London Gold Market: A Fractional System?

Includes: GLD, SGOL
by: Bron Suchecki

A reader has asked me to comment on these two recent GATA articles and, which claim that the London gold market is a fractional reserve system.

Adrian Douglas’ assertion is that there is at a minimum four owners for each ounce of unallocated metal held in London. His support for this is to apply the ratio of average daily share trading in GLD (11.9m) to its shares outstanding (325m), rounding to a ratio of 1:30, to an estimate of the daily trading in gold in London to derive the amount of gold London should have. This is then compared to an estimate of what London does have, resulting in the 1:4 fractional ratio.

For his estimates of the London market, Douglas relies on a report by Paul Mylchreest. His report is fact based and takes a very logical approach to estimating of the amount of gold in London. His conclusion is that there is:

an aggregate pool of gold of just over 16,866 tonnes of gold to support an average of 2,134 tonnes of daily spot gold trade. On this basis, 12.7% of the pool of available gold is being turned over every day on average. … And the entire pool is turned over every 7.9 working days. In my opinion, this level of trade relative to the estimated pool of gold liquidity is excessive and doesn’t pass the smell test.

Firstly, he makes a series of assumptions to get to his figures. For example, his 16,866t figure relies on World Gold Council/industry estimates of above ground gold and the percentage that is investment. Being a trade organisation representing miners who want a high gold price one should expect that “stock” numbers will be estimated on the downside. When estimating what the real trading volume of gold is, then he steps into a more rubbery area because he is relying on only two guesses from some industry people - we need more than that.

As a result, one must consider his 12.7% turnover figure to have a fair margin of error considering all the assumptions and estimations used to derive it. This is not to say that it is wrong, just that it is not a “hard” proven figure.

Secondly, even if 12.7% is correct, I don’t think it logically follows that this “doesn’t pass the smell test", a conclusion he comes to by comparing gold to equities, other commodities and fiat currencies. The last one is probably the most relevant. In this he has to again make some assumptions about currency trading turnover to come to a figure of 2.6% for Sterling, conceding that when including forwards and swaps “daily Sterling turnover is only equivalent to 8.4% of UK broad money”.

If one does the same calculations for the Australian dollar, you get 4.1% for spot and 13.3% including forwards and swaps. Does gold’s 12.7% (which could be lower if some of Mylchreest’s assumptions are changed) now appear as an “excessive amount of gold trading relative to the likely pool of available gold”?

Mylchreest’s final conclusion is that there is either 1. “more than one ownership claim on each gold bar” or 2. “far more gold bullion held in private hands than is acknowledged by current industry estimates”.

I would suggest that there is another OR that Mylchreest has not considered: the very fact that gold is no one’s liability and cannot be printed means it attracts a disproportionate amount of trading and speculation. Why is it assumed that 12.7% is excessive and unreasonable? Could not the 12.7% figure be proof of the special monetary nature of gold, proof that it is the King of Currencies?

I have spent a bit of time on Mylchreest’s report because it is the key input into Adrian Douglas’ calculations. Before I move on to his numbers, I would like to say that I have a lot of respect for Mylchreest’s report and look forward to it being improved with more accurate data.

On that, I note Mylchreest’s statement on page 25 that “I haven’t a clue what COMEX inventories were in 1997, but let’s assume 200 tonnes …” That information is available at going back to 1975. A subscription is required but would be worthwhile as Sharelynx has a lot of other data that would be very useful for Mylchreest’s analysis.

Now on to Adrian Douglas’ calculations. He applies GLD's turnover of 3.66% to Mylchreest’s turnover figure of 2,134t to come to an implied stock holding that London should have of 64,000t. This is then contrasted to Mylchreest’s estimate of 15,000t of non-leased physical to derive the 1:4 fractional ratio.

This analysis assumes that the behaviour of over-the-counter (OTC) players is/must be the same as those trading GLD. Let us consider each of Douglas’ statements in support of this.

The purpose of buying investment gold is for it to store wealth. This necessarily implies that it is held for a long time.

This is a very broad statement and one that I don’t think can be supported. Investors have all sorts of different time horizons. Remember we are talking about trading in unallocated and whether that is backed. The fact that it is unallocated rather than allocated bars would imply, if anything, that the investors have shorter time frames rather than long.

If gold is bought and traded quickly it would destroy wealth, not store it, because there would be a large loss due to transactional fees.

It is actually the other way around. The quicker you can trade something the less risk you have to changes in its price. Bullion banks have a spread between their bid and ask prices – they MAKE money from quickly trading gold. For those dealing with bullion banks in the OTC market, the tightness of those spreads combined with the volatility of gold mean it is entire reasonable that they can make money day trading gold.

Considering these limitations [minimum trade limit of 1,000 ounces] it is likely that OTC participants would turn over a lot less than 1/30th of the inventory in a day.

I do not see how the $1 million trade size must mean a lower turnover. That is not a big figure for wholesale market participants. With bullion bank spreads of $0.50 to $1.00, a 1000oz deal only means $500 to $1000 profit. This would mean that a spot gold trader would need to do a lot of trading to make a decent return on the capital employed, which means they would trade more frequently, rather than less.

As with Mylchreest’s comparisions to currency trading, I don’t think Douglas’ comparisions to GLD make any conclusive case that London gold turnover is suspicious.

For further support, Douglas notes that:

In the last 14 years the supply of dollars has increased from $4 trillion to $15 trillion (+275 percent) while the gold price has risen from $400 in 1995 to $1,000 in 2009 (+150 percent). How could this happen? … There has to be an alternative massive supply of gold to make the price rise slower than the influx of dollars.

How it could happen is that those extra dollars were diverted into equities and house prices, rather than gold, pushing up their price more instead.

He also says that “If the OTC market traded only gold that was in the vaults on a 100 percent reserve ratio, there could never be a lack of liquidity.”

I would assert that a lack of liquidity has nothing to do with stocks, backed or not - it has to do with a depth of buyers and sellers. If you have 100% backed unallocated, but few of the holders want to sell, then you have a lack of liquidity as well.

For Australians interested in this topic, I will be speaking on the London market on November 1st at the Gold Standard Institute's free investor day in Canberra and look forward to meeting Australian gold "enthusiasts" to discuss this and other issues.

In closing, Lawrence Williams from Mineweb sums it up best:

The big problem, though, with much of this kind of analysis is that the analysts and observers are working with a mixture of real and assumed figures. It thus tends to rely on statistics being manipulated, perhaps subconsciously, to support pre-conceived theories.

Disclosure: Long gold via ASX:ZAUWBA