Renewed estimates of Chinese gold demand

Editor’s note: This article was previously published here at It is republished with thanks.

Geopolitical and market background

I have been revisiting estimates of the quantities of gold being absorbed by China, and yet again I have had to revise them upwards. Analysis of the detail discovered in historic information in the context of China’s gold strategy has allowed me for the first time to make reasonable estimates of vaulted gold, comprised of gold accounts at commercial banks, mine output and scrap. There is also compelling evidence mine output and scrap are being accumulated by the government in its own vaults, and not being delivered to satisfy public demand.

The impact of these revelations on estimates of total identified demand and the drain on bullion stocks from outside China is likely to be dramatic, but confirms what some of us have suspected but been unable to prove. Western analysts have always lagged in their understanding of Chinese demand and there is now evidence China is deliberately concealing the scale of it from us. Instead, China is happy to let us accept the lower estimates of western analysts, which by identifying gold demand from the retail end of the supply chain give significantly lower figures.

Before 2012 the Shanghai Gold Exchange was keen to advertise its ambitions to become a major gold trading hub. This is no longer the case. The last SGE Annual Report in English was for 2010, and the last Gold Market Report was for 2011. 2013 was a watershed year. Following the Cyprus debacle, western central banks, seemingly unaware of latent Chinese demand embarked on a policy of supplying large quantities of bullion to break the bull market and suppress the price. The resulting expansion in both global and Chinese demand was so rapid that analysts in western capital markets have been caught unawares.

I started following China’s gold strategy over two years ago and was more or less on my own, having been tipped off by a contact that the Chinese government had already accumulated large amounts of gold before actively promoting gold ownership for private individuals. I took the view that the Chinese government acted for good reasons and that it is a mistake to ignore their actions, particularly when gold is involved.

Since then, Koos Jansen of has taken a specialised interest in the SGE and Hong Kong’s trade statistics, and his dedication to the issue has helped spread interest and knowledge in the subject. He has been particularly successful in broadcasting market statistics published in Chinese to a western audience, overcoming the lack of information available in English.

I believe that China is well on the way to having gained control of the international gold market, thanks to western central banks suppression of the gold price, which accelerated last year. The basic reasons behind China’s policy are entirely logical:

  • China knew at the outset that gold is the west’s weak spot, with actual monetary reserves massively overstated. For all I know their intelligence services may have had an accurate assessment of how much gold there is left in western vaults, and if they had not, their allies, the Russians, probably did. Representatives of the People’s Bank of China will have attended meetings at the Bank for International Settlements where these issues are presumably openly discussed by central bankers.
  • China has significant currency surpluses under US control. By controlling the gold market China can flip value from US Treasuries into gold as and when it wishes. This gives China ultimate financial leverage over the west if required.
  • By encouraging its population to invest in gold China reduces the need to acquire dollars to control the renminbi/dollar rate. Put another way, gold purchases by the public have helped absorb her trade surplus. Furthermore gold ownership insulates her middle classes from external currency instability which has become an increasing concern since the Lehman crisis.

For its geopolitical strategy to work China must accumulate large quantities of bullion. To this end China has encouraged mine production, making the country the largest producer in the world. It must also have control over the global market for physical gold, and by rapidly developing the SGE and its sister the Shanghai Gold Futures Exchange the groundwork has been completed. If western markets, starved of physical metal, are forced at a future date to declare force majeure when settlements fail, the SGE and SGFE will be in a position to become the world’s market for gold. Interestingly, Arab holders have recently been recasting some of their old gold holdings from the LBMA’s 400 ounce 995 standard into the Chinese one kilo 9999 standard, which insures them against this potential risk.

China appears in a few years to have achieved dominance of the physical gold market. Since January 2008 turnover on the SGE has increased from a quarterly average of 362 tonnes per month to 1,100 tonnes, and deliveries from 44 tonnes per month to 212 tonnes. It is noticeable how activity increased rapidly from April 2013, in the wake of the dramatic fall in the gold price. From January 2008, the SGE has delivered from its vaults into public hands a total of 6,776 tonnes. This is illustrated in the chart below.

This is only part of the story, the part that is in the public domain. In addition there is gold imported through Hong Kong and fabricated for the Chinese retail market bypassing the SGE, changes of stock levels within the SGE’s network of vaults, the destination of domestic mine output and scrap, government purchases of gold in London and elsewhere, and purchases stored abroad by the wealthy. Furthermore the Chinese diaspora throughout South East Asia competes with China for global gold stocks, and its demand is in addition to that of China’s Mainland and Hong Kong.

The Shanghai Gold Exchange (SGE)

The SGE, which is the government-owned and controlled gold exchange monopoly, runs a vaulting system with which westerners will be familiar. Gold in the vaults is fungible, but when it leaves the SGE’s vaults it is no longer so, and in order to re-enter them it is treated as scrap and recast. In 2011 there were 49 vaults in the SGE’s system, and bars and ingots are supplied to SGE specifications by a number of foreign and Chinese refiners. Besides commercial banks, SGE members include refiners, jewellery manufacturers, mines, and investment companies. The SGE’s 2010 Annual Report, the last published in English, states there were 25 commercial banks included in 163 members of the exchange, 6,751 institutional clients accounting for 81% of gold traded, and 1,778,500 clients of the commercial banks with gold accounts. The 2011 Gold Report, the most recent available, stated that the number of commercial bank members had increased to 29 with 2,353,600 clients, and given the rapid expansion of demand since, the number of gold account holders is likely to be considerably greater today.

About 75% of the SGE’s gold turnover is for forward settlement and the balance is for spot delivery. Standard bars are Au99.95 3 kilos (roughly 100 ounces), Au99.99 1 kilo, Au100g and Au50g. The institutional standard has become Au99.99 1 kilo bars, most of which are sourced from Swiss refiners, with the old Au99.95 standard less than 15% of turnover today compared with 65% five years ago. The smaller 100g and 50g bars are generally for retail demand and a very small proportion of the total traded. Public demand for smaller bars is satisfied mainly through branded products provided by commercial banks and other retail entities instead of from SGE-authorised refiners.

Overall volumes on the SGE are a tiny fraction of those recorded in London, and the market is relatively illiquid, so much so that opportunities for price arbitrage are often apparent rather than real. The obvious difference between the two markets is the large amounts of gold delivered to China’s public. This has fuelled the rapid growth of the Chinese market leading to a parallel increase in vaulted bullion stocks, which for 2013 is likely to have been substantial.

By way of contrast the LBMA is not a regulated market but is overseen by the Bank of England, while the SGE is both controlled and regulated by the People’s Bank of China. The PBOC is also a member of both its own exchange and of the LBMA, and deals actively in non-monetary gold. While the LBMA is at arm’s length from the BoE, the SGE is effectively a department of the PBOC. This allows the Chinese government to control the gold market for its own strategic objectives.

Quantifying demand

Identifiable demand is the sum of deliveries to the public withdrawn from SGE vaults, plus the residual gold left in Hong Kong, being the net balance between imports and exports. To this total must be added an estimate of changes in vaulted bullion stocks.

SGE gold deliveries

Gold deliveries from SGE vaults to the general public are listed both weekly and monthly in Chinese. The following chart shows how they have grown on a monthly basis.

Growth in public demand for physical gold is a reflection of the increased wealth and savings of Chinese citizens, and also reflects advertising campaigns that have encouraged ordinary people to invest in gold. Advertising the attractions of gold investment is consistent with a deliberate government policy of absorbing as much gold as possible from western vaults, including those of central banks.

Hong Kong

Hong Kong provides import, export and re-export figures for gold. All gold is imported, exports refer to gold that has been materially altered in form, and re-exports are of gold transited more or less unaltered. Thus, exports refer mainly to jewellery which in China’s case is sold directly into the Mainland without going through the SGE, and re-exports refer to gold in bar form which we can assume is delivered to the SGE. Some imported gold remains on the island, and some is re-exported from China back to Hong Kong. This gold is either vaulted in Hong Kong or alternatively turned into jewellery and sold mostly to visitors from the Mainland buying tax-free gold.

The mainstream media has reported on the large quantities of gold flowing from Switzerland to Hong Kong, but this is only part of the story. In 2013, Hong Kong imported 916 tonnes from Switzerland, 190 tonnes from the US, 176 tonnes from Australia and 150 tonnes from South Africa as well as significant tonnages from eight other countries, including the UK. She also imported 337 tonnes from Mainland China and exported 211 tonnes of it back to China as fabricated gold.

Hong Kong is not the sole entry port for gold destined for the Mainland. The table below illustrates how Hong Kong’s gold trade with China has grown, and its purpose is to identify gold additional to that supplied via Hong Kong to the SGE. Included in the bottom line, but not separately itemised, is fabricated gold trade with China (both ways), as well as the balance of all imports and exports accruing to Hong Kong.

The bottom line, “Additional supply from HK” should be added to SGE deliveries and changes in SGE vaulted gold to create known demand for China and Hong Kong.

SGE vaulted gold

The increase in SGE vaulted gold in recent years can only be estimated. However, it was reported in earlier SGE Annual Reports to amount to 519.55 tonnes in 2008, 582.6 tonnes in 2009, and 841.8 tonnes in 2010. There have been no reported vault figures since.

The closest and most logical relationship for vaulted gold is with actual deliveries. After all, public demand is likely to be split between clients maintaining gold accounts at member banks, and clients taking physical possession. The ratios of delivered to vaulted gold were remarkably stable at 1.05, 1.03, and 0.99 for 2008, 2009 and 2010 respectively. On this basis it seems reasonable to assume that vaulted gold has continued to increase at approximately the same amount as delivered gold on a one-to-one basis. The estimated annual increase in vaulted gold is shown in the table below.

The benefits of vault storage, ranging from security from theft to the ability to use it as collateral, seem certain to encourage gold account holders to continue to accumulate vaulted metal rather than take personal possession.


Supply consists of scrap, domestically mined and imported gold


Scrap is almost entirely gold bars, originally delivered from the SGE’s vaults into public hands, and subsequently sold and resubmitted for refining. Consequently scrap supplies tend to increase when gold can be profitably sold by individuals in a rising market, and they decrease on falling prices. There is very little old jewellery scrap and industrial recycling is not relatively significant. Official scrap figures are only available for 2009-2011: 244.5, 256.3 and 405.8 tonnes respectively. I shall therefore assume scrap supplies for 2012 at 430 tonnes and 2013 at 350 tonnes, reflecting gold price movements during those two years.

Scrap is refined entirely by Chinese refiners, and as stated in the discussion concerning mine supply below, the absence of SGE standard kilo bars in Hong Kong is strong evidence that they are withheld from circulation. It is therefore reasonable to assume that scrap should be regarded as vaulted, probably held separately on behalf of the government or its agencies.

Mine supply

China mines more gold than any other nation and it is generally assumed mine supply is sold through the SGE. That is what one would expect, and it is worth noting that a number of mines are members of the SGE and do indeed trade on it. They act as both buyers and sellers, which suggests they frequently use the market for hedging purposes, if nothing else.

Typically, a mine will produce doré which has to be assessed and paid for before it is forwarded to a refinery. Only when it is refined and cast into standard bars can gold be delivered to the market. Broadly, one of the following procedures between doré and the sale of gold bars will occur:

  • The refiner acts on commission from the mine, and the mine sells the finished product on the market. This is inefficient management of cash-flow, though footnotes in the accounts of some mine companies suggest this happens.
  • The refiner buys doré from the mine, refines it and sells it through the SGE. This is inefficient for the refiner, which has to find the capital to buy the doré.
  • A commercial bank, being a member of the SGE, finances the mine from doré to the sale of deliverable gold, paying the mine up-front. This is the way the global mining industry often works.
  • The government, which ultimately directs the mines, refiners and the SGE, buys the mine output at pre-agreed prices and may or may not put the transaction through the market.

I believe the government acquires all mine output, because it is consistent with the geopolitical strategy outlined at the beginning of this article. Furthermore, two of my contacts, one a Swiss refiner with facilities in Hong Kong and the other a vault operator in Hong Kong, tell me they have never seen a Chinese-refined one kilo bar. Admittedly, most one kilo bars in existence bear the stamp of Swiss and other foreign refiners, but nonetheless there must be over two million Chinese-refined kilo bars in existence. Either Chinese customs are completely successful in stopping all ex-vault Chinese-refined one kilo bars leaving the Mainland, or the government takes all domestically refined production for itself, with the exception perhaps of some 100 and 50 gram bars. Logic suggests the latter is true rather than the former.

Since the SGE is effectively a department of the PBOC, it must be at the government’s discretion if domestic mine production is put through the market by the PBOC. Whether or not Chinese mine supply is put through the market is impossible to establish from the available statistics, and is unimportant: no bars end up in circulation because they all remain vaulted. It is material however to the overall supply and demand picture, because global mine supply last year drops to about 2,490 tonnes assuming Chinese production is not available to the market.

Geopolitics suggests that China acquires most, if not all of its own mine and scrap production, which accumulates in the vaulting system. This throws the emphasis back on the figures for vaulted gold, which I have estimated at one-for-one with delivered gold due to gold account holder demand. To this estimate we should now add both Chinese scrap and mine supply. This would explain why vaulted gold is no longer reported, and it would underwrite my estimates of vaulted gold from 2011 onwards.

Further comments on vaulted gold

From the above it can be seen there are three elements to vaulted gold: gold held on behalf of accountholders with the commercial banks, scrap gold and mine supply. The absence of Chinese one kilo bars in circulation leads us to suppose scrap and mine supply accumulate, inflating SGE vault figures, but a moment’s reflection shows this is too simplistic. If it was included in total vaulted gold, then the quantity of gold held by accountholders with the commercial banks, as reported in 2009-11, would have fallen substantially to compensate. This cannot have been the case, as the number of accountholders increased substantially over the period, as did interest in gold investment.

Therefore, scrap and mined gold must be allocated into other vaults not included in the SGE network, and these vaults can only be under the control of the government. It will have been from these vaults that China’s sudden increase in monetary gold of 444 tonnes in the first quarter of 2009 was drawn, which explains why the total recorded in SGE vaults was obviously unaffected. So for the purpose of determining the quantity of vaulted gold, scrap and mined gold must be added to the gold recorded in SGE vaults.

Though it is beyond the scope of this analysis, the existence of government vaults not in the SGE network should be noted, and given cumulative mine production over the last thirty years, scrap supply and possibly other purchases of gold from abroad, the bullion stocks in these government vaults are likely to be very substantial.

Western gold flows to China

We are now in a position to estimate Chinese demand and supply factors in a global context. The result is summarised in the table below.

Chinese demand before 2013 had arrived at a plateau, admittedly higher than generally realised, before expanding dramatically following last April’s price drop. Taking the WGC’s figures for the Rest of the World gives us new global demand figures, which throw up a shortfall amounting to 9,461 tonnes since the Lehman crisis, satisfied from existing above-ground stocks.

This figure, though shocking to those unaware of these stock flows, could well be conservative, because we have only been able to address SGE deliveries, vaulted gold and Hong Kong net flows. Missing from our calculations is Chinese government purchases in London, demand from the ultra-rich not routed through the SGE, and gold held by Chinese nationals abroad. It is also likely that demand from the Chinese diaspora in SE Asia and Asian is also underestimated by western analysts.

There are assumptions in this analysis that should be clear to all. But if it only serves to expose the futility of attempts in western capital markets to manage the gold price, the exercise has been worthwhile. For much of 2013 commentators routinely stated that Asian demand was satisfied from ETF redemptions. But as can be seen, ETF sales totalling 881 tonnes covered only one quarter of the west’s shortfall against China, the rest coming mostly from central bank vaults. Anecdotal evidence from Switzerland is that the four major refiners have been working round-the-clock turning LBMA 400 ounce bars into one kilo 9999 bars for China. They are even working with gold bars that are battered and dusty, which suggests the west is not only digging into deep storage to satisfy Chinese demand at current prices, but digging a hole for itself as well.

1 Comment

  • Corrigan says:

    All very well and good, but all of this is to assume (a) that trade statistics are genuine (and we know that,especially between HK & the Mainland they are decidedly not so); (b) that gold bought by China is being tightly held for its wondrous mystical, Once and Future King properties as money-in-waiting (highly unlikely given what we know goes on in the use of copper, iron ore, soybeans, and petrochemicals as collateral for onshore borrowing); (c) that a good portion of the stuff has not been smuggled back out and sold surreptitiously as a way to effect capital flight; and (d) that a lot of the last year or so’s supposed excess demand was not related to the Iranians’ attempts to evade US imperial sanctions as was the case with Turkey (see #c above).

    Quod erat nil demonstrandum, I’m afraid!

Comments are closed.