Gold Etf Holdings And Flows Data: Dataset Structure and Field Coverage

Three different records get labelled “gold ETF data,” and they answer different questions. Holdings are the stock of metal inside a trust on a given day - tonnes or troy ounces, plus the NAV that metal backs. Flows are the day-over-day change in that stock across many funds at once, the number headline writers quote when money moves into or out of the asset class. Positioning is a third thing entirely: futures commitments on COMEX, not metal in a vault.

Within Datadory's gold slice - 6 primary datasets plus 11 related ones - only a handful publish the holdings-and-flows pair directly. SPDR Gold Shares (GLD) Fund Data is the fund-level anchor, quality score 9, and the World Gold Council GoldHub data hub (quality score 9) carries the industry-wide aggregation among its 14 named series. Everything else in this guide exists to contextualise those two.

SPDR Gold Shares (GLD) Fund Data is the primary source because it comes from the issuer rather than an aggregator. GLD launched 18 November 2004 as the first US-listed ETF backed by a physical asset, tracks the LBMA Gold Price PM less a 0.40% expense ratio, and the trust's own archive covers every trading day since inception: 5,676 daily rows through 20 August 2026, in a single ~537 KB XLSX.

Read the table as three layers: one fund audited to the bar (SPDR), the whole listed market (Goldhub), and the derivatives book around both (CFTC). Only the first two are holdings-and-flows records in the strict sense.

Tonnes-held tells you how much; the bar lists tell you what exactly. SPDR discloses backing at the individual bar level through an HSBC list - 5,944 allocated bars totalling roughly 2.41 million fine weight ounces as of 21 August 2026, served as a ~120-page PDF - plus a parallel JPMorgan list hosted separately. Every row carries bar number, refiner, gross weight, fine weight, assay, year of manufacture and vault name, with metal held in London, Zurich and New York.

A repeatable pipeline takes under an hour once the files land:

Macro strategists read Goldhub's aggregate flows against the quarterly demand balance to separate ETF investment from bar-and-coin hoarding. Structurers monitor the premium/discount column for dislocation windows. The persona mapping is strongest for quantitative roles: Datadory's best data for investors and quants page orders these sources as a workflow rather than a list.