2026 (315 pairs at zero lead), and the partial correlation with Global Liquidity’s six-week change held constant. The gold line peaks at 13 weeks (0.11, shift-test p 0.85); with Global Liquidity held constant the peak is at 0 weeks (0.09, p 0.92). Derived statistics only. Model estimate as of 28 October 2026; past results do not indicate future results.Source: BlockPhi analysis of the BlockPhi Global Liquidity Model, PAX Gold (Binance) and Bitcoin weekly closes (Bitstamp), 28 October 2026.

Open gold and Bitcoin in the terminal’s Chart Desk (opens in a new tab)

On the Global Liquidity Model: Correlation by lead, gold’s six-week change against Bitcoin’s six-week change k weeks later: the full window (4 September 2020 to 23 October 2026), the window from 18 February 2022, and the full window with Global Liquidity’s six-week change, dated 11 weeks before Bitcoin’s window ends, held constant. Model estimate as of 28 October 2026; past results do not indicate future results.

Lead (weeks)	Gold	Gold, from February 2022	Gold, Global Liquidity held constant
0	0.008	0.134	0.089
1	−0.035	0.106	0.025
2	−0.075	0.068	−0.034
3	−0.105	0.036	−0.079
4	−0.133	0.005	−0.122
5	−0.151	−0.012	−0.155
6	−0.141	−0.016	−0.153
7	−0.108	−0.010	−0.132
8	−0.064	0.018	−0.113
9	−0.016	0.037	−0.081
10	0.041	0.076	−0.038
11	0.092	0.101	0.004
12	0.103	0.090	−0.001
13	0.114	0.090	0.021
14	0.096	0.048	0.033
15	0.065	0.014	0.019
16	−0.002	−0.059	−0.038
17	−0.063	−0.133	−0.092
18	−0.101	−0.192	−0.109
19	−0.145	−0.263	−0.148
20	−0.171	−0.297	−0.176
21	−0.191	−0.343	−0.199
22	−0.184	−0.363	−0.181
23	−0.176	−0.362	−0.161
24	−0.155	−0.322	−0.141
25	−0.119	−0.272	−0.113
26	−0.095	−0.236	−0.107
27	−0.047	−0.157	−0.067
28	−0.006	−0.082	−0.037
29	0.057	0.007	0.009
30	0.087	0.052	0.029
31	0.085	0.084	0.029
32	0.069	0.091	0.026
33	0.031	0.047	0.004
34	−0.004	−0.002	−0.019
35	−0.081	−0.098	−0.068
36	−0.144	−0.188	−0.102
37	−0.173	−0.253	−0.110
38	−0.158	−0.278	−0.079
39	−0.155	−0.295	−0.080

6 Does the lead survive once Global Liquidity is held constant?

There is no stable lead to survive, and holding the Global Liquidity Model constant moves what little there is to zero weeks. The partial correlation in Figure 5 removes from gold and from Bitcoin the part associated with Global Liquidity’s six-week change dated 11 weeks before Bitcoin’s window ends; its peak falls to 0 weeks, at 0.09, with a shift-test p of 0.92. This is a model estimate as of 28 October 2026, the model’s weeks before 31 July 2026 are reconstructed, and past results do not indicate future results.

A regression puts the two side by side. Over 297 weeks of gold data from 16 October 2020 to 19 June 2026, Bitcoin’s six-week change 18 weeks later was regressed on gold’s six-week change and on Global Liquidity’s, the latter dated seven weeks after gold’s and so 11 weeks before Bitcoin’s window ends, with Newey-West standard errors. Gold’s t-statistic was −1.24 and Global Liquidity’s 5.50, with standardised coefficients of −0.10 and 0.40. On this sample, gold’s change carried no association with Bitcoin’s beyond what Global Liquidity’s already carried. This is a model estimate as of 28 October 2026, and past results do not indicate future results.

The alignment the control uses is one the terminal draws. On its Bitcoin Overview page, GL (+11W) & BTC 6W Changes, Figure 6, moves Global Liquidity’s six-week change forward 11 weeks against Bitcoin’s six-week change. Across the quarter ends in its table, the two lines agree in direction in most and part in several: in the six weeks to 24 June 2022 Bitcoin fell 27.4% after a Global Liquidity change of −1.35% in the six weeks to 8 April 2022, the same direction at a very different size, while in the six weeks to 29 December 2023 Bitcoin rose 14.9% after a change of −1.62% to 13 October 2023. How often the two agree across every week is what the flagship article measures. For Premium members the Global Liquidity line runs on past the latest Bitcoin week; the capture here stops at the last Bitcoin week, 23 October 2026. This is a model estimate as of 28 October 2026, and past results do not indicate future results.

Note: Global Liquidity’s six-week change moved forward 11 weeks, against Bitcoin’s six-week change, to 23 October 2026; the capture stops at the last Bitcoin week. In a regression of Bitcoin’s six-week change 18 weeks after gold’s on gold’s six-week change and on Global Liquidity’s, the latter dated 11 weeks before Bitcoin’s window ends, 297 weeks of gold data from 16 October 2020 to 19 June 2026, Global Liquidity’s t-statistic was 5.50 and gold’s −1.24. Weeks before 31 July 2026 are reconstructed. Model estimate as of 28 October 2026; past results do not indicate future results.Source: BlockPhi terminal, Bitcoin Overview > GL (+11W) & BTC 6W Changes, captured 29 October 2026.

In the terminal for Premium members·Open this chart in the terminal (opens in a new tab)

The card’s two lines at each quarter’s last week, December 2020 to 23 October 2026: Bitcoin’s six-week change and Global Liquidity’s six-week change 11 weeks earlier, percent, recomputed by BlockPhi from the Global Liquidity Model and Bitstamp closes (the card’s own Bitcoin feed can differ by a fraction of a point). Weeks before 31 July 2026 are reconstructed. Model estimate as of 28 October 2026; past results do not indicate future results.

Bitcoin week	Bitcoin, six-week change	Global Liquidity, six-week change	Global Liquidity week
25 December 2020	+51.2%	+0.71%	9 October 2020
26 March 2021	+16.2%	+1.61%	8 January 2021
25 June 2021	−36.7%	−1.08%	9 April 2021
24 September 2021	−10.5%	−0.39%	9 July 2021
31 December 2021	−20.5%	−0.91%	15 October 2021
25 March 2022	+4.5%	+0.05%	7 January 2022
24 June 2022	−27.4%	−1.35%	8 April 2022
30 September 2022	−6.8%	−3.11%	15 July 2022
30 December 2022	−0.5%	−3.00%	14 October 2022
31 March 2023	+15.9%	+1.38%	13 January 2023
30 June 2023	+13.3%	+2.22%	14 April 2023
29 September 2023	+3.3%	−0.40%	14 July 2023
29 December 2023	+14.9%	−1.62%	13 October 2023
29 March 2024	+34.0%	+0.86%	12 January 2024
28 June 2024	−10.0%	−0.74%	12 April 2024
27 September 2024	+11.7%	−0.77%	12 July 2024
27 December 2024	+3.4%	−0.81%	11 October 2024
28 March 2025	−13.5%	−1.13%	10 January 2025
27 June 2025	+3.5%	+1.93%	11 April 2025
26 September 2025	−6.6%	+0.79%	11 July 2025
26 December 2025	−7.6%	+0.41%	10 October 2025
27 March 2026	−3.6%	+1.06%	9 January 2026
26 June 2026	−24.1%	−0.67%	10 April 2026
25 September 2026	+33.5%	−0.67%	10 July 2026
23 October 2026	+5.0%	+0.09%	7 August 2026

A free account sees the same 11-week shift on the Lead & Forecast page. GL vs BTC Price, Figure 7, sets the Global Liquidity Model’s level against Bitcoin’s weekly price, with its lag slider at 11 weeks; the free line ends 11 weeks before the latest week, so moved forward it stops at the last Bitcoin week, 23 October 2026. Two rising series look aligned over a wide range of shifts, so the slider is a way of reading the lead, not a test of it; the tests are the ones above. This is a model estimate as of 28 October 2026, the model’s weeks before 31 July 2026 are reconstructed, and past results do not indicate future results.

Note: On the Lead & Forecast page, GL vs BTC Price with its lag slider at 11 weeks: the Global Liquidity Model’s level moved forward 11 weeks against Bitcoin’s weekly price, as a free account sees it, so the shifted line ends at the last Bitcoin week, 23 October 2026. Bitcoin closed that week at $81,076, against Global Liquidity of $194.89 trillion on 7 August 2026. Two rising series look aligned over a wide range of shifts, so the slider is a way of reading the lead, not a test of it. Weeks before 31 July 2026 are reconstructed. Model estimate as of 28 October 2026; past results do not indicate future results.Source: BlockPhi terminal, Lead & Forecast > GL vs BTC Price, captured 28 October 2026.

Free with an account·Open this chart in the terminal (opens in a new tab)

Bitcoin’s weekly close and the Global Liquidity Model 11 weeks earlier, at the last Bitcoin week of each quarter, 25 March 2022 to 23 October 2026, from the terminal’s Lead & Forecast series as published on 23 October 2026 (its own Bitcoin feed). Weeks before 31 July 2026 are reconstructed. Model estimate as of 28 October 2026; past results do not indicate future results.

Bitcoin week	Bitcoin close (US$)	Global Liquidity week (11 weeks earlier)	Global Liquidity (US$ trillion)
25 March 2022	44,321	7 January 2022	167.40
24 June 2022	21,215	8 April 2022	166.86
30 September 2022	19,425	15 July 2022	160.50
30 December 2022	16,603	14 October 2022	155.93
31 March 2023	28,476	13 January 2023	163.63
30 June 2023	30,469	14 April 2023	166.09
29 September 2023	26,911	14 July 2023	162.62
29 December 2023	42,072	13 October 2023	160.99
29 March 2024	69,903	12 January 2024	167.33
28 June 2024	60,315	12 April 2024	167.47
27 September 2024	65,781	12 July 2024	167.56
27 December 2024	94,168	11 October 2024	174.08
28 March 2025	84,379	10 January 2025	170.15
27 June 2025	107,100	11 April 2025	178.39
26 September 2025	109,696	11 July 2025	184.75
26 December 2025	87,296	10 October 2025	187.66
27 March 2026	66,359	9 January 2026	188.33
26 June 2026	60,019	10 April 2026	192.87
25 September 2026	84,090	10 July 2026	194.31
23 October 2026	81,076	7 August 2026	194.89

7 Has the link between gold and Bitcoin changed over time?

It has drifted, but inside a narrow band around zero. The 52-week rolling correlation of weekly returns in Figure 8 ranged from −0.14 on 24 October 2025 to 0.32 on 10 November 2023, and stood at 0.19 on 23 October 2026; across the whole sample the same-week correlation was 0.10. In no 52-week window of this sample did it reach the +0.5 or more that the May 2025 version described for periods of stress. PAX Gold has traded only since August 2020, so this run cannot test claims about the years before. Past results do not indicate future results.

Note: Correlation of gold’s and Bitcoin’s weekly log returns over the trailing 52 weeks, same week, from 3 September 2021 to 23 October 2026. It ranged from −0.14 on 24 October 2025 to 0.32 on 10 November 2023 and was 0.19 on 23 October 2026. Derived statistics only. Past results do not indicate future results.Source: BlockPhi analysis of PAX Gold (Binance) and Bitcoin weekly closes (Bitstamp), 28 October 2026.

Open gold and Bitcoin in the terminal’s Chart Desk (opens in a new tab)

The 52-week rolling correlation of weekly log returns at each quarter’s last week, September 2021 to 23 October 2026; its low was −0.14 on 24 October 2025 and its high 0.32 on 10 November 2023. Past results do not indicate future results.

Week	52-week correlation
24 September 2021	0.07
31 December 2021	−0.05
25 March 2022	0.01
24 June 2022	0.24
30 September 2022	0.27
30 December 2022	0.08
31 March 2023	0.29
30 June 2023	0.16
29 September 2023	0.11
29 December 2023	0.24
29 March 2024	0.04
28 June 2024	0.04
27 September 2024	−0.01
27 December 2024	0.01
28 March 2025	−0.01
27 June 2025	−0.01
26 September 2025	−0.06
26 December 2025	0.01
27 March 2026	−0.10
26 June 2026	0.03
25 September 2026	0.15
23 October 2026	0.19

Gold moving first, then, is a pattern the eye finds in particular stretches, not one this sample holds across its six years.