Gold Beyond Real Yields: Why the Old Model Is No Longer Enough

Editorial thesis

Gold is often explained through a single variable: US real interest rates.

That relationship is important, but incomplete.

The classic model says that higher real yields increase the opportunity cost of owning a zero-cash-flow asset. Yet recent gold cycles have shown that the metal can remain resilient — or even rise — while real yields are elevated.

The correct conclusion is not that real yields no longer matter. It is that gold increasingly trades as a multi-factor monetary asset whose price reflects real rates, the dollar, reserve diversification, fiscal confidence, geopolitical demand, market positioning and optionality.

The central argument is that the real-yield framework still provides the baseline, but no longer explains enough of the price action on its own. Gold is increasingly behaving as a monetary hedge against the quality of the sovereign benchmark itself, not merely as an asset competing with it.

1. Start with the analogy — then know where it breaks

A useful analogy is to think of gold as a very long-duration monetary asset with no contractual cash flow. But the analogy should not be taken literally.

A Treasury bond has coupons, principal and a maturity date. An equity has expected dividends or distributable cash flows. Gold has none of these.

Contractual cash flow from physical gold = 0

That means gold cannot be valued with a conventional discounted-cash-flow model.

Instead, its market value is determined through relative opportunity cost and monetary demand. The main competing asset is not another productive claim, but interest-bearing money and sovereign debt.

That is why real interest rates are so important.

Key distinction: gold behaves like a long-duration monetary asset, but conventional bond duration is not formally defined for an asset with no promised cash flows.

2. Real yields still anchor the baseline

The standard starting point is the real Treasury yield.

Real yield ≈ Nominal Treasury yield − Inflation breakeven

For example, if a nominal 10-year Treasury yields 5% and the 10-year breakeven inflation rate is 2.5%, the approximate real yield is:

5.0% − 2.5% = 2.5%

Physical gold generates no coupon. Therefore, as the real return available on safe government bonds rises, holding gold becomes more expensive in opportunity-cost terms.

Real yield rises
Safe inflation-adjusted income becomes more attractive.

→

Gold opportunity cost rises
Demand for zero-yielding bullion can weaken.

The baseline relationship is therefore:

Real yields ↑ → Gold ↓All else equal.

The phrase all else equal is essential. Gold is not driven by real yields alone.

3. The old quantitative model still matters

A quantitative desk can represent the relationship using a log-linear regression:

ln(Gold_t) = α − β₁RealYield_t − β₂ln(DXY_t) + ε_t

This model contains two familiar drivers.

Real yield coefficient, β₁. If estimated β₁ is positive inside the negative sign, higher real yields reduce the fitted gold price.

Dollar coefficient, β₂. Gold is quoted globally in US dollars, and a stronger dollar frequently coincides with weaker gold prices for non-US buyers.

Suppose an estimated model produces:

β₁ = 8

If the real yield increases by 100 basis points, or 0.01 in decimal form:

Δln(Gold) ≈ −8 × 0.01 = −0.08

The model would therefore imply an approximate 8% decline in gold, conditional on the other explanatory variables remaining unchanged.

This is an empirical sensitivity, not a law of nature. The coefficient depends on the sample period, data frequency, specification and monetary regime. A coefficient estimated over one decade should not automatically be carried into another.

4. The residual is becoming the story

The regression produces a fitted price:

Gold̂_t = f(RealYield_t, DXY_t)

The residual is the difference between the observed and fitted value:

ε_t = ln(Gold_t) − ln(Gold̂_t)

If gold repeatedly trades above the level implied by real yields and the dollar, the residual is not necessarily random noise. It may indicate that the model is omitting a persistent structural variable.

A richer specification might therefore be:

ln(Gold_t) = α − β₁RealYield_t − β₂ln(DXY_t) + β₃CB_t + β₄FiscalRisk_t + ε_t

where:

CB represents central-bank accumulation or reserve demand;
FiscalRisk represents a proxy for sovereign-risk or fiscal-confidence stress.

This does not guarantee a better model. But it provides a framework for testing whether the old two-factor relationship has become incomplete.

5. Central banks are changing the demand function

Central banks have become a major source of physical gold demand.

World Gold Council data show that central banks purchased approximately 863 tonnes in 2025, well above the 2010–2021 annual average. Buying remained strong in 2026, with net purchases of roughly 289 tonnes in Q2 alone.

The institutional motive differs from that of a short-horizon investor.

A reserve manager may value gold for:

reserve diversification;
absence of issuer default risk;
liquidity across jurisdictions;
reduced exposure to sanctions or reserve freezes;
currency diversification;
strategic resilience during geopolitical stress.

That means central-bank demand may be much less sensitive to a small change in real yields than leveraged financial demand.

In the 2026 World Gold Council reserve survey, 89% of surveyed reserve managers expected global central-bank gold holdings to increase over the following 12 months, while 45% expected their own institutions to increase holdings.

Model implication: persistent official-sector buying can shift the intercept of a gold regression or generate sustained positive residuals relative to a model based only on TIPS and DXY.

6. Not all yield shocks mean the same thing for gold

A Treasury yield can rise for different reasons.

That distinction is central to understanding gold.

A useful decomposition is:

Long Treasury yield ≈ Expected future short rates + Term premium

Now compare two regimes.

Yield shock	Typical interpretation	Potential gold effect
Growth-driven	Stronger real activity, tighter monetary policy, higher expected real rates	Usually negative
Inflation-driven	Higher inflation expectations without equivalent policy credibility	Mixed to positive
Fiscal / term-premium driven	Heavy sovereign supply, fiscal uncertainty, weaker confidence in long-duration government debt	Ambiguous; can become positive
Systemic-risk driven	Banking, geopolitical or sovereign stress	Potential safe-haven demand

If long yields rise because the economy is strong and expected real short rates increase, the opportunity-cost channel is straightforwardly negative for gold.

If yields rise because investors demand more compensation for holding sovereign duration, the signal is different. Higher yields are then partly reflecting risk in the benchmark asset itself.

In that environment, gold may retain demand as a non-sovereign reserve asset even while nominal and real yields remain elevated.

This is not proof that sovereign-risk variables “overwhelm” real yields. It means the sign of the total gold response becomes regime-dependent because several forces are moving simultaneously.

7. Options markets show where the tail risk is priced

Spot prices tell us where gold trades. Options tell us how investors price the distribution of future outcomes.

A common measure is the 25-delta risk reversal:

25Δ RR = σ_{25Δ Call} − σ_{25Δ Put}

If the risk reversal is positive, out-of-the-money calls trade at a higher implied volatility than comparable puts.

That means the market is paying more for upside convexity than downside convexity.

25Δ RR > 0 → upside implied volatility premium

Possible explanations include:

demand for protection against a sharp gold rally;
geopolitical event risk;
reserve or central-bank demand uncertainty;
fiscal or monetary-policy tail risk;
dealer positioning and inventory effects.

But the signal should not be over-interpreted.

A positive risk reversal does not uniquely imply “fiat debasement.” It tells us that upside tail protection is relatively expensive. The macro explanation must be inferred from the broader regime.

8. The market now requires a regime dashboard

The advantage of the multi-factor framework is that it prevents a single variable from dominating the interpretation.

Variable	Gold-positive signal	Gold-negative signal
10Y real TIPS yield	Falling	Rising
DXY	Falling	Rising
Central-bank purchases	Strong / accelerating	Weak / net selling
Sovereign fiscal risk	Rising	Stable / improving
Term premium	Rising because of fiscal-risk repricing	Falling
25Δ risk reversal	Positive / rising	Negative / falling
Gold ETF flows	Inflows	Outflows

The model can therefore be summarized as:

Gold = f(Real Yields, USD, Central-Bank Demand, Fiscal Risk, Options Skew, Flows)

The objective is not to assign a permanent coefficient to every variable. It is to identify which factor is setting the marginal price in the current regime.

The same yield move can carry a different message

Consider two periods in which the 10-year real yield rises by the same amount.

Regime A: growth surprises higher, the Federal Reserve becomes more hawkish, the dollar strengthens, central-bank purchases are moderate and options skew is neutral.

Most signals point in the same direction:

Real yields ↑ + DXY ↑ → Gold pressure

Regime B: real yields rise because the Treasury term premium expands, fiscal concerns intensify, the dollar fails to strengthen, central banks continue accumulating gold and call skew becomes more expensive.

Now the signal set is mixed:

Real yields ↑ but Fiscal risk ↑ + CB demand ↑ + Upside skew ↑

Gold may therefore remain resilient or rise despite the higher real yield.

The same numerical change in yields can generate a different gold response because the economic cause of the move is different.

Bottom line: real yields remain the baseline opportunity-cost variable for gold, but they are not a complete valuation model.

Gold has no contractual cash flow, so its value is fundamentally relative: relative to real interest rates, currencies, sovereign credibility, reserve preferences and the price investors are willing to pay for tail protection.

The important analytical shift is therefore from a one-factor model:

“Real yields up, gold down.”

to a regime-dependent framework:

“What is driving real yields, what is the dollar doing, who is buying physical gold, and how is the options market pricing asymmetric risk?”

That is the framework required to understand why gold can trade above the level implied by traditional real-yield models for extended periods.

Sources & further reading

World Gold Council — Central-bank demand in 2025.

World Gold Council — Central Banks, Q2 2026.

World Gold Council — 2026 Central Bank Gold Reserves Survey.

World Gold Council — Gold, rates and the US dollar.

Regression coefficients cited here are illustrative. They should be estimated for the chosen sample, frequency and regime rather than treated as fixed structural parameters.