Why Most Premium Readings Are Noise
On any given trading day, the premium of a regional gold market against COMEX will move.
It will almost always move.
If you check the Shanghai Gold Exchange premium at 9:00 AM Beijing and again at 3:00 PM Beijing, you will get two different numbers, and neither number is telling you much about the actual state of Chinese gold demand.
The reason is that a premium reading is a ratio of two prices that were captured at slightly different moments, in different currencies, on venues with different microstructures.
Every time either denominator or numerator ticks, the premium ticks too.
Most of what you see when you watch a premium chart hour-by-hour is exchange-rate wobble, contract-roll effects, and end-of-day settlement artifacts.
It is not a signal about supply and demand.
This piece is about separating the signal from the noise.
It is written for a reader who is willing to look at more than one snapshot before drawing a conclusion.
The Three Kinds of Premium Moves
In our experience monitoring COMEX, SGE, MCX, and LBMA daily, premium moves fall into three categories.
Getting the category right is the whole game.
1. Mechanical moves
These are the largest category and the least informative.
They come from things that are not about metal at all.
A 1.5% depreciation of the Indian rupee against the dollar, holding rupee-denominated MCX gold prices constant, will show up as a 1.5% widening of the MCX premium.
Nothing has changed about Indian gold demand.
What has changed is the currency the premium is denominated against.
Similarly, when COMEX rolls from its February contract to April, the active-month price jumps by the calendar spread — often a few dollars per ounce for gold, more for silver.
Every regional premium calculated against that futures contract would shift by the same spread on that day. This is a historical COMEX example, not a description of the site's current USD spot benchmark.
Again, no information about physical demand anywhere.
The tell for mechanical moves is that they are shared across markets.
If SGE, MCX, and LBMA premiums all tighten by roughly the same amount on the same day, first check whether the shared USD spot benchmark or exchange rates moved before attributing the change to regional demand.
2. Flow moves
These are the moves that matter for anyone actually trading physical metal.
They come from real supply and demand hitting the specific market.
Two examples make this concrete.
In the weeks before Diwali or Akshaya Tritiya, India's two largest gold-buying festivals, MCX premiums tend to widen.
Retail jewelers stock up, importers front-run demand, and the physical premium above the international benchmark reflects the queue of buyers.
This can persist for weeks.
The signature is a sustained widening in one market without a matching widening elsewhere.
SGE will not care about Diwali.
The 2020 COVID episode is another example.
When western refineries in Switzerland shut down in March 2020, physical gold delivery in New York became suddenly difficult.
COMEX EFP (Exchange for Physical) blew out.
This was a flow-driven move: real demand for physical metal exceeded the deliverable supply pipeline.
It normalized in weeks, not months, but while it was live it registered as an unusual COMEX-to-London basis.
3. Regime moves
These are the rarest and most important.
A regime move is a change in the underlying rules that link a local market to the global benchmark.
India's 2013 import duty hike from 4% to 10% was a regime move.
So was China's 2015 tightening of gold import licensing.
When a regime move lands, the premium doesn't just spike — it moves to a new equilibrium and stays there.
The tell for a regime move is that the premium's typical range shifts.
Before India's 2013 duty change, MCX premiums might oscillate between -1% and +3%.
After, the same premium might spend most of its time between +6% and +10%.
The whole distribution moved, not just one day's reading.
Practical Rules for Reading a Single Chart
When you look at a premium chart on this site or anywhere else, we recommend running through this checklist before deciding whether the chart is telling you anything.
📊 The Four Questions
- Is the move shared across markets? If SGE, MCX, and LBMA all moved by roughly the same amount on the same day, it's mechanical. Move on.
- Did the local currency move? Cross-reference the FX rate on the same date. A rupee move of 1% will fully explain a 1% MCX premium shift by itself.
- Has the move persisted for a week? A one-day 2% spike is almost always noise. A 2% shift that holds for five trading days is a flow move worth investigating.
- Does the historical distribution look different? If the current reading is inside the past year's range, it's probably not a regime move. If it's outside — three sigma or more — pay attention.
Illustrative Example: Reading a Short SGE Gold Premium Spike
To keep this concrete, here is an illustrative scenario — not a claim about a specific dated observation in the site's history, just a shape that shows how to apply the four questions above.
Suppose SGE gold sits around +1% versus the USD spot benchmark for a January, spikes to something in the +3% to +4% range for three days in mid-February, and comes back down to the +1% area by early March.
Applying the four questions:
- Shared across markets? Assume MCX and LBMA premiums do not move meaningfully on those February days. So the spike was SGE-specific.
- Currency? If CNY was stable against USD across that window, the move was not FX-driven.
- Persisted for a week? No — three days. This puts it in the noise-to-flow border.
- Historical distribution? A three-day SGE spike in that range sits inside the historical band Chinese gold premiums have moved in over recent years. Not a regime move.
Read on this template it looks like a short flow event, plausibly a physical demand pulse related to Chinese New Year post-holiday inventory rebuild — interesting but not signal-worthy for a long-horizon reader.
If it had persisted for three weeks at that level and been accompanied by public reports of tighter Chinese gold import quotas, that would have been a regime move.
What This Site Does Not Try to Do
We do not smooth, average, or otherwise denoise the premium series we display.
This is deliberate.
A tracker that hides day-to-day volatility hides the very thing a reader might want to see.
Our position is that the reader should have access to the raw series and the interpretive framework, and should decide for themselves whether what they are looking at is signal or noise.
What this means in practice: any given day's premium on our dashboard can move by a percentage point without meaning much.
Do not build a trading decision on a single day's reading.
Do build one on a sustained shift across a week or more, in a market that isn't sharing the move with its peers.
When Premium Watching Actually Helps
Premium tracking is a lagging indicator of physical demand, not a leading indicator of price.
It will not tell you whether gold is about to rally in dollar terms.
What it will tell you, when read carefully, is where the physical market is tight and where it is slack.
That is a different kind of information than a price chart provides, and for a physical-metal-focused reader, it can be more useful than yet another COMEX close.
📌 Key Takeaways
- Most day-to-day premium moves are mechanical noise driven by currency wobble and contract rolls, not by supply and demand.
- A move that is shared across SGE, MCX, and LBMA on the same day is almost always mechanical — ignore it.
- Sustained one-market moves that persist for a week or more are flow signals worth investigating.
- Regime moves change the whole distribution of the premium series, not just one day's reading.
- Premium data is a lagging indicator of physical market tightness, not a leading indicator of price direction.
Related Posts
Why We Don't Smooth the Premium Series
A short editorial note on why our daily premium chart shows every wobble rather than a nicer-looking moving average.
Central Bank Gold Holdings by Country: Who Owns What, and What the Trends Show
A tour through the top central bank gold holders as of mid-2026, the buyers and sellers of the past decade, and what these numbers do and do not tell you about future prices.
Gold and Inflation: What Fifty Years of Data Actually Shows
The "gold is an inflation hedge" story is repeated so often it feels like established fact. The historical record is more nuanced. Here is what the data actually says, and how to use it.