Why gold for later usually costs more than gold for now

The curve is a price of time, not a forecast

A forward or futures price for gold is not the market guessing where the price will be. It is today's price plus the cost of holding metal until that date. The distinction sounds pedantic and it is the single most useful idea in this whole area. If a far month trades above the near month, nobody is predicting a rally. They are quoting the cost of carry.

Contango describes a curve that slopes upwards, with later delivery more expensive than earlier delivery. Backwardation describes the opposite, where metal for immediate delivery commands a premium over metal later. For most commodities, the normal state depends on harvests, storage capacity and consumption. For gold, which does not rot, is cheap to store relative to its value and exists in enormous above ground quantity, the normal state is contango, and it is close to mechanical.

Every so often it is not. Those episodes are informative precisely because the arithmetic is usually so reliable.

What the cost of carry is made of

Imagine financing a position in physical gold. You borrow dollars, buy metal today, store and insure it, and simultaneously sell a forward contract to lock in the price at which you will hand it over. Your profit is the forward price minus the spot price minus your financing and storage costs. If that is positive, the trade is free money and will be done until it is not. That arbitrage is what pins the curve.

So the slope is built from three pieces:

• The dollar interest rate you pay to finance the metal. Higher rates mean a steeper curve.
• The gold lending rate you can earn by lending the metal out while you hold it, which offsets part of your financing cost and flattens the curve.
• Storage and insurance, small for gold but not zero, which steepens it slightly.

This is why the curve responds to policy. The connection traced in interest rates and gold shows up in the curve far more directly and immediately than it shows up in the spot price, because the spot price is fighting a dozen other influences and the curve is mostly just doing maths.

Why the basis must collapse to nothing

A futures contract and physical metal converge by construction. On the day a contract can be settled with delivery, holding the contract and holding the bar are the same thing, so any gap between the two prices is an immediate arbitrage. The gap, called the basis, therefore shrinks towards zero as expiry approaches, not because traders agree it should but because anyone can collect the difference if it does not.

This has a practical consequence that catches people out. The carry embedded in a futures price is continuously bleeding away. A holder of a far month contract in a steep contango is paying for that carry whether they think about it or not, because the price they bought must decay towards spot over the life of the contract if spot stays still. It is not a fee anybody charges. It is the geometry of convergence.

For anyone who has only ever traded the spot pair, the gap between the two instruments is the main thing to internalise, and spot versus futures in gold covers the rest of the differences.

What backwardation in gold actually implies

Because contango is the arithmetic default, a flip to backwardation is a statement that the arbitrage has stopped working. Someone is willing to pay more for metal now than the carry maths says they should, and in a market with vast above ground stocks that only happens when the metal that exists is not where it is needed, or cannot be mobilised quickly enough.

So backwardation in gold is best read as a scarcity of immediately available, deliverable metal rather than a scarcity of gold. The bars might be in the wrong city, in the wrong form, locked in long term holdings, or held by owners unwilling to lend. It can also reflect a reluctance to extend balance sheet: the carry trade requires borrowing dollars and warehousing an asset, and in a credit squeeze that capacity shrinks regardless of how attractive the spread looks on paper.

The honest summary is that backwardation flags friction in the physical and funding plumbing. It is a reliable sign that something is strained. It is not a timing device and not a price target.

What makes the curve flip

The usual causes are logistical and financial rather than speculative.

• Location mismatch. Deliverable metal sits in specific vaults in specific cities, and moving it takes flights, insurance and sometimes recasting into a different bar size. A sudden demand for delivery in one location can price metal there above metal elsewhere.
• Trade friction. Tariff threats, export restrictions or disrupted freight can make the cost of getting a bar from one jurisdiction to another jump, and the curve absorbs that cost.
• Funding stress. When short term dollar funding becomes expensive or simply unavailable, nobody can run the carry trade at any price, and the curve loses its anchor.
• Lending withdrawal. If the large holders who normally lend metal stop doing so, the lending rate climbs and the forward slope compresses or inverts.

Notice that none of these is a view on gold. Each is a constraint. That is why the curve can dislocate while spot does very little, and why it sometimes dislocates in the opposite direction to what a headline writer would expect.

Reading the spread between exchange and over-the-counter prices

Most of the time the futures price and the London over-the-counter price differ by exactly the carry, and the spread between them is traded as a product in its own right. When that spread widens sharply, it is telling you the two pools of metal have become harder to arbitrage against each other, usually because physical delivery into one of them has become slow, expensive or uncertain.

For a retail trader this is mostly a warning light rather than an opportunity. A dislocated basis is accompanied by wider quoted spreads, thinner resting liquidity and fills that land further from where you aimed. The sensible response is smaller size and more patience, not a clever trade on the spread itself, which requires balance sheet, vault access and delivery capability that no retail account has.

It is also a reminder that the price on your screen is one venue's price. The relationship between real yields and the metal, laid out in real yields and gold, operates on the whole complex, but the plumbing decides which venue feels it first.

The limits of reading the curve

Here is where curve analysis earns its reputation for being overrated. Contango is the default state, so its presence tells you nothing at all. A steepening curve usually just means financing costs rose. Most of the time the shape is a restatement of the interest rate environment with no gold specific content whatsoever.

Backwardation is more informative and still not directional. Episodes can be brief, mechanical and resolve with no lasting price effect once metal is shipped or a funding problem passes. They can also appear during sharp selling, which ruins the lazy reading that tight physical metal means price must rise. And the data itself is awkward for retail traders, since forward and lending rates are dealer quoted rather than publicly posted in any convenient series.

Treat the curve as a diagnostic of market condition, in the same family as spreads and funding stress. Note it, let it adjust your position size and your expectations for slippage, then go back to reading price on the live chart. Nothing about the shape of the curve tells you which way the next move goes.