Does the dollar lead Bitcoin?

Green, Jack

BlockPhi Insights No. 2

Study29 October 2026Version 1

Does the dollar lead Bitcoin?

Jack Green, founder

BlockPhi

Abstract

The dollar’s six-week direction, inverted, matched Bitcoin’s 13 weeks later in 228 of 343 weeks from January 2020 (66%, against about 50% by chance), on a basket BlockPhi builds from Federal Reserve exchange rates. A lag scan of the same weekly data, to 23 October 2026, peaks near 12 weeks, but a block bootstrap places it anywhere from 0 to 26 weeks. This is a model estimate as of 28 October 2026, and past results do not indicate future results.

Key findings

Direction. The dollar’s six-week direction, inverted, matched Bitcoin’s 13 weeks later in 228 of 343 weeks from January 2020 (66%, against about 50% by chance), on a basket BlockPhi builds from Federal Reserve exchange rates (model estimates as of 28 October 2026).
Section 4 Table 3
Lead. A lag scan of the same weekly data, to 23 October 2026, peaks near 12 weeks, but a block bootstrap places it anywhere from 0 to 26 weeks (model estimates as of 28 October 2026).
Section 4 Figure 2 Table 1 Table 2
Translation. The weekly translation term accounts for 49% of the variance of Global Liquidity’s weekly change, because the model counts every component in US dollars (model estimates as of 28 October 2026).
Section 5 Figure 3 Table 4
Beyond Global Liquidity. With Global Liquidity as published held constant, the dollar’s coefficient is not statistically significant (t 1.66), while Global Liquidity’s is (t 4.94; model estimates as of 28 October 2026).
Section 6 Table 5

Model estimates as of 28 October 2026; past results do not indicate future results.

Data to: 23 October 2026
Measures: BlockPhi Global Liquidity Model; third-party data, as each figure credits it
Exhibits: 5 figures, 5 tables; 4 figures open in the terminal, 1 with a free account
JEL: C32, E44, F31, G15
Keywords: US dollar, DXY, Bitcoin, Global Liquidity, currency translation, lead-lag

The terminal’s Macro Regime card on the Bitcoin Overview page: Bitcoin’s z-score line against the inverted dollar basket line, 90 days earlier, inside a shaded band two standard deviations wide on each side, daily from 1 January 2017 to 23 October 2026. Bitcoin stays inside the band through 2022, sits above it for much of December 2023 to July 2024 and of January to June 2025, and below it for much of November 2025 to July 2026. — Figure 1 Bitcoin sat inside the band around the inverted dollar on about three days in four

1 What does the inverted dollar chart in the terminal show?

It shows Bitcoin against the dollar turned upside down and shifted a quarter forward, so a weaker dollar draws a rising line. The card, Macro Regime: BTC/USD vs Inverted Dollar Basket, sits on the terminal’s Bitcoin Overview page and opens with a free account (Figure 1, above).

Bitcoin’s price, in logarithms, is expressed as a z-score against its own trailing 365 days; the dollar basket is inverted, shifted 90 days forward and scored the same way; a band two standard deviations either side of the dollar line marks where Bitcoin is consistent with the dollar a quarter earlier. From 1 January 2017 to 23 October 2026, Bitcoin sat inside the band on 2,632 of 3,583 days (73%), above it on 817 and below it on 134, as the card’s own rows read on 29 October 2026. The band is four standard deviations wide, so the chart says more about direction than about distance. Model estimates; past results do not indicate future results.

The chart is a lens; the sections below test it on the same basket.

2 What is the DXY, and how does BlockPhi measure the dollar?

The DXY is ICE’s US Dollar Index: the dollar against six currencies, with the euro carrying more than half of the weight (glossary). BlockPhi does not redraw ICE’s series: the terminal rebuilds a basket to the same formula and weights from the Federal Reserve’s noon exchange rates, and that basket drives the Macro Regime card and every test here.

The Chart Desk carries a different measure, the Federal Reserve’s trade-weighted broad dollar index (FRED DTWEXBGS, listed there as Dollar (broad)), which a reader can set against weekly Bitcoin there, as quoted or on an inverted scale.

Quarter by quarter, the same-week link is loose. In the quarter to 30 June 2022 the broad dollar rose 5.0% and Bitcoin fell 56.1%, but in the quarter to 31 December 2024 both rose, by 6.6% and 47.5%. Federal Reserve and Coinbase data to 23 October 2026; past results do not indicate future results.

3 Why would a stronger dollar tighten global funding?

Because much of the world borrows in dollars: a stronger dollar raises the burden of that debt and lowers the value of the collateral behind it, so lending capacity shrinks. Bruno and Shin (2015) link a weaker dollar to faster cross-border bank lending, and Avdjiev, Du, Koch and Shin (2019) link a stronger one to costlier dollar funding and slower lending in dollars.

The Rate gaps with the US card on the terminal’s Global Rates page shows one reason the dollar was strong from 2022: the US 2-year yield over Japan’s widened from 22 basis points in February 2021 to 513 in October 2023, then narrowed to 228 by the end of 2025. Wide gaps pay investors to hold dollars. Data to 23 October 2026; past results do not indicate future results.

4 Does the dollar lead Bitcoin, and by how long?

On six-week changes, a weaker dollar was associated with a stronger Bitcoin about a quarter later, though the lead is a range, not a fixed delay.

A lag scan by the terminal’s method for Global Liquidity, a cross-correlation of the inverted basket’s six-week change with Bitcoin’s six-week change 0 to 26 weeks later, peaks at 12 weeks (0.36) on weekly data from 3 January 2020 to 23 October 2026, at 14 weeks from 2017 and at 11 from 2022 (Figure 2, Tables 1 and 2). None of 304 circular shifts of Bitcoin’s series reached the peak’s height (p below 0.01), but in every window a block bootstrap places the peak anywhere from 0 to 26 weeks, and the method’s own stability check names no single lag. Model estimates as of 28 October 2026; past results do not indicate future results.

A line chart of correlation against lead, 0 to 26 weeks, one line for each start date. Each line dips to a first low between 4 and 6 weeks, rises to a peak between 11 and 14 weeks, of 0.31 to 0.41, and falls back; the line from February 2022 turns negative from 21 weeks and ends at minus 0.27 at 26 weeks. — Figure 2 The inverted dollar’s correlation peaks at 11 to 14 weeks by window, though no single lag is fixed

Table 1 The inverted dollar’s six-week lead profile, by window Correlations, by lead in weeks

Lead, weeks	From 6 January 2017	From 3 January 2020	From 18 February 2022
0	0.16	0.17	0.15
2	0.12	0.10	0.10
4	0.11	0.06	0.03
6	0.06	0.07	0.08
8	0.12	0.21	0.23
10	0.20	0.32	0.38
11	0.25	0.35	0.41
12	0.29	0.36	0.40
13	0.30	0.32	0.36
14	0.31	0.31	0.35
16	0.26	0.22	0.24
18	0.20	0.15	0.16
20	0.13	0.10	0.05
22	0.09	0.06	−0.09
24	0.04	−0.01	−0.22
26	0.03	−0.04	−0.27

Note: Correlation of the inverted dollar basket’s six-week change with Bitcoin’s six-week change the given number of weeks later, weekly data to 23 October 2026. Model estimates as of 28 October 2026; past results do not indicate future results.

Table 2 The lag scan, by window Leads and intervals in weeks; correlations; p from a circular-shift test of the maximum over 27 leads

Window	Weeks	Peak lead, weeks	Correlation at the peak	Correlation at 13 weeks	Interval of the peak, weeks	Peak, first 70%	Peak, last 30%	Circular-shift p
From 6 January 2017	512	14	0.31	0.30	0 to 26	14	12	below 0.01 (none of 460 shifts)
From 3 January 2020	356	12	0.36	0.32	0 to 26	11	14	below 0.01 (none of 304 shifts)
From 18 February 2022	245	11	0.41	0.36	0 to 26	11	3	below 0.01 (none of 193 shifts)

Note: The peak lead and its correlation, the correlation at the registered 13-week lead, the 95% block-bootstrap interval of the peak, and the peaks in the first 70% and the last 30% of the weeks, weekly data to 23 October 2026. Model estimates as of 28 October 2026; past results do not indicate future results.

The direction test is simpler. For dollar weeks from 3 January 2020 to 24 July 2026, against Bitcoin to 23 October 2026, the inverted dollar’s six-week change pointed the same way as Bitcoin’s 13 weeks later in 228 of 343 weeks (66.5%, against 49.5% by chance); from 2017, in 322 of 499 (64.5%); from 2022, in 160 of 232 (69.0%) (Table 3). The 13-week lead, the terminal’s 90-day shift in weeks, was fixed before the data were fetched. The six-week windows overlap, and a circular-shift test still puts each result below a 1% chance. Model estimates as of 28 October 2026; past results do not indicate future results.

Table 3 Direction tests: the inverted dollar and Global Liquidity against Bitcoin Leads in weeks; rates in per cent; p from a binomial and a circular-shift test

Leader	Lead, weeks	Leader weeks	Matches	Match rate	Chance rate	Binomial p	Circular-shift p
Inverted dollar, from 3 January 2020	13	3 January 2020 to 24 July 2026	228 of 343	66.5%	49.5%	below 0.0001	below 0.01
Inverted dollar, from 6 January 2017	13	6 January 2017 to 24 July 2026	322 of 499	64.5%	49.8%	below 0.0001	below 0.01
Inverted dollar, from 18 February 2022	13	18 February 2022 to 24 July 2026	160 of 232	69.0%	49.6%	below 0.0001	below 0.01
Global Liquidity as published	11	3 January 2020 to 7 August 2026	224 of 345	64.9%	51.6%	below 0.0001	0.003
Global Liquidity net of FX translation	11	14 February 2020 to 7 August 2026	182 of 339	53.7%	53.2%	0.45	0.47

Note: The sign of the leader’s six-week change against the sign of Bitcoin’s six-week change at the lead shown, leader weeks as listed, Bitcoin to 23 October 2026. The Global Liquidity rows read the BlockPhi Global Liquidity Model, whose weeks before 31 July 2026 are reconstructed. Model estimates as of 28 October 2026; past results do not indicate future results.

Week to week, the lead fades: the largest forward correlation of weekly changes, 0.12 at 11 weeks, does not clear a circular-shift test (p 0.16), and the same-week correlation is larger (0.19). Weekly data from 3 January 2020 to 23 October 2026; model estimates; past results do not indicate future results.

5 How much of Global Liquidity’s move is the dollar?

About half of its weekly variation: the BlockPhi Global Liquidity Model counts every component in US dollars, so currency moves change its level mechanically.

That is a translation effect, not a change in any balance sheet. The model is a composite of five central bank balance sheets and 17 government debt markets, revalued every Friday in US dollars and multiplied by a collateral multiplier (methodology), so a euro of debt counts for more dollars when the euro rises (coverage). Any global aggregate stated in dollars carries the same effect, global M2 included (Global Liquidity and M2); the terminal splits it out (what moves Global Liquidity).

The Global Liquidity page’s Dollar translation effect card adds up what currency moves alone have done to the collateral base since 3 January 2020: plus $3.48 trillion by the week of 8 January 2021, minus $9.83 trillion by the week of 17 January 2025 and minus $5.21 trillion on 23 October 2026, or minus $8.93 trillion of Global Liquidity once carried through the collateral multiplier (Figure 3). The weekly translation term accounts for 49% of the variance of Global Liquidity’s weekly change, and in 2022 it took $9.48 trillion off a total that fell $5.51 trillion (Table 4). Model estimates; past results do not indicate future results.

A shaded line of cumulative FX translation on every non-dollar collateral base component since 3 January 2020, in US$ trillions, weekly to 23 October 2026: up to plus $3.48 trillion in January 2021, down to minus $9.83 trillion in January 2025, and minus $5.21 trillion on 23 October 2026. — Figure 3 Currency moves alone added to the collateral base into 2021 and took from it into 2025

Table 4 FX translation and the change in Global Liquidity, by calendar year US$ trillions at Global Liquidity scale

Year	FX translation, US$ trillions	Change in Global Liquidity, US$ trillions
2020	+5.65	+31.66
2021	−5.76	+12.85
2022	−9.48	−5.51
2023	−0.24	+6.35
2024	−6.22	+2.17
2025	+6.58	+16.48
2026, to 23 October	+0.54	+11.88

Note: By calendar year from 3 January 2020 to 23 October 2026 (2026 to date). Weeks before 31 July 2026 are reconstructed. Model estimates as of 28 October 2026; past results do not indicate future results.

The explorer’s Drivers view splits a recent window the same way. Of Global Liquidity’s rise of $4.11 trillion in the 11 weeks to 23 October 2026, balance sheets and debt gave plus $3.05 trillion, the collateral multiplier plus $1.97 trillion, FX translation minus $0.59 trillion and bond prices minus $0.33 trillion (Figure 4). The Weekly change in the collateral base card shows it a Friday at a time: in the week of 25 September 2026 the base fell $532 billion, and by the Drivers split currency moves alone took $613 billion off it (Figure 5). Model estimates; past results do not indicate future results. BlockPhi’s Global Liquidity page hangs the translation card too, frozen at an earlier week.

The terminal’s Global Collateral Base explorer with all components in view: on the left the 22 components ranked by their change over the 11 weeks to 23 October 2026, China’s government debt first at plus $699 billion and the European Central Bank last at minus $178 billion; on the right the Drivers bars, balance sheets and debt plus $3.05 trillion, FX translation minus $0.59 trillion, bond prices minus $0.33 trillion and the collateral multiplier plus $1.97 trillion, summing to plus $4.11 trillion of Global Liquidity. — Figure 4 Over the 11 weeks, balance sheets, debt and the multiplier added; currency and bond prices took a little off

Bars of the weekly change in the collateral base in US$ billions, rises in green and falls in red, for the 26 weeks from 1 May to 23 October 2026, from minus $829 billion in the week of 2 October 2026 to plus $1,861 billion in the week of 18 September 2026. — Figure 5 In the week of 25 September 2026 currency moves took more off the base than its whole fall

6 Does the dollar add anything beyond Global Liquidity?

Beyond Global Liquidity as published, little; beyond its non-dollar part, yes.

Because the model counts in dollars, its six-week change and the inverted dollar’s correlate at 0.61 on weeks from 3 January 2020, and at 0.10 once translation is removed. With Global Liquidity net of translation held constant, the dollar stays associated with Bitcoin’s six-week change 13 weeks later (Newey-West t 4.41), and so does the non-dollar part (t 2.55). With Global Liquidity as published held constant, the dollar’s coefficient is not statistically significant (t 1.66), while Global Liquidity’s is (t 4.94) (Table 5). Weeks to 24 July 2026 against Bitcoin to 23 October 2026; model estimates as of 28 October 2026; past results do not indicate future results.

Table 5 The dollar and Bitcoin 13 weeks later, alone and with Global Liquidity held constant Weeks of data; t from Newey-West standard errors with 12 lags

Model	Regressor	Weeks	Coefficient	Newey-West standard error	t	Standardised coefficient	R-squared
Dollar alone	Inverted dollar, six-week change	343	3.10	0.74	4.20	0.32	0.11
Dollar, Global Liquidity net of FX translation held constant	Inverted dollar, six-week change	339	3.07	0.70	4.41	0.33	0.15
Dollar, Global Liquidity net of FX translation held constant	Global Liquidity net of FX translation, six-week change ending 11 weeks before Bitcoin’s	339	3.16	1.24	2.55	0.18	0.15
Dollar, Global Liquidity as published held constant	Inverted dollar, six-week change	343	1.23	0.74	1.66	0.13	0.18
Dollar, Global Liquidity as published held constant	Global Liquidity, six-week change ending 11 weeks before Bitcoin’s	343	3.81	0.77	4.94	0.33	0.18

Note: Bitcoin’s six-week change 13 weeks ahead regressed on the inverted dollar’s six-week change, alone and with Global Liquidity’s six-week change held constant, Newey-West standard errors with 12 lags, leader weeks from 3 January 2020 (the net-of-translation rows from 31 January 2020) to 24 July 2026 against Bitcoin to 23 October 2026. Model estimates as of 28 October 2026; past results do not indicate future results.

The same split bears on Global Liquidity’s own lead. Its six-week direction, advanced 11 weeks, matched Bitcoin’s in 224 of 345 weeks from 3 January 2020 (64.9%, against 51.6% by chance); with translation removed it matched in 182 of 339 (53.7%, against 53.2%, binomial p 0.45). The reading registered before the run, that Global Liquidity still leads with the dollar held constant, therefore fails on the direction test, although a lag scan of the non-dollar part still peaks, at 7 weeks (correlation 0.26, circular-shift p 0.007). Liquidity weeks to 7 August 2026 against Bitcoin to 23 October 2026; model estimates as of 28 October 2026; past results do not indicate future results.

The terminal keeps the three apart: Global Liquidity on its own page, the Drivers split in the explorer, the dollar on Bitcoin Overview. Global Liquidity’s own lead is tested in Does Global Liquidity lead Bitcoin?

7 Does the link hold through crypto-specific shocks?

Not always: the chart held through the crypto failures of 2022 but came apart after October 2025.

The dollar basket rose to a high on 26 September 2022, and the band, carrying that rise 90 days on, sat low through 2022. Bitcoin stayed inside it through the collapse of the Terra stablecoin in May and of the FTX exchange in November; its deepest point, on 18 June 2022, was 3.51 standard deviations below its trailing mean, against a band edge of 3.94. From December 2023 to July 2024, around the launch of US spot Bitcoin exchange-traded funds, it sat above the band on 133 of 244 days. Model estimates; past results do not indicate future results.

After 10 October 2025 the two came apart. The basket had fallen about 12% from 13 January to 16 September 2025, lifting the band into the autumn, yet Bitcoin fell: from 14 November 2025 it dropped below the band, and from November 2025 to July 2026 it sat below on 128 of 273 days, reaching 3.63 standard deviations below its trailing mean on 5 February 2026. Model estimates; past results do not indicate future results. The funding conditions into that week are read in What caused the 10 October 2025 crypto liquidation?

References

Ang, B. W. (2005). “The LMDI approach to decomposition analysis: a practical guide.” Energy Policy 33(7). doi.org/10.1016/j.enpol.2003.10.010 (opens in a new tab)
Avdjiev, S., Du, W., Koch, C. and Shin, H. S. (2019). “The dollar, bank leverage, and deviations from covered interest parity.” American Economic Review: Insights 1(2). doi.org/10.1257/aeri.20180322 (opens in a new tab)
Bruno, V. and Shin, H. S. (2015). “Cross-border banking and global liquidity.” The Review of Economic Studies 82(2). doi.org/10.1093/restud/rdu042 (opens in a new tab)
Künsch, H. R. (1989). “The jackknife and the bootstrap for general stationary observations.” The Annals of Statistics 17(3). doi.org/10.1214/aos/1176347265 (opens in a new tab)
Newey, W. K. and West, K. D. (1987). “A simple, positive semi-definite, heteroskedasticity and autocorrelation consistent covariance matrix.” Econometrica 55(3). doi.org/10.2307/1913610 (opens in a new tab)
Seabold, S. and Perktold, J. (2010). “Statsmodels: econometric and statistical modeling with Python.” Proceedings of the 9th Python in Science Conference. doi.org/10.25080/Majora-92bf1922-011 (opens in a new tab)

Appendix A

How was it estimated?

On the terminal’s own dollar basket and by its lag-scan method, with the tests, the leads and the four readings they were judged on written down and hashed before the data were fetched on 28 October 2026; the chart readings above are descriptive.

The basket follows the DXY’s geometric weights (euro 57.6%, yen 13.6%, pound 11.9%, Canadian dollar 9.1%, krona 4.2%, Swiss franc 3.6%) on the Federal Reserve’s H.10 noon rates, as the terminal builds it for its Macro Regime card, taking the last print of each week to Friday; Bitcoin is Bitstamp’s daily close on the same Fridays, both to 23 October 2026. Changes are six-week; the interval of the peak comes from a 13-week moving-block bootstrap of 2,000 draws (Künsch 1989), and every result is tested against a circular-shift null because the windows overlap. Model estimates as of 28 October 2026; past results do not indicate future results.

Translation is the currency term of a log-mean Divisia split of each component’s weekly change into quantity, exchange rate and bond price (Ang 2005), the terminal’s Drivers method, summed over the non-dollar components and carried to Global Liquidity with the prior week’s multiplier. The regressions use Newey-West standard errors with 12 lags (Newey and West 1987), computed in statsmodels (Seabold and Perktold 2010). Weeks of the model before 31 July 2026 are reconstructed (the note under Figure 3). Model estimates as of 28 October 2026; past results do not indicate future results.

Appendix B

Data and charts

The charts in this article, live in the terminal: every figure here is a chart in the BlockPhi terminal or a BlockPhi calculation on a series the terminal draws, and Table B1 says where each one lives.

Table B1 Where each figure lives in the terminal

Figure	In the terminal	Access	Open
1	Bitcoin Overview > Macro Regime: BTC/USD vs Inverted Dollar Basket	Free with an account	Open Figure 1 in the terminal (opens in a new tab)
2	BlockPhi analysis of Federal Reserve H.10 exchange rates (the terminal’s dollar basket) and Bitstamp Bitcoin closes, by the lag-scan method the terminal uses for Global Liquidity	BlockPhi calculation	Open Figure 2 (Open the dollar and Bitcoin in the terminal’s Bitcoin Overview) (opens in a new tab)
3	Global Liquidity > Dollar translation effect	In the terminal for Premium members	Open Figure 3 in the terminal (opens in a new tab)
4	Global Liquidity > Global Collateral Base explorer (Drivers, 11W)	In the terminal for Premium members	Open Figure 4 in the terminal (opens in a new tab)
5	Global Liquidity > Weekly change in the collateral base	In the terminal for Premium members	Open Figure 5 in the terminal (opens in a new tab)

Federal Reserve Board, H.10 foreign exchange rates: the six noon rates of the dollar basket (FRED DEXUSEU, DEXJPUS, DEXUSUK, DEXCAUS, DEXSDUS, DEXSZUS)28 October 2026
Federal Reserve Board, nominal broad US dollar index (FRED DTWEXBGS)28 October 2026
Federal Reserve Board, H.15 selected interest rates: the 2-year Treasury yield (FRED DGS2)28 October 2026
European Central Bank, euro area AAA government bond yield curve, 2-year spot rate28 October 2026
Ministry of Finance, Japan, JGB interest rates, 2-year28 October 2026
Bitstamp, BTC/USD daily close (the tests and the Macro Regime card)28 October 2026
Coinbase Exchange, BTC-USD (the Chart Desk dashboard)28 October 2026
ICE Futures U.S., US Dollar Index FAQ: the six currencies and their weights28 October 2026
BlockPhi Global Liquidity Model: methodology23 October 2026

Author

Jack Green is the founder of BlockPhi.

Versions

Version 129 October 2026 First published. An earlier version, on different data, appeared on Medium on 26 December 2025.

Date	Bitcoin, 365-day log z-score	Inverted dollar basket, 90 days earlier, z-score	Band, lower edge	Band, upper edge
31 March 2017	1.51	−2.25	−4.25	−0.25
30 June 2017	2.10	−0.87	−2.87	1.13
30 September 2017	1.73	1.28	−0.72	3.28
31 December 2017	1.93	1.40	−0.60	3.40
31 March 2018	0.48	1.33	−0.67	3.33
30 June 2018	−0.14	1.19	−0.81	3.19
30 September 2018	−0.71	−1.27	−3.27	0.73
31 December 2018	−2.06	−1.24	−3.24	0.76
31 March 2019	−0.97	−0.98	−2.98	1.02
30 June 2019	2.11	−1.13	−3.13	0.87
30 September 2019	0.73	−0.42	−2.42	1.58
31 December 2019	0.12	−2.23	−4.23	−0.23
31 March 2020	−1.13	1.19	−0.81	3.19
30 June 2020	0.29	−1.78	−3.78	0.22
30 September 2020	1.18	0.81	−1.19	2.81
31 December 2020	3.08	1.48	−0.52	3.48
31 March 2021	1.98	1.86	−0.14	3.86
30 June 2021	0.56	0.33	−1.67	2.33
30 September 2021	0.52	−0.11	−2.11	1.89
31 December 2021	−0.02	−1.91	−3.91	0.09
31 March 2022	−0.01	−1.66	−3.66	0.34
30 June 2022	−2.84	−1.91	−3.91	0.09
30 September 2022	−1.53	−2.14	−4.14	−0.14
31 December 2022	−1.32	−2.08	−4.08	−0.08
31 March 2023	0.86	0.09	−1.91	2.09
30 June 2023	1.72	0.89	−1.11	2.89
30 September 2023	0.61	0.79	−1.21	2.79
31 December 2023	2.02	−0.87	−2.87	1.13
31 March 2024	2.46	1.31	−0.69	3.31
30 June 2024	1.05	−1.00	−3.00	1.00
30 September 2024	0.70	−1.02	−3.02	0.98
31 December 2024	1.72	1.39	−0.61	3.39
31 March 2025	0.52	−2.41	−4.41	−0.41
30 June 2025	1.21	0.33	−1.67	2.33
30 September 2025	1.02	2.07	0.07	4.07
31 December 2025	−1.22	1.12	−0.88	3.12
31 March 2026	−1.88	0.68	−1.32	2.68
30 June 2026	−1.89	−0.76	−2.76	1.24
30 September 2026	0.35	−2.13	−4.13	−0.13
23 October 2026	0.35	−2.29	−4.29	−0.29

Week	Cumulative FX translation, US$ trillions
27 March 2020	−2.15
26 June 2020	−0.47
25 September 2020	+1.76
25 December 2020	+3.38
26 March 2021	+2.03
25 June 2021	+1.91
24 September 2021	+1.51
31 December 2021	−0.01
25 March 2022	−1.05
24 June 2022	−5.37
30 September 2022	−9.09
30 December 2022	−5.51
31 March 2023	−4.75
30 June 2023	−6.12
29 September 2023	−7.43
29 December 2023	−5.66
29 March 2024	−7.19
28 June 2024	−8.35
27 September 2024	−5.46
27 December 2024	−9.08
28 March 2025	−7.51
27 June 2025	−5.31
26 September 2025	−4.77
26 December 2025	−5.50
27 March 2026	−5.43
26 June 2026	−5.38
25 September 2026	−4.68
23 October 2026	−5.21

Pane	Row	Change over the 11 weeks, US$ billions	Counted on
Drivers	Balance sheets & debt	+3,053	Global Liquidity
Drivers	FX translation	−587	Global Liquidity
Drivers	Bond prices	−327	Global Liquidity
Drivers	Collateral multiplier	+1,967	Global Liquidity
Drivers	Change in Global Liquidity	+4,107	Global Liquidity
Movers	China	+699	Collateral base
Movers	People’s Bank of China	+290	Collateral base
Movers	United States	+263	Collateral base
Movers	South Korea	+81	Collateral base
Movers	Bank of Japan	+62	Collateral base
Movers	Japan	+37	Collateral base
Movers	Australia	+13	Collateral base
Movers	Singapore	+9	Collateral base
Movers	Hong Kong	+9	Collateral base
Movers	Malaysia	+5	Collateral base
Movers	Federal Reserve	+5	Collateral base
Movers	Bank of England	+3	Collateral base
Movers	Canada	+2	Collateral base
Movers	Norway	+1	Collateral base
Movers	Brazil	0	Collateral base
Movers	Switzerland	0	Collateral base
Movers	Denmark	−3	Collateral base
Movers	South Africa	−11	Collateral base
Movers	Mexico	−16	Collateral base
Movers	Euro area	−35	Collateral base
Movers	United Kingdom	−94	Collateral base
Movers	European Central Bank	−178	Collateral base

Week	Change in the collateral base, US$ billions	Of which FX translation, US$ billions
1 May 2026	−702	−353
8 May 2026	+169	+351
15 May 2026	+115	+172
22 May 2026	+36	−680
29 May 2026	+273	+8
5 June 2026	+284	+249
12 June 2026	−772	−507
19 June 2026	+433	+168
26 June 2026	−536	−450
3 July 2026	+64	−253
10 July 2026	−241	+200
17 July 2026	−219	+79
24 July 2026	−60	−47
31 July 2026	−453	−178
7 August 2026	+615	+694
14 August 2026	+346	+251
21 August 2026	−157	−86
28 August 2026	+418	+405
4 September 2026	−472	−334
11 September 2026	+474	+437
18 September 2026	+1,861	+150
25 September 2026	−532	−613
2 October 2026	−829	−271
9 October 2026	−361	−354
16 October 2026	+122	−20
23 October 2026	+272	+114

Does the dollar lead Bitcoin?

Abstract

Key findings

1 What does the inverted dollar chart in the terminal show?

2 What is the DXY, and how does BlockPhi measure the dollar?

3 Why would a stronger dollar tighten global funding?

4 Does the dollar lead Bitcoin, and by how long?

5 How much of Global Liquidity’s move is the dollar?

6 Does the dollar add anything beyond Global Liquidity?

7 Does the link hold through crypto-specific shocks?

References

Appendix A

How was it estimated?

Appendix B

Data and charts

Author

Versions

Also in BlockPhi Insights

Does Global Liquidity lead Bitcoin?

Does gold lead Bitcoin?

What caused the 10 October 2025 crypto crash?

The model, live in the terminal.