| Metric | Reading | Basis / as of |
|---|---|---|
| Loading live readings… | ||
When discussing central bank gold purchases, the most frequently cited source is the international reserve composition data that countries report to the International Monetary Fund (IMF). Under this methodology, global central banks made net purchases of +311 tonnes over the trailing 12 months, with the statistics running through August 2026. The raw source is reserve asset detail self-reported by each country's monetary authority, with changes in gold holdings aggregated month by month. This is important to understand: it is not a real-time record of transactions observed in the market, but a summary of official reserve reports, and therefore inherently carries reporting and aggregation lags. For ordinary readers, the first thing to check when seeing the "+311 tonnes" figure is its cutoff date — through 2026-08 — meaning it reflects the reserve structure as of two months ago, not buying and selling happening right now.
The reason the +311 tonnes figure (trailing 12 months, through 2026-08) is "old" lies in three delays built into the data chain. The first is reporting delay: central banks submit reserve data to the IMF on different schedules — some economies report monthly, others quarterly or even less frequently — and revisions may follow. The second is verification delay: after receiving the data, the IMF must reconcile methodologies and standardize formats before publishing. The third is a feature of the methodology itself: net purchases are a rolling 12-month cumulative figure, so any single month's additions or reductions are smoothed across the twelve-month window; even a notable change in one month gets diluted in the rolling reading. Combined, these three delays mean that when you see this monthly data at the end of October, it describes the state of affairs through August — an observation gap of roughly two months.
There is another commonly cited methodology for central bank gold purchases: demand data compiled by the World Gold Council (WGC) based on public statements, central bank annual reports and industry surveys. The differences between the two methodologies are mainly threefold. First, coverage: the IMF methodology relies on official declarations, so unreported or incompletely reported purchases are not counted; the WGC methodology supplements with some public information, so total readings for the same period may differ. Second, update cadence: the IMF methodology updates on official reporting cycles and lags more visibly; the WGC publishes on a relatively fixed schedule, but it too is not real-time data. Third, nature: the IMF methodology reflects changes in holdings on official reserve statements, while the WGC methodology is closer to a demand-side estimate. The +311 tonnes figure cited in this article (trailing 12 months, through 2026-08) belongs to the former. When comparing purchase figures from different sources, the first step should be to verify the methodology and cutoff date, not to compare the numbers directly.
First, check the cutoff date. Any purchase data should state the statistical cutoff month — for example, this article's reading is explicitly "through 2026-08". If an article or chart does not state its cutoff, it should not be used to judge current market forces. Second, distinguish stocks from flows. The +311 tonnes over the trailing 12 months is a rolling cumulative figure reflecting a trend level; it cannot be used to infer marginal changes in any given month. To see marginal changes, you must wait for the single-month detail — which itself requires waiting another reporting cycle. Third, place the purchase data in the context of the current market. As of now, COMEX gold futures are at $3,884.4/oz (2026-10-30 close, daily change -3.69%), SHFE AU is at 831.68 yuan/gram (2026-10-31 night session close, daily change -3.51%), and the driver readings show fund flows at -53, oil price transmission at -50, and Treasury yield trend at +38. These are currently observable market readings, whereas the purchase data describes official reserve behavior from two months ago. The two operate on different time scales and cannot be stitched together into a causal story. We will continue to provide in-depth reviews and updates of the purchase data on this page.
It is a net purchase figure aggregated from the official international reserve composition reported by countries to the International Monetary Fund (IMF), on a trailing 12-month rolling cumulative basis, with statistics through August 2026. It reflects changes in gold holdings on official reserve statements, not real-time market transaction records.
Because the data chain contains three delays: central banks report reserve data to the IMF on different cycles and with revisions; the IMF must verify and aggregate before publishing; and the trailing 12-month rolling cumulative basis itself smooths out single-month changes. Combined, the reading seen at the end of October describes the state through August — an observation gap of roughly two months.
The IMF methodology relies on official declarations by countries; unreported purchases are not counted, and updates follow official reporting cycles, making the lag more visible. The WGC methodology is compiled from public statements, annual reports and surveys, and is closer to a demand-side estimate, with different coverage and publication cadence. Total readings for the same period may therefore differ between the two; before comparing, you must first align the methodology and cutoff date.
No. The reading runs through August 2026 and is a trailing 12-month rolling cumulative figure, reflecting a trend level rather than the current month's marginal change. To judge the current market, refer to real-time readings such as COMEX gold at $3,884.4/oz (2026-10-30 close) and SHFE AU at 831.68 yuan/gram (2026-10-31 night session close).
Rolling 12-month readings (such as +311 tonnes) are suited to viewing medium- to long-term trends, since single-month additions or reductions are diluted by the twelve-month window. To see a given month's marginal change, you must wait for the single-month detail — which itself requires waiting another reporting cycle. The two operate on different time scales and should not be mixed.
First verify three things: the data source methodology (IMF-reported basis or WGC demand-side estimate), the statistical cutoff date, and whether the figure is a rolling cumulative or a single-month value. Only after aligning methodology and timing are the numbers comparable; otherwise, comparing magnitudes directly can easily lead to wrong conclusions.
Yes, but for the right purpose. Purchase data is suited to assessing the official sector's medium- to long-term allocation trend, not to explaining short-term price moves. The current COMEX gold daily change is -3.69% (2026-10-30 close); short-term volatility is mainly driven by current-period factors such as fund flows (reading -53), oil price transmission (-50) and the Treasury yield trend (+38), which are not on the same time scale as reserve data from two months ago.
The conversion constants used on this site are: 1 troy ounce = 31.1035 grams; 1 gram = 0.0321507 ounces; 1 kilogram = 32.1507 ounces. Multiplying a USD/oz price by 31.1035 and dividing by the CNY exchange rate gives an approximate yuan/gram price.
Usually there is no clear one-to-one reaction. On one hand, the data itself lags by about two months, during which the market has already digested a large amount of current information; on the other, rolling cumulative figures change gradually and carry weak marginal signals. It is better treated as a medium- to long-term background variable than as a short-term trading signal.
You can continuously track updates to readings such as gold prices, rates, the dollar index and central bank purchases on the xaudaily.com homepage; this page will also provide in-depth reviews and follow-up updates on the purchase data.