"Ted pointed out something that I'd missed in Tuesday's column on Comex gold inventories---and that was the fact that the 595,102 troy ounces that the report showed withdrawn from JPMorgan on Friday was, with the exception of a few ounces, totally reversed in Monday's report from the Comex."
This big adjustment was also noted by "Pirocco" on the SilverStacker forum here. When questioned by Pirocco, Comex’s response was that "the adjustment column allows the depository or warehouse to make changes to their inventory in either of the categories for various circumstances as needed", which as Pirocco notes, just avoids answering the question because "various circumstances" says nothing.
Well I’ve worked out what one of those "various circumstances" is. If you subtract the 594,506.898 Monday adjustment from the Friday figure of 595,102.000 you get 595.102. That is not a coincidence. In other words, the Friday figure was a fat finger keying error, where someone at JPM keyed in "595 comma 102" instead of "595 point 102"! Hence JPM had to do an adjustment of 594,506.898 to turn 595,102.000 into the correct figure of 595.102.
I do wonder if the fat finger extended to the paperwork sent to the receiver of the erroneous 595,102 ounce withdrawal - perhaps a temporary bit of excitement that they were getting an unexpected 18+ tonnes?
What is surprising is that this sort of data transfer between warehouses and the CFTC isn't automated. Worryingly, this is not an isolated incident:
- CFTC charging JP Morgan $650k for repeatedly submitting "large trader reports that contained hundreds of errors"
- Screwtapefile's Warren finding discrepancies between the numbers in the GLD trade settlement spreadsheet and the GLD bar list
- Rand Refinery losing 87,000 ounces ($113 million) of gold
- Canadian Mint's 2009 $15 million gold inventory "discrepancy"