Use access key #2 to skip to page content.

TheDumbMoney (42.99)

Updating the Gold Model



December 13, 2011 – Comments (10) | RELATED TICKERS: GLD , AUY

I have blogged quite extensively, for example here, about Eddy Elfenbein's model for gold, and posted about it on multiple threads of Fool and other random bloggers, since at least July or August of 2011.  Peter Brandt digs it, and now I note another respected financial/investment blogger, David Merkel, of Alephblog, has updated and highlighted the model.  His post was also highlighted today on AbnormalReturns.  My post above says more about the model than Merkel's post does, but it is a symptom of the broadening of the appeal of this model.  Merkel thinks gold keeps going up because of this.  I do not, in the short term.  I think if you see signs of deflation without further Federal easing (and/or liquidity problems), as we have been seeing, then gold goes down.  I don't see gold having another major spike until and if either: a) the economy improves but without the Fed taking its foot off of it easing gas; or b) the Fed implements another round of quantitive easing.  But that's just one non-expert's opinion.

I view gold as essentially a currency speculation that people conflate with being an alternative to stocks.  As I have stated before, including here, posting as DTAF (though now that I check that link, it looks like all of the comments got deleted), I view gold miners as underperforming because they are just companies, and thus are essentially claims to 20-years of free cash flows or so.  Thus, since the stock market does not expect current gold prices to last for anything remotely approaching 20-years, they will never appreciate to the level one would otherwise expect, given current gold prices.  (Whether the market is correct or incorrect is a separate issue.)

Disclosure:  I am neither long nor short any gold, silver, gold miners, silver miners, other metals, related ETFs, leveraged ETFs, etc., etc. 

10 Comments – Post Your Own

#1) On December 13, 2011 at 9:50 PM, tekennedy (71.14) wrote:

Thanks for the links; those are very interesting articles.  My problem with gold investing is the lack of ability to value gold and although this doesn't solve that, it gives some method to evaluate its attractiveness relative to a cash position. 

Keep posting stuff like this!

Report this comment
#2) On December 13, 2011 at 10:56 PM, TheDumbMoney (42.99) wrote:

Thanks, te, I'm just the bringer of good models.  I agree, I think gold is as equally impossible to value as a currency is.  It's value depends upon a million different variables.  What is interesting is this model's ability to show, concretely, how its value varies in relation to real interest rates.

Report this comment
#3) On December 13, 2011 at 11:43 PM, FleaBagger (29.37) wrote:

The price of gold can almost always be priced as a function of the money supply times the expectation of the future rate of inflation. I myself don't have a precise formula for this, but I know that as money supply sharply increases (as it has in the recent past) and some experts fear inflation, the price of gold will, generally, go up until rates are raised, with temporary pullbacks. I use this as a broad, imprecise measure.

Report this comment
#4) On December 14, 2011 at 9:18 AM, XMFSinchiruna (27.97) wrote:


Thanks for posting those. 

As I've said before, I find it interesting as one tool within the massive tool bag that anyone hoping to assess or forecast gold prices must lug around, but truthfully its primary value is as a means to assess the current and historical prices relative to the price drivers considered by the model. Since Bernanke has frozen the Fed's rate through 2013, extrapolating those inputs into future expectations of real rates has at least grown a bit easier.

The problem I have with reductionist models of the sort is that they have a tendancy to work neatly until they don't. I will always have much more confidence in my comprehensive analytical process than I could have with any single model for a currency with as many driving factors as gold.

For example, while gold seems to be tracking the model pretty closely right now, what happened during the span between 1983 and 1990, when the gold price did not correlate as neatly? Without a thorough accounting by the proponents of the model for the cause of that rather dramatic 7-year departure, then reliability of future correlation irrespective of complicating circumstances can not be rationally anticipated.

I had wondered previously why Elfenbein had selected 1989 as the starting point for his chart here, but perhaps now I know the answer.

More importantly, what about the factors that are ignored by the model? Are Fools to assume that factors not considered by the model have no capacity to impact the price of gold? Of course not, and the moment one or more of those factors external to the model's parameters makes its/their presence known in the gold price, the modelers will be left wondering where their approach failed them.

To his credit, Elfenbein concedes this point, stating: "Let me make this clear that this is just a model and I’m not trying to explain 100% of gold’s movement."

As an example, the model ignores gold supply as a determining factor of the gold price. Bummer ... seems like that could be important. Now, it may be that over the period from 1989 to present, supply and demand were not sufficiently out of whack from each other to force a departure from the real-rate correlation. But let's say as a hypothetical that China decides to aggressively accelerate its official gold purchases and its population follows suit. With a mining industry that lacks capacity and reserves necessary to meaningfully add production, such a development could of course lead to a marked increase in the gold price irrespective of the corresponding trend in real interest rates in the U.S. The model, in that scenario, could break down in a heartbeat.

So while I do want to encourage Fools to explore and consider the model, I also wish to caution them against holding it aloft like some sort of elegant replacement of the exhaustive fundamental analysis that gold will always require.

Report this comment
#5) On December 14, 2011 at 12:22 PM, TheDumbMoney (42.99) wrote:


Thanks for your thoughts.  I linked to the link you link above in my original post.  As I have also noted, it doesn't explain everything.  Looking at Alephblog's 1970-Today representation, the sheer scale makes recent history look like a closer fit than it really is.  Actually Eddy's 1989-Today model more clearly shows recent divergences.  (And Eddy's 8/18 post, which I also linked to in my original post, takes the model all the way back to 1951.)

But there are no simple answers.  For example, the model actually suggests gold prices should have been rising over the last few months.  They have not been (and that in the face of, if I'm not mistaken) all-time record Chinese demand.  Even if the 'dominant fundamental' of gold is in fact as a currency savings alternative to the world's reserve currency when real interest rates are low (which can happen either in times like this, or in times of high interest rates where inflation is also extremely high), huge divergences are not only possible, but as you (and Eddy and David) point out, have been observed. 

I think it is helpful, however, to have a general trend-line that is somewhat backed up by forty or more years of data, and then to start from there when asking why the divergences do occur, and whether a divergence may be a short or a long-term one.  (For example, the late 2008 divergence was ultimately a short-term divergence likely caused by liquidity needs.)  In that regard, I actually think the 1983-1990 divergence, after which the fit much more closely returns, strengthens my view of the model.  Perhaps, as that was the height of the Cold War (during which my mother, bless her, and many others, actually bought space a private bomb shelter), fear of a nuclear end to the world overtook what was otherwise the dominant fundamental.  But who knows.  Not me.  I can only speculate.

Finally, if Peter Brandt is correct, the more people that become aware of this model, the more it will actually break down anyway. 


Report this comment
#6) On December 14, 2011 at 1:32 PM, TheDumbMoney (42.99) wrote:

Here is another attempted optimization of Eddy's model, from another financial blog, profitimes, published a few days ago:


Report this comment
#7) On December 14, 2011 at 2:13 PM, TheDumbMoney (42.99) wrote:

Apropos my comment above about the Cold War, in the 1980s, gold was something like 5% of the average investor's portfolio.  Today it's around 1%.  That 5% level in the 1980's I suspect goes a long way to explaining the divergence seen in the charts.  My Cold War theory is a plausible explanation, but so is the theory that investors remained enamored of gold as an inflation hedge (and terrified of +8% inflation) long after Mr. Volcker crushed the worst inflation; hence its high percentage in 1980's portfolios. 

I should note that this model actually made me much, much less of a gold bear than I was before I discovered it, because (however imperfect) it provides a numerically-based "explanation" for gold's recent meteoric rise.  It also potentially supports much higher gold prices, at least in the short term (1-3 years out), assuming interest rates and economic growth do not return to "Great Moderation" levels during that period, as I suspect they will not.

But as for me personally, I still view it essentially as currency and/or commodity speculation.  Just as I do not speculate on the Yen-to-dollar exchange rate, or buy cotton futures, I don't really do gold.  It's just not my bag.  But the model seems to support short-term (1-3 years) bullishness on gold at these levels (barring another liquidity crunch).


Report this comment
#8) On December 14, 2011 at 5:01 PM, TheDumbMoney (42.99) wrote:

I made some conflicting bullish and bearish statements above.

The bearish sentiment in the original post is based on the Fed's statement that it will continue the "Operation Twist."  This should raise short term real yields, hurting gold in the very short term, as discussed more below.  (If I'm right.)

The bullish sentiment is based on the fact that, nonetheless, short term real yields are plumbing all-time lows, similar to those reached at the August 2011 gold high (maybe this is in fact the cause of the Fed's recent statement), and the fact that I don't see short term real yields returning to 2% or so any time in the next three years. 

As discussed, my bullish stance is tempered by my perception of what the Fed is doing.  In other words, my view is that without Operation Twist, the price of gold would be much, much higher.  (Because short term real yields would be much lower, and long term real yields would be higher.)  If I'm right, it goes to show, in my view, the extent to which gold is held hostage by Fed policy.  (Not that that's necessarily a bad thing.) 

Theoretically, the Fed could keep "Twisting" for the next two years, putting more and more pressure on the price of gold.  (Assuming my understanding of the effects is correct.)  It is probably no accident that the price of gold took such a major tumble, truly breaking at least its very-short term trend, in the days immediately following the original 9/21/2011 Operation Twist announcement, which was explicitly about raising short-term real yields and lowering long-term real yields!  I think most people, even those who follow this model, don't quite understand what Operation Twist did, and how significant it was/is.

But that said, I just don't see how, no matter WHAT the Fed does, short term real interest rates reach a more "normal" level (and gold a much lower level) within the next few years.  Thus I think the more likely scenario is that an upward trend continues at some point during this time, and for a significant amount of time.  

And of course it's also possible, even likely, that investors/ savers will continue to increase gold holdings even after present issues are resolved, thus creating a 1983-1990-like divergence, even if the model is correct and short term real yields rise dramatically in a (some day) improving economy and under a more normal Fed interest rate policy.  And it's also possible the model breaks down more fundamentally and permanently.

It sure is an interesting story to watch.


Report this comment
#9) On December 14, 2011 at 7:07 PM, rfaramir (29.57) wrote:

"There’s an old joke that the price of gold is understood by exactly two people in the entire world. They both work for the Bank of England and they disagree."

The truth behind the joke is that the 'price' of gold is measured in fiat currency (probably not totally fiat when the joke first came out, but allowing fractional-reserve banking fraud is enough), and the value of the currency is up to the whims of central bankers. Trying to predict the price of gold is near-equivalent to predicting Bernanke's future actions. (Hmmm, predicting past actions is much easier.) Those two BofE employees simply disagree on the future actions of the BofE.

It's good to try to get your head around the price of gold, but don't expect math to help much. It is Human Action--supply of dollars by central bankers, demand for dollars by savers, producers of gold and purchasers of gold (both speculative and useful consumers)--that work together in a free market to set prices. (And remember it's not free when manipulators of this small market have government backing.)

Eddy's focus on real rates is a pretty good start, for those insistent on using math to predict human choices. It summarizes the quantity of confiscation of purchasing power the central bankers are getting away with, and therefore the *likely* reactions of savers, gold producers, and gold investors.

He makes some good points in his blog: "there’s no reason for there to be a relationship between equity prices and gold," "the TIPs yield curve indicates that low real rates may last for a few more years," and "the price of gold is essentially political."

(Long gold, wishing the currency he was paid in was as good as gold once again.)

Report this comment
#10) On December 14, 2011 at 7:25 PM, TheDumbMoney (42.99) wrote:

rfaramir, to extend on what you say in your fourth paragraph, this model does not exist in a vaccum.  Because of all of the macro factors that go into both interest rates and inflation (or deflation), I view it really as a rough attempt at actual quantification of a lot of the broader macro themes that people discuss.  Maybe not all of them, but many (if not most) of them. 

Ray Dalio (Bridgewater), alternatively, just uses the ratio of above-ground gold stocks to the quantity of reserve currency money currently in existence.


Report this comment

Featured Broker Partners