Precious Metals Investing: The 5-Minute Crash Course

The Allocation Math

5% to 15%. That is the standard allocation band for precious metals inside a diversified portfolio. The number is not arbitrary. It is a function of correlation to risk assets and contribution to portfolio variance.

Gold posts a 0.14 correlation coefficient with global equities over a 20-year window. At 0.14, a metals sleeve behaves nearly independently from an equity sleeve. Variance contribution compresses. Sharpe ratio improves marginally without sacrificing expected return. The number is structural. It persists across rate regimes, inflation regimes, and credit cycles. It is not a function of the current cycle. It is a function of the asset class.

Decompose the band. At 5%, the metals position functions as a tail hedge. At 15%, the sleeve begins to dominate portfolio variance. Beyond 15%, concentration risk compounds and the Sharpe contribution inverts. Miners, royalty companies, and streaming vehicles add further operational leverage on top of the underlying metal price. Treat them as a separate sleeve in the correlation matrix. Their beta to spot runs higher. Their correlation to broad equities also runs higher. They are not a substitute for the metal.

For allocators extending metals exposure into the broader real-asset bucket, alternative investment vehicles carry a different return profile and lockup structure. The metals sleeve is liquid. Private capital is not. Position size accordingly. The correlation matrix runs deeper than public-market beta alone.

Gold at 0.14 equity correlation is the textbook definition of a diversifier — a quiet variance sink by design.

Gold: Central Bank Flow and the Macro Hedge

Central banks absorbed 863 tonnes of gold in 2025. That is the fourth-largest annual expansion of official reserves on record. The flow is persistent and price-insensitive in the short term. Reserve managers accumulate on quarterly schedules driven by diversification mandates. They do not react to weekly closes. They do not mark to market daily. The bid is structural.

This bid matters for positioning. It sits below spot. It absorbs supply on down days. It compresses downside skew on the curve. For a quant reading the order book, the central bank bid is the floor under the gold tape. The market knows this. That knowledge is priced in.

The macro hedge function operates on a longer timeframe. Gold's correlation to monthly CPI prints is inconsistent. Over short windows, the relationship breaks. Over multi-year windows, it stabilizes. Allocators who size gold as an inflation hedge on monthly data will see whipsaw. Allocators who size it as a multi-year real-rate hedge see the function hold. The distinction matters.

For a portfolio construction view, term structure is a signal worth tracking but not a trade in itself. The shape of the gold curve reflects carrying cost, lease rates, and the funding cost of a long position. In normal conditions the curve sits in modest contango, with storage and insurance embedded in the futures price. Episodes of backwardation can signal tight physical availability and aggressive official-sector or jewelry demand absorbing nearby supply. The signal is useful for context but secondary to the structural allocation case.

Open interest and managed-money positioning are feeds that matter at extremes. They tell you whether the trade is crowded. Crowded trades unwind violently. Uncrowded trades absorb flow without dislocation. Neither extreme is comfortable. The middle is where Sharpe lives. Implied volatility on near-dated gold options carries information too — elevated readings signal hedging demand or stress, compressed readings signal complacency. Neither is a reason to override the allocation band.

Silver, Platinum, Palladium: The Industrial Sleeve

Silver prints $57–$58/oz. The metal behaves differently from gold. 58% of annual silver demand originates in industrial applications: solar photovoltaic panels, electronics, brazing alloys, photographic film (residual). The remaining 42% is monetary and jewelry demand.

The split drives the volatility profile. Industrial demand links silver to manufacturing cycles, green-energy capex, and electronics build rates. Monetary demand links it to gold's tape. The two legs do not always correlate. Silver carries higher beta than gold on both upside and downside. Weekly return standard deviation runs roughly 1.6x gold's. The curve reflects this. Silver futures trade in shallow contango in the front months. Roll yield is mildly negative. The cost of carry is real.

Platinum trades $1,630–$1,670/oz. Palladium trades $1,310–$1,340/oz. Both are heavily exposed to automotive catalytic converter demand. Internal combustion production schedules and emission regulation timelines drive the flow. The platinum-palladium spread has compressed materially from prior decade averages as autocatalyst recipes shift toward platinum substitution. Position this as a structural shift in auto chemistry, not a directional trade. Demand destruction in palladium is a multi-year story.

For an allocator, the industrial sleeve functions as a cyclical hedge within the metals book. When industrial PMIs accelerate, silver and platinum outperform gold. When industrial activity contracts, they underperform. The cycle leg adds variance. It also adds return optionality. The trade is not a static allocation. It is a regime-dependent bet.

Supply-side dynamics matter here as well. Silver is largely a byproduct metal — a majority of mine supply comes from lead, zinc, copper, and gold operations where silver is a secondary credit. That structure means silver supply does not respond elastically to price. A price spike does not pull new ounces out of the ground on a 12-month timeline. It pulls existing scrap and incentivizes byproduct credit optimization. The same dynamic applies to platinum, where production concentrates with a handful of South African producers plus Russian supply. Palladium is structurally tighter given Russian export constraints and auto-catalyst substitution. These supply inelasticities backstop the floor under industrial metals in drawdowns.

Silver is gold with a manufacturing beta. Platinum and palladium are autocatalyst spreads wrapped in a precious-metals ticker.

Physical vs Paper: Liquidity, Storage, Carry

Two structures exist: physical metal and paper exposure. The trade-off is mechanical. The choice depends on position size, holding period, and liquidity tolerance.

Physical bullion: direct ownership, no counterparty (if allocated). Costs include dealer spreads of 2%–5% over spot for retail coins and small bars, allocated vaulting at 0.5%–1.0% annual, and insurance at 0.1%–0.3% annual. Liquidity is lower. Bid-ask on retail bullion runs wide. Settlement is slower. The metal sits in a vault or safe-deposit box. It is not earning yield. It is not marking to market.

Paper alternatives: ETFs, futures, and exchange-traded certificates. GLD carries a 0.40% annual expense ratio. SLV carries 0.50%. Futures expose the holder to roll yield, contango, or backwardation in the curve. Bid-ask is tight. Liquidity is deep. Settlement is T+1 or T+2 for ETFs, daily mark-to-market for futures.

For institutional allocators, futures dominate on cost and liquidity. For high-net-worth allocators seeking direct ownership, allocated physical in a third-party vault clears the custody question. For retail, ETFs strike the balance. The fee differential versus the spread differential determines the break-even holding period. At a 10-year hold, storage drag on physical exceeds the cumulative ETF expense ratio. At a 1-year hold, the spread on physical exceeds the ETF fee. Time horizon is the variable.

A separate allocation leg runs through mining equities and royalty companies. These instruments carry operational leverage on the underlying metal price. A 20% move in gold does not translate to a 20% move in a producer's equity. Beta runs 1.5x–2.5x depending on cost-curve position, hedge book, and capex structure. Mining stocks are not metals. They are equities with metal-price exposure. The distinction matters for correlation math and for downside behavior in a producer drawdown.

Risk Parameters

Three structural risks run through the metals sleeve. Quantify each.

First: no yield. Precious metals do not pay dividends, coupons, or interest. Return is price appreciation only. In a high-real-rate environment, the opportunity cost of holding metals rises. Cash and short-duration Treasuries compete directly with the position. The cost is not theoretical. It is the foregone yield on allocated capital. At a 5% real rate, a 10% metals allocation carries 50 basis points of annual opportunity cost before any price appreciation. That cost compounds.

Second: operational risk in mining exposure. Miners face cost inflation, grade decline, jurisdictional risk, labor disruption, capex overruns, and environmental remediation liabilities. A mining equity can underperform the underlying metal even when the metal price rises. Treat miners as a separate sleeve from physical metal and ETFs. Do not blend them in the correlation matrix. They do not behave the same.

Third: storage and insurance drag on physical positions. At 0.6%–1.3% annual, the cost compounds. Over a 10-year hold, total drag runs 6%–13% of position value before any price appreciation. The math is straightforward. It is also why institutional flows concentrate in ETFs and futures rather than allocated physical.

Unknowns for the remainder of 2026: rate path, dollar trajectory, central bank accumulation cadence, industrial demand recovery in silver and platinum. Analyst price targets diverge. J.P. Morgan and Goldman Sachs have revised 2026 averages lower. Other desks remain constructive. Variance in forecasts is itself a signal. Consensus has not converged. The trade is not crowded.

A fourth risk worth flagging: regulatory and tax treatment. Collectible coins can carry higher long-term capital gains rates than bullion bars in certain jurisdictions. ETF structures add a layer of tax efficiency for some accounts but not all. Futures trigger 60/40 treatment under Section 1256 in the US. These structural frictions do not change the allocation case but they do change the after-tax return profile. Allocators should size positions in after-tax terms, not pre-tax.

Closing the Position

Spot marks for mid-July 2026: gold $4,030/oz, silver $57/oz, platinum $1,650/oz, palladium $1,325/oz. Allocation band: 5%–15%. Equity correlation for gold: 0.14 over 20 years. Industrial demand share for silver: 58%. Central bank gold accumulation 2025: 863 tonnes. ETF expense ratios: 0.40%–0.50%. Storage drag for physical: 0.6%–1.3%.

The allocation is a variance calculation. It is not a directional bet. Position size follows correlation math, not conviction. The metals sleeve reduces portfolio variance. It does not generate income. It does not compound. It absorbs shocks and rerates against fiat debasement over multi-year windows.

The setup entering the second half of 2026 is neutral. Spot marks sit well above prior consolidation ranges. The structural bid from official-sector accumulation remains intact. Industrial demand for silver and platinum tracks the green-energy and auto cycles, both of which face uncertain near-term trajectories. Implied volatility has compressed from its 2024 peak but remains above the multi-year floor. The trade is size, not timing.