Gold price manipulation proof, M3 update
Note: this article has been completely superceded by a revision that incorporates the addendum, etc.
January 22, 2007
Addendum June 3, 2007
There's really not much to say, since the raw charted data is so obvious, but there are some facts.
- The correlation between the European Central Bank's gold purchase and sale activity and the gold price in Euros is .97. That means that 97% of the gold price changes since 1999 can be explained by ECB actions alone.
- The left hand scale on both charts is in metric tonnes. One metric tonne is about 32,150 troy ounces of gold.
- As just one example of the very large size of ECB gold operations, during the gold price increase from early April 2006 to the peak in mid May 2006, the ECB sold 1,800 metric tonnes of gold. That translates to about 58 million troy ounces.During the fall from mid May 2006 to a bottom in June 2006, they bought back 2,345 metric tonnes, which is approximately 75 million troy ounces.
- Just for reference, the total supply of gold per the World Gold Council in 2005 was 3,953 metric tonnes or about 127 million troy ounces.
- The total gold price increase from about 325 to 500 Euros caused the average ECB gold inventory to drop about 1500 metric tonnes or about 48 million troy ounces.
- The quantity of gold trading by the ECB dwarfs the total 500 metric tonne sale limit of the Washington Agreement.
- For those who might point out that futures trading quantities are far larger, that's true. But not only are they only paper trades and physical gold is seldom delivered, but also there are thousands of different traders and none consistently trade in quantities similar to the ECB.
- The data is published by the ECB anywhere from 1-3 weeks behind the events, so it's not useful for trading. Here is the weekly data link (added 8/2007) for 2007.
- Some might prefer to look at the data as evidence of control, as opposed to manipulation.
- Addendum 1/31/2007: The huge majority of the ECB trades are in the "receivables", or "paper gold" area. For proof, simply repeat the search in the first point above and drop the word receivables.
The gold trading record of the ECB, since 1999
The close-up gold trading record of the ECB
"So the question is, having very consciously and purposely tried to break the bubble and upset the markets in order to sort of break the cocoon of capital gains speculation, we are now in a position—having done that and in a sense succeeded perhaps more than we had intended—to try to restore some degree of confidence in the System."
-- Alan Greenspan, Chairman of the Federal Reserve.
Source: Federal Open Market Committee (FOMC) meeting minutes from March 22, 1994 (page 41)(note that the source is a large pdf file).
"Central banks stand ready to lease gold in increasing quantities should the price rise."
-- Alan Greenspan, testimony to Congress on July 24, 1998
Gold Market Lending, an excellent treatment of many issues in this area from Blanchard and Company, Inc.
Note for researchers and the curious:
The ECB data for gold and receivables purchases and sales is both here and is also linked in the first bullet above. It brings up a search page in the ECB's database. Just check/select the first option ( the key is "ILM.W.U2.C.A010.Z5.Z0Z" as of late January 2007), and then click the "show selected data" button just above it. That will bring up the actual weekly data going back to 1999, which is in millions of Euros.
The only adjustments we've made to that raw data is to convert the values into troy ounces using the Euro gold price on the week ending date, and then converting to metric tonnes using a factor of 32,151 troy ounces per metric tonne. There are undoubtedly some minor discrepancies due to the conversion process, but not enough to materially change the results.
Also, Empirically observed covariation is a necessary but not sufficient condition for causality.
Gold price manipulation proof, addendum
June 3, 2007
There has been a question raised by a number of readers about our calculation methodology, as specified above in the second paragraph under "Note for researchers and the curious:". Their point, and there's some evidence of it, is that the raw numbers are expressed in a gold price which is only adjusted roughly quarterly by the ECB, as opposed to weekly as we oroiginally calculated.
We have therefore added a light grey line to our two charts showing our best guess of the value of tonnes on hand every week, assuming that a fixed gold value was used. The quarterly dates chosen for that fixed gold value were the closest week ending dates to the 15th of January, April, July and October. (August 2007 - switched the gray and black lines to reflect that we were incorrect in our initial assumptions, and thanks again to alert and knowledgable readers who point out our inadvertent errors.)
- It drops the correlation from .97 to about .71.
- During the gold price increase from early April 2006 to the peak in mid May 2006, the ECB gold sales drop to about 1,000 (was 1,800) metric tonnes of gold. That translates to about 31 (was 58) million troy ounces.
- The total gold price increase from about 325 to 500 Euros caused the average ECB gold inventory to drop about 1200 (was 1500) metric tonnes or about 39 (was 48) million troy ounces.
- Our sincere apologies to anyone who was misled by our inadvertent possible error, but it still appears that our basic conclusions are and were correct. The ECB does apparently manipulate (or control if you prefer) the gold price in Euros.
M3 updateJust last week, we broke above a 12% annual rate of change which is the first time that has happened since May 2001. For what its worth, the official recession started in March 2001.