Wednesday, February 1, 2012

Fun With PM Prices And Premiums (January 2012)

In this article I will present whatever looks interesting from the data set I am collecting from the eBay / Widget showing whatever patterns or other information that we can get from it.  This data from just January 2012 is not much, and so the below comments probably not wort that much either, but as I get more data and work with it more, I will learn stuff.  And pass it along to you in future articles.

All of the raw data comes from the eBay / JV widget at, the widget is found at the bottom of their Home Page.  You can view the data I took each time at:

This first graph shows gold price from January 1 - 31.  I took 39 observations (meaning I noted price of gold twice on certain days, and maybe skipped one or two days, but these are in chronological order), but it is clear enough how gold moved in January.

"Click on any image for a better view."

Notice the nice big jump in late January, the day the Fed announced ZIRP until at least late 2014.

The next chart shows the premium of American Gold Eagles vs. "Spot" gold price ("paper gold" price), note that this is in chronological order as well:

The average premium for January 2012 was 9.8%.  The standard deviation was 2.9% (meaning that approximately 68% of the premiums observed were between 6.9% - 12.7% (9.8% +/- 2.9%).  I would note for you that having looked at this widget for a couple of years now that the 9.8% average premium is fairly high, the more typical range I have observed is 5.5% - 8.5%.  Note very high premiums around January 20 or so.

This next graph shows the premiums of Gold Eagles vs. Silver Eagles:

The blue diamonds are Gold Eagle Premiums, the red squares are Silver Eagle premiums.  The average Silver Eagle Premium for January was 39% with a standard deviation of 14%.  Note three things:
  1. Silver premiums are much higher than gold (we all knew that though).  Why?  Part of this can be explained by the sheer bulk of silver, meaning it is more expensive to move and handle.
  2. Relatively speaking ASE standard deviation / average and AGE standard deviation / average are roughly comparable (ASE: 0.36, AGE: 0.30).  That means on a relative basis that silver premiums do not vary much more than gold premiums.
  3. The correlation coefficient between the AGE premiums and the ASE premiums is only 0.26, usually a 0.30 correlation coefficient is the minimum (that "Social Scientists" use) to show a significant relationship between two variables.  Here, this means that AGE premiums and ASE premiums "move together" only a little bit.
Lastly, I calculated the correlation coefficient of price of gold vs. the AGE premium: -0.21.  That is, as the price goes UP, the premium goes down (and ESPECIALLY vice-versa, when they take down the price of gold, the premiums get higher as sellers do not want to let their gold go "too cheaply".

As I get more up to speed on long-forgotten statistics and I gather more data, we may find more interesting relationships than what I have mentioned above.

Hey, any of you readers who are GOOD at statistics, please review my work above (and email me!), I already know that I am not using certain terms correctly...  I know, I know, this is really sleazy statistics...  Help me learn...


  1. "Silver premiums are much higher than gold (we all knew that though). Why?"

    Part of the reason is that the ebay reporting widget captures sales of graded numismatics along with the bullion sales. It's hard to completely filter them out of search results.

  2. When looked at on a linear graph, silver prices have been extremely volatile, and the premiums even more so. I wonder how much longer paper silver will be able to hold back the tidal wave of buying, because at some point, demand simply overwhelms the supply and paper breaks off from physical.

  3. Did you know that you can shorten your long urls with Shortest and get dollars from every visitor to your shortened urls.


Note: Only a member of this blog may post a comment.