Gold is usually explained with one sentence too many.

The standard story is tidy. Lower rates reduce the opportunity cost of holding a metal that pays no coupon, so gold rises when the market expects easier policy. That story is not wrong. It is worse than wrong. It is incomplete in the most convenient direction.

This note tests a narrower claim: monthly gold returns are more closely associated with changes in inflation compensation than with changes in expected fed funds. The result is not cinematic. It is better than that. It is modest, measurable, and harder to abuse.

The data set is monthly. Gold is measured in U.S. dollars per troy ounce. Breakeven inflation is the 10-year expected inflation compensation series. Fed funds expectations are proxied from the front federal funds futures contract. The regression uses monthly log gold returns as the dependent variable and monthly changes in breakevens and implied fed funds as explanatory variables.

The model is deliberately small. No twenty-factor shrine. No macro confetti cannon. Just the parts of the story people keep claiming to understand.

1. The price history has a regime break hiding in plain sight

The gold price series reaches back to 1833, which sounds impressive until you remember that much of that history is a policy artifact. For long periods the price was fixed or heavily constrained. Treating that as market behavior is how bad backtests dress themselves for dinner.

A trend fitted across the full sample would give the early fixed-price regime too much authority. It would make the modern market look more explosive than it is because the first part of the chart is not really a market. It is a ruler nailed to the wall.

So the trend sample begins in April 1968, after the fixed-price architecture starts to break. That is not perfect. It is simply the cleaner empirical cut.

Across the modern sample, the fitted log trend grows at roughly 5.93 percent per year, with an R-squared near 0.81. Nominal gold has had a strong drift since the market was allowed to breathe.

That matters because it keeps the regression honest. We are not asking whether gold went up over fifty-eight years. It did. The question is what helps explain the month-to-month changes around that long upward drift.

2. Inflation compensation is not inflation

A breakeven rate is not a clean forecast from some all-seeing bond oracle. It mixes expected inflation, inflation risk premia, liquidity effects, Treasury Inflation-Protected Securities market plumbing, and whatever else the bond market spilled on the table that month.

That is fine.

Financial variables are rarely pure. Purity is usually where useful analysis goes to die wearing a lab coat.

Breakevens are useful because they measure what the market is willing to pay for inflation protection. They are not a prophecy. They are a price. And prices matter because people can argue with narratives forever, but they have to trade prices.

The chart puts gold beside the two expectation variables. The shared history is uneven, which is the point. Gold is not a mechanical derivative of either series. It is a financial asset with its own flows, shocks, storage logic, crisis premium, and central-bank bid.

But when inflation compensation reprices, the relationship is visible enough to test. Not worship. Test.

3. The regression result

The regression is intentionally plain:

gold monthly log return = alpha
                        + beta_1 * change in 10Y breakeven inflation
                        + beta_2 * change in implied fed funds rate
                        + error

The sample runs from February 2003 to May 2026 and contains 280 monthly observations. Standard errors are HAC robust, because monthly macro-finance data is not a clean marble countertop. It has autocorrelation, heteroskedasticity, and fingerprints.

The coefficient on breakeven changes is positive and statistically significant. The estimate is about 0.0597 with a p-value near 0.001. In plain English: months with rising inflation compensation have tended to be stronger months for gold, after controlling for changes in implied fed funds.

The coefficient on implied fed funds changes is negative, which is directionally consistent with the popular rate story, but it is not statistically significant in this specification. The estimate is about -0.0164 with a p-value near 0.217.

That does not mean rates do not matter. It means this simple monthly model does not give the fed funds expectations variable the starring role people keep handing it.

The R-squared is 0.043. That number is low, and it should be low. Gold is not a savings account with a costume budget. A two-variable monthly model should not explain most of it. If it did, the model would deserve suspicion, not applause.

4. The scatterplots say the quiet part

The scatterplots are useful because they refuse to perform sophistication. Each dot is a month. Each line is the fitted relationship. No grand theory. No macro cosplay. Just slope and noise.

The breakeven relationship slopes upward. It is not clean. It is not meant to be clean. Markets do not hand you laboratory charts unless someone has already tortured the data.

The signal is that positive breakeven shocks and gold returns have tended to move together. This supports the idea that gold responds to inflation compensation, not merely to the ritual chanting of future rate cuts.

The fed funds expectations scatter is flatter. The sign points the expected way, but the evidence is weaker. In this sample, changes in expected policy rates do not carry the same explanatory weight as changes in inflation compensation.

That distinction matters. A lower expected policy rate can be bullish for gold if it lowers real yields or weakens the dollar. But if the rate cut is already priced, or if it arrives with collapsing growth expectations, the effect can vanish into the machinery. The rate itself is not the mechanism. The mechanism is the pressure it creates across real returns, currency confidence, and inflation compensation.

5. Interpretation

The empirical story is simple.

Gold is more sensitive to the market repricing inflation protection than to the headline movement in expected fed funds. The rate story is a shortcut. Sometimes useful. Often lazy.

Breakevens are closer to the nerve. They sit near the place where the market translates inflation fear, policy credibility, liquidity, and risk premia into a traded number. Gold listens to that number because gold is not just an anti-rate asset. It is an anti-complacency asset.

This is not a claim that breakevens explain gold. They explain a slice. A small slice. The remaining movement belongs to real yields, dollar strength, central-bank purchases, positioning, crisis demand, ETF flows, geopolitics, and the recurring human desire to own something that cannot be printed by a committee with a press release.

That is the useful conclusion. Not that gold is mysterious. Not that gold is simple. Gold is a market price for distrust, and breakevens are one of the cleaner gauges of when that distrust starts getting marked up.

Conclusion

The popular rate-cut story is too neat. It compresses a monetary asset into a one-line explainer and then acts surprised when the world refuses to fit inside it.

In this sample, breakeven inflation changes carry the stronger evidence. Fed funds expectations have the right sign but weaker statistical backing. The difference is not cosmetic. It changes what you should watch if you are trying to understand gold rather than narrate it.

Do not ask only whether the Fed is expected to cut.

Ask whether the market is repricing inflation compensation.

That is where the cleaner signal lives.