Ronan Manly
BullionStar Blogs
Ronan Manly
Posted on 23 Nov 2015 by

From Good Delivery bars to Kilobars – The Swiss Refineries, the GFMS data, and the LBMA

In early September 2015, I wrote an article titled “Moving the goalposts….The LBMA’s shifting stance on gold refinery production statistics”, in which I explained how the London Bullion Market Association (LBMA) had, on Wednesday 5 August, substantially lowered its 2013 gold and silver refinery production statistics literally a few days after I had commented on the sizeable figure of 6601 tonnes of 2013 refined gold production that the LBMA had previously published in May 2015.


  • On 5 August, the LBMA substantially altered and republished Good Delivery List gold and silver refinery production statistics in two of its published files: LBMA Brochure Final 20120501.pdf and LBMA Overview Brochure.pdf
  • For gold, the alterations were most pronounced in the 2013 refined production figure which was reduced from 6601 tonnes to 4600 tonnes, i.e. a 2001 tonne reduction
  • Other years’ figures for refined gold refinery output (2010-2012) were also reduced, with the 2008-2009 figures being increased
  • As part of the update, the LBMA linked its amended figures solely to GFMS estimates of gold mining and scrap output,  adding the words ‘estimated to be‘ in front of the 4,600 tonnes figure, and the words ‘owing to recycling of scrap material‘, thereby framing the revised figure solely in terms of scrap gold in excess of 2013 gold mining supply. This use of GFMS data is bizarre because all refiners on the LBMA’s Good Delivery List provide exact refinery production statistics to the LBMA Executive as part of the LBMA Pro-Active Monitoring programme, so there are no need to reference estimates from external data providers
  • In the updated versions of the brochures, the LBMA made no reference to why the gold figures had been reduced, nor what the original figures referred to, particularly for the huge difference of 2,000 tonnes of gold refinery output in 2013 between its two sets of figures
  • By 12 August, the LBMA had again updated its 2013 gold refinery output figure to 4579 tonnes

In my Part 1 article, I had concluded that:

“There are 2,300 tonnes of 2013 gold refining output in excess of combined mine production and scrap recycling being signalled within the  6,601 tonnes figure which was removed from the LBMA’s reports on 5 August 2015.

Could it be that this 6,601 tonne figure included refinery throughput for the huge number of London Good Delivery gold bars extracted from gold ETFs and LBMA and Bank of England vaults and converted into smaller gold bars in 2013, mainly using LBMA Good Delivery Swiss gold refineries? And that maybe this 6,601 tonne figure stood out as a statistical outlier for 2013 which no one wanted to talk about?”

Note that for 2013, Gold Field Mineral Services (GFMS) estimated gold mining production to be 3,022 tonnes, and gold scrap supply to be 1,280 tonnes for 2013, so in total GFMS estimated gold mining + scrap supply at 4,302 tonnes in 2013. Therefore, the LBMA’s original figure for 2013 gold refinery production of 6,601 tonnes exceeded the combined GFMS mine and scrap supply by 2,300 tonnes.

Whose interests are served by replacing actual refinery output figures with far lower estimates comprising GFMS gold mine production and scrap recycling data? What happened to the third major source of gold supply to refineries during 2013, i.e. London Good Delivery gold bars, and why won’t the LBMA reference this? Why would the LBMA go to great lengths to de-emphasise the huge volume of Good Delivery gold bars being sent to gold refineries (especially in 2013) for conversion into 9999 fine kilobars, when its obvious for all to see that this huge migration of bars happened?

This article, which is Part 2 of the analysis into the LBMA’s 2013 gold refinery statistics, looks into this 6,601 tonne number and the 2,300 tonne delta compared to GFMS estimates, specifically examining the mountain of evidence that highlights the huge volume of Good Delivery bars that were processed through the Swiss gold refineries in 2013, and the huge associated shipments of gold from the UK to Switzerland, and onward from Switzerland to Asia.

Part 2 also looks at the extent to which GFMS and the World Gold Council, through their report text and data, addressed, and did not address, the non-stop processing of Good Delivery gold bars into smaller finer kilobars during 2013.

When Part 1 was written, I had also planned that Part 2 would examine the 2013 gold withdrawals from the London-based gold ETFs, and the 2013 withdrawal of gold from the Bank of England, however, these topics were subsequently addressed in a separate piece titled “How many Good Delivery gold bars are in all the London Vaults?….including the Bank of England vaults“.

That article itself had found a lot of interesting information including:

  • that the entire London LBMA vault network (including the Bank of England) lost 1,500 tonnes (120,000 bars) between 2011 and early 2014, shrinking from 9,000 tonnes to 7,500 tonnes
  • Between the end of February 2013 and the end of February 2014, the amount of gold in custody at the Bank of England fell by 755 tonnes
  • In 2013, the large physically-backed gold ETFs which store their gold in London saw a 720 tonne outflow of gold (GLD 561, IAU 60, ETF Securities’ PHAU 52, ETS Securities GBS 42, ‘Source’ Gold 31)
  • The full set of gold ETFs storing their gold in London can, nearly down to the exact tonne, account for all of the LBMA vaulted gold held outside the Bank of England vaults (See  start of my article titled “Central bank gold at the Bank of England” for an explanation of this)

Note: Deutsche Bank gold ETFs and an ABSA gold ETF also store their gold in London, and during 2013, these 2 sets of ETFs lost approximately a combined 12 tonnes of gold (~9 tonnes from Deutsche and ~3 tonnes from ABSA, so this would increase the 720 tonne ETF loss above, to about 732 tonnes.


Yet another Change to the LBMA Brochure in September 2015

On 29 September 2015, the LBMA made a further alteration to the 4-page LBMA Overview Brochure, the brochure that had featured the shifting gold and silver refinery output statistics.

On this occasion, although the data in the table remained unchanged, some unusual footnotes were added underneath the table of refining statistics. The text, table and the new footnotes are as follows:

LBMA brochure refining Sept 2015 text

The footnotes are highlighted as per yellow box:

LBMA brochure refining Sept 2015 footnotes and table

Let’s look at these 3 footnotes one by one.

Note 1): The data for 2008-2013 contains estimates which will be updated when actual data becomes available.

This note is illogical, since the LBMA already has all of the exact data of gold and silver output per refinery. This was stated in the previous versions, and it’s all detailed in my previous article.

Also, specifying ‘Figures correct as at September 2015’ is illogical since the LBMA states that the data is ‘estimates’ and not ‘actual data’. Correct relative to what? How can ‘estimates’ be deemed to be correct if the ‘actual data’ is not published?

That would also explain the bizarre note number 2.

Note “2) Refined production should include only the refinery’s output that has gone through a refining process”.

Footnotes to tables are normally used to explain data, not to justify the data. This Note 2 sounds more like a pronouncement or a direction from a LBMA communication to the refineries rather than an explanatory footnote.

In English grammar, ‘Should‘ means to give advice, a recommendation or a suggestion, and to express obligation or expectation. This footnote looks like it has been lifted out of a directive from the LBMA to the member refineries.

Converting a 995 fine Good Delivery ~400oz bar into a series of 999 kilo bars does involve a a chemical refining process in addition to melting and pouring. The transformation by the refineries of large bars into smaller bars is still throughput, and is a refinery process (as you will see below).

Also problematic to the LBMA’s footnote is that converting 9999 fine scrap (in the form of old bars) to new 9999 bars, which sometimes happens, would not necessarily be captured in the above LBMA footnote, so this approach to seemingly attempt to tie in the LBMA data to GFMS mining and scrap refining data opens up a can of worms.

Note 3): the production of newly accredited refiners excludes production in the years prior to accreditation.

Note 3 should be obvious, and besides, it wouldn’t change much in terms of the huge gaps in the numbers between 6601 tonnes in 2013, and the GFMS figure of 4302 tonnes.


Macquarie 2013 – Where has the ETF gold gone

In August 2013, Macquarie Commodities Research, in its report “Where has the ETF gold gone” commented that:

“over 1H 2013 it [the UK] has exported 797 tonnes [of gold], equivalent to 30% of annual gold mine production”

“…gold bars from ETFs have gone to Switzerland, where most of the world‟s gold refining capacity is, to be remelted into different size bars and coins and then sold on end consumers, predominantly in Asia, specifically China and India.

“Trade data also backs up this movement of gold – Hong Kong customs reported imports of gold from Switzerland of 370t in 1H 2013, up 284t on 1H 2103 (fig 3), while Indian imports from Switzerland appear to have risen by more than 100t YoY.

It is not really very surprising that the gold has found its way from vaults in London (and most likely the US and Switzerland) to Asia via Swiss refineries. We have repeatedly noted that gold ETFs are part of the physical gold market and if investors don’t want the gold it has to go somewhere else.”

Since the four large Swiss gold refineries account for the lions share of worldwide annual gold refinery output (See my article “Swiss Gold Refineries and the sale of Valcambi“), its important to examine what the Swiss gold refineries had to say about the smelting of London Good Delivery gold bars into smaller bars in 2013, as well as their comments about the dramatic reduction in gold scrap coming into the refineries during that time.

Note that London Good Delivery gold bars are variable weight bars that weigh about 400oz each (12.5kgs). These are the standard type of gold bars stored in central bank vaults and held in physically backed gold Exchange Traded Funds (ETFs) such as the SPDR Gold Trust (GLD).

 Swiss Gold Imports from the UK: 2013

In 2013, Switzerland imported more than 2,600 tonnes of gold and exported approximately 2,800 tonnes of gold. That year’s gold import and export totals were the highest ever annual totals recorded for Switzerland. See chart below from Nick Laird’s Sharelynx.

Although Switzerland doesn’t possess any major gold mines, it does host one of the largest physical gold markets in the world, which regarding investment gold, primarily comprises the large Swiss gold refineries along with some bullion banks (including UBS and Credit Suisse), the Swiss National Bank and the Bank for International Settlements (BIS), and the Swiss wealth management and private banking sector. But the throughput and precious metal processing of the four large gold refineries accounts for nearly all the country’s gold imports and exports.


The UK is consistently the largest import source of gold into Switzerland. In 2013, Switzerland imported nearly 1,400 tonnes of gold from the UK during the year, with hardly any gold moving back in the opposite direction. Notwithstanding the fact that the UK does not have any producing gold mines, 1,400 tonnes is 46% of GFMS 2013 global gold mining production estimate of 3022 tonnes. And despite the fact that GFMS itself stated that the UK only contributed 41 tonnes of gold scrap to the 1280 tonne global gold scrap total in 2013, 1400 tonnes of UK gold exports to Switzerland is 109% of GFMS’s 2013 global gold scrap estimates.

So why is the LBMA not including all of this 1400 tonnes of UK to Switzerland gold exports in its 2013 gold refinery production statistics?


Even Swiss gold imports from the United States in 2013, at 267 tonnes, paled into comparison compared to Switzerland’s imports of 1,373 tonnes of gold from the UK, and left all other import sources such as Italy and France in a distant third.


Thanks to Nick Laird of Sharelynx for permission to use the above 3 charts.

Swiss Refineries – From the Horses’ Mouths

Let’s look at what the Swiss gold refineries had to say about the conversion of Good Delivery gold bars into smaller bars during 2013. You will see that the large Swiss refining companies treat Good Delivery bars as one of three sources of supply coming in to their refineries.

It’s important to note that the transformation of London Good Delivery bars of 995 fineness into, for example, kilobars of 9999 fineness, still involves the use of chemicals in reactions, albeit smaller amounts than when refining mining ore, and is not just a simple melting and re-casting exercise.

Argor-Heraeus’s perspective on Good Delivery bars in 2013

In its 2013 Corporate Sustainability Report, Argor-Heraeus had the following comments to say about the 400oz bar to smaller bar transformations:

In 2013, we consumed 3,120,603 kg of chemicals, 4% less than in 2012, despite a slight increase in precious metals processing. This decrease derives from the fact that a large percentage of gold processing involved the re-smelting of metal already in circulation (Good Delivery) to obtain high-fineness ingots, which are in great demand. The processing of a metal that is already pure requires smaller amounts of chemicals in reactions, as opposed to the refining of raw materials from mines.”

Argor-Heraeus even divides the gold inputs that go into its refining process into three distinct categories, namely, a) Scrap, b) Mines and c) Good Delivery, such is the importance of the Good Delivery refining activity to the refinery. See the following graphic from the Argor-Heraeus 2013 Sustainability report, complete with descriptive icons of the three input sources inputs of metal:

AH flow


The text box from the left-hand corner of the above graphic has been zoomed in and magnified below to aid readability:

AH text

Elsewhere in the same report, Argor-Heraeus reiterates the same 3 sources of gold supply that come in to its refineries for ‘Transformation and Processing‘.

AH graphic

Argor-Heraeus picks up the Good Delivery bar theme again in its 2014 Corporate Sustainability Report, where it produces a similar but slightly more detailed graphic, complete with the icons, and which explains that the Good Delivery bars can be either ‘grandfathered or non-grandfathered‘ and that the materials are ‘already certified Good Delivery, or already high-quality‘. High quality but not good delivery could be signifying gold bar brands on the former London Good Delivery list, or else lower grade coin bars, that had originally been made from melting down and casting into bars the gold coins that were previously  in circulation. Coin bars were at one time on the London Good Delivery list up until 1954.

Grandfathered is a term used by the LBMA in its discussions of ‘Responsible Gold Guidance‘ and is defined as:

Grandfathered Stocks: Gold investment products (ingots, bars, coins and grain in sealed containers) held in bullion bank vaults, central bank vaults, exchanges and refineries, with a verifiable date prior to 1 January 2012, which will not require a determination of origin. This includes stocks held by a third-party on behalf of the listed entities.

 The Argor-Heraeus 2014 graphic referencing Good Delivery bars is as follows:

AH 2014 graphic1


Metalor’s information on Good Delivery bars in 2013

In its 2013 Annual Report (large file 3.4 MBs), within the review of 2013 performance section, large Swiss based gold refinery Metalor Technologies highlights a steady demand for ‘recasting of gold bars for banks’:

“Full-year net sales in the Refining business unit declined by 16 percent as precious metal prices remained low, reflecting a weak global economy. The drop in prices negatively impacted the price/volume mix, as reduced quantities were retained at lower prices. This was partly offset by steady demand in less profitable activities, such as the recasting of gold bars for banks.”

Metalor also provided a host of pertinent insights into other drivers of the 2013 gold market:

“The spot-price of gold and silver declined by more than 30 percent over a six-month period, and this prompted sharp sell-offs of the gold stored in ETF (Exchange-Traded Funds) vaults. The consensus is that this surplus was absorbed by strong China based bullion purchases, while price-dependent scrap flow fell rapidly.”

High grade precious metal bearing scrap flows worldwide dropped sharply due to sustained price erosion. This market development created an overhang in refining capacity, and a much more competitive pricing environment, although some of the volume reduction in scrap flows was offset by new mining doré contracts. The drop in price led to strong bullion purchases, mainly driven by China.”

The Refining business unit saw a challenging 2013, due to reduced gold prices. This resulted in a continuous slowdown in the scrap market. …….a decreasing volume of mining doré coming from abroad, due to changes in country regulations.”

“In Asia, the Hong Kong refinery was able to sustain a high level of activity due to strong demand and a high premium on bullion products.”


Valcambi on refining of Banks’ gold

Valcambi has an annual refining capacity “in excess of 1,200 tons for gold and 400 tons for silver“, so is known for having potentially unused refining capacity.

Following the July 2015 Valcambi acquisition by Indian company Rajesh Exports, the acquirer clarified to Indian newspaper ‘Business Standard’ that it was a regular activity for Valcambi to use its excess capacity to meet “emergency” refining requirements for gold held by bullion banks.

In fact, on the recently updated Valcambi website, an entire web page is now devoted to describing how transportation works for banker clients, in addition to clients that are miners, scrap dealers, other refineries, and watch makers. See ‘Transportation for Bankers‘ web page which details the import and exports procedures which the Valcambi refinery offers its banker clients.

Valcambi bankers

Valcambi 1

Under its Assaying web page, Valcambi even sees fit to specifically explain the process for the incoming ‘shipments of Good Delivery (GD) bars‘ which are merely checked to confirm that they haven’t been tampered with, as opposed to the shipments of ‘Non Good Delivery (NGD) precious metals‘, which are subjected to homogeneity checking, sampling and analysis. This shows that the volume of Good Delivery bar shipments into Valcambi is significant enough to warrant specific coverage on its website.

Valcambi Assaying

Valcambi good delivery

Under its Refining web page, Valcambi again details its ‘3’ sources of incoming gold, namely “primary doré supplied by mines”,  “industrial scrap and recycling“, and “metals invested and owned by financial and governmental institutions“, i.e. London Good Delivery Bars.

Valcambi refin

Valcambi refining

On the phenomenon of a low gold price leading to a decline of gold scrap coming into Valcambi, the CEO, Michael Mesaric, recently had the following to say while talking with Indian publication Bullion Bulletin at the India International Gold Convention (IIGC) 2015 in Goa:

Bullion Bulletin: The gold price is coming down continuously, is there any impact on the refinery segment?

Michael Mesaric: There is a small impact as well because if the gold price is very low there is very little scarp coming in.”


Argor-Heraeus interviews –They’re bringing in good delivery bars”

On 4 December 2013, Alex Stanczyk from Anglo Far-East group, in an interview with Koos Jansen published on his BullionStar blog, said that he (Stanczyk) and colleague Philip Judge, accompanied by Jim Rickards, had just returned from a visit to Switzerland where they had met with the managing director of one of the large Swiss refineries. Although the identity of  the refinery was not revealed, Alex Stanczyk said that the refinery MD informed them that there was huge demand for fabrication at his refinery and that:

“They put on three shifts, they’re working 24 hours a day, and originally he (the MD) thought that would wind down at some point. Well, they’ve been doing it all year. Every time he thinks its going to slow down, he gets more orders, more orders, more orders. They have expanded the plant to where it almost doubles their capacity. 70 % of their kilobar fabrication is going to China, at a pace of 10 tons a week.”

They’re bringing in good delivery bars, scrap and doré from the mines, basically all they can get their hands on.”

“…sometimes when they get gold in, it’s coming from the back corners of the vaults. He knew this because these were good delivery bars marked in the (nineteen) sixties.”

The same Swiss gold refinery executive was interviewed by Jon Ward of the Physical Gold Fund in September 2015, with the interview published as a podcast and as a transcript.

Jon Ward: In 2013, I recall you commented on the tightening of physical supply in the gold market and even the difficulties you were having in sourcing material. In fact, as I remember, you remarked that in 30 years, you’d never seen anything like it.”

The exact identify of the Swiss refinery executive was also not revealed in the September 2015 interview, however the executive is most certainly from the Argor-Heraeus refinery. Why? Because, the introduction to the 2015 interview states that:

“The gentleman we are interviewing  is part of senior management of one of the largest Swiss refineries.  His refinery is one of only 5 global LBMA referees…”

The LBMA appoints 5 refinery assay laboratories to help it to maintain the Good Delivery system. These appointees are known as ‘Good Delivery Referees’ and they meet on a quarterly basis at the LBMA. The 5 Good Delivery Referees are Argor-Heraeus, Metalor Technologies and PAMP (all from Switzerland), Rand Refinery (South Africa), and Tanaka Kikinzoku Kogyo (Japan).

Therefore, the interviewee has to be from one of three Swiss refineries, namely, Argor-Heraeus, Metalor or PAMP.

Furthermore, and this is the critical point, during the interview, the refinery executive states that his company has just opened in Santiago, Chile.

“Head of Refinery: ..looking at mining partnerships, we are expanding in Latin America. We have just opened in Santiago, Chile, and are trying to provide even more competitive services for the Latin American mining industry.”

Out of the short-list of Argor-Heraeus, Metalor, and PAMP, the only one of the three to open an operation in Santiago, Chile in 2015 (and the only one of the three to even have an operation in Chile) is Argor-Heraeus. See Argor-Heraeus new item below from the news page of its website dated 16 September 2015:

AH Santiago

The press release for the above Chilean plant announcement is only in Italian, but can be read here.

Lets look at what the Argor-Heraeus refinery executive says about conversion of Good Delivery bars to kilobars, both in 2015 and during the few years prior to that. From his September 2015 interview:

Jon Ward: Over the last couple of years, has this meant that you actually had to melt down and re-refine a whole lot of 400-ounce bars for China? If you have, I’d like to know where the bars come from.

Head of Refinery: The bars are coming from what you could call “the market.” Looking back, there were all these ETF liquidations, and the ETFs were holding bars in the form of 400-ounce bars. At that time a lot of the physical liquidity maintained in the London gold market was actually in 400-ounce large bars. The final customers were not interested in 400-ounce bars, so it was one of our jobs to take these bars, melt them down, refine them up to the 999.9 standard, and cast them into kilo bars.

Jon Ward: Were a whole lot of these bars coming from London?

Head of Refinery: Regarding the ETF liquidations, this gold had to go somewhere, and that was all converted. This is a thing you see every year. You also see some liquidations of physical gold held with COMEX and NYMEX. More or less, these are the sources of gold other than newly mined.

 PAMP – Three Shifts and Full Capacity – Barkhordar

In January 2014, in an article titled “Gold Flows East as Bars Recast for Chinese Defying Slump“, Bloomberg highlighted that the PAMP refinery, owned by MKS (Switzerland) SA, was at full capacity during parts of 2013,  and the article quoted PAMP Managing Director Mehdi Barkhordar as saying that they had to add production shifts to cope with processing demand:

“Gold’s biggest slump in three decades has been a boon for MKS (Switzerland) SA’s PAMP refinery near the Italian border in Castel San Pietro, whose bullion sales to China surged to a record as demand rose for coins, bars and jewelry.”

To keep up with orders, MKS added shifts at the PAMP refinery, located about 4 miles (6.4 kilometers) from the Italian border, Barkhordar said in November…”

Furnaces that can process more than 450 tons a year were at full capacity from April to June, melting mined metal, scrap jewelry and ingots at 1,000 degrees Celsius (1,832 degrees Fahrenheit) into the higher purities and smaller sizes favored by Asian buyers.”

“The surge in orders meant some parts of the refinery worked three shifts instead of the usual two, Barkhordar said.”

Again, you can see that there were three sources of supply for the PAMP refinery in 2013, i.e. mining, scrap and ingots (bars). According to GFMS, global scrap gold supply fell by 354 tonnes (21%) from 1634 tonnes in 2012 to 1280 tonnes in 2013, so this did not account for the ‘surge in orders’ and the need to add extra refinery shifts. Likewise, global gold mining output only increased by 160 tonnes (5%) from 2860 tonnes in 2012 to 3022 tonnes in 2013, and much of this increase was in China, Russia, Australia, Kyrgyzstan, and Indonesia which refine their own gold domestically, so this would also not explain the surge in orders, which therefore can only be attributable to recasting existing large gold bars into “smaller sizes favored by Asian buyers“.

Therefore, all 4 of the 4 large Swiss gold refineries are on the record that London Good Delivery gold bars were a very significant source of gold supply into their refineries during 2013 and even since then. So why did the LBMA amend its 2013 gold refining production statistics and seek to purely link its revised ‘estimate’ numbers to GFMS estimates of gold mine supply and gold scrap supply? There is an entire third source of gold supply to the refiners being overlooked because the LBMA dramatically reduced its 2013 gold refining production figure of 6,601 tonnes. Classifying Good Delivery bars as a supply source for refining is as legitimate as classifying gold scrap as a supply source for refining, and both come from above ground gold stocks.


GFMS and the World Gold Council

The well-known gold research consultancy GFMS, as well as gold mining lobby group the World Gold Council, between them produce a number of gold supply and demand reports each year. [Note: GFMS, formerly known as Gold Fields Mineral Services, is now part of Thomson Reuters].

Each year GFMS publishes a gold survey and related update reports later in the year. In 2013, this GFMS gold survey included two update reports. The 2013 survey and its updates were sponsored by Swiss refiner Valcambi and Japanese refiner Tanaka, with ‘generous support‘ from a selection of entities including Swiss refiner PAMP (part of the MKS Group),  South African refiner Rand Refinery, US gold mining companies Barrick and Goldcorp, bullion bank Standard Bank, US futures exchange CME Group, and the gold mining sector backed World Gold Council.

Its notable that the GFMS reports are ‘sponsored’ by some of the large Swiss gold refiners, yet there is nothing in the GFMS reports that puts cold hard factual numbers on the amount of Good Delivery bars processed through the refineries. As you will see below, GFMS mentions the good delivery bar processing in passing in its text, but not in its 2013 gold supply-demand ‘model’.

What, if anything, did GFMS have to say about conversion of London Good Delivery gold bars into smaller gold bars, such as kilobars, during 2013?

In its GFMS Gold Survey for 2013 – Update 1 (large file 11MBs) report, published in September 2013, the report states that:

“Strong trade flows were recorded between the UK and Switzerland, where Good Delivery metal was refined to smaller bars and shipped to India and China.”

The GFMS Gold Survey for 2013 – Update 2 (large file 9.8 MBs), published in January 2014, reiterated this point about large bar to small bar refining. On page 5 of the Update 2 report it states:

The duality of disinvestment in the developed world and an increase in physical demand from Asia was witnessed by the largest movement of gold, by value, in history as bars were shipped to Asia, often being melted down into smaller bars en route.

Notice that not all Good Delivery bars were converted to smaller bars before shipment to Asia. Some shipments went straight to Asia without being melted and converted.

And on page 9 of the same Update 2 report, the source of some of these smaller bars is given, i.e. the source was UK ETF gold holdings:

As a consequence, UK-led ETF outflows found their way to Switzerland, where refiners melted the metal into smaller bars, and shipped them East, in order to satisfy the surge in demand.”

The World Gold Council (WGC), regularly issues its own gold supply demand reports called ‘Gold Demand Trends‘, and publishes these reports in the form of an annual version, followed by shorter quarterly updates. In ‘Gold Demand Trends Q3 2013’, published in November 2013, the WGC said:

“Gold continued to work its way through the supply chain, to be converted from London Good Delivery bar form, via the refiners, into smaller Asian consumer-friendly kilo bars and below. This process is borne out by recent trade statistics. Data from Eurostat show exports of gold from the UK to Switzerland for the January – August period grew more than 10 fold to 1016.3 tonnes. This compares to a total of just 85 tonnes for the same period in 2012.”

In its Full Year 2013 edition of ‘Gold Demand Trends’, published in February 2014, the World Gold Council had this to say about the London Good Delivery bar shipments going to refineries, being transformed into smaller bars, and then recommencing their onward journey to the East:

No review of 2013 would be complete without a mention of the unprecedented flow of gold from western vaults to eastern markets, via refiners in North America, Switzerlandand Dubai.”

These shifts resulted in the shipment and transformation – on an epic scale – of 400oz London Good Delivery (LGD) bars into smaller denominations more suitable for consumers’ pockets.”

Notice the reference to refiners in North America and Dubai also, in addition to Switzerland.

In its ‘Gold Demand Trends Q1 2014‘ published in May 2014, the WGC stated that:

As illustrated last year when gold flowed out of western ETFs, through refineries in Switzerland and to consumers in the East, official trade data can provide insights into global gold flows.”

The full GFMS Gold Survey for 2013 (large file 6.2 MBs), i.e the report before the 2 updates, was originally published in April 2013, and was written too early in 2013 (probably written in March 2013) to really capture the flows of Good Delivery gold bars from the UK to Switzerland that were smelted into smaller bars. This was before the massive gold price smash of April 2013 that got the ETF gold sales going. That report mentions ETF gold outflows of 148 tonnes up to 11th March 2013, including 111 tonnes from the SPDR Gold Trust (GLD), but the 2 GFMS update reports from September 2013 and January 2014 were written at a later date, with a better vantage point, when the 400oz bar to smaller bar trend had gathered momentum.

Where was the Swiss refinery output going to in 2013?

On the outbound export route, Swiss gold exports of 2,800 tonnes in 2013 went primarily to Hong Kong (939 tonnes), India (520 tonnes), China (254 tonnes), Singapore (179 tonnes), Thailand (149 tonnes), Turkey (147 tonnes) and the United Arab Emirates (125 tonnes), with the residual 500 tonnes going to other destinations as detailed in the below chart from Nick Laird’s Sharelynx.




GFMS – Masking the Swiss refining of Good Delivery Bars?

Given that the LBMA decided to compare its amended gold (and silver) refinery production statistics against GFMS ‘estimates’ of gold supply (especially out of sync for 2013), then its important to look at what GFMS claimed gold supply and demand to be in 2013. This may help in determining a possible rationale the LBMA had for reducing its refinery output figures.

So, does the 2013 GFMS gold supply and demand data model show this “largest movement of gold, by value, in history” “on an epic scale” phenomenon from the UK to Swiss refiners to Asia? The answer is explicitly NO, neither in 2013, nor in any prior year, but to a limited extent yes, but only after drilling down into the sub-components of an obscure GFMS balancing items within the GFMS supply-demand equation.

But GFMS precious metals supply data and the way it’s presented does not seem to want to highlight the ‘largest movement of gold, by value, in history‘. So even though GFMS mentions (in passing – see above) the historically important 2013 movement of 400oz bars to refineries through places like Switzerland and their transformation into smaller bars by the large gold refineries, the GFMS gold supply statistics keep some of the relevant numbers locked away and jumbled up within a rather odd rolled up figure that it calls “implied net (dis)investment“. Other relevant data, such as OTC demand data, is not even detailed by GFMS, it’s just assumed.

 GFMS gold supply – Disaggregating the implied figure

Here is how GFMS gold supply statistics looked for 2013, taken from the GFMS Update 2 2013 report published in January 2014. In 2013 GFMS used 4 supply categories, namely, ‘Mine production‘, ‘Old gold scrap‘, ‘Net producer hedging‘ and ‘Implied net disinvestment‘.

GFMS-style gold supply and demand figures, 2013 - from GFMS Update #2 report
GFMS-style gold supply and demand figures, 2013 – from GFMS Update #2 report

The first thing to notice is that there is no GFMS supply category called ‘Good Delivery bars’, unlike the large Swiss gold refiners themselves which actually list Good Delivery bars as a distinct gold supply category, such is the importance of that supply source.

Neither is there any category for Gold ETF outflows. So even though 6,600 tonnes of gold came out of LBMA gold refineries in 2013, if you looked at a GFMS supply demand model from 2013, you would never know this. Apart from gold mine production of 2,982 tonnes and old scrap supply of 1,371 tonnes (which together totalled 4,353 tonnes), the only other non-zero supply figure in the GFMS model was ‘implied net investment’ of 383 tonnes.

On the demand side in 2013, GFMS listed jewellery fabrication (2,198 tonnes), other fabrication (792 tonnes), central bank purchases (359 tonnes), physical bar investment (1338 tonnes), and producer de-hedging of 50 tonnes. Again, looking at this demand side, you would not know that gold refinery output in 2013 reached 6,600 tonnes, and that this figure was 2,300 tonnes more than combined mine production and scrap recycling.

There was also a footnote to the above GFMS supply and demand summary table which defines the GFMS definitions of ‘Net producer dehedging‘ and ‘Implied net disinvestment/investment‘.

GFMS defines ‘Implied net disinvesment‘ or “Implied net investment‘ as a residual figure in its supply-demand table (i.e. a plug figure), and states that this “captures the net physical impact of all transactions not covered by the other supply/demand variables“, So basically, it’s a catch-all plug figure. GFMS says that “the implied net (dis)investment  figure is not independently calculated, but derived as the item which brings gold supply and demand into balance.” See full GFMS explanation below:

GFMS disclaimer

This ‘Implied net’ (investment/disinvestment)’ figure is where the 2013 GFMS supply and demand figures become, in my view, completely convoluted and opaque. GFMS says, in both its 2013 Update 1 and Update 2 reports that:

“It is interesting to examine how the implied figure compares to information on activity within the different arenas of investment over the year, (although given aforementioned limitations in this information, it is not possible to dis-aggregate accurately the implied figure into these components)”.

How GFMS exactly makes sense of its ‘Implied net’ (investment/disinvestment)” figures is hard to fathom because there is no proper explanation of the ‘aforementioned limitations‘ that GFMS alludes to except the fact that it doesn’t seem to be able to offer estimates for physical bar movements in Comex nor physical bar movements in OTC activity, part of which it considers the bar shipments to Switzerland to be.

GFMS could also maybe ask the gold refineries in Switzerland and elsewhere for the throughput figures on what they refined in 2013, be it gold mine doré, scrap metal, or Good Delivery bars, and then use that data also. And GFMS could also ask the SPDR Gold Trust Authorised Participants how much gold each of them took out of the GLD in 2013 and how this gold made its way to Switzerland and elsewhere, did the banks send the gold to Switzerland themselves using secure transporters such as Brinks, or did they sell it to other parties who then sent it to the refineries etc etc. The same question could be asked of the Bank of England and the amount of gold withdrawn from its gold vaults and the bullion bank identities of who withdrew it.

In the GFMS world, demand has to equal supply, so whichever side of the equation is greater, the other side has to have a plug figure. In 2013, GFMS put the above items into the demand side, and arrived at an estimate of 4,737 tonnes for demand. It then did an estimate for supply using only 2 components (mining and scrap), and arrived at 4,353 tonnes for supply. Since demand did not equal supply, GFMS then said that implied dis-investment was 383 tonnes. (The figures are 1 tonne out due to what must be a rounding error).

Here is my quick and easier to read version of the GFMS 2013 gold supply – demand table:

gfms 2013 reformat

GFMS then takes the plug figure of 383 tonnes and thinks about an explanation for it.

In its 2013 gold surveys, GFMS also produced another figure which it called ‘World Investment‘, which it defined as “the sum of implied net investment, physical bar investment, and all coins“. It provided this ‘world investment’ figure for both H1 and H2 2013.

This ‘world investment’ figure includes investment demand for physical gold bars and coins, gold medallions, and imitation coins (made of gold), but it also includes investment in products such as gold-backed ETFs. So if there is a huge outflow of gold from the gold ETFs, as there was in 2013, GFMS did not consider this to be gold supply, but rather, GFMS considered it to be negative demand, that it then buries in the implied net investment category.

Since the Authorised Participants of the large gold ETFs redeemed huge amounts of gold from these ETFs in 2013, especially in the first half of 2013, GFMS refers to this as gold ETF ‘investors’ redeeming gold from the ETFs. This is not entirely true because only large investors can redeem from an ETF such as GLD. Small investors just sell their shares in GLD. GFMS calls these 2013 ETF redemptions ‘implied disinvestment’, and it is this phenomenon that caused the GFMS ‘implied disinvestment’ category to be negative in the first half of 2013, but not in the second half of 2013, when GFMS insists that there was positive ‘implied net investment’.

GFMS calculated that there were 550 tonnes of gold outflows from ETFs in the first half of 2013, and 330 tonnes of gold outflows from the same ETFs in the second half of 2013, making a total outflow of 880 tonnes for 2013. Somehow, although the 550 tonnes of gold that left ETFs in H1 2013 caused the H1 implied net investment to be a negative 613 tonnes (as would be expected), the 330 tonnes of outflow from gold ETFs in H2 2013 did not, in GFMS’s eyes, have the same effect, and GFMS’s implied net investment in H2 2013 was a positive 230 tonnes, meaning that although ETFs had an 880 tonne outflow for the full year 2013, the GFMS implied net investment was only -383 tonnes. This then creates another residual number which would have to have been a positive 497 tonnes from some other type of investment demand.

gfms world inv 2013

What else is buried in this GFMS implied net investment apart from ETF flows? It seems to have been Comex exchange activity and OTC activity that is within this implied figure, but GFMS avoids putting numbers on it, hence the confusion.

The reason given by GFMS for a positive net investment of 230 tonnes in the second half of 2013, which cancelled out approximately 500 tonnes of the ETF gold outflows, was what it calls  “significant net buying” in the OTC market.

GFMS refers to its implied net investment figure as “a proxy for institutional investor activity” and said that it “shifted to negative territory” in H1 2013. I’ve included the GFMS 2013 discussion below, just to should how convoluted and unsatisfactory this GFMS logic was. Firstly, the GFMS Update 1 report discussion on ‘implied net investment’:

“The implied net (dis)investment figure is not independently calculated, but derived as the item which brings gold supply and demand into balance. The figure should therefore not be seen as an exact tonnage equivalent but instead an indication of investment activity separate from retail bar and coin demand. Additionally, although a substantial majority of this tonnage will reflect such activity, implied net (dis)investment could also include other flows that, technically, are outside the definition of investment. One example is the impact of any central bank activity that is not being picked up in our official sector figures and that would, as a result, be absorbed within our implied net (dis)investment category.”

“Despite this caveat, implied net (dis)investment typically does provide a clear indication of the overall impact of investor activity on the market for the period discussed. Furthermore, using information collected through field research and publicly available data, Thomson Reuters GFMS performs a ‘reality check’ on these values.”

“It is interesting to examine how the implied figure compares with information on activity within the different arenas of gold investment (although given aforementioned limitations in this information, it is not possible to disaggregate accurately the implied figure into these components).

Due to the nature of gold ETFs and other similar products, we are certain that the near 580-tonne decline in ETF holdings had a one-to-one impact on the volume of investment. The picture is somewhat more opaque when it comes to the futures and OTC markets. As for the former, at end-June, noncommercial and non-reportable net positions in Comex futures were 477 tonnes lower than the end-2012 figure. Turning to the OTC market, however, the first half-year saw robust volumes of investment.

 “As a shortage of bullion rapidly developed in many regional markets and local premia jumped, transactions that were related to physical gold transfer jumped in the London market. Feedback from our contacts, gold trade data and clearing statistics published by the LBMA indicate that a substantial amount of large gold bars (from redemptions of ETFs and sales from unallocated accounts) were shipped to Switzerland from mid-April to be converted to small bars for markets in Asia and the Middle East”.

In its 2013 Update 2 report, GFMS then stated the following. Notice how a lot of the text is copied over from the previous Update 1 report. Update 2:

GFMS update 2 implied

Therefore, GFMS throws a number of items into its OTC category but steers clear from committing itself to really explaining what it means by OTC activity. It states that “the OTC
market is dominated by institutional investors“. It states that  “a substantial amount of large gold bars (from redemptions of ETFs and sales from unallocated accounts) were shipped to Switzerland from mid-April to be converted to small bars for markets in Asia and the Middle East“.

It alludes to “direct shipments, albeit more restrained, from the United Kingdom to the Far East also jumped, as refineries reached full capacity.”

GFMS hazily refers to ‘metal accounts’, which I would consider to be unallocated accounts, and not directly related to absorbing physical ETF gold outflows. GFMS says in its 2013 Update 1 report that “Metal accounts held by western high-net-worth investors also posted a net rise, largely reflecting gold’s traditional role as a means of wealth preservation. This was also partly related to the ongoing shift out of gold ETFs, as metal accounts offered lower fees, while transactions in the OTC market were less transparent than in ETFs.

By the time it wrote its Update 2 report for 2013, GFMS had concluded that:

GFMS update 2 otc

So an 880 tonne outflow of gold from the large ETFs (which are predominantly based in London), as well as hundreds of tonnes of gold outflows from the Bank of England, that led to 1373 tonnes of gold being exported from the UK to Switzerland in 2013, the lions share of which were transformed into kilobars and then shipped to the Asian markets, somehow, according to GFMS, turned into only a negative 383 tonne implied net investment due to “significant net buying for the year as a whole” in the OTC market. There is no attempt to explain the 1373 tonnes of gold exported from the UK to Switzerland in 2013.

If you classify gold ETF outflows as a distinct supply category of gold, which seems logical to me and which the large Swiss gold refineries also consider it to be, then a GFMS supply-demand model would look like this:

gfms 2013 reformat ETFs

The trouble (for GFMS) then is, that the model doesn’t balance, and they are left with a 496 (or 497 tonne) item on the demand side that they can’t explicitly explain what it refers to.

World Gold Council version of GFMS 2013 data

The World Gold Council (WGC) also publishes gold supply and demand data in its annual and quarterly ‘Gold Demand Trends‘ publication. Until 2015, the WGC used GFMS data as a data source, after which it switched to using gold supply and demand data from the Metals Focus consultancy (see below for discussion of the WGC – Metals Focus switch). The WGC uses a different (and easier to understand) layout format for presenting the gold supply and demand data, but for the 2013 format, it still subscribed to the approach of putting ETF withdrawals in the demand category as a negative number.

In its ‘Gold Demand Trends – Full Year 2014’ report, which has the most complete data for 2013, the WGC states in a footnote that the source is

“Source: GFMS, Thomson Reuters; The London Gold Market Fixing Ltd; World Gold Council. Data in the table are consistent with those published by GFMS, Thomson Reuters in their Gold Survey but adapted to the World Gold Council’s presentation

WGC 2013 table

The above WGC model puts gold ETF outflows (Good Delivery bars) into its own line item, but instead of including it as Supply, the WGC puts this in a negative demand. There is also another line item under demand that the WGC calls ‘OTC investment and stock flows‘, which it defines as “Partly a statistical residual, this data is largely reflective of demand in the opaque over-the-counter (OTC) market, with an additional contribution occasionally from changes to fabrication inventories.

GFMS changes its Supply-Demand Methodology in 2014

When the GFMS 2014 Gold Survey was published in April 2014,  GFMS had surprisingly altered the methodology and formatting of its supply-demand data model to include gold ETF outflows as an explicit line item. GFMS also ditched the implied investment concept, but came up with a physical surplus /deficit plug figure instead. I say surprisingly because GFMS had used its previous supply-demand model for a long number of years. GFMS did not dwell on why this had not been done earlier, choosing instead to highlight the benefits of such a change:

GFMS 2014 methodology

Could it be that GFMS subscribers questioned as to why the huge ETF withdrawals were not explicitly listed in the 2013 GFMS supply-demand model, that forced the change? Perhaps.

The inclusion of ETF gold flows (and gold flows from gold futures exchanges) were explained as follows. The OTC category continued to seem to cause problems to GFMS. See below:


gfms meth 3

The actual re-gigged GFMS supply-demand model, redone for 2013 was as follows. The figures for 2013 are slightly different from the ones that GFMS published during 2013, since the table below was published in April 2014 when GFMS probably had updated data about 2013 compared to the reports it published during 2013:

gfms 2014

The above GFMS revised model can also be reformatted as below, moving ETF and Exchange ‘build’ to the supply side, since they are supply and not demand:

gfms 2013 using its 2014 formatting

How the 99 tonnes of Exchange Inventory supply is calculated is not clear. Net Balance of 277 became 276 due to rounding differences. Even including ETFs and Exchange Inventory, there is no explanation by GFMS of what the Net Balance referred to beyond a vague reference to OTC activity.

This GFMS 2014 Survey report was sponsored by Swiss refiner Valcambi, and Japanese refiner Tanaka, with support from Swiss refiner PAMP, the CME Group, the World Gold Council, German refiner Heraeus, Italian refiner Italpreziosi (Italy), Rand Refinery of South Africa, and Istanbul Gold Refinery. Again my question would be why not ask all of these refiners (especially the Swiss refiners) what their throughput of Good Delivery bars was during 2013.

Instead, GFMS still seemed to struggle with explaining what it calls ‘OTC trade’. It even discussed (with a straight face) the huge London gold market clearing volumes of paper gold in 2013, seemingly trying to use this as some sort of vague connection to physical bar movements:

gfms otc

As to GFMS’ assessment (on page 26) of OTC activity, there is nothing concrete offered by GFMS as to what the OTC investment consists of. It mentions bars being shipped to Switzerland and on to Asia, but why is this activity not captured in physical demand?

However, GFMS does have a section in its 2014 (discussing 2013) titled “Supply from Above-Ground’ Stocks”.

“If we include the sales of ETF holdings, then the visible supply of gold to the market from above-ground stocks was 2,160 tonnes, equivalent to 42% of total demand in 2013. The figure comprises 1,280 tonnes of scrapped fabricated products and 880 tonnes of sales from ETF stockpiles.”

And it also included a table of ‘Visible Supply’ in which it did add ETF withdrawals of 880 tonnes to the ‘SUPPLY’ side for 2013, which created a total of 5,182 tonnes of gold supply for 2013. So this is further proof that the amended LBMA gold refinery figures for 2013 are completely out of sync with reality, since even GFMS now includes this ETF supply.

GFMS Visible Supply 2013

But still, 5182 tonnes of supply does not explain 6600 tonnes of gold refining output for 2013. What about all the gold that was withdrawn from the Bank of England in 2013 and shipped to Switzerland? Does GFMS capture this central bank related flow?I can’t see anywhere in the GFMS model where these type of gold flows are captured.

GFMS claims that for official sector transactions, it uses sources such as the IMF and central bank websites, and also “our own proprietary data on undeclared central bank activity, compiled using information collected through field research“. Then why does it not capture all the gold at the Bank of England that has been lent by central banks to bullion banks which has then been withdrawn from the vaults of the Bank of England and flown to Zurich during 2013?

And even for some central bank purchases that it has learned about, GFMS won’t reveal who the purchasers were due to ‘respect of confidentiality’. What does this say for accuracy of a supply-demand model if the nontransparency of central bank transactions prohibits gold transactions being publicised? See example from GFMS Update 1 report 2013:

“South Korea raised its bullion holdings by 20 tonnes in March. The balance of gross buying in the public domain consisted of small gains in gold reserves in a handful of countries. The overwhelming majority of these purchases were made by Asian countries, including Nepal, Mongolia, Brunei and Indonesia. Apart from the aforementioned buyers, over 40% of gross purchases or some 80 tonnes were accounted for by undeclared transactions, details of which cannot be released in respect of confidentiality. In some cases, gold was added quietly in the local market.”

 By the time it wrote its Update 2 report for 2013, GFMS listed some additional central bank buyers during 2013, and then stated that:

“Apart from the aforementioned buyers, over 60% of gross purchases or some 225 tonnes were accounted for by undeclared transactions, details of which cannot be released in respect of confidentiality. In some cases, gold was added quietly in the local market.”

That’s more than 135 tonnes of central bank purchases during 2013 that were not captured in the GFMS model.


Borrowing Gold in London

In my 7 September article “How many Good Delivery gold bars are in all the London Vaults?….including the Bank of England vaults“, I included a quotation from the Financial Times on 2nd September 2015 which stated:

“The cost of borrowing physical gold in London has risen sharply in recent weeks. That has been driven by dealers needing gold to deliver to refineries in Switzerland before it is melted down and sent to places such as India, according to market participants.”

And I concluded that:

“it begs the question, why do the dealers need to borrow, and who are they borrowing from. And if the gold is being borrowed and sent to Swiss refineries, and then shipped onward to India (and China), then when will the gold lenders get their gold back?”

Scotia Mocatta, a bullion bank which is very active in the Indian and Hong Kong/Chinese gold markets, vindicated this point in its ‘Metals Monthly September 2015‘  (page 3):

“The recent low Gold price has spurred physical buying interest to the extent that lease rates have climbed as metal is borrowed and delivered to refineries to be melted into the required bar sizes (such as kilobars) before being shipped to its final destination.

 So, where in the GFMS and World Gold Council data models is this “metal that is borrowed and delivered to refineries to be melted into the required bar sizes (such as kilobars)” being reflected? It appears that these gold bar movements are not being reflected at all.


World Gold Council switch from GFMS to ‘Metals Focus’

Earlier this year, the World Gold Council (WGC) switched from using GFMS as a data provider of gold supply and demand data. In an announcement, the WGC said:

“Starting in May 2015, we will be publishing gold supply and demand data provided by Metals Focus, a leading precious metals consultancy. These data will feature in Gold Demand Trends First Quarter 2015 onwards. Previously, we sourced gold supply and demand data from GFMS Thomson Reuters. The decision to change data providers was based on rigorous market research and a competitive pitch process. For more information, please see the focus box in Gold Demand Trends First Quarter 2015”

The focus box in Gold Demand Trends First Quarter 2015 states:

When new data sets become available and new methodologies are developed, we review how these might complement and advance our own methods. To that end, in 2014 we conducted a rigorous assessment of the gold market data landscape – a process which involved an in-depth review of a number of leading data providers. Following this review we appointed Metals Focus as the provider of our core demand and supply statistics.

“The World Gold Council is committed to publishing the most accurate gold demand data available. We are confident that the move to Metals Focus supports this aim.”

What the WGC didn’t mention in its press release nor in its Gold demand Trends Q1 2015  report is that in October 2013, the WGC purchased a 50% shareholding in Metals Focus Data Limited via its subsidiary WGC (UK) Ltd. The other 50% is owned by Metals Focus Limited. Surely this 50% shareholding is material information that should have been divulged by the WGC in its ‘focus box’ statement above? With its recent emphasis on costs savings, the WGC may have opted for switching from GFMS to Metals Focus partially because it may save money by using a data provider that it has an ownership interest in.

From the WGC 2014 financial statements:

WGC Metals Focus

WGC (UK) Ltd (Company No. 07867682) is a fully owned subsidiary of the World Gold Council, operating out of the same address as the parent company, 10 Old Bailey, London.

Metals Focus Data Limited is a joint venture for “the collection of data relating to the supply and demand for precious metals and licensing of data to third parties”.

What is Metals Focus Limited?

Metals Focus Ltd (Company No 08316950) was incorporated in December 2012, and was founded by Nikos Kavalis, Charles de Meester and Philip Newman, all of whom have previously worked at GFMS. Kavalis (through Premier Metals Consulting Ltd), de Meester and Newman each own a 28.87% shareholding in Metals Focus according to CompanyCheck. Metals Focus 2013 accounts can be seen here.

Metals Focus Data Limited, the 50-50 joint venture between the World Gold Council and Metals Focus Ltd, whose latest accounts can be seen here, has the following directors: Nikos Kavalis, Philip Newman and Lisa Mitchell of Metals Focus, and Terry Heymann, an MD at the World Gold Council.

Some of the sponsors of Metals Focus and its reports include Swiss refiners Valcambi and PAMP/MKS PAMP, other refiners Asahi Refining,  TCA (Italian precious metals refining), the World Gold Council (obviously), Brady Commodity Software Solutions, the CME Group, and G4S. So Metals Focus could also obtain very direct data from at least these Swiss refineries as to their throughput of Good Delivery gold bars.

Although the World Gold Council has now switched data suppliers to Metals Focus since earlier this year, in its 2015 Q1 Gold Demand Trends, it still uses the same supply-demand presentation format as previously,  with ETFs in 2013 being classified as negative demand and not supply. Interestingly, in the Metals Focus data, the ETF line item for 2013 has now risen to 916 tonnes.

wgc metals focus pres



With 6,600 tonnes of Good delivery refinery gold refining production confirmed by the LBMA to have taken place during 2013 (before the LBMA altered its data), you can see in the above analysis that this is problematic for the models of GFMS, the World Gold Council and possibly the model of Metals Focus too. Since the LBMA is sent refining data by its members, then, if it chose to, the LBMA could generate very accurate data for gold and silver refinery output for all of 2014 and nearly all of 2015.

Almost all other industries are able to publish accurate industry production figures with a minimal lag of maybe 2-3 months that provide an up-to-date snapshot of that industry’s activity. This is also true of economic data such as labour statistics and housing starts. Why then is it so hard for the LBMA to publish full and comprehensive gold refinery output data on a quarterly basis?

If this reporting procedure was put in place, the global gold industry would have far more clarify and insight into the huge flows of kilobar gold that are, on a daily and weekly basis, now being flown from Switzerland into Delhi, Ahmedabad, Chennai, Bangalore, Hyderabad and Kolcata in India, and that are also flowing at a torrential rate through Brinks vaults in Hong Kong and on into China.


Ronan Manly
E-mail Ronan Manly on:

  • rowingboat

    How about the situation when a former investor sells to another investor and the gold bar remains in the vault? How does anyone measure that transaction to determine from “supply/demand” data whether collective selling pressure overwhelms collective buying pressure (or vice versa) when the bullion doesn’t move and stays where it is?

    We can help glean it from national trade data, which fortunately has become available in recent years. During the bull market Western nations like the UK, Switzerland and America cumulatively and consistently imported thousands of tonnes. How could anyone measure this demand by the London Bullion Market, as one example, without examining the trade data? ETFs and conversion into kilo bars for the Chinese provide insights, but these are only subsets of a larger dynamic. E.g. gold is no doubt in demand and changing ownership in these Western countries but their net imports have halted and/or reversed into exporters. How do you include all of this in a supply-demand table?

    What is great about the Swiss trade data is it goes back to 1982. This is shortly after the invention of new CIP/CIL processing technology making large-scale open pit mining of gold deposits possible. Combined with the opening up of new mining frontiers around the world, I’d estimate that the amount of non-monetary gold in the world has doubled and perhaps tripled since 1982. Examining the trade data will help understand the
    drivers of bull and bear markets, particularly the UK-Swiss axis and large swings into and out of London.

  • festina

    It seems that the research of Koos Jansen and Ronan Manly of
    BullionStar is showing up huge deficiencies in the big-budget World Gold
    Council research publications. The king indeed has no clothes!

    Research on the gold market is mostly read to project future
    trends in the gold price. The WGC pieces do not explain

    How much gold has flowed from West to East

    The source of this gold

    The source of and the financial arrangements
    behind the massive volumes of paper gold overwhelming the gold price from time
    to time

    Whether the bullion banks are short of gold and
    whether the central banks will be backstopping them in the case of a short

    Etc, etc

    • rowingboat

      How much gold has flowed from West to East is irrelevant because these flows are dynamic and reversible and have proven so. If you look at the Swiss trade data, countries driving demand in the 1990s progressively shrunk from view and eventually became suppliers to the market in the 2000s as gold prices rose with Western demand and imports. Hong Kong in 2015 has flipped from a net importer to exporter; a couple of years ago HK’s net imports were 600 tonnes. Turkey has also turned net exporter to Switzerland during this year, which is extremely rare since 1982. When Western demand does eventually return, gold flows east will further slow and maybe reverse as imports to Switzerland, USA, UK and other nations increase again. The evidence is there from the morphing of bear market to bull market at the turn of the century. By the way, this is all about plain vanilla non-monetary flow and nothing to do with central banks back-stopping short bullion banks, a theory debunked many times over during the past decade or two.

      • JanNieuwenhuijs

        The US Treasury has zero proof there is gold in Fort Knox. Although they claim they have.

        • rowingboat

          Central banks have long been accused of losing their gold when really it’s just the acceleration of global gold production after 1980, which we should be focusing on because non-monetary gold and its flow have increased so much.

          I’ve been on holiday for a month and keenly noticed Keith Weiner’s falling cobasis last night during the month of October.
          Why did this happen? When I checked the Swiss data, net imports from the UK were only 3 tonnes in October, which is the lowest level since the great transfer to China from 2013. Global demand/exports out of Switzerland were quite modest in October but there was also increased supply to Switzerland from Uzbekistan, Turkey and UAE (Swiss net imports have been essentially zero with the UK feeding the gap between exports and imports).

          What’s interesting is that Keith’s cobasis has been rising steeply in November so I bet UK’s exports to Switzerland have needed to rise as well this month.

          • JanNieuwenhuijs

            I’m not talking about the supply of gold, I’m talking about the US Treasury has no proof there is gold in Fort Knox. They lost the audit reports.

      • Who debunked theories that central banks back-stopped short bullion banks?

      • festina

        Thanks. However, your response leaves me none the wiser re the information missing from the WGC pieces that I mentioned. For example, can you explain who was involved in the massive takedowns of gold during April 2015 on huge paper volumes? There are many equities that would also collapse under such an assault – but apparently the rules applying to equities do not allow such attacks?

        (London pm fixes): Thursday, 11 April 2013 $1565.00, Monday, 15 April 2013 $1372.75, down 13%.
        Wednesday, 19 June 2013 $1370.00, Friday, 28 June 2013 $1192.00, down 13%

        • rowingboat

          Check out the archives on Keith Weiner’s website who demonstrated in real time as it was happening by the way, that it was physical supply driving price those two days in April (also read his material debunking manipulation and conspiracy theories one hears so much about in the gold community).

          Particularly in the years following the GFC in 2008 a large amount of gold had been imported into Western countries. In America, including local production, I estimate 900mt from 2009-11, into the London bullion market 4,100mt from 2006-12, and into Switzerland 1,900mt from 2006-12. This demand was driving price higher but from 2013 their imports on a net basis stopped and is continuing.
          In fact these flows that were driving price higher during the bull market have turned to supply… London (UK) for example has exported approximately half of its imports, 2200mt, from 2013-15.

          Why April 2013 as the turning point? I recall Pater Tanebrarum’s blog publishing a chart showing a collapse of inflationary expectations and several economic reports were released in April claiming QE tapering was for real and that corporate borrowing was gaining a foothold to replace QE. Investors in the West sold bullion and switched asset classes, which was later proven to be the case from the trade data.

          I agree, the daily moves you quote are rare and historic for gold. But look at the daily moves at the peak in 1980. ‘sb’ posted a paper here describing someone in Dubai that day in 1980, and everyone on the streets who would normally be buying were cashing in.

          • festina

            Thanks, I checked into Keith Weiner who said on April 19, 2013 “Unfortunately for the speculators, sentiment changed on Monday and there was heavy selling of physical metal leading the sales of futures.” Doesn’t explain why someone would ditch >100tons in a few minutes / hours take it much further does it? Normally a seller would feed stock into the market gradually.

          • rowingboat

            Markets are a white-knuckled fight to the death and it was no doubt designed to take out major support at 1500… brutally achieved and extraordinarily effective when examining the trade flows since

          • festina

            We’re now into poetry before I got anything resembling an answer. Let’s leave it there on the basis of “we’ll meet again”!

          • rowingboat

            “Doesn’t explain why someone would ditch >100tons in a few minutes / hours does it?”

            The answer to your question is clear… to decisively break 1500 support thereby breaking the back of the bull market to the point that several years later investors still aren’t importing gold into the US, UK and Switzerland on a net basis. It was a brilliant move. When these same people start buying again, the WGC or anyone else will be none the wiser as it is impossible to measure, but the signs will be there in the trade data just as they were at the end of the last bear market.

          • festina

            “white knuckled” is when you kayak over a high waterfall. Do you mean “bare knuckled”? In which case my remark would be that if one of the fighters have an undisclosed gun and the other has not because he was told it would be a fist fight, it is quite reasonable to predict that the guy who plays by the rules will not survive. So, once again, with whose balance sheet did the shorts come to the fight? Was it that of the Exchange Stabilisation Fund? Who would have settled the losses if the Chinese took the other side of the shorts and then aggressively bid up the gold price?

          • CPDLC

            I find it difficult to understand your contention that the balance of US physical gold flows was positive over the 2009 – 2011 period, let alone by your quoted 900t.

            According to USGS data US production was as follows:
            2009 = 233t, 2010 = 231t, 2011 = 234t. Total = 688t.

            Consumption was as follows:
            2009 = 201t, 2010 = 225t, 2011 = 211t. Total = 636t

            Imports were as follows:
            2009 = 283t, 2010 = 333t, 2011 = 316t. Total = 932t

            Exports were as follows:
            2009 = 436t, 2010 = 437t, 2011 = 631t. Total = 1,504t

            I make that a net negative balance of -520t of physical gold flow.

          • rowingboat

            We agree on the production numbers. However I have imports of 604mt and exports of 383mt in 2010 etc. The historic data I have dates back to 1991 and was from USGS I had thought but I need to recheck as I did this in 2013. Since 2013 I’ve been extracting the data from the monthly USGS reports.Thanks for posting.

          • rowingboat

            Here we go from the USGS website:

            Unless I’m cross-eyed and getting the lines mixed up, for 2010:

            Production 231mt
            Imports 604mt
            Exports 383mt

            So if I’m reading this right, the gold stock in America increased by 452mt in 2010, of which reported consumption was 180mt. This reported consumption is probably what the WGC is able to measure but what about the remaining 272mt? I would contend this is investment demand, opaque, impossible to measure and mainly held within the financial system.

            I would also suggest this is the type of bullion that has been sold to feed local consumption because America has exported all of its imports and mine production (plus more) in the last four years. How long this will continue is a very good question as America is an important supplier to Switzerland and to lesser degree UK.

          • CPDLC

            I can understand your confusion!
            The figures are reported in Kg and US$ but the Kg amounts for dore, ore & concentrates, waste & scrap, powder, compounds, are GROSS weight and not gold content.
            As an example, look at Ore & Concentrates for 2010 imports in 2010. It is given as 257t yet the Value is only $58m. When you calculate how many oz of gold (converted to tonnes) that the $58m related to at the prevailing gold price, it is only c. 1.5t.
            I have done all the gold flows since 1950 but when I reached 1981 I had to calculate all import & export data from Value rather than Quantity in order to keep strictly to physical gold quantities.

          • rowingboat

            I thought that as a possibility too, that I’d made a mistake, but at the top of the link I posted is this statement:
            “Data in metric tons of gold content unless otherwise noted”

          • CPDLC

            It is not a ‘possibility’ it is reality!
            The headings are quite clear. Ore & Concentrates are clearly NOT metric tonnes of gold. Only Refined Bullion fits that description.
            As a rough check look at pp 31.16 of the 2010 report.
            Exports, total value = $14.7B
            Compare with pp 31.18
            Imports, total value = $11.6B
            Difference = $3.1B which equates to c. 78t of net exports at the average gold price of $1,228/oz over 2010.

          • rowingboat

            we’re not looking at the same document, the one I’m referencing is only 2 pages:

          • CPDLC

            I have used all the USGS Yearbook data.

          • rowingboat

            Would you mind cross-checking your data in the USGS yearbook with the link I posted of their annual summary statement please? There shouldn’t be such a large discrepancy given our data source is exactly the same.
            I’ve noticed that the current year’s data is “estimated” in their annual summary so maybe significant post revisions occur after the yearbook has been released.

          • CPDLC

            Sorry about the delay – I was called away.
            Your link correctly mirrors the total tonnage as detailed in the full Yearbook tables. However, it is the tonnage itself which I am questioning (as per my prior posts).
            Without detailing multiple years I would continue to point at the 2010 quantity of ores & concentrates as a prime example of how misleading the totals become when 257t of material containing only 1.5t of gold by value is used as part of a combined yearly total which is supposedly all 100% gold – when patently it is not.
            257t of gold would have been valued at c. $10.15B using the average of $,1,228 for 2010. Whereas the table reports it’s value as $58m – that is an error of 17,396%.
            When I do the same calculations for the Dore figures I get a difference of 29% (about right for the usual grade of Dore).
            When I do the same for refined gold in the same table I get a difference of just 12% (likely due to price volatility over the year).

          • CPDLC

            In my previous post it would have been better to have calculated the implied grade of the ore & concentrate (based on it’s reported value) rather than calculating the ‘error’ as a percentage.
            The implied grade works out at 0.57% which would appear reasonable for concentrate (about 184 oz per tonne).

          • rowingboat

            I completely agree with you. The annual summaries report bulk tonnage
            while the yearbooks have more detail including refined bullion, which I should be using in the import/export analysis. Thanks for pointing out my error.

          • rowingboat

            I’ve recalculated imports/exports after converting the bulk volumes for ores, dore and scrap into fine-equivalent gold. Then added this to mine production to determine annual changes of the US gold stock from 1994-2015.

            From 2009-11 the US gold stock increased by 336mt, which has been unwound/exported by 236mt in the four years since.

            From 2002-06 the US gold stock increased at a slightly faster rate by my calculation, 686mt over this period; it then declined by 408mt in 2007/08.

            Annual consumer demand of 150-180mt is currently being met by recycled gold. Interestingly this was also the case from 1994-2001
            when the US gold stock didn’t change much at all despite very healthy consumer demand, begging the question how much did the stock rise during the 1970s bull market and then sold?
            All of the data appears available in the USGS archives to answer such questions.

          • CPDLC

            Frankly, I am far less sanguine than yourself regarding the grossly misleading bulk tonnages passed off as fine gold in the USGS summary reports that you were using. I had not been aware of this fact until we entered into this discussion as I had only used the full yearbook tables for all of my own analysis.

            You still appear to be achieving a different outcome from myself but at least you are now aware of the pitfalls of purely relying on tonnage data.

            I have had to assume that the US Census Bureau data, from which the Import/Export tables are compiled, is accurate. However it is worth keeping in mind that the valuation figures in the Ores & Concentrates category are most likely reflecting the value of ALL minerals contained. As concentrates are typically produced by flotation of copper rich ore it is highly likely that the value of contained copper may also be a significant part of the ascribed value.

            My own estimate for solely the Import/Export gold balance over the 1981-2013 period is a net export of -7,319 tonnes. However, given the possible unreliable nature of the translations required to get from bulk tonnages to actual fine gold I also simply added all the US$ import/export valuations over that period, which resulted in a net export value of -US$127.7B. Applying the average gold price over the period of $538/oz then results in -7,380 tonnes – so it does appear to confirm the earlier result.

            Over the 23 year period only 5 years showed a net import tonnage. These were 1982-1986.

          • rowingboat

            I missed the footnote which was careless of me but perhaps there’s been a change for the better in 2013, which is their latest yearbook. E.g. imported ores & concentrates are reported as 369 kg whereas in 2012 they were 11,800 kg. Comparing my old 2013 data from the annual Summary they nearly match the data derived from the yearbook.

            From your earlier post you appear to be including reported consumption in your import/export gold balance. If so, I disagree with this and consider changes to the gold stock level in-country more important because of internal recycling.

            For years 2009-14, the USGS made this statement: “XXX tons of new and old scrap were recycled, more than the reported consumption.” This is significant because from 2011 America is exporting all of its mine production as well as production passing through it from Central & South America. Previously and for most of the bull market it was absorbing some of it.

            I’ve also done this for Switzerland by the way, converted imported tonnages into fine gold equivalent since 1981. For most of the bull market gold was being pulled into Switzerland. Now I am sure people are still buying and storing gold in that
            country, but since 2013 the overall net flow has stopped and reversed slightly… similar to what we have seen in the US and especially UK.

          • CPDLC

            Footnote or not, the summary data you (and presumably many others) were using was obviously completely useless and misleading through the addition of bulk tonnages – so why were those summaries even released?

            You have completely ignored my point regarding the Import/Export balance in VALUE terms over the 1981-2013 period, namely a net OUTFLOW of US$127.7 billion (c. 7,380t) as reported by the Census Bureau. Do you not recognise the significance of that?

            The ‘statement’ by USGS re Scrap/Consumption is far less significant. A simple check shows a NEGATIVE balance of 1,228t over the wider 1981-2013 period.

            I find your view that internal consumption is irrelevant very surprising. Surely it is very important to try and ascertain WHO actually owns the gold rather than just how much may or may not be within a country. A good example being India where their official holdings of 557t are completely swamped by private holdings and where that government are trying extreme measures to attempt to correct the imbalance.

            Worldwide, the percentage of gold held as official reserves compared to private holdings has fallen to 18%. In 1954 it was 59%.

            My reason for undertaking this research was to try and test the credibility of the Official US holdings since 1972. Clearly, I have to therefore also take an educated view on ‘non-official’ holdings within the USA, so consumption data has to form an essential part of the overall net balance. If you do not recognise that then there is really little point in continuing this exchange.

          • rowingboat

            Respectfully I suggest that you’re underestimating recycling. In the year gold peaked, recycling also peaked (263mt) so in effect consumer demand was actually negative in 2011 and has been for six years.

            Summing up the USGS data from 1991-2014 I have:

            Recycling supply of 3,357mt
            Consumer demand of 3,791mt
            Mine supply of 6,825mt

            This would suggest America exported 6,391mt of its production that it didn’t need, which of course is a net outflow. That the Census Bureau data suggests a net outflow of 7,380mt from 1981-2013 is consistent with this and therefore plausible to me.

          • CPDLC

            Unless you are remarkably prescient there is an obvious error in your opening statement, eg it is only 5 years since 2011 and only 4 years of reasonably full data – so your quote of ‘6 years’ is incorrect!

            Regarding your second summation: using your preferred timescale of 1991-2014 I get:-
            Recycled supply of 3,206t
            Consumption of 4,186t (coins and industry)
            Mine supply of 6,859t

            However, the problem with selectively using such a subset of the overall gold flow can be misleading, particularly the use of US refinery figures. Mine supply may well have been 6,859t but the actual reported figures for Concentrate & Dore refined in the US was only 4,735t (ie 2,124t less than production) so clearly quite a substantial amount of US internal production is refined outside the US and would therefore be recorded as an export. Likewise, as there are significant imports of Dore & Concentrates it is highly probable that a significant percentage are destined for refining in the US. This renders the refinery figures unusable in terms of overall gold flows into and out of the US because of the strong likelihood of double-counting.

            From the outset I decided to purely focus on Mine Production + Imports – Exports – Internal Consumption in order to try and assess the net gold flow of the US since the gold window closed in 1971. I did this as an impartial exercise using the figures as reported in the USGS yearbooks. My cumulative results for the 1972-2013 period are as follows (note that the 2014 year book is not yet available).

            US Mine Supply 8,276t
            Total Imports 14,226t (note: this includes 4,869t of earmarked foreign gold sold into the market from the Federal Reserve Vaults).
            Total Consumption 13,397t (incl the above foreign gold)
            Total Exports 14,860t

            This results in a net export of 5,754t of gold since 1972 which begs the question as to where this all came from given that ‘US official holdings’ have not changed substantially over the same period?

            Further, the reported non-monetary holdings reported for 2013 were:
            Industrial 4.94t
            COMEX 243t
            US ETF 1,810t

            I can only make the assumption that the COMEX & ETF holdings form part of the total consumption over the period – otherwise the net situation is even worse!

            Finally, having learned of the inaccuracies within the USGS summary data compared to the Yearbook data through our ongoing interchange on this forum, I decided to check the differences for all years over the period 1992-2014. Results are as follows:

            Mine production: effectively zero difference.
            Imports: overstated by 517t (ignoring foreign gold from Fed)
            Exports: understated by 1,085t
            Consumption: understated by 855t (ignoring foreign gold)

            Therefore a total misrepresentation of 2,456t over the period.

          • CPDLC


            I completely forgot to follow up on the Recycling/Consumption balance that you were referring to in your opening remarks. Here are my comments:

            You are correct, over the 4 years (2011- 2014) the net balance has been positive, eg 55t, 42t, 18t & 13t.

            However the previous 14 years they were negative to the tune of 1,270t. There was no refining or consumption data made available for 1995 & 1996. From 1992-1994 the balance was positive 136t.

            For what it is worth, the net balance over the wider 1981-2014 period was negative 1,215t (ie Consumption (Mint coin sales + Industrial) was higher than all gold recovered from recycling) .

          • rowingboat

            Please ignore my remarks… half way through the data extraction I realised that US scrap exports were significant and started to include them… then forgot to fill in gaps for the missing years, e.g. scrap export of 149mt (mainly to Switzerland) in 2011.

            Would you mind providing a detailed balance for one of the years please? For example, in 2013 I calculate a large negative balance:

            Imports: 0.4mt (ore) + 211.3mt (dore) + 98.6mt (fine) + 26.8mt (scrap) = 337.1mt
            Exports: 6.9mt (ore) + 184.5mt (dore) + 489mt (fine) + 20.9mt (scrap) = 701.3mt
            Mine production = 227mt

            Negative balance = 137.2mt, i.e. the overall stock level reduced by this much. It excludes consumer demand, and therefore differs to your definition, so I would like to understand how you include this to achieve 18mt positive balance in 2013.

            Summing up over nine years from 1993-2001, I calculate a decline in stock of 163mt then a significant rise from 2002-06 (754mt) as recycling supply plummeted according to the USGS summaries. From 2007-13 there was an overall decline of 475mt.

            I see parallels with 2002. As the US stock has now only increased by net 279mt since the start of the bull market, 200mt annual recycling surely can’t continue for much longer (not to mention declining US production and falling US imports in 2015). At some point the stock will need to rise again, choking supply to Switzerland and other nations… the Swiss trade data shows that Switzerland’s net imports from USA more than halved in 2002 from 2000/01 levels.

          • CPDLC

            I am afraid you are mixing up completely different terms here!

            The 18t positive balance I calculated for 2013 was purely for Recycled Scrap minus Consumption (ie 210t – 192t) because you had introduced the topic of Refined Scrap less Consumption based on the USGS statement that you had quoted to me. It certainly does NOT represent the actual overall gold flows of the US in 2013 which I estimate to have been negative 328t.

            That last figure comes from the overall flows which were:
            Production +230t
            Consumption -186t
            Imports +332t
            Exports -704t

            Please note that I have included a net export of 5.16t of bullion that was bought into foreign gold stocks held at the Fed during 2013. I recorded this as a corresponding negative 5.16t within the Consumption total because it was clearly not domestic consumption.

            I hope this clarifies!

          • rowingboat

            Great, my production import and export numbers match yours more or less.

            So then how are you accounting for recycling in your analysis using the yearbook data?

            Summing up consumer demand over a 20-30-40 year period needs to be corrected in order to account of former buyers who have sold, re-buy etc.

            Your earlier post stated for 1972-2013 cumulatively: US mine supply 8,276t; total imports 14,226t; total consumption 13,397t; total exports 14,860t… resulting in a net export of 5,754t of gold since 1972.

          • CPDLC

            I am not accounting for internal refining data at all.

            I agree that some proportion of recycled internal scrap is likely to be part of new consumption (and therefore may well be double-counted) but I have no way of coming up with a meaningful figure Once the refined gold is back on the market it might also have been part of the Export figures as that of consumption.

            Obviously one has to deduct the gold content of the annual Imported Scrap to get a measure of how much internal scrap provided the annual refined output. But how old is the scrap? How much of the current Industrial or Jewelry Scrap was originally consumed prior to the year range that is being measured? And given the high proportion of Exported Bullion compared to Imported Bullion, how much of the Exports are made up from Refined Scrap output?

            These are all questions that I cannot answer given the available data and therefore I cannot put any meaningful value on the impact on current consumption.

            Perhaps further detailed work might throw more light on this area but I doubt it would significantly change my own conclusions as to the overall trend of US gold flows.

          • rowingboat

            CPDLC, focusing on changes to the US gold stock (production + imports – exports) the yearbook data suggests there has been very little change over 21 years 1993-2013… flat until 2001, a rise to 2006 then an overall fall since 2007. Interpreting the data you’ve posted it therefore suggests a significant rise over the preceding 21 years (1972-1992) of approximately 8,000mt. Do I have that right?

          • CPDLC

            Now you are just going round the same circle again by ignoring US consumption!

            I am not trying to convince you of anything – you must make your own conclusions. I was sucked into this exchange because of the inaccurate data you had gleaned by using the USGS summary sheets which, at least, I appear to have have conclusively proved to you.

            Now you want to return to just the Production+Exports-Imports which will only give you an estimate of total +/- holdings over any specific period – whereas, as I have stated, I was testing the credibility of the longstanding ‘official’ holdings by checking ALL the flows, including those into non-official hands (like the ETFs, COMEX, Banks, Private Individuals, Industry, et al) and for that I have had to also take a view on consumption.

            Back in 1950 the US Monetary Stock was reported as 20,178 tonnes. By 1971 it was down to 8,354 tonnes. Yet the net flows (excl consumption) were only -1,658 tonnes. So clearly the net flow data was extremely unreliable as a predictor of drain on official reserves. Even adding-in the reported consumption of 2,512 tonnes still only covered some 33% of the official shortfall.

            Since then, official movement reporting has ceased completely, along with much useful commentary and data that used to be in the USGS reports. So I have had to work with a source of diminishing information over time.

            I obviously do not expect to come up with anything like a definitive figure – chances are that I am some way out. But I wanted to see just how feasible it would be for the US to still have official reserves of c. 8,140 tonnes 42 years after the closing of the ‘gold window’.

            On the balance of probability, given the work that I have done, I find it unlikely that those reserves could have remained at that level. But that is a personal view and people must do their own research and come to their own conclusion – as I am sure you will! Good luck with it.

          • rowingboat

            “Back in 1950 the US Monetary Stock was reported as 20,178 tonnes. By 1971 it was down to 8,354 tonnes. Yet the net flows (excl consumption) were only -1,658 tonnes. So clearly the net flow data was extremely unreliable as a predictor of drain on official reserves”

            Was the USGS only measuring / reporting “non-monetary” gold flow?

          • rowingboat

            Just quickly before I leave the house this morning, I meant six years 2009-14 with peak recycling occurring in 2011. I’ll read your detailed post later but to summarise and to be clear, this is what I’ve done to date:

            When I calculated the import/export balance for the period 1993-2013 using the yearbooks (without needing to delve into the archives for earlier years, which I intend doing eventually) I noted there has been very little change in the overall US gold stock level, which I calculated from mine production plus net imports/exports of ores, dore, scrap and refined gold.

            This therefore implies that most US demand has been met by the existing internal stock over this period. Then when I turn to the annual summary statements, the USGS actually provides direct recycled data supporting this, which also explains the large net outflows of gold you’ve observed from the Census Bureau data.

  • Pingback: From Good Delivery bars to Kilobars – The Swiss Refineries, the GFMS data, and the LBMA | NewZSentinel()

  • 59LesPaul

    Look Ronan it really is pretty simple. The US Treasury/Federal Reserve, BOE, LBMA, WGC, GFMS, etc. are trying to perpetuate a ridiculous highly leveraged paper gold system that is unravelling due to massive physical demand in the east. Unfortunately for them, the more that they suppress the price, the more that physical demand in the east increases and available physical supply will inevitably decrease.

    As Claude Juncker was quoted as saying, “When it becomes serious, you have to lie”.
    Well, things are very serious in London in the paper gold/silver world. Therefore the paper gold merchants have resorted to Orwellian (1984) tactics of changing or rewriting history by altering supply/demand data. In other words flat out deceit by “homogenising” the data just like with “climate science”.

  • Pingback: Gold Investment » Ronan Manly: LBMA fudges data to obscure gold shipments to Asia()

  • Pingback: « From Good Delivery bars to Kilobars – The Swiss Refineries, the GFMS data, and the LBMA – Ronan Manly()

  • Pingback: « From Good Delivery bars to Kilobars – The Swiss Refineries, the GFMS data, and the LBMA()

  • Pingback: « From Good Delivery bars to Kilobars – The Swiss Refineries, the GFMS data, and the LBMA – Ronan Manly()

  • Pingback: Brinks, Loomis, Malca Amit, G4S, CME Group, Hong Kong gold market, HKIA, Chep Lap Kok, CGSE, Hong Kong International Airport()

  • Pingback: The Gold Vaults of Hong Kong: Brinks, Malca Amit, Loomis |

  • Pingback: The Gold Vaults of Hong Kong: Brinks, Malca Amit, Loomis – Real News Now()

  • gold sales Uganda

    I would like to introduce you Gold sales Uganda limited a professional mining company licensed and legally operating in Uganda dealing in metals. In this regards, we are looking for a reliable and potential buyers for gold.

    Kindly please confirm your order requirements enabling us send you our best offer.

    Awaiting for your swift response towards establishing with serious buyers a long term relationship with their prestigious companies




    SKYPE ibradrogba20

Copyright Information: BullionStar permits you to copy and publicize blog posts or quotes and charts from blog posts provided that a link to the blog post's URL or to is included in your introduction of the blog post together with the name BullionStar. The link must be target="_blank" without rel="nofollow". All other rights are reserved. BullionStar reserves the right to withdraw the permission to copy content for any or all websites at any time.