Dust flux, Vostok ice core

Dust flux, Vostok ice core
Two dimensional phase space reconstruction of dust flux from the Vostok core over the period 186-4 ka using the time derivative method. Dust flux on the x-axis, rate of change is on the y-axis. From Gipp (2001).

Friday, February 26, 2016

Pop in gold x USDX reaching a critical point

Here at the World Complex I have been using gold x USDX (i.e., the gold price in US dollars per ounce multiplied by the US dollar index, divided by 100) as a proxy for the value of gold mined by companies not operating in the US. Assuming that their expenses are in some local currency, the cash flow of such companies can be improved either by a rising dollar (gold remaining constant) or a rising gold price. In fact, a rising dollar may be preferable, as when the gold price rises sharply, such companies are often hit with special "windfall taxes"--something I have yet to see when it is the dollar which rises (hopefully nobody gets any ideas about that).

There is a lot of excitement in the gold space in the past few weeks. As we saw over a year ago, there has been a breakout of the gold x USDX from a sizeable triangle.


The above chart lends itself to a couple of investment theses. One is that a lot of people seem to make a New Year's resolution to buy a lot of gold, as there is a notable move in the index at the beginning of each of the last three years.

With all the excitement of the last few weeks, it is time to take a closer look. We are at an important point in at least three important respects. At present, gold x USDX is at 1203.79. The peak in the index hit during the move last year was 1229.93. I would submit that the present peak has to exceed last year's level, or else it is just another lower high.


Here is a comparison of the performance of gold vs copper over the past six years (both are multiplied by USDX, gold is US$ per oz, copper in US$ per pound). The way this is plotted, wherever the two curves intersect, the gold-copper ratio is 400. Right now, gold is about 600x the price of copper. Historically, a ratio this high is uncommon--we last saw it briefly in 2009.

Ordinarily, I would say the above chart is a little scary for goldbugs, as it would seem to predict a drop in the price of gold (or a rise in copper, which seems a little unlikely in this economy). Your expectations will vary depending on your overall investment thesis. If deflationary forces grow stronger, this ratio could very well rise further, just as we are seeing in the gold-silver ratio. While Americans don't seem to think of gold as money, it looks as though someone does. If your hypothesis is that central banks are going to pull out all the stops to fight deflation, your future predictions depend on whether you believe they will be successful.


It's been awhile since I posted one of these. The idea here is that the gold price is behaves as do many other complex systems in nature--it spends long periods of time in certain areas of equilibrium, punctuated by rapid moves to some other area of equilibrium. There are three areas of relative equilibrium in the above figure. The gold x USDX index has been confined to the middle equilibrium since mid-2013, except for the hopeful little pop last year. Sadly, it didn't reach escape velocity, and fell back into the middle equilibrium.

Our current situation bears very close watching. Once again, we are at a possible breakout point. If we are to see a significant move in gold, we need to see a move towards the upper equilibrium. If the US dollar were to remain strong during such a move, this would suggest a gold price approaching $1400. The next eight weeks or so should tell the tale--either we will be well on the way to the next equilibrium, or we will fall back to the present one.

17 comments:

  1. Very interesting. There is an other way to plot products and quotients on log-log. Here are some examples. Hopefully the first two are up you alley and the last is one of my favorite demos of the usefulness of loglog plots.

    http://earthscience.stackexchange.com/questions/875/are-richter-magnitude-10-earthquakes-possible
    http://rds26.tripod.com/ResponseSpectrum.html (civil engineering earthquake engineering dynamic analysis for normally static structures in a dynamic world.)

    http://zingale.github.io/astro_animations/
    https://raw.githubusercontent.com/zingale/astro_animations/master/solar_system_motion/harmonic_law_fig/harmonic_law.png
    (slope is 2/3 or 3/2 instead of 1)
    https://raw.githubusercontent.com/zingale/astro_animations/master/solar_system_motion/harmonic_law_fig/harmonic_law.png

    ReplyDelete
    Replies
    1. Thanks for the heads up. I have used log-log plots, especially in discussing scale invariance of deposit sizes.
      http://www.worldcomplex.blogspot.ca/2011/05/scaling-laws-of-life-and-death.html
      http://www.worldcomplex.blogspot.ca/2012/01/gold-part-2-is-there-maximum-size-for.html
      http://www.worldcomplex.blogspot.ca/2012/01/scale-invariance-and-scaling-laws-of.html
      http://www.worldcomplex.blogspot.ca/2011/02/scale-invariance-in-geological.html
      Earthquakes and other natural hazards were the first system I ever applied these methods to.

      Delete
    2. I`m going to look over those intersting articles of yours.

      To clarify: The diagonal lines on the graph can each have a numerical value.

      I tried to find an engineering graphical calcular with diagonals labled but could not rember its name.

      Ok, here is one, Fig. 5.

      http://www.fluke.com/fluke/sgen/community/fluke-news-plus/articlecategories/electronics-news/555-timer

      So if

      Delete
    3. So if you have a path, the value of diagonal lines are isoquants.

      Delete
  2. log log lets you see both variables and which ones are moving.

    ReplyDelete
  3. OK I made the graph of the path. The diagonal lines are where are drawn through the points (ln x,ln y) where ln x+ln y= ln xy=constant.


    From 2012 to present:
    https://research.stlouisfed.org/fred2/graph/?g=3Du1

    From 1995 to present:
    https://research.stlouisfed.org/fred2/graph/?g=3DtY

    https://research.stlouisfed.org/fred2/graph/?g=3DtV

    Line graph:
    https://research.stlouisfed.org/fred2/graph/?g=3Duj

    ReplyDelete
    Replies
    1. I realize now you were referring to postings like this:
      http://www.worldcomplex.blogspot.ca/2015/08/usdx-index-vs-gold-once-more-with.html

      . . . especially with regards to your comment on isoquants above. I had considered plotting these on a log-log graph, in which case the isoquants would appear as diagonal lines.

      Delete
    2. Do you have any thoughts to share about the significance of changes in slope of the different segments of the graphs? Do you think they reflect changes in policy? Or "market confidence"?

      Delete
    3. RE: leap year at 9:46

      Exactly... Diagonal slope is +/- 1 for isoquant of xy or y/x.

      A strait line is Y=mX + B.

      If stait line in log log graph then Y=ln y etc.

      ln y= m ln x + ln b
      ln y^1/x^m = ln b
      or


      Delete
    4. RE: leap year at 9:53.

      I wondered if I should think about it. Of course it can be anilised purely mathematically.

      Now thinking about your question... I would guess there are many more variables at play than the just A or B.

      A = changes in policy(t) or policy(t)
      B = change or level of market confidentce(t) (sounds like econo word)
      C = gold commodity cycle

      Contrarian view: On B, in 2008 GFC gold dropped. Sart of Syria bad news, gold droped. Even J.R. said it might go up over Syria bad news.

      Short run: Knowing that would be golden. As changes per shorter time makes for higher rates and higher capital gain taxes.

      But here is what I have thought previously years ago.

      Long run:

      In short I think for each commodity there are long term cycles. Like near 30 years. On top of that now we have currency changes.

      I have had differnt long term view than many. If history repeats itself than gold could go down oscillating for about 15 more years. Or, if there are a long commodity cycles. Gold is now part commodity. Of course that is not exciting short term. But valuable info if on wanted to buy and hold for 15 years.

      In 2011, I vaugley rememberd from childhood gold does not always go up. And, it would be easy to empicly test just by graphing the gold price history. From 1980 to 1999 gold oscillated going from $800 to $240 over 20 years. Thus that idea that gold always goes up was proven quite false.

      Over that time for 20 years it would have been better to hold us dollars than gold! But, obviously over 20 years the dollar lost significant purchasing power! So, the question is not hold dollars or gold but hold what from a huge menue? Or there is a better question. Or trade from a huge menue. Now I´m at a loss for what is best. Looking back owning other things would have produced great gains.

      But, you have expertiese in geology, exploration, and mining industry. Successful speculators and investors say do what you know. --- J.R. and Buffet.

      My memory. In the 70s we had loss of purchading power of dollars and the problems with it. Sophisticated and history minded adults gave children the silver coins that were going out of circulation. $1, 1/2, 1/4, and .1. We were told to save them and why and to grab the few we see. And we did. Adults should still do this.

      Now, I remember being shown gold coins and people getting in at the end and being a little disapointed that the price droped significantly. How could that hapen when the currency was loose? But, they knew correctly with out gold being connected to US dollars after 1971 that the value of the dollar would exponetially decrease, ie history and the 70s experience. I also remember hearing very little about gold for a long time after. I did not get in on the run up. I analised to late for gold.

      I think it looks like about 30 year commodity cycle. Commodities seem to have long term cycles in addition to short term.

      So, in the early 80s Interst rates were raised. Policy changed supporting the dollar but not preventing printing. And, in the 60s and 70s market confidence was erroded due to policy.

      I might think more on your question.

      Delete
    5. A, 300 odd contries have Z odd policies. And a Zillion market participants are doing W(t) different things at differnt times.

      I am not in this market right now.

      Delete
    6. On slope. Mathematically it is simple.

      Causauly or what they reflect is harder.

      Delete
    7. When abs(slope)<>1 then Au !exactly = anti-dolllar douring that period.

      Delete
    8. Just by the graphs:


      The gold run up 2000-2012, monthlty (or supper anti-dollar)
      https://research.stlouisfed.org/fred2/graph/?g=3DtY

      So first period and last period kinda look atnti-dollar; with gold going down and dollar going up? Or is it anti-gold?
      https://research.stlouisfed.org/fred2/graph/?g=3DtY
      https://research.stlouisfed.org/fred2/graph/?g=3DtY

      Delete
    9. I interpret the first graph as more significant gold price run up.

      Delete
    10. Do you have any thoughts on your questions at February 29, 2016 at 9:53 PM?

      Delete
  4. That said, if some other dynamic comes with a magnetude greater that increased supply form increased mines that will be worked for more than 15 years, I guess gold could go up.

    Those would be the usual things that the gold sellers and holders talk about and atempt to predict. If the most of the emerging market prints or most of the emerged market prints...


    ReplyDelete