Mining the XAU for Gold

Adam Hamilton    June 8, 2001    4987 Words


As we all continue our collective journey through these hyper-fascinating early years of the Information Age, we continue to observe the gold market action with alternating interest, incredulity, and excitement.  In light of gold’s ancient role as the ultimate financial asset, virtually everything affects the gold markets making them extremely complex and challenging to attempt to analyze.  Everything from global physical gold supply and demand, to “paper gold” derivatives, to events in exogenous markets such as equities and bonds, to government policies on fiat currency creation, to wars and rumors of wars, can have a myriad of direct, indirect, immediate, and lagging impacts on the price of gold.


While slogging through the market melee and trying to wrap one’s mind around the gargantuan information flows that can potentially affect gold, it can be enlightening and valuable to take a step back and re-examine the relationship between the ancient metal of kings and the stocks of the companies that almost magically separate it from its rocky tombs.  The interaction between gold and the gold stocks has been analyzed innumerable times in the past, and will likewise be interrogated ad infinitum as we plunge into the future.  With great respect and deference for the brilliant gold analysts who have examined this relationship in the past, we offer another perspective on the relationship between gold and gold equities.


One substantial hurdle involved in analyzing the interaction of gold and gold stocks arises right away due to the lack of good tools available.  Unfortunately, the most widely followed gold stock index is the Philadelphia Stock Exchange Gold and Silver Index, commonly known by its symbol, XAU.


The XAU has been in widespread use since late 1983, when it was constructed as a batch of gold miners with a little bit of silver thrown in to attempt to track the precious metals sector of the equity markets.  The “100 mark”, the index value of the XAU, was set equal to 100 at 1979 levels.  Options on the XAU began trading in December 1983 on the Philadelphia Stock Exchange.  The XAU is market-capitalization weighted, so companies worth more dollars in absolute terms in the equities market have a higher impact on the movements of the XAU.  Since the XAU has been carefully monitored by the markets for almost 18 years, it has carved out a role as the de facto standard for measuring equity levels in the precious metals sector.


This would be great, but the XAU is so watered down that it is hurtling towards worthlessness as a measure of the activity of quality gold stocks.  By quality gold stocks, we simply mean unhedged or lightly hedged companies in the business of mining and selling gold that have the potential to rake in huge profits when the gold price soars.  Of the nine stocks in the XAU today, the two largest that account for 39% of the value of the index have little relationship to the price of gold and greatly damage the credibility of the index.


The largest XAU component by market capitalization is Barrick Gold (ABX-NYSE).  Barrick is around the third or fourth largest producer of gold in the world, but it is far from a typical gold mining operation.  Barrick relies incredibly heavily on hedges, where it has aggressively bet that the gold price is NOT going to rise for the foreseeable future.  In the latest numbers we have seen, Barrick has locked in a selling price for the equivalent of 100% of the gold it is likely to mine in the next four years.  The company has sold vast amounts of gold forward dumping it on the market and is a major factor in the great damage that wanton hedging practices have wreaked on the price of gold and the entire gold mining industry.


In his landmark lawsuit against the Bank for International Settlements and bullion banks allegedly involved in suppressing the gold price, Reg Howe asserts that Barrick had material inside knowledge of the decisions made by a few Western governments and bullion banks to try and stealthily place a price cap on the international gold price through strategically liquidating large amounts of physical gold from central bank vaults whenever necessary to satiate demand and slaughter any fledgling rallies.  According to Howe v. BIS, Barrick was in a position to use this information to snatch up gold properties at firesale prices and unfairly benefit from the gold market manipulation scheme. 


Most hardcore gold investors we have had the pleasure of meeting around the world would not buy Barrick stock if it were the last gold mining company on earth, as it has evolved into a derivatives-based hedge fund-like monstrosity.  Often called “a hedge fund in drag”, Barrick would be in serious trouble in a mega-gold rally as its enormous hedgebook could rapidly implode in an Ashanti-style debacle as gold raced through $400 to new peaks.  If you are speculating or investing in gold stocks to leverage the coming gold rally, it is probably wise to avoid heavy hedger companies like the plague.  This hedge fund masquerading as a gold mine makes up 24% of the XAU.


Now it is not too hard to understand how a hedged-to-the-hilt company like Barrick, which has sold-off virtually its entire upside for the inevitable end of the gold bear market, has been granted a position in the XAU.  After all, Barrick DOES actually mine gold, it just sells it before it is mined at a hedged price that is FAR lower than potential price levels in a neo-gold rally.  On the other hand, the wizards controlling the XAU must have had a little too much to drink before they decided to add Phelps Dodge (PD-NYSE).


Phelps Dodge is a great company that produces BILLIONS of pounds of copper a year, and is a world-class primary copper miner.  Perplexingly even to the company itself, the Philadelphia Stock Exchange added it to the XAU, a gold and silver index, even though gold and silver mining is NOT Phelps Dodge’s business and it only produces trivial amounts of gold and silver as byproducts of its huge copper mining operations.  In 2000, the company reported producing 148,000 ounces of gold as a byproduct.  If placed up against real gold mining companies, Phelps Dodge would BARELY make the top 50 list in global annual gold production!


Claiming Phelps Dodge is a gold mining company that should be included in the XAU is similar to trying to put Hershey Foods in a semiconductor producers index.  Intel makes chips, Advanced Micro Devices makes chips, Hershey’s makes chocolate chips (yummm!), so they should all be in a “chips” index, right?  Goofy XAU custodians!  Since Phelps Dodge is a miner, and gold companies are miners, there is apparently no reason in the twisted fantasyland of the XAU-keepers’ minds why they cannot be thrown into the same precious metals index.  Strange world!  The copper giant Phelps Dodge accounts for 15% of the weight of the XAU.  So far, with Barrick and Phelps Dodge, 39% of the Philadelphia Stock Exchange Gold and Silver Index is barely even related to classical gold mining.


Finally, the third largest component of the XAU is another heavily-hedged beast like Barrick.  Anglogold (AU-NYSE) is the largest gold mining operation in the world by far, but is also heavily hedged.  The South Africa based company has already sold and locked in prices for many years worth of its future production.  If gold rallies above those hedge prices in the next few years, Anglogold shareholders will not reap the benefits of the gold rally as their company is shackled in voluminous oppressive hedge contracts.  Anglogold is also mentioned in Howe v. BIS as having material inside knowledge of the official gold suppression scheme, and one of the US bullion bank defendants in Howe’s landmark lawsuit even sits on the Board of Directors of Anglogold.  One big, happy gold-dumping family!  This company comprises another 15% of the XAU.


When hedge-fund-in-drag Barrick, copper miner Phelps Dodge, and heavily hedged Barrick-wannabe Anglogold are added together, they make up a whopping 53% of the XAU!  Phelps Dodge’s earnings will see effectively ZERO impact from movements in the gold price and Barrick and Anglogold will both face breath-taking “mark to market” losses in their hedgebooks and possibly even solvency problems like Ashanti in a major gold rally.  Ashanti Goldfields, of course, was the African miner that was driven to near-bankruptcy by its hedgebook alone when gold had its spectacular Washington Agreement RALLY in autumn 1999.  Aren’t gold mines supposed to BENEFIT from rising gold prices?!?  With over half of the XAU comprised of companies that have fundamentals not closely aligned with the gold price, it is no wonder most serious gold market analysts take the XAU with a truckload of salt!


There ARE better gold stock indices out there, such as the American Stock Exchange Gold BUGS (Basket of Unhedged Gold Stocks) Index (symbol HUI), but they have not been in existence as long and do not YET have as wide of following as the XAU.  Someday the XAU will be eclipsed by an honest gold stock index like the HUI, but for know it unfortunately remains the best long-term tool we have.


A couple weeks ago I received an e-mail from an esteemed member of the Gold-Eagle Gold Forum (  The gentleman, known by his cyber-handle of PMTrader, suggested that it would be an interesting research undertaking to do an analysis of the spread between gold and the XAU.  By taking the gold price (say $300), and subtracting the XAU level (say 60), the gold XAU spread is obtained (240 in this example). 


PMTrader suggested that a current analysis of the gold XAU spread could be useful and revealing on two fronts, potentially addressing two common perceptions of gold investors.  First, there is the age-old question pondering how good of leading indicator gold stocks as measured by the XAU prove to be for the spot price of gold itself.  Second, there is a widespread perception that the gold XAU spread “should be” around 220.


The 220 spread idea is important as many today use it as a quick and dirty valuation proxy for gold stocks.  For instance, at a gold price of $270, the proponents of the 220 spread would claim the XAU should be around 50 ($270 gold less 220 spread equals 50 on the XAU).  For traders using this model, they would tend to buy the XAU (either buying the underlying stocks themselves or purchasing options on the index) if it was trading below the predicted level of 50, as the spread would be greater than 220 indicating the XAU was undervalued relative to physical gold.  On the other hand, proponents of the 220 spread would tend to sell the XAU if the spread narrowed below 220, indicating the XAU was overvalued relative to the price of gold.  For instance, if the gold price was trading at $265 and the XAU was at 60, the spread would be 205, too narrow using the 220 rule.  A trading decision would be made based on this information to sell the XAU in anticipation of a fall unless an imminent rally in gold is expected.


So, looking for enlightenment on the question of whether or not the XAU is a good leading indicator of gold prices and also for validation or refutation of the 220 spread idea, PMTrader and I saddled up and attacked the raw data.  PMTrader and I each worked independently analyzing the datasets, and compared notes throughout the process.  He even wrote a really impressive FORTRAN program to allow deeper analysis of the data than is possible in spreadsheet programs like Excel.  Both PMTrader and myself thought the initial results of our research were quite interesting.


The following graphs in this essay are an initial glance at some of the discourse we shared and an initial peek into the gold and XAU relationship.  We may present further conclusions in the future in additional publications, but this discussion covers an initial foray into the intriguing relationship of gold and the XAU.  And so we begin mining the XAU for gold!


Our first graph simply presents the raw data thrown in a graph.  Since 1984 was the first full year for the trading of XAU options, which helped lead to the XAU’s widespread acceptance as the premier gold stock index, we started there.  1984 is not a bad year to kickoff a modern gold analysis anyway, as it eliminates the spectacular bubble blow-off in gold that occurred in early 1980 and its resulting aftershocks.  In ALL the graphs in this essay, the dates on the X-axis are marked every 200 trading days, leaving about ten months between each X-axis tick on all graphs.  The first graph of the raw daily gold and XAU closes since 1984 provides a great strategic overview of the dataset.



One of the first observations that leaps out of this graph is expected, the enormous volatility of the XAU as compared to the sedate volatility of the spot gold price itself.  When the gold price goes up, world-class UNhedged gold producers see their cashflows and earnings increase dramatically.  The reason is easy to illustrate and understand.


Imagine a gold company that is unhedged and can produce gold at a cash cost of $290 per ounce.  If gold is trading at $300, the company can make an operating profit of $10 per ounce for the gold it mines ($300 spot price less $290 cash cost equals $10 per ounce profit).  Now, if gold goes up 10% to $330, what kind of gain in cash operating profits will our hypothetical company realize?


Even if the gold price moves to $330, the company STILL can produce it for $290.  The new cash operating profits after a 10% rise in the gold price to $330 are $40 ($330 gold price less $290 cost equals $40 profit).  So, cash profits rocket by 300% ($40 new profits less $10 old profits divided by $10 old profits equals 300%) even though there was only a 10% rise in the gold price for our unhedged mining company!  With this fantastic operating leverage to the price of gold, it is easy to see why gold stocks are much more volatile than the gold price.  Even a relatively small movement in the price of gold can greatly affect cashflows, and hence share prices, of gold mining companies.


We can also gain some initial insights from this graph into the question of whether the XAU is a good leading indicator of gold prices.  We looked at XAU rallies from the three major bottoms in the graph above as an initial foray into the puzzles of this realm.


At rally “1” above, the XAU actually fell dramatically as gold began to rally.  It wasn’t until gold had a very sharp spike that the XAU galloped northwards and showed some signs of life.  Gold continued higher, making a few intermediate tops, but the XAU lagged a little while longer.  Finally, right before gold reached its highest levels since 1984, the XAU roared up to its high on the graph.  Overall, from this 100,000 foot macro-perspective, the XAU did not appear to lead a gold rally, but rose more or less concurrently with the ancient metal of kings.


As we move on to the second bottom and huge XAU rally marked by “2”, the XAU did not break north out of the heavy resistance at the top of its trend channel (bound by the red dashed lines) until gold once again leapt towards the heavens in a large spike.  Following this XAU rally, the index remained high and traded in an exciting range for years.  Once again, however, from this broad perspective, the XAU did not lead gold but rallied dramatically the moment gold showed some fire in its belly.


After trading in a high, volatile range mirroring gold, both the XAU and gold suddenly fell off a cliff around early 1995, and the plunge is marked with the white arrow above.  Provocatively, this is the time that the gold carry trade was believed to have really kicked into high gear, this was the time that Alan Greenspan usurped his authority and took a seat on the Bank for International Settlements contrary to the original intentions of the United States of America, and this is the time when GATA notes all kinds of warning bells starting to ring in the gold markets.  Something fundamentally changed around 1995, as central banks began dumping their citizen’s gold hoards like mad and the gold price plummeted, along with the XAU.  From this point to today, all gold rallies were very short-lived and were beaten back down with aggressive physical gold selling within hours or days after they reared their heads.


That brings us to the final “3” XAU bottom, which occurred in 2000.  This latest rally is really interesting as the XAU began to rally well before gold had its latest rally culminating in the anomalous May 18, 2001 post-London close trading spike.  We discussed that extensively in our earlier essay “Gold Prepares to Erupt”.  Unfortunately, however, even though the XAU has led gold in 2001 it has not shattered the top of its descending trend channel established through much of the 1990s.  Also, alas, gold has yet to exhibit any substantial follow-through to the May 18 spike, so we remain in a state of limbo regarding the ultimate outcome of this recent exciting XAU rally we have witnessed.


So, from a big-picture strategic perspective, the evidence for the idea of the XAU leading the spot gold price in rallies appears to be inconclusive, at least in this raw dataset.  Granted, much more research needs to be done on this that zooms into tactical gold trading rallies on a much smaller time scale, but the big-picture strategic view from the above graph certainly appears like the XAU makes its really spectacular moves right alongside of gold, not in readily apparent anticipation of future gold rallies.  It will be interesting to do further research to see if the XAU truly does indeed lead gold in short-term rallies, and if so, what percent of the time the leading indicator relationship proves valid.


Moving on for now, let’s zero in on “the spread”.  Once again, the spread is simply the dollar price of gold minus the XAU.  It doesn’t mean anything per se, but it can be interesting when viewed over a time series to establish a meaningful trend.  In the graph below, which we scaled back to only include 1990 to present to provide better resolution on the 220 debate, the purple series represents the spread computed on daily closing data.  Once again, the X-axis tickmarks each enclose 200 trading days.



Right away this graph illuminates much regarding the famous 220 spread used by many short-term traders today as a valuation proxy on the XAU.  First, we notice that the spread DOES indeed cluster around 220 from 1997 or so until the present.  As we move back in time, however, the fairly solid 220 relationship of recent history rapidly vaporizes.


If the spread is calculated and averaged from January 1984 to the end of May 2001, it weights in at 269.  From 1990 to the present, we get 254.  From 1995, that drops again to 238.  If an average is calculated from January 1998 to 2001, then we obtain 221, from which the 220 “rule” was derived.


It is apparent the 220 rule HAS held fairly constant in the recent few years, but before that it wasn’t even close.  The question must be asked whether the time period from 1997 to 2001 where the 220 spread holds is a normally trading representation of the gold market.


We believe, for a lot of fundamental reasons, that the gold market between 1997 and 2001 is NOT typical.  While some believe the manipulation hypothesis and some do not, GATA has collected enough evidence of anomalous gold market trading activity defying logic, reason, and free markets in the last few years to make the producer of the X-Files Chris Carter giddy with excitement.  Even mainstream commentators are beginning to make more and more noise about abnormal activity in the gold market.


Because of GATA’s revelations that all is not as it seems in gold today, and also because of the HUGE supply/demand deficit overhanging the world gold market each year, we believe there is a very low probability that the gold market of the next three years will trade like the gold market since 1997.  As such, the 220 “rule” should be used with great caution and only by short-term speculators.  It is not appropriate as a long-term buy and sell valuation indicator on the XAU or the price of gold.


While investigating the leading indicator plausibility of the XAU and the 220 spread, PMTrader and I observed a really interesting relationship that may prove to be a much more reliable indicator of macro buy/sell timing on the XAU index.  In the remainder of this essay, we will attempt to step you through our logic in arriving at that conclusion.


To begin, the following graph shows gold and a synthetic data series created by dividing the XAU level by the price of gold on a daily basis.  The orange line shows the result of this calculation.  This and the remaining graphs run from 1984, so they encompass virtually the entire useful life of the XAU. 



Note that since 1984, the average value of the XAU divided by gold has been 0.256.  Even if we use shorter periods to draw our average over, the number still revolves around 0.25.  From 1990 to the present it weighs in at 0.262 averaged, and 0.221 is derived from 1998 to the present.  In general, with over 4400 trading days of closing data analyzed, the “golden ratio” for gold and the XAU appears to be around 0.25.


Stated another way, an almost 18 year average XAU/gold ratio of 0.25 means that the XAU, on average, revolves around a value of 25% of the spot price of gold.  If gold is trading at $300, for instance, the XAU could be expected to be near 75 ($300 multiplied by 25% equals 75).  Of course, employing an average simply creates an approximation that may be right or wrong at any particular moment in time, but this approximation is useful as it gives us a ballpark target for the XAU given a certain gold price based on 4400+ trading days of historical precedent.


Because of its much longer time horizon stretching back to 1984, this XAU/gold ratio of 0.25 existed well before the market manipulators of the Clinton Administration began pulling their levers behind the curtain and mucking around with the free markets.  This 0.25 ratio, unlike the much more recent 220 spread, existed before the United States capital markets were subjected to the silly socialist experiments like the President’s Working Group on Financial Markets (aka “The Plunge Protection Team”).  The 0.25 ratio also existed before Alan Greenspan scoffed at his authority and joined the Bank for International Settlements, and well before GATA believes the gold markets began to be corralled and ambushed by well-planned tactical sales of central bank gold.

In short, the 0.25 ratio has some history behind it well before the present messy gold market situation, unlike the 220 spread rule that was created in the midst of the anti-free market anti-gold maelstrom.


Using this 0.25 ratio, we can construct an “implied XAU”.  The implied XAU is simply the price of gold multiplied by 0.25.  Using our earlier example of $300 gold, we get an implied XAU of 75.  The graph below shows the actual XAU in light blue and the implied XAU in dark green.  The results of this analytical construct are quite provocative.



As the implied XAU line is a synthetic construct, it is much less volatile than the real XAU and tracks gold with a 100% perfect correlation.  It gets really interesting because it appears to have important valuation ramifications for the XAU.  The majority of the time the actual XAU deviates far from the implied XAU, a countertrend move is imminent to regress the actual XAU back towards the implied XAU.  Is this relationship a reliable valuation gauge for the actual XAU?


Note the red arrows in the graph above.  Each red arrow marks a time when the actual XAU moved high above the implied XAU determined by multiplying the closing gold price by 25%.  Virtually every single time that phenomenon occurred, the actual XAU was apparently overvalued and quickly collapsed into a sharp bearish slump.  Note that each red arrow was followed by an aggressive plunge in the XAU.  It is not too big a stretch of the imagination to posit that the XAU was racing ahead of the implied XAU as defined by our 0.25 ratio, so in the absence of a large enough gold rally it had to collapse to move back down into a more normal valuation range.


A warning sign on potential overvaluation of the XAU relative to the price of gold is certainly useful, but the XAU versus the implied XAU also appears to be a valid buy indicator as well.  Each time the actual XAU sunk far below the implied XAU, indicating low gold stock valuations and probably negative gold investor sentiment, the actual XAU soon launched a material and decisive rally, each marked by the green arrows.  The latest excellent XAU rally, marked by the “3”, also began when the actual XAU was far below the implied XAU, “implying” that the XAU was undervalued in relationship to the actual gold price, gold investor sentiment was too low, and the XAU was ripe for a substantial rally.


Interesting, eh?  Obviously more study needs to be done on this, but at first glance it looks like the implied XAU of 25% of the spot gold price appears to be a reliable valuation and maybe sentiment gauge for the actual XAU itself!  The plot thickens…


Our final graph elaborates a bit on the concept of an implied XAU of 25% of the gold price.  In this graph, we take the actual XAU and subtract the implied XAU from it.  We take that difference and divide it by the implied XAU.  The result is yet another synthetic data series, but this one gives us a percentage representation of the valuation levels of the actual XAU in relation to the implied XAU.  For instance, if the actual XAU was 75, but the implied XAU was only 50, this equation would yield a result of 50%, indicating that the actual XAU is 50% overvalued compared to where the implied XAU based on 25% of the spot gold price would place it.


In the following graph, this equation that shows the percentage over or under valuation of the actual XAU versus the implied XAU is represented by the blue lines.  The light sky blue line in the background is the actual data, and the bold dark blue line is a 50 day moving average to smooth out some of the extreme volatility.  When the equation implies the actual XAU is overvalued relative to the price of gold, the background is shaded red.  When the equation implies the actual XAU is undervalued relative to the price of gold, the background is shaded green.  Gold is the usual bold yellow series behind the XAU valuation data. 



Interestingly, this representation of the data brings us full circle back to our original question… is the XAU a reliable leading indicator for the spot price of gold?  Although we concluded above that the strategic perspective was inconclusive while looking at the raw data graphed, this latest permutation with the relative valuation of the XAU appears to be a fairly GOOD indicator of future spot price activity.


We focused on the 20% zones in the graph above.  When the XAU was 20% overvalued relative to the implied XAU, marked by the top white dashed line, it was a good time to SELL gold AND the XAU.  In four of the five spikes of the dark blue 50 day moving average of XAU valuation shooting over the +20% mark, the price of gold fell soon afterwards, and the XAU typically followed.  So, in the last 18 years, an XAU trading 20% above the implied XAU in 50 day moving average terms has usually been a valid SELL signal.


On the buy side, we witness a similar phenomenon.  When the actual XAU approaches the 20% undervalued mark relative to the implied XAU, a substantial rally usually follows in the spot gold price and the XAU.  This -20% undervalued level is marked by the lower white dashed line in the graph.  This apparent relationship is very interesting and provocative!


One further point to note.  The recent XAU lows of 2000 were marked by the most undervalued XAU in the entire 18 years it has been widely followed.  Using the above methodology and graph, it is apparent that this point in time represented a screaming buy.  Also, it implies that not only the XAU but also the spot gold price will recommence their rallies and continue marching north.  It will be most interesting to watch near-future events unfold and see if this latest buy signal below a 20% undervaluation proves to be a valid physical gold buy signal as it has in the past 18 years.


The bottom line is the XAU still appears undervalued relative to gold, the XAU may indeed be a valid leading indicator for gold if the data is massaged or tortured enough, and this current gold and XAU rally probably still has legs if past precedent is to prevail!  (And this prose was penned before June 8’s awesome post-London close gold spike!)


Definitely some encouraging food for thought for gold investors from mining the XAU for golden information!


Adam Hamilton, CPA     June 8, 2001     Subscribe