Gold pays no coupon. Everything about how it responds to macro conditions follows from that one sentence, and the series that captures it best is the real yield on inflation-protected government debt. This edition is a working explanation of the mechanism rather than a forecast — we do not publish price targets.
The mechanism
Holding gold costs you the return you could have earned on a safe asset instead. If a ten-year Treasury Inflation-Protected Security yields two per cent above inflation, the opportunity cost of holding a non-yielding asset for a decade is compounding two per cent a year. If that same real yield is negative, the safe alternative is guaranteeing you a loss of purchasing power, and the cost of holding gold instead falls to roughly nothing.
That is the entire argument. It is unusually clean for a macro relationship because it does not require anyone to be irrational, to panic, or to hold a view about inflation. It only requires investors to compare two assets on the return they actually keep.
Nominal rates do not do this job. A nominal yield of six per cent alongside eight per cent inflation is a negative real yield and a historically supportive environment for gold, which is why commentary anchored on the policy rate keeps getting the direction wrong.
Where to find it
The Federal Reserve Bank of St. Louis publishes the ten-year TIPS constant-maturity yield free of charge as series DFII10, updated daily. The ten-year break-even inflation rate — the difference between the nominal and inflation-protected yields, and therefore the market's implied inflation expectation — is series T10YIE on the same site. Between those two you have both the level and its decomposition, at no cost, with a full history.
Two habits make the series more useful. Watch the direction of change as well as the level: gold has historically responded more to a real yield that is falling than to one that is simply low. And watch the dispersion between regions — euro-area and Japanese real yields have spent long periods well below US levels, which matters for a metal priced in dollars but held globally.
What the break-even tells you that the real yield does not
The nominal yield decomposes into a real yield plus a break-even inflation rate, and the two components carry different information for gold. A falling real yield driven by a rising break-even — inflation expectations climbing faster than nominal rates — is the configuration in which gold has historically performed best, because the metal is being paid for both the lower opportunity cost and the loss of confidence in the currency. A falling real yield driven by a falling nominal yield with stable break-evens is a growth-scare configuration, and gold's response tends to be real but smaller and shorter.
This is why two months with identical real yields can produce entirely different gold outcomes. Always read DFII10 alongside T10YIE, and note which component did the moving. The distinction takes ten seconds and removes a large share of the confusion in monthly commentary.
Levels, changes and the second derivative
Three readings of the same series answer three different questions. The level tells you the regime: deeply negative real yields have historically coincided with the strongest gold environments, and real yields above roughly two per cent with the weakest. The change tells you the direction of pressure, and it is the reading that correlates best with gold's month-to-month behaviour. The rate of change tells you about stress: a real yield moving thirty or forty basis points in a fortnight is a liquidity event, and in liquidity events gold frequently falls first with everything else before it recovers.
That last point catches people out repeatedly. In an acute stress episode, gold is sold because it is liquid and because margin calls are denominated in currency, not because the hedging case failed. The hedge does its work over the subsequent quarters, not in the first week.
When the relationship broke
An honest framework names its failures. This one has at least three well-documented ones.
2013. The taper tantrum drove real yields sharply higher and gold fell hard — the relationship worked, but the magnitude on the gold side far exceeded what the rate move alone implied. ETF liquidation and positioning unwinds did the rest.
2018–2019. Real yields rose through much of 2018 while gold declined only modestly, then rallied strongly in 2019 as the Federal Reserve reversed course. The lag between the rate signal and the gold response was long enough to bankrupt anyone trading it mechanically.
2022–2023. The clearest breakdown of the modern era. Real yields rose to levels last seen before the financial crisis, and the textbook response would have been a severe gold bear market. Gold instead held its range and then broke higher. The widely accepted explanation is that a different demand bucket — official-sector buying, discussed in Edition 08 — absorbed the pressure. Whatever the explanation, the honest conclusion is that a single-variable model of a multi-variable market will periodically be wrong, and it will be wrong at exactly the moments that matter most.
Two more caveats worth carrying
First, the TIPS market is a market, not a measurement. Its real yield embeds a liquidity premium and an inflation risk premium, both of which move, so the series is a decent estimate of the real return investors demand rather than an objective reading of the real cost of money. In March 2020 the TIPS market dislocated badly enough that the printed series briefly described its own plumbing rather than the economy.
Second, the dollar is doing work that gets attributed to real yields. Real yields and the trade-weighted dollar are correlated, so a single-variable regression against gold quietly credits the yield with part of the currency effect. If you want to separate them, look at gold priced in euro, yen and rupee terms as well: when the dollar price is falling and the local-currency price is not, the story is currency, not gold.
How to use it without over-using it
Treat real yields as a regime identifier rather than a timing tool. The question it answers well is what kind of environment is this? — is the opportunity cost of holding gold rising or falling, and from what level. The question it answers badly is should I buy this week?
Pair it with a second and a third input before drawing a conclusion. The Commitments of Traders report tells you whether speculative positioning is already crowded on the same side of the trade. Official-sector reserve data tells you whether a price-insensitive buyer is present. Physical premiums in retail markets tell you whether the metal itself is tight. When those disagree with the rate signal, the disagreement is the information.
A worked example
Take a ten-year real yield moving from minus one per cent to plus two over eighteen months, as happened in 2022. Mechanically, the opportunity cost of holding a non-yielding asset goes from negative — the safe alternative guarantees a loss — to three percentage points of compounding forgone return. Every textbook says gold should fall materially.
It did not, and the useful discipline is to write down, at the time, which of the other drivers must be doing the work: sovereign accumulation, a currency event, or physical tightness. Then check whether the candidate driver is visible in its own data series rather than only in the story. In 2022 it was — official-sector net purchases exceeded a thousand tonnes on published reporting.
That is what a framework is for. Not to be right every quarter, but to tell you precisely which assumption has stopped holding, so the next question is a research question rather than a shrug.
Positioning discipline
We publish frameworks and parameters, not names and not price targets. Our monthly mining screen filters on production scale, all-in sustaining cost relative to spot, balance-sheet leverage and jurisdiction ranking; we publish those parameters so readers can run their own version. A screen is the start of research, not a conclusion.
Reading list
Three primary sources worth an hour each: the World Gold Council's quarterly Gold Demand Trends, the BIS Quarterly Review's coverage of monetary architecture, and the IMF working paper series on reserve currency composition. All are published free by their issuers, and all are better than the commentary written about them.