What an auction average can tell you

A published auction result makes part of the emerald trade visible. The difficulty begins when a useful observation is asked to do more than it can: price a different stone, describe an entire market or predict a future sale.

Start with one dated result

For its higher-quality rough emerald auction held from 25 August to 11 September 2025, Gemfields reported revenue of USD 32.0 million, all 38 offered lots sold and an average realised price of USD 160.78 per carat. Those are figures from one identified auction, published on 12 September 2025, not a current quotation for emeralds generally. Gemfields’ auction announcement.

Notice how much information belongs beside the average: the date, the seller, the material category and the fact that the emeralds were rough. Remove that context and the number becomes much easier to misuse.

An average is not a price list

Consider a deliberately fictional example. One group of stones weighs 100 carats and sells for 10,000 units of currency. Another weighs 10 carats and sells for 5,000. Together, the average is about 136.36 per carat: 15,000 divided by 110. Yet the two groups realised 100 and 500 per carat respectively. Neither sold at the combined average.

This is arithmetic, not an emerald valuation model. It simply shows why an aggregate figure cannot tell you what every component was worth. It also shows why changing the mixture of goods can change an average without establishing that every comparable item became more expensive.

Read for questions, not shortcuts

Before comparing two releases, our suggested reading note records:

• Date and scope: when the sale took place, the seller and the stated material category.
• Goods offered and sold: both figures, keeping lots distinct from carats.
• Average and currency: the stated price basis, not just the largest number in the headline.
• Missing context: differences in the selections or information that prevents a like-for-like comparison.

For the release above, the note should say higher-quality rough emeralds and September 2025, not simply “emerald prices”. A second release belongs alongside it only after you have compared what each actually measures.

Write down what a release does not tell you as carefully as what it does. A gap in the public information should stay a gap; there is no need to turn it into a confident market narrative.

For an individual emerald, return to evidence about that emerald. Do not multiply its weight by a headline auction average and call the result a valuation. Nor does a successful auction establish a buyer or a sale timetable for another lot. The useful insight is narrower and more durable: understand what was measured before deciding what the number means.