Dollar Index vs Gold

The ICE US Dollar Index (DXY) measures the dollar against a basket of six currencies with fixed weights set in 1973: Euro 57.6%, Japanese Yen 13.6%, British Pound 11.9%, Canadian Dollar 9.1%, Swedish Krona 4.2%, and Swiss Franc 3.6%. It is the most-quoted dollar index in financial media and the underlying benchmark for most dollar-linked futures and options contracts.

The Federal Reserve publishes a different dollar index called DTWEXBGS (the Broad Trade-Weighted Dollar Index), which covers 26 currencies weighted by US goods and services trade. DTWEXBGS includes the Chinese yuan, Mexican peso, and other emerging-market currencies absent from DXY, making it a more economically meaningful measure but a less-quoted one. This page uses DTWEXBGS because it captures the dollar's value against actual trading partners, not just a mid-20th-century basket. DXY and DTWEXBGS are highly correlated (above 0.9 typically) but can diverge during emerging-market stress or when China's yuan policy matters.

Gold is priced in dollars globally. When the dollar strengthens, a given dollar price of gold buys more in other currencies, which reduces foreign demand and puts downward pressure on the dollar price of gold. When the dollar weakens, the reverse happens. This is the mechanical reason most observers expect a negative correlation between DXY and gold.

The relationship has been real and persistent over long periods. Rolling 60-day correlations typically run near negative 0.45, and rolling 12-month correlations generally fall in the negative 0.4 to negative 0.6 range. The correlation is strongest when the dollar is the dominant macro driver, typically during Fed tightening cycles, flight-to-quality episodes, or when US growth diverges sharply from the rest of the world.

The gold-dollar correlation has moved through distinct regimes. In 2014-2016, a strong dollar cycle (DXY rallied from 80 to 103) compressed gold from $1,900 toward $1,050, producing a tight inverse correlation around negative 0.7. In 2020-2021, a weakening dollar (DXY from 103 to 89) helped lift gold from $1,500 to $2,075.

The correlation broke most visibly twice since 2020. From mid-2022 through 2023, the Fed hiked rates aggressively and DXY surged from 96 to above 114, yet gold held the $1,600-$1,800 range rather than collapsing as the textbook model predicted. Then in 2023-2024 both assets rallied together: gold moved from $1,800 to $2,700 while DXY held near 103-107. The common driver was central bank buying, which was largely price-insensitive and operated outside the normal dollar-channel mechanics.

The most important macro story in this pair over the past three years is the central bank gold bid. Following the February 2022 freezing of roughly $300 billion of Russia's dollar-denominated reserves by G7 nations, many non-Western central banks concluded that gold reserves were structurally more attractive than Treasury holdings because gold cannot be frozen by sanctions.

World Gold Council data shows central banks bought 1,082 tonnes in 2022 (a modern record), 1,037 tonnes in 2023 (second-largest year on record), and continued heavy buying through 2024 and 2025. China's PBOC alone added 225 tonnes in 2023, its largest single-year increase since 1977. This bid has been concentrated in China, Turkey, India, Poland, and Singapore. Because central bank buying does not respond to dollar strength the way private investor demand does, it essentially disconnected gold from its normal dollar-channel behavior during the 2022-2024 window.

Beneath the dollar-gold mechanical correlation sits a more fundamental relationship with real yields. Gold pays no interest, so its opportunity cost rises when real (inflation-adjusted) yields rise. Historically gold has had a rolling correlation near negative 0.7 with the 10-year TIPS yield, stronger than its correlation with DXY.

Since 2022, the gold-real yield relationship has partially decoupled for the same reason the gold-dollar relationship has: central bank buying. Through 2022 and into 2023, real yields rose from negative territory to above positive 2%, which would ordinarily have crushed gold. Instead gold rose. The structural bid overwhelmed the real-yield channel. This is an important warning that gold's single most reliable historical driver (real yields) can be dominated by flows when those flows are large enough.

The 2025 dollar decline was the most severe in over two decades. DXY fell from 110 in mid-January 2025 to 96.35 by July 2025, a roughly 12% drop. Gold rallied concurrently to new all-time highs above $3,000 by mid-year and above $4,000 by late 2025. The 2025 regime restored the textbook inverse correlation, with gold taking the dollar decline plus the persistent central bank bid and compounding into a 60%+ rally for the year.

Entering 2026, DXY has stabilized in the 96-100 range and gold has pushed higher to roughly $4,700 by late April 2026 on the Middle East geopolitical premium from the 2026 Iran war. The relationship has held inverse on short horizons but gold has continued to outperform what dollar moves alone would suggest, indicating the structural bid remains operative even as the acute 2025 dollar decline has passed.

Professional traders express dollar-gold views through direct spread positions (long gold, short DXY futures) or through options-based structures that isolate the correlation bet. Retail traders more commonly use gold ETFs (GLD) alongside dollar-bullish ETFs (UUP) or currency ETFs for individual legs.

Spread trades are carry-sensitive. Long gold positions have negative carry (no yield, small storage/insurance cost). Short DXY positions typically have small positive carry when US rates are above foreign rates. Net carry on a long gold, short DXY spread has been modestly negative for most of the past decade, requiring a directional view on the relationship resolving within a reasonable horizon. Historical mean reversion in the 60-day rolling correlation from extremes (below negative 0.7 or above zero) typically plays out over 6 to 18 months.

The dollar-gold relationship filters through to gold mining equities with amplified magnitude. Major gold miners (Newmont, Barrick, Agnico-Eagle) typically show 1.5 to 2x beta to the gold price itself, because their cost structures are relatively fixed (labor, energy, royalties) while their revenue moves with gold. A 10% rise in gold can produce a 20%+ rise in mining equities, and the reverse is true on the downside.

Mining equities also carry operational risk (mine-specific issues, grade changes, permitting) that pure gold exposure does not. For views on the dollar-gold macro relationship, spot gold or gold ETFs (GLD, IAU) are usually the cleaner expression. Mining equities make sense for additional conviction on gold upside with tolerance for operational dispersion between names.

Gold is often described as an inflation hedge, but the evidence is regime-dependent. Gold performed well during the 1970s inflation (when both the dollar and real yields were weak), during the 2008-2011 stimulus period (same pattern), and during 2020-2023 (central bank buying plus initial COVID fiscal-monetary stimulus).

Gold performed poorly during the 1980-2000 inflation-fighting era (Volcker tightening, then disinflation) and during the 2012-2018 low-inflation expansion (rising real yields). The pattern suggests gold works best when inflation is accompanied by policy accommodation or by structural dollar weakness, not during inflation-fighting tightening. The 2022-2024 regime was unusual because gold rallied despite both inflation and Fed tightening; the structural central bank bid overrode the normal inflation-hedge math.

The primary signal to watch is whether central bank gold buying continues at the 1,000-tonne annual pace set in 2022-2023 or decelerates. Early 2025 data showed 254 tonnes bought in the first 10 months, which annualizes to roughly 300 tonnes, well below the 2022-2023 pace. If the deceleration continues, gold's structural bid weakens and the dollar channel should reassert dominance.

The secondary signal is DXY level. Sustained DXY above 105 with stable or rising real yields would be the cleanest setup for gold to correct, which has not occurred in the current cycle because the 2025 dollar decline took DXY into the 96-100 zone and it has not recovered decisively. The 2026 Iran war and Hormuz closure have kept geopolitical premium elevated on top of that. A durable ceasefire that reverses DXY weakness and slows the central bank gold bid would be the test of whether gold remains supported in the $4,000+ range or reverts toward the dollar-implied fair value in the $3,000-$3,500 zone.