My Book Arrived


Delivered in this morning's mail, my most recent purchase: "Gold Refiners and Bars Worldwide" by Nigel Desebrock, and I am quite pleased.

The database work continues and now includes the GLD bar list (which is published daily) and I confess my attention has been turning to gold recently, simply because the signficance of the stock to flow ratio is starting to hit home for me.

The book was published in 1991, so although 20 years old - it contains a wealth of information on 50 major refiners, and specifically their numbering systems for their bullion bars. Most of these bars described will still be in circulation possibly with a whole chunk in the ETF's, so this gives us the opportunity to match approximate DATE OF MANUFACTURE RANGES for the bars, as well as identify any changes to the numbering sequence. In theory this can be mixed and matched with more recent information, but the upshot is that this information will be used to tag the numbering sequences in the database. Because there is so much freaking data to sift through, we simply flag the sequences which look legitimate and then focus on the areas where the numbering doesn't make sense - getting a kind of 'rating' in the database (between 1 and 100) for the legitimacy of each bar. It may take a while, but any discrepancies should then show up and we'll be able to then chart the percentages involved. A low percentage of suspect serial numbers is hoped for and means that all is well, but a high percentage will warrant special investigation.
This is based on my theory that the easiest way to introduce 'imaginary' inventory into the system would be to falsify a production run of serial numbers and then 'hide' it in the ETF, making sure that the items are never in any danger of being delivered (like those plentiful Russian and Chinese bars .. perhaps?).
Here is what the book looks like. The book is quite large at 476 pages with quite a lot on each of the 50 refiners. It is normally purchased by gold dealers [link to publisher]. I do realise I'm probably overdoing it, but it's fun. By the way, JPMorgan - I am happy to halt my research if you pay out my home mortgage for me ;p.







3 comments:

Kid Dynamite said...

nice, Warren. Did you add PHYS/PSLV to the database yet?

also, keep in mind that some of these refiners may have changed their methodology in the last 20 years...

Warren James said...

Hi KD, yes I've got the data converted but it's not yet in the database tables. I underestimated the amount of work involved in collating, normalizing and coding. Each PDF source has its own proprietary layout and hence requires its own 'adapter' in the upload software.

Yes the book is a bit dated but I'm hoping it will provide a foundation for analysis of at least 50% of the ETF inventory for gold. A bit of pattern analysis can fill the gaps for modern sequences as well. Nigel (the author) has also kindly offered to help with any questions I might have as well.

I also have to decide on a format for a primary key which will be flexible enough to work with all data sources. I hope to talk about this in my next database article. :)

Louis Cypher said...

That's pretty cool Warren and very nice of the author to help out. You should probably ask for payment in the precious rather than paper from JPM.