Gold ETF Impact

Adam Hamilton     February 10, 2006     3733 Words


On November 18th, 2004 the first gold exchange-traded fund in the United States started trading.  Known as GLD, this trust spearheaded by the World Gold Council granted American stock investors the opportunity to buy a stock-like asset designed “to track the price of gold”.


If you follow the gold world closely, you will remember the enormous controversy the launch of this ETF generated.  Some folks were thrilled that this gold ETF would open up gold to broader investor participation while others were convinced that it was a Trojan horse designed to suppress gold by shunting capital away from physical gold into yet another form of paper.  GLD divided the gold camp like few other events, and very intelligent hardcore pro-gold people joined both sides of the raging debate.


Actually this GLD background and the surrounding controversy is really valuable to understand and it puts much into perspective.  If you subscribe to our Zeal Intelligence newsletter, you can use the logon information on Page 8 to log in to our archives containing past letters.  The 12/04 ZI published less than two weeks after GLD’s launch goes into great detail discussing the mechanics of this trust and its implications.


Today almost a year and a quarter later, the intense controversy surrounding the GLD launch has been largely forgotten.  GLD is apparently growing more widely accepted.  Contrarians on the gold forums talk of using it to park stock capital in between gold-stock trades.  And when discussing gold prices commentators on financial television sometimes bring up GLD as an easy way for the average mainstream investor to gain exposure to the gold price.


Since GLD launched I have been very interested in observing its performance relative to gold technically.  This encompasses a variety of perspectives.  Do GLD’s underlying gold holdings change in direct relation to the price of gold?  Does GLD’s trading volume react to gold prices in a typical manner?  And does GLD really accurately and consistently track the underlying price of gold?


With the first full calendar year of GLD trading now under its belt, I think we finally have enough data to make some initial observations on GLD’s technical behavior and impact.  Due to gold’s behavior last year, spending the first half in a consolidation and the second half in its biggest bull-to-date upleg, the 2005 data is an exceptionally good sample to use to evaluate GLD.


We’ll start by examining GLD’s gold holdings over the life of the ETF.  This holdings data is courtesy of GLD’s custodians.  GLD’s reported gold held in London vaults in metric tonnes is charted under the spot price of gold.  This chart is the most fascinating of these three charts for me as it shows the GLD ETF getting larger and larger even as gold drifted listlessly for nearly 10 months.  I certainly had not expected this.



Prior to its launch, GLD proponents including me expected the gold trust to add significant volatility to gold.  The broader the array of investors and speculators with easy access to gold, the more capital can come into play to drive its price around.  I had expected this to be a double-edged sword though, stock capital shunted into gold via GLD would drive uplegs up higher and corrections down faster.  So far the latter hasn’t been the case though.


In the first half of 2005 gold sentiment was pretty rotten.  The metal eclipsed $450 briefly in late 2004, incidentally just after the GLD ETF launch which was probably a contributing factor, then it ground lower for a couple months.  It rallied last February and March but soon started collapsing again trapping a lot of bulls.  For the next six months gold did nothing exciting, half-heartedly rallying occasionally before grinding back down to its baseline 200-day moving average.


I am amazed to see that the GLD gold holdings did not follow gold’s five significant slides lower last year.  Even while gold dipped GLD holdings remained stable to growing on balance.  I had expected a different outcome, that physical gold dips would lead to enough GLD selling that the ETF would have to sell some of its own gold to keep GLD’s price tracking gold.  This is very encouraging but requires some theory to understand.


ETFs are explicitly designed to track the price of their underlying assets.  This is not an easy undertaking though.  The underlying asset in this case, gold, has its own unique supply and demand profile that is independent from GLD’s totally separate supply and demand dynamics.  The only way that GLD can track gold is if the trust has an active connection to the physical gold world to stabilize supply and demand differentials between GLD and gold.


GLD’s physical connection to equalize supply and demand is its ability to add to or shrink its gold holdings.  Depending on the scenario, demand and supply are shunted from GLD to gold or vice versa as the ETF trust buys and sells gold occasionally to keep GLD tightly tracking gold prices.  This is accomplished by the trust buying and selling gold “baskets” in the spot gold market.  Each basket represents 10k oz of physical gold and 100k shares of GLD since GLD is supposed to track the price of 1/10th of an ounce of gold.


There are two primary scenarios to consider, GLD demand outstripping gold’s and GLD supply exceeding gold’s.  In the first, American stock investors get excited about GLD and start bidding up its price.  GLD starts to climb but gold may not be following if there isn’t parallel physical gold demand.  Soon the GLD price gets too high relative to gold and starts decoupling.  The GLD ETF needs to shunt this stock demand into physical gold to equalize prices and maintain tracking.


In order to accomplish this it issues new shares of GLD to soak up the excess demand and stabilize prices.  With the capital that selling these new GLD shares nets, the trust goes out in the physical market and buys gold in 10k oz baskets for its London vaults.  This process takes GLD demand and shunts it into the physical world by using this newly-issued stock capital to buy real gold.


If you examine the last couple months or so in this chart above, this has just happened and it contributed to the most impressive gold upleg of this entire bull to date.  As soon as gold broke $500 American investors naturally got excited and started pouring so much capital into GLD that it started to rise far faster than gold and threatened to decouple.  The ETF trust had to issue more shares to keep GLD tracking gold and it used this capital to buy gold and radically ramp up the trust’s vaulted London gold holdings.


While I certainly believe that the dawn of Stage Two global gold investment demand was the primary driver by far in this latest massive gold upleg, after studying this chart I believe that GLD was a peripheral factor too.  From December 1st when gold first closed above $500 to its latest interim high of $572 on February 2nd, GLD’s reported holdings rocketed up an enormous 114 tonnes, a massive 50%!


According to the World Gold Council, only 29 nations now report individual official gold reserves larger than this two-month GLD increase!  And only 16 nations now hold more physical gold reserves than GLD!


And this is indeed the magic of the gold ETF, why I have been so excited about it for years now.  Stock trading is the easiest and most developed form of trading in the world.  And it is also a vastly larger market in capital terms than hard commodities futures.  GLD enables American stock investors’ demand and capital to be shunted into the physical gold world in a quick and painless way.  It creates a conduit that allows vastly more capital to flood into gold than would trickle in via traditional futures or physical coins alone.


But this shunting should work both ways in theory, a double-edged sword.  What if something, like a poorly performing gold price, made American GLD investors skittish and they started selling en masse.  GLD prices would start to fall but the underlying physical gold market may not have similar selling pressure.  This creates a problem for the ETF with GLD threatening to decouple to the downside.  The only way to rectify this situation is for the trust to shunt the GLD supply into physical gold.


In order to arrest a unilateral GLD decline, the trust has to start buying back its own shares.  But it generally doesn’t have any cash so the only way it can raise the capital to purchase its shares is via selling 10k oz baskets of its physical gold.  So it starts selling physical gold in London which puts pressure on spot gold prices.  Then it takes this capital and buys back GLD shares, absorbing the GLD selling pressure.  The GLD price soon stabilizes and tracks gold once again when the GLD supply pressure is equalized into physical.


Now during gold corrections I had expected this scenario to happen.  GLD sales would be so high that the trust would have no choice but to sell physical gold and therefore increase downside gold volatility.  Yet in the chart above this generally didn’t happen.  GLD’s holdings were on a solid growth track with only occasional slight declines despite gold’s lackluster and slumping behavior in the first half of 2005.


This behavior is fascinating.  It implies that GLD selling so far, when gold is weak, is remaining roughly proportional to actual physical gold selling.  GLD investors are not getting more scared than gold investors and are not dumping their holdings at a significantly faster rate than those in the physical market.  This also suggests that GLD is being picked up more by long-term investors, who aren’t prone to sell, than gunslinging speculators.  This is fantastic news as GLD’s greatest promise was opening up gold to non-traditional gold investors.


Ultimately this could have an enormous impact.  The American GLD ETF is but one of a handful of gold ETFs trading in major markets worldwide.  Together all these ETFs are creating conduits for global mainstream stock investors to take a small stake in a gold-tracking asset.  Small stakes times hundreds of millions of investors equals enormous amounts of capital.


By shunting stock demand and capital into gold, gold’s bull market ought to grow even more powerful with broader participation.  The more investors join us in bidding up gold, the better our gains will prove before this bull gives up its ghost.


A final caveat is important here.  Just because GLD has so far seemed to only help gold by increasing its holdings when gold is surging doesn’t mean this will always be the case.  Less than a year and quarter of data really isn’t much in the grand scheme of things.  If GLD investors ever get more scared than gold investors and start dumping their shares too fast, the trust will have no choice but to sell gold in the open market to buy back its shares which will exacerbate downside gold volatility.


Next I wanted to take a look at GLD volume relative to the price of its underlying gold.  At Zeal we have done many studies on volume in various markets and I’ve been wondering how GLD stacked up.  Not surprisingly it is tending to trade more like a stock than a commodity since stock-investor capital is driving it.



In the stock markets trading volume rises on excitement, both positive and negative.  But when speculators are not excited, which usually happens when prices are moving lethargically sideways with no action, trading volume tends to be low.  We see this above in GLD as its volume was pretty modest back in the first half of 2005 when gold was struggling and gold sentiment was rotten.


But when things got exciting investors and speculators dramatically ramped up their trading volume in GLD.  This happened near both strong surges and pullback lows.  During strong surges, such as last September’s, rising gold prices drive dramatically increased interest in GLD and its volume rockets higher.  This is greed driven, of course, as nothing excites investors more than rapidly rising prices.


The other side of this coin is fear-driven excitement.  Near major interim lows, such as last February, gold prices had been falling for long enough to really spook GLD investors.  Investors are notorious for their herd behavior and near bottoms they all tend to get scared at once and bolt for the exits in unison.  This is very typical and expected behavior for stock-market traders and it is interesting that GLD is following the stock volume model.


Since late last year GLD volume has actually been trending higher, as shown by the red dotted lines framing its recent uptrend.  The higher markets rise, the more investors become interested in them.  Nothing begets interest and seduces capital into a particular sector like persistently rising prices.  Interestingly secular rising open-interest trends mirroring secular rising price trends are common in commodities as well, as I outlined in my studies on gold and silver futures open interest last year.


With GLD’s volume profile lining up with expectations for stock trading and even mirroring commodities in some ways, I find the most fascinating aspect of this chart from reading between the lines.  Last February, May, and July GLD volume spiked on fear near interim lows in gold.  While GLD investors were spooked into selling to some degree, thankfully their selling was not disproportionate to that in the underlying gold market.


This inference is supported by the fact from our first chart above that GLD gold holdings remained stable during all three of these high-volume GLD selling sprees.  And with the GLD gold holdings stable and GLD still tracking gold, it reveals that there was not a significant supply and demand disconnect between the ETF on the stock markets and gold in the physical markets.  This is really intriguing and is once again not what I expected when gold was weak.


Since the GLD ETF was explicitly designed with the stated goal of tracking the gold price, this is really the most important factor in its ultimate success or failure.  If GLD can accurately track gold it will win support from mainstreamers and contrarians.  But if GLD is not actively managed rendering it unable to track gold tightly, then it will be deserted and avoided like the plague.  Thankfully it has marvelously succeeded in tracking gold so far in its young life.



While this chart looks busy, it is really pretty simple.  Spot physical gold is tracked in dark blue on the right axis and GLD is superimposed on top of it in yellow on the same axis.  Since each GLD share represents 1/10th of an ounce of gold, the GLD closing price is multiplied by 10 to get it into common one-ounce units.  As you can see visually, the GLD and gold price lines are extremely tight.  You could interchange one for the other and even a diligent student of the markets would have no idea which was which.


They actually run a stellar correlation of 0.9955 which yields a magnificent 99.1% r-square value.  This means that 99.1% of the daily price action in GLD is directly attributable to and explainable by the underlying action in gold.  This seems very acceptable to me and I suspect it is in line with the correlations achieved by far larger ETFs like the famous QQQQ NASDAQ 100 tracking stock.


I ran some more variance analysis on the left axis.  The GLD closing price (times 10) is compared to both the spot gold price and the London PM gold fix on a daily basis and the percentage difference between GLD and the metal it tracks is charted.  Most of the time it seems like GLD is within a half percent or so of where it should be relative to actual closing gold prices on any particular day.  This is really pretty darned close.


On the red spot-gold variance, GLD tended to periodically have extremes where it was plus or minus 1.6% or so away from its target.  These may seem large, but they are acceptable for the way this trust has to interface with the physical gold market.  When GLD is threatening a disconnect from gold, the trust waits until the end of the day to see how GLD demand and supply stacks up against physical demand and supply.  If an adjustment in GLD’s physical gold holdings needs to be made, the underlying physical gold is usually traded the next morning.  This is why even the extremes tend to be short-lived.


But measuring GLD relative to spot gold, even though this is the most relevant comparison for investors, is not necessarily fair.  The reason is the trust bases GLD not off of spot, but off of the London PM fix.  Now regardless of how anachronistic or ludicrous it is in our Information Age for a price to be set in a smoky room by a few men in private, GLD ought to be measured by its own chosen yardstick.  As the white PM fix variance shows, GLD is even tighter to that particular gold measure than the spot price.  Extreme PM fix variances generally only run about plus or minus 1.2%.


So if you are a mainstream stock investor who is interested in having some gold exposure, GLD does indeed track gold well enough to be a viable alternative to all the hassles and large commissions involved in buying futures or physical gold coins.  GLD provides a quick and painless way to get some gold-like performance into a portfolio and the ETF is really becoming popular as evidenced by its soaring gold holdings.


Now personally I am a hardcore contrarian and the only way I want to own gold is physically in the form of national one-ounce coins in my own immediate physical possession.  While I think GLD is a great idea I have never owned any and probably never will.  For me gold is a physical asset I control and the ultimate insurance policy against the always possible (but quite improbable) total systemic meltdown.  In such a doomsday scenario all paper gold will be useless or inaccessible at best and physical will be all that matters.


But as I have argued for years, gold ETFs are not designed for established metals investors.  GLD’s target market is mainstream investors who do not already own gold.  Its mission is to broaden investor participation in the gold market by giving people an easy way to gain gold exposure in their already familiar stock accounts.  And it is brilliantly succeeding in this regard.


Many contrarians loathe the GLD ETF because they believe that its underlying physical gold holdings cannot be verified independently.  I think this is a weak argument against it though, as inability to verify plagues every paper asset on earth.  GLD is certainly not unique in this regard.


I have brokerage accounts and they send me statements monthly or quarterly that tell me what my stock holdings are.  Yet all I have is those pieces of paper.  I have no idea if my brokers really bought and are holding the stock they told me they bought.  For all I know some third grader with a color laser printer might have hijacked my accounts and be mailing me cunningly forged statements.


My banks also mail me statements monthly.  They tell me how much cash I am supposed to have but really all I know is the bank sent me a piece of paper that told me what I wanted to believe.  Until I go down to a bank and demand a 100% cash withdrawal on the spot, I have no way of knowing if my cash is really there or just a computer fiction.  My digital gold accounts are the same way, I log on and they tell me I have gold balances but who knows if the gold really exists.  Even as a CPA and ex-Big Six auditor I sure don’t!


GLD is the same way.  It is not physical gold and will never replace it.  It is a paper asset designed to track the gold price and nothing more.  It is not a substitute for gold in your own physical control if you are a gold investor.  It is not designed for you if you are already a gold investor, it is for your friends and family members who wouldn’t buy a gold coin if their life depended on it but they may put a few percent of their capital into a stock that tracks gold.  GLD is for the mainstreamers, not the contrarians.


At Zeal we started recommending buying physical gold coins in May 2001 when gold traded under $265 per ounce.  These physical coins form the core of our investment portfolio.  We then add investments and speculations in elite gold stocks to greatly leverage the gains in gold.  GLD will track gold’s gains one-for-one, but gold stocks can often multiply them many times over.


While we were just stopped out of our latest gold-stock speculations deployed last year at realized gains running up to 100%+ per trade, we are preparing to deploy more capital back into elite gold stocks once probabilities seem to be wildly in our favor again for the next major gold upleg launching.  Please subscribe to our acclaimed newsletter today if you want to join our coming redeployment into the most promising gold stocks when the opportunity again looks the most ripe.


The bottom line is the GLD ETF, despite all the controversy it has generated in contrarian circles, has fulfilled its mission so far.  It is accurately tracking the price of gold and it is providing a conduit for mainstream stock-investor capital to flow into the underlying physical gold market.  And surprisingly interest in GLD continues to grow even when gold is lethargic as evidenced by GLD’s stable to surging reported gold holdings.  Worldwide the various gold ETFs are really causing a major impact in the gold markets.


And while GLD is no substitute for physical gold in your own hands, GLD may be an option for any friends or relatives who cannot be swayed to buy physical.  Mainstreamers can still benefit greatly by diversifying some fraction of their assets into the GLD ETF to ride this powerful gold bull indirectly.


Adam Hamilton, CPA     February 10, 2006     Subscribe