On December 9, 2014, Albert Cheng, Managing Director Far East of the World Gold Council, was interviewed by the China Gold Network. The interview was published in Chinese only.
The Gold Demand Trends published each quarter by the World Gold Council (WGC), show aggregated Chinese consumer demand Q1 – Q3 2014 was 638.4 tonnes. But, in the interview Cheng notes that the chairman of the Shanghai Gold Exchange (SGE), Xu Luode, has stated the SGE was supplied by 1,100 tonnes of gold import in the first eleven months of 2014 and this number may reach 1,250 tonnes by year end. Supplemented by 450 tonnes of domestic mine production this year “total demand should reach about 1700 tonnes”, said Cheng.
I have been long disputing Chinese gold demand numbers from the WGC, as total supply (Chinese net import, domestic mining and scrap) persistently has been transcending the WGC numbers. The WGC has never been able to elucidate the difference between massive supply and their demand numbers. The aggregated difference from 2007 until present is about 3,000 tonnes.
The total supply can be tracked on a weekly basis by SGE withdrawals, which have proven to be the best proxy. Though the import (and scrap) composition of SGE withdrawals can only be estimated until confirmed by the SGE or China Gold Association.
And so I was surprised Cheng openly (in Chinese) elaborated on the SGE model and instead of solely talking about how much gold was sold at retail level, also expanded on how much gold is actually added to Chinese (non-government) gold reserves, measuring import and domestic mine production, that are prohibited from being exported from the mainland.
The next quote is translated by LK, gold investor from Hong Kong (who we all should be very thankful for his work!):
The China Gold Network: The recently published Q3 Gold Demand Trends report says that China’s gold demand is down year-on-year. How do you interpret this?
Albert Cheng: This year, the Q3 gold demand figure that we publish is down because last year’s gold demand was a special case. Given that last year compared to 2012 same time was up 40%, there really is nothing strange. However, if we compare this year’s Q3 figure to the average of the last 5 years at this time, we still find positive growth, and still up slightly compared to 2012.
Using the numbers supplied by Xu Luode, they in fact show that China imported about 1100 tonnes of gold in the first 11 months this year through the SGE, and may reach 1200 to 1300 tonnes by year end. Adding together domestic production, total demand should reach about 1700 tonnes. So, the energy of the China gold market hasn’t diminished; compared to last year, the development is still healthy.
For clarity: the translator is a native Chinese speaker and a financial expert. The translation has been confirmed by a second native Chinese speaker and financial expert, which severely limits the probability of the Chinese text being misinterpreted.
As my regular readers know I often make estimates of Chinese net import using SGE withdrawals as a proxy. Last week I estimated China in the first 11 months of 2014 imported 1,200 tonnes, we now know now I probably overestimated gold import by 9 % (1,200 tonnes vs 1,100 tonnes).
In hindsight it’s always more easy to analyze. I think what happened is that in mid 2014 (March and June) a part of the SGE withdrawals supply composition shifted from import to scrap. If scrap went up, import went down, as:
SGE withdrawals = import + mine + scarp (neglecting stock-carry over)
In the chart above we can see premiums going negative in March and June while withdrawals staying relatively strong. This means domestic supply (scrap) was increasing relative to import, hence the gold in China became cheaper than in London. The reason the discount isn’t immediately arbitraged is because gold in China is prohibited from being exported.
The dip in Chinese gold import was also reflected in export from Hong Kong to China mainland.
Because scrap apparently was more than I calculated in my model, I can now adjust the model. If from January until November import was 1,100 tonnes and mining was 413 tonnes, than scrap had to be 328 tonnes, as withdrawals were at least 1,841 tonnes in the mainland.
Additionally, elevated scrap means quite some Chinese have been selling physical gold that found its way to the SGE. Perhaps some were expecting the price to rise sooner. I don’t think elevated scrap signals leases were unwound; if a lease expires the lessee is most likely to buy gold on the SGE, it would not make sense for him to buy gold in the domestic market to bring to the SGE in order the repay the loan.
SGE withdrawals in week 49 have dropped by a whopping 28 % w/w, to 38 tonnes. Year to date withdrawals stand at 1,905 tonnes.
However, when corrected by SGEI trading volume withdrawals in week 49 could have been as low as 27 tonnes, which was not expected for the seasonally strong December month. Year to date withdrawals corrected for SGEI volume is 1867 tonnes.
Let’s wait what next week will bring to see if China will reach the 1,700 tonnes.