How Are Gold Prices and the US Dollar Really Related?

The mechanism: why a strong dollar usually weighs on gold

Gold has no yield of its own — it pays no interest or dividend. Its price is set globally, but quoted in US dollars. When the dollar strengthens against other currencies, gold becomes more expensive for a buyer in euros, yen or rupees, even if the dollar price hasn't moved, which tends to soften demand from outside the US. A stronger dollar is also frequently a symptom of tighter US monetary policy or higher US interest rates, and higher rates raise the "opportunity cost" of holding an asset that generates no income — money sitting in gold could otherwise be earning interest. That second channel, real interest rates (the yield on US government bonds after subtracting expected inflation), is closely tied to the dollar because both often move in response to the same Federal Reserve expectations. As of September 25, 2026, the 10-year real yield stands at 2.76%, a level MacroGold classifies as a "restrictive" regime, and it has been rising over the past three months alongside a dollar that has been comparatively strong on a historical basis.

What the numbers actually show right now

Comparing gold futures to the DXY dollar index over several time horizons, using the same reference dates for both series, gives a more concrete picture than a single correlation figure:

Read this table left to right: on every horizon from a week to six months, gold and the dollar moved in opposite directions, consistent with the usual inverse relationship. Stretch the window to a full year, and the pattern breaks down completely — gold gained nearly 15% while the dollar also gained. That single divergence is the clearest evidence that the "strong dollar, weak gold" rule is a tendency, not a law: other forces, which this data does not isolate, dominated over the past year.

Why "the dollar" isn't always the same number

Part of the confusion comes from which dollar gauge is being used. The DXY, the index most quoted in financial media, is heavily weighted toward the euro and a handful of other developed-market currencies. The Federal Reserve's own broad trade-weighted dollar index (DTWEXBGS), which includes a much wider basket of trading partners, tells a different story for the same period: as of September 18, 2026, it was roughly flat to slightly down over the past year (-0.12%) and sits at the 45th percentile of its one-year range — an unremarkable level. The DXY, by contrast, sits at the 94th percentile of its one-year range on September 25, 2026, meaning it has rarely been this strong over the past year by that specific measure. Two legitimate "dollar" readings, same day, pointing to different conclusions about how strong the dollar actually is. Anyone checking this relationship should note which index is being cited before drawing conclusions.

The correlation itself keeps changing

MacroGold tracks the rolling correlation between gold and the dollar — a statistic between -1 and +1 that measures how closely their day-to-day percentage moves have tracked each other over a given window, with -1 meaning they move in perfect opposition and 0 meaning no relationship. As of the most recent data, the picture shifts noticeably depending on the window used:

Read this as: the more negative the number, the more tightly gold and that dollar measure have moved in opposite directions over that stretch. The DXY relationship has been strongest recently (-0.60 over the last month) but weaker over the full year (-0.33). The broad dollar index tells a steadier story, sitting consistently between -0.47 and -0.51 across all three windows. Neither number is "the" correlation — they are three different snapshots that can and do disagree, which is itself the point: the link is a statistical tendency that strengthens and fades, not a fixed constant. Part of the instability also comes from real yields moving through the same channel — because the dollar and real interest rates often respond to the same Fed signals, it can be genuinely hard to tell which one, if either, is doing the work in a given stretch.

How to check whether the link is working right now

Rather than assuming the relationship holds, it can be checked directly with two comparisons, both of which use only public, dated data:

1. Compare the direction of gold and the dollar (whichever index) over the same recent horizon — a week, a month, a quarter. If they moved in opposite directions, as they did on every horizon up to six months in the table above, the inverse relationship was active. If they moved together, as over the past year, something else was likely driving gold.
2. Compare short-window correlation to long-window correlation. When the short window (21 days) is more negative than the long window (252 days), as it is now for DXY (-0.60 versus -0.33), the relationship has been tightening recently. When they're similar, as with the broad dollar index (-0.47 versus -0.51), the relationship has been more stable.

Because both the dollar and real yields tend to move around scheduled events — Fed decisions, inflation releases, employment reports — tracking those dates against gold's reaction is a practical way to see the mechanism in real time; MacroGold's economic calendar lists the releases most likely to move both.

Takeaway and its limits

A stronger dollar usually creates headwinds for gold, through both the currency-conversion effect for foreign buyers and its frequent link to higher real yields — but "usually" is doing real work in that sentence. Over the past year, gold and the dollar rose together, and the correlation between them ranges from -0.33 to -0.60 depending on which dollar index and which time window is chosen, meaning the relationship explains part of gold's moves but leaves much of it unexplained. This data does not identify what else was driving gold over the past year, does not distinguish the dollar's effect from the real-yield effect when both move together, and — like every statistic on this site — describes what already happened rather than what will happen next. MacroGold's own backtests have found no measurable trading edge from watching this or any other single relationship, so none of the figures above should be read as a signal to act on, only as a way to understand the mechanism and check, with real numbers, whether it is currently holding.