Before the headline, read the definition.

Separate flows from stocks

Mine output describes newly produced silver over a period. Recycling describes silver returning through recovery channels over a period. Above-ground holdings are a stock measured at a point in time. These categories answer different questions and should not be added indiscriminately. Build separate rows for each in a research dashboard. Record what the source includes, because refinery throughput, mine production and recovered scrap are not interchangeable measures of new supply.

Ask what the inventory includes

A vault total has a geographic and institutional boundary. LBMA’s London vault data describe the metal held within that reporting scope, not every ounce in the world. Stored silver also has owners and potentially existing commitments. The inference that all reported inventory is immediately offered for sale therefore does not follow from a holdings total. Distinguish a measured stock from a separate estimate of freely available supply, and keep any assumptions visible.

Keep units and dates attached

Record the original unit, whether a figure means silver content or gross material, the reference period and the publication date. Metric tonnes and troy ounces require a consistent conversion before comparison. An annual estimate and a month-end snapshot should not look like observations of the same interval. USGS data releases include supporting definitions and metadata; use them. Retain the source edition too, so later revisions do not silently overwrite what an earlier chart represented.

Label evidence by its status

Keep reported observations, estimates and forecasts in separate categories. A forecast for next year is not a measurement of supply already delivered. A revision may reflect improved information rather than a sudden physical event. Before sharing a silver headline, ask what changed, against which comparison, and whether the source actually supports the conclusion. This is a method for reading research, distinct from a transaction checklist: a clear dataset can still support several plausible interpretations.

FOLLOW THE EVIDENCE

Sources & further reading

Sources checked 14 November 2026. Reference organisations do not endorse this website. Price examples are hypothetical and educational.