Silver supply demand statistics in Moz: the balance, unit by unit

"Moz" is how silver keeps its books: million troy ounces, the unit the metal physically settles in. Two named products carry the balance in that unit, and they disagree on purpose - one holds the industry's current-year ledger, the other its seventy-five-year archive.

Silver Institute - Silver Supply & Demand statistics is the current layer, scored 7/10 on Datadory's rubric. The Silver Institute compiles it with Metals Focus as research partner, and it runs four panels - mine production, recycling, silver demand, jewelry and silverware - one survey year per cycle, every level carrying its year-over-year move beside it, closed out by top-20 producing-country and producing-company rankings led by Mexico.

Silver Institute Market Statistics (World Silver Survey) is the archive twin, scored 8/10: every Survey edition from 1991 through 2026 plus a combined volume covering 1950-1990, thirty-seven editions whose rows share one shape. Both arrive as typed rows rather than prose panels, which matters more here than anywhere else in metals - roughly fifty numeric values define a survey year, so the work is joining them honestly, not extracting them.

The demand-by-application split is the only one in Datadory's twelve-record silver pool: industrial, photovoltaic, jewelry, silverware, photography and physical investment arrive as separate lines, ready-made weights for market-sizing work.

Flat, typed, one row per supply or demand line item per survey year - exactly as a delivery lands:

# Supply & Demand summary - 2024 survey year, million troy ounces
supply_category=Mine production      volume_moz=819.7   yoy_change_pct=+0.9
supply_category=Recycling            volume_moz=193.9   yoy_change_pct=+6.0   note=12-year high
demand_category=Total demand         volume_boz=1.16    yoy_change_pct=-3.0
demand_category=Jewelry fabrication  volume_moz=208.7   yoy_change_pct=+3.0
demand_category=Silverware           volume_moz=54.2    yoy_change_pct=-2.0

# Leading producers - order carried as data, not prose
rank_1=Mexico   rank_2=China   rank_3=Peru   rank_4=Bolivia   rank_5=Chile

# The punchline row
market_balance_moz=-150   # total supply minus total demand; negative reads as a deficit

Seven fields carry the summary layer: supply_category for mine production, net hedging and recycling; demand_category for the fabrication and investment lines; volume_moz holding the quantity; yoy_change_pct riding beside it rather than derived later; country_or_company and producer_rank driving the two league tables; and market_balance_moz closing the ledger. Definitions are written against the printed table structure and verified against live records when a sample is cut. Finer segment cuts - photovoltaic and electronics detail, hedging lines, historical survey-year series - fold under additional fields on request rather than being promised blind.

To 1950, and the depth is the point. Thirty-seven editions sit in the archive - every World Silver Survey from 1991 through the 2026 edition plus a combined volume covering 1950-1990 - and because the rows share one shape across all of them, a seventy-five-year panel assembles as a union rather than a translation project. Each publication cycle reports the year just closed, so the current edition carries calendar-2024, while the surrounding narrative stretches back to roughly 3000 BCE.

A volume story wants a price axis under it. World Bank Commodity Markets (Pink Sheet) supplies it at monthly grain from January 1960 - 805 observations - in nominal dollars and constant-2010 dollars deflated through the MUV index, scored 10/10, the only perfect score in the silver pool. Pair the two and the series gains a real-price axis that survives comparison across decades without stitching vintages.

Because two compilers measure one market on different conventions. Federal mineral statistics state silver content in metric tons; the industry states it in million troy ounces because that is how the metal trades. One metric tonne equals 32,150.7 troy ounces, so USGS Mineral Commodity Summaries 2025 - Silver (scored 9/10) and the Institute ledger count the same ounces at different scales.

Converted at that rate, the two views agree with each other almost embarrassingly well:

Mexico's converted 202.5 Moz sits consistently with the ledger ranking it first against a world total of 819.7 Moz - about 25,500 tonnes. Three cautions keep the conversion honest. First, match like with like: federal figures are silver content, while some ledger lines are gross-weight categories. Second, the programs estimate independently - the chapter cites the World Silver Survey for global consumption but compiles its own country totals, so expect small gaps at the decimal level rather than identical figures. Third, treat estimate-flagged cells as soft: each revises in the following January edition, so pin the vintage before quoting a year.

A balance model rarely needs one product; it needs four, keyed honestly. The ledger owns demand by application, the archive owns history, the federal pair owns physical supply and reserves, and the trust owns what investors actually hold.

iShares Silver Trust (SLV) Daily Holdings deserves the emphasis: about 5,100 daily fund rows since April 21, 2006 - ounces in trust, NAV per share struck against each day's announced LBMA Silver Price, shares outstanding, premium/discount - plus a bar list enumerating roughly 503,600 individually vaulted bars across London and New York. The survey weighs physical investment once a year inside total demand; the trust measures it day by day, which makes the two the natural join for flow work. Scored 8/10 alongside a head-to-head SLV versus Silver Institute comparison.

• Investors and quant researchers work the deficit arithmetic directly - 1.16 billion ounces of demand against 819.7 mined and 193.9 recycled leaves a structural gap worth sizing against price and ETF flows. See silver data for investors and quants.

• Market researchers and consultants adopt the fabrication split as ready-made demand weights, quotable to one consistent series with the change column attached. See silver data for market researchers.

• Competitive-intelligence and product teams read producer concentration off the ranked tables: five countries account for the bulk of mine supply, and the company table beneath them shows who moves when a jurisdiction shifts.

• Journalists, academics and students cite figures the industry itself treats as canonical - compiled by its trade association with Metals Focus and attributed edition by edition, so every number traces to its source without archaeology.

The seams are the hard part, and they are handled before delivery. Units stay beside values so million troy ounces never masquerade as metric tons. Change columns ride their levels. Compiler and edition attribution attach to every row, so a figure cites its survey year without investigation. Each new cycle lands as new rows under the same fields - no re-integration project when the next survey prints.

Name the years, the line items and the channel when you request a sample and it arrives cut to exactly that scope - the current-year summary, a historical window, or the whole seventy-five-year panel - with the field dictionary, coverage statement and sample yours regardless of what happens next.

This page is the Moz-balance thread of a twelve-record silver shelf. Continue with:

1. Precious metals & minerals data guide - the pillar covering all 32 pooled records across the wider metals complex.
2. Silver supply demand data in Excel - the same balance from the spreadsheet angle, workbook-shaped.
3. LBMA silver price fix history - the noon print that prices what the balance quantifies.
4. Best silver datasets - the whole pool ranked on one quality rubric.
5. USGS Mineral Commodity Summaries vs USGS hub - when the annual chapter beats the monthly feed for production work.

For the structured index behind all of it, the silver data hub lists every record with coverage chips, granularity and use cases - samples ship first.