Gold

However, the former includes an estimated 754t of ASM, while only 66t was identified as ASM in the COO series. Metals Focus’ data also includes Russian gold mine supply of 331t. Although the country’s 2021 output occurred pre-invasion (and so was still treated by GDL-accredited operations in Russia), LBMA started to collect the COO data in 2022 after the Russian GDL refineries’ LBMA accreditation had been suspended1. As a result, the COO data for 2021 excludes virtually all Russian activity (for both gold and silver)¹.

Combining Metals Focus’ estimates of ASM and Russian activity for 2021 comes to 1,085t. This accounts for 81% of the difference between total Metals Focus and LBMA gold mine production, a gap of 260t. This metal will either consist of gold recovered from concentrates, was treated by non-GDL operations, most likely in Mainland China, or a portion of it may have been classified as recycling. (Metals Focus’ concentrate total for 2021 was round 500t, suggesting that an important share of this was total was still ultimately treated by GDL operations.) Using the above figures (i.e. excluding Metals Focus’ ASM and Russian data) reveals just a 0.6% rise year-on-year, against a 2% drop for GDL operations. The apparent underperformance of the COO dataset may partly reflect market share losses at the expense of non-GDL refiners, which may not be a surprise given that ASM is estimated to have jumped by 12% y/y in 2021 to a record high (as the market emerged from COVID).

Looking at some of the key gold producers can highlight the impact of concentrate and ASM supplies on the two datasets. Starting with South America, there is an apparent disparity for Brazil (MF: 93t; COO: 61t), in part due to Metals Focus’ estimate for 2021 of 18t of ASM and 11t of gold recovered from concentrate (although some of this concentrate may have been treated in Brazil). That said, Metals Focus’ Gold and Silver Doré Flows Service (and so excludes gold and silver recovered from concentrates), which looks at 49 mines and projects in Brazil, reports 55t of gold mine production in 2021, which comes close to the COO LSM figure of 61t.

Staying with South America, the gap in the headline Peruvian mining figures is sizeable (MF: 127t; COO 57t). However, this is explained by Metals Focus’ ASM estimate of 36t and the 31t of gold recovered from gold- bearing doré in 2021. There is also only a modest level of 11t of gold mined in concentrate, some of which may be recovered by non-GDL operations.

Turning to Africa, Burkina Faso also stands out, with marked differences in COO (68t) and Metals Focus (103t) reporting for 2021. However, Metals Focus’ ASM estimate of 34t immediately closes the gap, and it is revealing that no GDL refiners reported taking any ASM from Burkina Faso. On one level, that is not surprising, given ASM treated by GDL operations globally which is reported in the COO survey stood at just 66t in 2021. Even though this compares favourably to the extremely modest 16t reported for 2020, to put these figures into perspective, Metals Focus estimated global ASM gold supply at 754t in 2021.

Staying with supply and looking at mainland China, the world’s largest gold mine producer, the gap between the headline LSM figures is noticeable (MF: 332t; COO: 272t). This, however, may reflect gold that is treated by non-GDL refiners in China. The reported COO figure for recycling in China of 449t for 2021 is also noteworthy, far exceeding Metals Focus’ estimate of 168t. As noted in the Appendix on recycling, the former will reflect the melt of 400oz and kilobars bars, the treatment of process scrap collected from manufacturers, and the recovery of gold from jewellery that has been exchanged for a new piece; none of these three activities are included in Metals Focus’ recycling figures.

Staying with East Asia, the COO LSM data for Indonesia in 2021 stands out, at 55t against Metals Focus’ total of 111t. The latter, however, includes 30t of ASM and 50t of gold recovered from concentrates (mostly from Grasberg). Taken at face value, this suggests that 26t of this concentrate material is handled by non-GDL refiners. Turning to gold scrap supply, although perhaps reassuring that the COO and Metals Focus figures of 30t are almost identical, this is purely coincidental, given the estimated nature of recycling and the fact that a portion of Indonesian scrap supply will be processed by local non-GDL entities.

In the Middle East, there are sharp differences between the COO and Metals Focus datasets. This includes Iran where the latter reports 55t of recycling, none of which has been treated by GDL operations. This is to be expected, as jewellery sold back by consumers in Iran is melted locally and delivered to domestic jewellery and bar manufacturers.

With regards to Turkey, the gap between the COO data of 125t of scrap supply, against 52t reported by Metals Focus, offers a good example
of where the two definitions of recycling vary considerably. The latter captures the melt of jewellery either sold back by consumers or by the trade, which will be refined by the country’s two GDL refiners and also by a large number of small operations across the country (which therefore falls outside the COO dataset). By comparison, the COO data also captures the sizeable melt of old jewellery which is exchanged for a new piece (with the consumer only paying the labour charge). It also includes process scrap collected from jewellery, bar and coin manufacturers. Finally, the COO data will incorporate the melt of gold bars (mostly kilobars, but also some 400oz), to produce jewellery and investment products.

India also features an active gold recycling market, including a significant trade in the exchange of old for new jewellery (which as noted above would be included in the COO data, where it is treated by the country’s one GDL refiner, but not in the Metals Focus gold scrap supply data). In spite of Metals Focus’ far narrower definition of recycling, the consultancy estimated 75t of gold recycling compared with just 11t in the COO data. This is because India is home to a large number of non-GDL refiners, where the bulk of this material would be processed.