What you are paying for above the quoted price
The quoted gold price is a wholesale price for large, unallocated metal between institutions. A one ounce coin in your hand is a different product. Getting from one to the other involves refining to a higher purity, minting to a precise weight and finish, assaying, packaging, insured transport, a dealer holding inventory, and a retail transaction.
Every one of those steps costs money and someone has to be paid for it. The premium is the sum of those costs plus the margins of the parties involved. It is not a mark up in the sense of something arbitrary. Most of it is the genuine cost of turning a wholesale commodity into a small, verified, portable, resaleable object.
That is why a premium never goes to zero and should not be expected to. The useful question is not why there is a premium but why this premium, on this product, right now, and what the variable part of it is responding to. The same distinction between a quoted price and a transactable price shows up in trading costs too, as covered in spread cost reality.
The fixed components
Several parts of the premium barely move. Fabrication cost is set by the mint or refinery and reflects the labour, energy, tooling and quality control involved in producing a given product. A plain cast bar is cheap to make. A struck coin with a security feature and a mirror finish is not.
Distribution adds insured carriage, secure handling and the cost of financing inventory while it sits in a vault or a shop. A dealer carrying stock has capital tied up in metal whose price can move against them, so they hedge, and the hedge has a cost.
Then there is margin. A dealer quotes two prices, a buy and a sell, and lives on the difference along with volume. Reputable operations run thinner margins on large standard products where competition is fierce, and wider margins on collectible or unusual items where comparison is harder. None of this part of the premium tells you anything about the market. It is the cost of a service, and it is reasonably stable through time.
The variable component: capacity and queueing
The interesting part is what happens when demand for small physical units rises faster than anyone can make them. Minting capacity is a physical constraint. Presses run at a fixed rate, blanks have to be prepared, refineries process a set tonnage, and secure logistics has finite slots.
When retail demand surges, the system cannot answer with quantity in the short run. So it answers with price and with time. Premiums widen, delivery dates stretch, and popular products go to allocation or sell out entirely. This is the clearest example in the gold market of a supply constraint being expressed as a premium rather than as a shortage.
The disruption around the 2020 pandemic made the mechanism visible. Refineries and mints faced operating restrictions at the same moment as retail demand jumped and air freight capacity contracted, and premiums on small units widened sharply while the wholesale market continued to function normally. The metal existed. The capacity to turn it into coins, and to move those coins, did not.
Why smaller units cost more per ounce
This follows directly from the fabrication argument. The cost of making, assaying, packaging and shipping a unit does not scale down in proportion to its weight. A fractional coin takes almost as much handling as a full ounce one, so the same overhead is spread over less metal, and the premium per ounce rises.
The ordering is consistent across the market. Large cast bars carry the thinnest premium per ounce. Smaller bars are dearer. One ounce coins are dearer again, because they are struck rather than cast and carry more quality control. Fractional coins are the dearest of all per unit of metal.
There is a trade off rather than a right answer. Small units cost more per ounce but are easier to sell in part, which matters if you ever want to liquidate a portion rather than the whole holding. Large bars are efficient to acquire and awkward to divide. The choice is about divisibility against cost, and it is worth making deliberately rather than by default.
Premium spikes as a stress gauge
Because the variable component responds to small unit retail demand against fixed capacity, a widening premium is a reasonable measure of retail urgency. When ordinary buyers want physical metal badly enough to pay well above the screen price and to wait weeks for delivery, that is information about sentiment that no survey captures as honestly.
It is particularly informative when it diverges from the screen price. A wholesale market trading calmly while coin premiums blow out says the stress is located in retail demand and fabrication capacity, not in the metal itself. The reverse, a falling spot price with premiums collapsing too, says selling is broad based.
The caveat is severe and worth stating before anyone builds anything on it. Retail urgency is usually highest after a large move has already happened, not before. As a sentiment gauge this reads the present honestly and the future not at all, and the temptation to treat it as a contrarian timing tool should be resisted for the same reason described in why is gold crashing.
The asymmetry when you want to sell
Premiums are not symmetrical, and this is the part that surprises people. You pay the premium on the way in. You do not reliably collect it on the way out.
A dealer buying your coin quotes a bid that is typically below the screen price or only slightly above it, because they now have to hold, verify, possibly re assay, and resell it. If you bought during a premium spike and sell during a calm period, you can lose the entire variable component even though the metal price has not fallen at all.
Practically this means the real cost of physical ownership is the round trip, not the entry premium. Standard, widely recognised products from well known mints hold their resale premium better than unusual ones, because the buyer pool is larger and verification is easier. Products with high premiums justified by collectability carry the most resale risk, since that premium depends on a market in collectors rather than in metal.
Why premiums do not transmit to the screen price
It is tempting to reason that heavy retail buying should push the gold price up. Mostly it does not, and the reason is structural. The wholesale market deals in large unallocated metal between institutions. The retail market deals in small fabricated units. They are connected by refining and minting capacity, and that connection is a bottleneck rather than a conduit.
When retail demand surges, the pressure accumulates at the bottleneck. Premiums widen and queues form. The wholesale market sees a modest increase in fabricator buying spread over the weeks it takes to actually produce the product, which is nothing like the urgency visible at the retail counter.
So the two prices can move independently for extended periods, and the distinction between a wholesale and a retail market for the same metal is worth holding on to alongside the contract differences explained in spot versus futures gold. What you see on the live chart is the wholesale market. The coin shop is a different market with the same underlying.
The honest summary
Premiums are mostly a service charge and partly a capacity signal. The service charge part is stable, explicable and not worth thinking about beyond shopping sensibly. The capacity part is a real time reading of retail urgency that no other data source provides as cleanly.
The failure mode is treating it as predictive. Retail buyers crowd in after big moves, so a premium spike is usually a late signal about something already in the price. Using it as a contrarian trigger assumes retail is always wrong at extremes, which is an assumption rather than a finding, and one that has no reliable timing attached even when it holds.
The one genuinely practical lesson is about your own costs. If you buy physical metal, buy when premiums are normal rather than when they are wide, prefer standard products with deep resale markets, and treat the round trip cost as the real cost. That is a cost discipline rather than a market call, and cost discipline is the part you actually control.
FAQ
Why is physical gold more expensive than the spot price?
Because they are different products. The quoted price is wholesale, unallocated and in large sizes, while a coin or small bar has been refined, minted, assayed, packaged, insured, transported and held in inventory by a dealer. The premium is the sum of those costs plus the margins of everyone involved, so it never reaches zero.
What makes premiums widen sharply?
A surge in demand for small fabricated units against fixed minting, refining and secure logistics capacity. Production cannot rise quickly, so the system responds with higher prices and longer delivery times instead of more product. Premiums widen, popular items go to allocation, and the wholesale market can stay calm throughout.
Why do smaller coins cost more per ounce?
Because the cost of making, verifying, packaging and shipping a unit does not fall in proportion to its weight. A fractional coin requires nearly as much handling as a full ounce one, so the same overhead is spread across less metal. The gain is divisibility, which matters if you expect to sell part of a holding.
Do I get the premium back when I sell?
Usually not in full. Dealers bid below or only slightly above the wholesale price because they must hold, verify and resell the item. Buying during a premium spike and selling in calm conditions can cost you the entire variable component even with an unchanged metal price. Standard widely recognised products retain resale premium best.
Does a premium spike signal that the gold price will rise?
There is no dependable link. Retail urgency generally peaks after a large move rather than before one, and the bottleneck between retail and wholesale markets means the pressure shows up as queues rather than as buying pressure on the screen price. It reads current sentiment honestly and says nothing about what follows.
ⓘ See these ideas on real price: open the free XAUUSD live chart.