Monday, January 1, 2018

Happy New Year!

2018
Year of the POP



If you'd like to read "Year of the POP", I'm sure you can guess where it can be found. ;D

At the Speakeasy, we discuss things like bullion, bitcoin, bubbles, banking, and much much more. As I did in Nine, at the ninth anniversary of this blog, and back in May at the second anniversary of the new Speakeasy, I have started a new tradition of sharing samples from the Speakeasy here. For this one, I picked a fun post from back in August. It's about the coat-check room view of GLD, and it's called…

Just Another GLD Coat Check Room Post


With GLD once again on the potwatchers' radar, I thought it might be a good time to do another "coat-check room" post. Fortuitously, Rab sent me a pdf link from the SEC's website with a section on GLD. It's a lame assessment of GLD and how it works, but it's still very interesting in several respects. So what I want to do is to first give you the original article as it appears on the SEC's website. I think you will find it quite interesting. Then I will give it a thorough fisking.

First, the background. As I'm sure you are aware, there's basically an ETF for anything you can imagine. There are ETFs for stocks, bonds, commodities and currencies, there are leveraged ETFs, inverse ETFs, even leveraged inverse ETFs. For gold alone there are at least 16 different ETFs, with GLD being by far the largest. Yet even with their immense popularity, there are some serious concerns about them and their proliferation.

In the summer of 2015, the SEC put out a request for public comment on some questions it had about ETFs, or as it calls them ETPs (exchange-traded products). The primary questions revolved around arbitrage and pricing, but you can read the full list of 53 questions here, beginning on page 35.

The pdf from Rab was a 107 page response submitted to the SEC in late 2015. The section on GLD is only 7 pages out of the 107. The authors are anonymous, but "the SEC knows the authors." Here's what the paper says about that:

Author’s Comments

It is not important who supplied the information in this submission; the data comes from regulators, exchanges, ETP operators and ETP media sites promoted as researchers of ETP information. For the record, the SEC knows the authors of this submission and has had time to confirm the sources of the data.

We are taking a holistic view across various market metrics in this paper rather than just concentrating on a small subset of information, which is a consistent practice of ETP industry observers. In documents to regulators and investors, ETP operators consistently discuss the mechanics of the ETP itself, without talking about the secondary market trading for investors that ETPs create by design through their Authorized Participant business model. This model is flawed and investors in ETPs in the secondary markets are a significant focus of this submission.3

The market trading discussed herein (see data sources below the Summary Index) is being executed between investors and counterparties mostly consisting of Authorized Participants, market makers or clearing firms (which may be the same firms), which in many cases is not causing a net creation of shares (purchasing underlying assets) for certain important ETFs. In some ETPs, there is a conflict of interest between the investor and the contra parties in the secondary market.

Anyone that has been critical of ETPs has been immediately attacked by the industry, without any factual data from the industry to support their positions. The strategy has simply been ‘attack the messenger’, which does not address the underlying problems within ETPs.

This paper challenges the industry to actually address the data. It is designed to be read by regulators and financial industry experts familiar with ETPs. The data is based on long-term patterns and practices within the industry and shows the outcome of how ETPs are actually operating, not just a narrow industry view or study that distracts from serious flaws in ETPs.

The SEC, FSOC, CFTC, U.S. Treasury, the Federal Reserve and other global regulators are now concerned about different aspects of ETPs. These agencies appear to be ready for a serious discussion of the data. This will take input from the industry, not just simple responses to demean or deflect sophisticated ETP critics by saying they just do not understand how the products work. This industry answer alone raises significant red flags that ETPs may not be suitable for most investors.

The information below is not a 5 point bullet presentation because it is derived from data sets containing many millions of data points which are presented here in readable and hopefully understandable formats. As an example, we have assembled the short sale data from exchanges, which is so voluminous the SEC stated it was “unaware of the transaction-level data being widely used by any group other than academics,” and “data vendors informed the Division that they had not created products utilizing this data.” The SEC also stated “the Commission and other regulators lack direct access to the data necessary to quickly identify short sellers and short position holders.”

The data and the systemic risks revealed from a holistic view of available information should be discussed before the next financial crisis. ETPs are currently golden eggs benefitting Authorized Participants, ETF operators, clearing firms and short selling, which could shatter the U.S. economy again. Temporary profits are not worth risking the soundness of the global financial system.

3 Generally, Authorized Participants are required to trade with the ETF only in large blocks of typically 50,000 share units. Authorized Participants are not obligated to create shares/assets. ETF operators do not have any authority to cause the creation of shares/assets by Authorized Participants.



I'll tell you right up front that one of the aspects I found most interesting about this paper is they refer to the public market for ETF shares as the "secondary market". The Authorized Participants and ETF operators are the primary market.

This is an important concept to understand. Understand it, and the latitude necessary to operate GLD as a coat-check room becomes a no-brainer. I have been trying to explain this concept for years, and this is the first semiofficial paper I've seen which clarifies it.


As an investor in GLD, you already know you don't actually own the physical gold, but did you also know that you can never ever own even the shares outright? Because they are "in street name only," all you can ever own is a liability denominated in "GLD shares" from one of the AP banks. Just like an unallocated gold credit, or even a dollar in a Bank of America account, all you actually own is a bank liability. The same goes for GLD, and that's why the authors of this paper call the market that investors can participate in "the secondary market."

I explained this in a Speakeasy post back in 2013, titled Another Angle on GLD. Here's an excerpt:

"Did you catch that ZH article recently about Goldman Sachs holding 4.4 million shares of GLD? Of course you can completely ignore their analysis as they don’t know what they’re talking about, but the reason I bring it up is that bullion banks can hold these shares as reserves against their bullion liabilities. ZH thinks it’s a positional bet by GS to milk their muppets, but that’s just ridiculous. We don’t and can’t know how many of those shares are offset by GLD liabilities to GS clients nor how many are offset by XAU (unallocated ounce-denominated) liabilities. Those shares are simply reserves and GS has liabilities to its clients, some denominated in GLD shares and others denominated in gold ounces.


http://www.zerohedge.com/news/2013-08-30/guess-which-bearish-bank-bought-record-amount-gld-q2

That’s the definition of a bullion bank — a bank that also carries ounce-denominated liabilities on its books.

[…]

Notice that while Goldman Sachs is an LBMA market making member, it is neither a clearing member nor an approved LBMA custodian. But it is an LBMA bullion bank. So it has ounce-denominated liabilities on its books, but its physical reserves are in one of those other 10 vaults.

GS is also a broker dealer, market maker and AP for GLD, so it also carries GLD share-denominated liabilities. It should have one share of GLD for every GLD share-denominated liability it has to its clients, but that is not the case. It could have more or fewer actual GLD shares than it has GLD share-denominated liabilities.

Because GLD shares are issued as “global certificates” rather than “individual certificates” and because short selling is allowed, there are always more people on this planet that think they own a share of GLD than there are shares in existence. Another way to understand this is that GLD shares are listed in “street name only” which means that when someone short sells a share, the bank doesn’t need someone to specifically lend that share. It just lends a generic share from its reserve of GLD shares since all shares are registered with the DTC in the name of the bank itself. So there’s no one who is told that his share has been lent to a short seller. The bank itself lends that share out of its “GLD share reserves” and we end up with more people who think they own shares than there are shares in existence.

Here’s just a quick example of what I’m talking about:


“Take the SPDR S&P Retail ETF (NYSE: XRT) as an example. The number of shares short was nearly 95 million at the end of June, while the shares outstanding of the ETF were just 17 million. The ETF was over 500% net short! Or to look at it from another perspective, the ETF’s operator, State Street Global Advisors, believed that there were 17 million shares of the SPDR S&P Retail ETF in existence and owned shares in the S&P Retail Index portfolio to underlie those 17 million ETF shares. But, in the marketplace there were another 95 million shares of the ETF owned by investors who had purchased them (unknowingly) from short sellers.”
Source: www.businessinsider.com/run-on-an-etf-2010-9


That’s how it’s possible that a bank has fewer actual shares of GLD in reserve than it has GLD share-denominate liabilities outstanding. Some of the bank’s GLD share liabilities are offset by other assets, i.e. “GLD shares owed to the bank” by the short sellers. This is perfectly normal and it is not fraud. It is also possible that a bank, like Goldman Sachs, could have more GLD shares in reserve than it has outstanding GLD share-denominated liabilities. And that’s because GLD shares can also be used by the bank as reserves for its XAU or ounce-denominated liabilities.

[…]

So at the moment of creation of those shares or, on the other end, if GS buys them back from its clients and holds them as reserves against its unallocated liabilities, it could be holding more GLD shares than it has GLD share-denominated liabilities. The point of all of this was to show you that any one of these APs, at any given time, has an amount of GLD shares registered in its name at the DTC that is different from the amount its clients think they own. It could be more or it could be less, but it’s probably never exactly the same.

If you can wrap your head around this concept, then you will see how GLD is part of the reserves of the bullion banks. “GLD shares plus physical reserves combined” stand as fractional reserves behind “GLD share-denominated liabilities plus unallocated gold ounce-denominated liabilities combined”. Simple as that!

[…]

What do you think makes GLD so special among the gold ETFs? Why do you think it is so much bigger than the others? Do you think it’s just because it was first? It wasn’t. Perth Mint had a gold ETF in 2003 called ZAUWBA, now called PMGOLD, a full year earlier than GLD. Do you think it’s because of marketing, or brand? Sprott PHYS is a pretty good brand with heavy marketing in the gold bug community, yet PHYS is tiny compared to GLD. Back when I wrote ‘Who is Draining GLD’, PHYS only had 25 tonnes while GLD had 1,224 tonnes. Today PHYS has almost 50 tonnes (it has doubled in size) while GLD still has 911 tonnes, 18 times more than Sprott. So why is GLD so much larger?

I’ll tell you why I think it’s so much larger than any of the other gold ETFs. It’s because the size of the ETF has absolutely nothing to do with demand for the ETF. It simply has to do with how much gold the operator has available, and wants, to put into the ETF…

[…]

GLD had 1,353.35 tonnes at its peak less than a year ago, and still has 911.12 tonnes today, after a hell of a draining! So if popularity or demand was the driver of size, then GLD must have been really popular before and really hated now. Or maybe it has nothing to do with relative popularity and specific demand in the case of GLD. Maybe GLD is the coat-check room for the LBMA bullion banks’ reserves."


Now, with that in mind, here are pages 44-50, the section on GLD, in the Response to SEC Questions Regarding Exchange Traded Products - File Number S7-11-15 [none of the emphasis is mine, it's all just as it appears in the SEC paper]:

The State Street SPDR Gold Shares ETF (Symbol: GLD)

The State Street SPDR Gold Shares ETF (Symbol: GLD), holds only 1 commodity asset, physical gold. The GLD is a straightforward and simplistic ETF example that demonstrates the supply and demand for shares and values of assets are not showing the expected natural economic relationships between the trading of the ETF and its gold assets.

The GLD is an ETF that has experienced both periods when shares should have been created but were not and times when massive amounts of shares were redeemed (decreasing GLD assets at more than twice the rate of changes in the price of gold).

Each type of creation/redemption period shows unexpected risk for investors in the GLD. By any measure, the GLD exemplifies asset risk to investors outside of the fundamental movement in the price of gold.

Because the only underlying asset is gold, new investment in the GLD should result in the ETF purchasing more gold when gold is in favor as an investment. In fact, the GLD for 2 years reviewed did not purchase any significant amounts of gold to match the enormous trading of the GLD by investors.

The ETF operators know underlying assets are not adjusting to incoming net investments, but have not disclosed the fact that creations are not happening as the SEC expects for many systemically important ETFs.

Chart 2 illustrates the creation/redemption process along with the GLD price action (changes in investor sentiment for physical gold). The GLD price is reset each day to approximately 10% of the price of an ounce of gold.55 In Chart 2, the daily shares outstanding and closing prices for each month were averaged in order to keep the chart understandable for the long trading period.

55 This price fluctuates at a .5% premium or .5% discount.

44


Chart 2 – GLD Average Monthly Shares Outstanding vs. Average Monthly Closing Price from November 2010 through September 2014 (985 Trading Days – Almost 4 Years)


Change in GLD Assets Over This 552 Trading Day Period

On April 15, 2013, the closing price of the GLD was $131.31. The last time the GLD was valued at this level was 552 trading days prior, on February 2, 2011 when the closing price was $130.45.

For most of the 552 days, the GLD was a relatively ‘hot’ ETF and the data suggests significant incoming investment into the GLD occurred.

During the 552 trading days from February 2, 2011 through April 15, 2013, the reporting markets data showed more than 1 of every 2 GLD shares sold were the product of a short sale;

7.38 billion GLD shares were traded worth $1.17 trillion, but by April 15, 2013, the underlying physical gold value held by the ETF had declined by $784 million from the beginning of the period.

For the majority of this time, gold was ‘in-favor’ as an investment moving from $1,300 to $1,800 (in the first 150 days with flat or decreasing GLD shares outstanding). For the next 362 days the average price of gold remained above $1,600. It then took 40 trading days for gold to decline to $1,300 (February 18 through April 15, 2013), with the majority of the price decrease occurring within the final 3 days. Investment in physical gold and the GLD during this period had little economic influence on the GLD’s net purchasing of physical gold.

45


Table 23 shows the asset values and the total volume traded in the GLD during the 552-day period.

Table 23 – Asset Values and Volume Traded in the GLD from February 2, 2011 through April 15, 2013 (Two Years - 552 Trading Days)

Value

Asset Value February 2, 2011

$52,730,385,538

Asset Value April 15, 2013

$51,946,541,141

Highest Asset Value During Period (August 22, 2011)

$77,511,602,759

Lowest Asset Value During Period (April 15, 2013)

$51,946,541,141

Average Asset Value During Period

$66,500,114,530

Value of Shares Traded During Period

$1,170,756,995,770

By investing in the GLD, the investor pooled monies into an ETF that did not gain any new net holdings of physical gold (underlying assets of the GLD actually declined by $784 million) and investors only received the change in gold value (a GLD increase of $0.86 per share).

It appears that the pooled investor interests in the GLD were diluted with short sales that created large synthetic unreported short positions (producing excess leverage and undisclosed liabilities). The short sales were greater than half a trillion dollars when the average asset value was only $66.5 billion.

Simply put, despite investments, on February 2, 2011 and two years later on April 15, 2013, the ‘pot of gold’ owned by the GLD was the same pot of gold.

At the end of this 26-month period, an investor in the GLD did not have any more representative ownership of gold assets under management by its ETF operator than it started with and the investor asset value was diluted from high levels of short selling.

These findings of static assets (inadequate net creations during positive investment periods) and sustained short selling have continued in multiple large ETFs we have reviewed.

GLD No Net Creation – Strong Positive Investor Sentiment in 2011

It takes net creation of new shares for the GLD to purchase its’ underlying assets of physical gold.

As shown in Chart 2, from February 1st through September 6, 2011 (151 trading days) the price of the GLD increased from $129.92 to $184.49 or by 42%, while there was essentially no net creation of ETF shares to purchase the underlying physical gold. There were 404 million shares outstanding on February 1st and on September 6th, there were 406.8 million shares outstanding; an increase of less than 1% (virtually no net change in the amount of gold in the pot).

46


During the 151-day period, 2.7 billion GLD shares traded worth $412 billion. Reporting markets/SROs data showed on average 52% of the shares were sold short, equating to approximately 1.4 billion shares sold short at a value of $215 billion. The only change in underlying asset value is attributable to the GLD price change (+$54.57 reflecting the price increase of gold), not because of new purchases of gold by the ETF operators/Authorized Participants. This data is shown in Table 24.

Table 24 – GLD Data February 1, 2011 through September 6, 2011 (151 Trading Days)

2/1/2011

9/6/2011

Per Share NAV

$129.92

$184.49

Shares Outstanding

404,200,000

406,800,000

Total Assets Under Management

$52,513,965,320

$75,048,750,790

Volume

Value

February 1 - September 6, 2011

Between Dates

Traded Between Dates

Shares Executed (Consolidated Tape)

2,683,600,600

$412,372,915,050

Short Sales (Based on U.S. Reporting Exchanges Percent)

1,399,586,669

$214,968,361,274

Essentially, the increase in the total assets under management mirrors only the price increase of gold. Purchasers’ investment into the GLD did not generate net purchases of new gold assets. The data suggests products operating like this provide an efficient way to siphon investors’ monies through short sales while bypassing the purchasing or borrowing of assets. This is akin to selling fictitious financial instruments (thin air), which is illegal in the U.S.

Large GLD Redemptions Caused Multiplying Losses of Physical Gold Assets – 2013

From March 1st through December 31, 2013 (212 trading days), the price of the GLD declined by $37.25 or 24%, aligning with the $378 drop in the price of gold. At the same time, there was a large amount of GLD redemptions that decreased the shares outstanding from 416 million to 266 million.

Due to the combined price decrease in the underlying asset (gold) and the net redemptions in the GLD, the value of assets under management dropped by $33 billion or 52% during the period. Simply put, the price of gold and the GLD declined by 24% and the value of the GLD assets fell by 52%.

This is very important. The 212-day period of net redemptions shows ETF underlying assets became unhinged from the underlying gold. Assets decreased by more than double the rate of gold's decline in price. Investors wishing to participate in the gold market would not buy the GLD if they knew that a price decline in gold could result in twice as much underlying asset decline for the GLD.

The summary of this period is shown in Table 25.

47


Table 25 – GLD NAV, Shares Outstanding and Assets Under Management March 31 through December 31, 2013 (212 Trading Days)

Date

Per Share NAV

Shares Outstanding

Total Assets Under Management

3/1/2013

$153.12

416,400,000

$63,758,194,679

12/31/13

$115.87

266,000,000

$30,822,044,650

Change

($37.25)

(150,400,000)

($32,936,150,030)

Percent Change

-24%

-36%

-52%

Chart 3 shows the decline in shares outstanding and the price of the GLD during the 212-day period.

Chart 3 – GLD Shares Outstanding vs. Per Share NAV March 1, 2013 through December 31, 2013 (212 Trading Days)


During the 212-day period, there were over 2.2 billion shares worth $298 billion traded for the GLD. The 51% short sale percentage on all reporting markets equates to approximately 1.1 billion shares sold short at a value of $151 billion.

When the GLD assets under management declined at a significantly different rate than the price of the underlying gold, red flags should have been/were triggered at ETF operators and the GLD auditor that serious fundamental flaws and asset risks in the GLD exist that have not been properly disclosed to regulators and investors.

48


Again, A Period of No Net Creation for the GLD – 2014

On January 2, 2014 there were 264.8 million shares outstanding and on September 3, 2014 (169 trading days later) there were 264.1 million shares outstanding; virtually no net change. The underlying asset value (amount of gold holdings) remained flat at the reduced levels of over $30 billion less than when gold declined by 24% and the GLD gold holdings declined by 52% in the previous discussed period.

During this 169 days, over 1.1 billion GLD shares traded worth $143 billion with sales of the GLD averaging 62% short each day on reporting SROs/exchanges. Using the reporting markets percentage indicates approximately 704 million shares were sold short worth $88 billion, while there was again no net creation of shares. The summary of this period is shown in Table 26.

Table 26 – GLD NAV, Shares Outstanding and Assets Under Management January 2, 2014 through September 3, 2014 (169 Trading Days)

1/2/2014

9/3/2014

Per Share NAV

$118.14

$121.72

Shares Outstanding

264,800,000

264,100,000

Total Assets Under Management

$31,424,456,594

$32,146,670,696

Volume

Value

January 2 - September 3, 2014

Between Dates

Traded Between Dates

Shares Executed (Consolidated Tape)

1,145,395,200

$143,114,573,074

Short Sales (Based on SRO Reporting Markets Percent)

703,746,799

$87,997,877,090

Most investors believe ETFs perform like a type of mutual fund.56 For the GLD, the assets under management are deviating from what would be expected from a ‘mutual fund type investment’. This is caused by the creation/redemption process implemented by the ETF operators and Authorized Participants, but not disclosed to regulators or investors. The data shows assets are not created despite incoming investment and synthetic shares increase the number of actual shares trading in the marketplace, which can exacerbate the downfall of the ETF assets under management during large redemption periods in stressed markets or over longer time periods, creating a slow insidious decline in asset value before the risks in the investment are discovered.

56 In testimony before the Senate Subcommittee on Securities, Insurance and Investment in October 2011, Eileen Rominger, the director of the SEC’s Division of Investment Management explained the SEC’s understanding of physical ETFs; “ETFs offer investors an undivided interest in a pool of securities and other assets.” “Apart from the fact that ETFs trade intraday, most ETFs are similar to mutual funds in that they both translate investor purchases and sales in the fund (and changes in investor sentiment) into purchases and sales of underlying holdings.” Eileen Rominger, Director, Division of Investment Management, Testimony on Market Micro-Structure: An Examination of ETFs, October 19, 2011 http://www.sec.gov/news/testimony/2011/ts101911er.htm

49


GLD Assets vs. Gold in 2015

When gold was at its high of approximately $1,895 per ounce in September 2011, the GLD was priced at $185. At the end of July 2015, the price of the GLD was fluctuating between $103 and $105. In both cases, whether it is the purchase of physical gold or the GLD, the price has declined since September 2011 by approximately the same percentage; 42%.

However, the investor in the GLD also suffered losses from the decline in physical gold assets held by the GLD. In September 2011, at gold’s high price, the GLD held approximately 39.6 million troy ounces of gold. By the end of July 2015, the GLD held approximately 21.8 million troy ounces. Not only did the GLD investor lose at the price level from gold’s decline, but the physical assets held by the GLD were almost cut in half.

This has occurred because the process of creation has not been happening on a regular basis, but redemptions are being executed and the physical gold held by the GLD is being liquidated.

Since the 2008/2009 financial crisis, the GLD and other ETFs have diverged from their expected relationship with their underlying assets. In this case, it has created a lose-lose for GLD investors, along with a potential collapse of GLD held assets. Moreover, the excessive short selling indicates that there is massive over-leveraging of GLD shares sold that do not actually exist (there are multiple owners for each issued share of the GLD).

When these factors are taken into consideration, it appears that the value of the assets held by the GLD are seriously diluted and over-leveraged, which has created a potentially toxic ETF. Again, these facts are not being disclosed by the ETF operators. The various metrics suggest that another $200 drop in the price of gold could cause another run in the GLD, which could leave the GLD with little assets.

State Street has not addressed these changes in the makeup of the GLD. Who would invest in the GLD if the above information was clearly disclosed?

50





According to Wikipedia, "Fisking is a blogosphere slang describing a point-by-point criticism that highlights perceived errors, or disputes the analysis in a statement, article, or essay… adding paragraph-by-paragraph commentary, dissecting and claiming to debunk… assertions and opinions. According to The Guardian, "fisking" has come to denote the practice of "savaging an argument and scattering the tattered remnants to the four corners of the internet".

Now for the fisking. Expect some repetition from me, because the article repeats some of the same mistakes over and over again:


The State Street SPDR Gold Shares ETF (Symbol: GLD)

The State Street SPDR Gold Shares ETF (Symbol: GLD), holds only 1 commodity asset, physical gold. The GLD is a straightforward and simplistic ETF example that demonstrates the supply and demand for shares and values of assets are not showing the expected natural economic relationships between the trading of the ETF and its gold assets.



That's a brain twister of a sentence, but let me make it simple for you. Right off the bat, they are stating that GLD and "gold" are not exhibiting the expected relationship. In reality, however, GLD and "gold" are exhibiting precisely the correct, the stated, the intended and expected relationship, perhaps better than any other ETF in the world, and entirely automatically (without the need for management). That is, GLD is tracking the price of "gold" more closely than any other gold ETF, and perhaps more closely than any other unmanaged ETF. What's wrong, here, is the authors' expectation.

This is actually kind of funny and ironic, because people that don't get (or don't want to get) the coat-check room view, always try to claim that the GLD inventory changes in response to investor interest in "gold", or GLD, or the relative interest between the two, exactly what the authors expect to be the "natural economic relationships between the trading of the ETF and its gold assets," but which they explain, in great detail, is actually not happening.

So, in a way, they're almost inadvertently making the case for "coat-check". Like I said, it's kind of funny when you think about it. :D


The GLD is an ETF that has experienced both periods when shares should have been created but were not and times when massive amounts of shares were redeemed (decreasing GLD assets at more than twice the rate of changes in the price of gold).



So, sometimes GLD's inventory doesn't change in concert with changes in the price of "gold", and sometimes it changes too much. Duh! ;D That's because, A) it's not supposed to change with the price of "gold"—it could theoretically never ever change at all, and still fulfill its objective of tracking the price of "gold", and B) any change in GLD's inventory is a choice made by an AP, not driven by investor interest in GLD, but for some other reason related to the bank's or the LBMA's management of its gold reserves.


Each type of creation/redemption period shows unexpected risk for investors in the GLD. By any measure, the GLD exemplifies asset risk to investors outside of the fundamental movement in the price of gold.



They're going to show us different periods of time when GLD behaved differently than they erroneously think it should have behaved, and even though it did its job perfectly (almost perfectly tracking the price of "gold"), they see risk for the investor. They are, of course, correct, but for entirely the wrong reason. ;D


Because the only underlying asset is gold, new investment in the GLD should result in the ETF purchasing more gold when gold is in favor as an investment. In fact, the GLD for 2 years reviewed did not purchase any significant amounts of gold to match the enormous trading of the GLD by investors.



Here's where they start confusing themselves, thinking of GLD as more like an open-end mutual fund. Ari said it best back in 2011:

"I’ve said it before and I’ll say it again now, the reporters are getting it wrong when they equate outflows of gold from the ETFs with “sour” investor sentiment. What they need to work harder to understand is that these are NOT actively managed funds whose gold inventory is tweaked to ebb and flow based on public sentiment in the shares. Instead, the ETFs are more like a central coat-check room in which the various bullion banks have temporarily hung out their own inventories (i.e., meaning, their unallocated stock which they hold loosely on behalf of their depositors). And whereas the claim tickets (ETF shares) may freely circulate on the open market, any significant outflow of physical inventory is simply and primarily indicative of a bullion bank reclaiming the original inventory based on a heightened need or desire for physical metal in a tightening market — for example, to meet the demands emerging from Asia."

Some mutual funds and actively managed/low volume ETFs do in fact buy and sell their underlying assets in response to investment inflows and outflows. They need to, in order for their NAV to track the price of the underlying asset(s). But GLD trades in volumes so high that only a high speed trading computer could profit from the arbitrage between the difference in prices. It is not actively managed because it tracks the price of "gold" better than any actively managed ETF could, and it does so because of the high volume in trading.

Any significant inflow of investor funds that drives the price of GLD up relative to the price of "gold" will be quickly arbed away by computers, leaving no need to add more of the underlying asset.

Now, I should note here that I suspect they did add inventory due to incoming investor money back in 2016, but I think they did this for a purpose. That purpose was to give someone (BlackRock) a giant positon in GLD (13% of the fund) off-market. You can read more about that in Eight.


The ETF operators know underlying assets are not adjusting to incoming net investments, but have not disclosed the fact that creations are not happening as the SEC expects for many systemically important ETFs.



This is just a ridiculous statement. GLD discloses inventory data every day. The charge of "incoming net investments" is simply wrong, and is based on a misunderstanding of the arbitrage that keeps GLD tracking the price of gold. As you'll see in a moment, they are getting that "net investment" idea from the short sale data they have collected.

For every buyer there is a seller, and vice versa. The buyer or seller could be another investor, a trading computer, or even the bank (the market maker or AP). In the authors' view, every buyer is a real buyer, i.e., "incoming investment." A seller who previously bought a share, on the other hand, is "outgoing investment." Then there are those who sell before buying (sell first, buy later). Those are the "short sellers." So, if all buying is "incoming investment", and selling is, say, half "outgoing investment" and half "short selling," then you might think you have "incoming net investment" at any given moment in time.

But notice that those short sellers have to buy later. So not all buying is "incoming investment." Some is closing out a short position. And as you'll see in a moment, the authors' conclusion of "incoming net investments" is totally unsubstantiated.


Chart 2 illustrates the creation/redemption process along with the GLD price action (changes in investor sentiment for physical gold). The GLD price is reset each day to approximately 10% of the price of an ounce of gold.55 In Chart 2, the daily shares outstanding and closing prices for each month were averaged in order to keep the chart understandable for the long trading period.

55 This price fluctuates at a .5% premium or .5% discount.



In the first sentence, it is entirely unclear what they mean, even though they felt it was important enough to be bolded. Do they mean the GLD price action shows changes in investor sentiment for physical gold? Or do they mean the creation/redemption process relative to the GLD price action shows changes in investor sentiment for physical gold? Do they mean changes in investor sentiment for physical gold relative to paper gold, or relative to other assets? And finally, are they suggesting a direct or inverse correlation in the two? In other words, are they considering GLD to be physical gold?

I point out all these different ways to read that sentence because it's actually not so difficult to be clear about what you mean. Again, Ari said it well back in 2011, share redemptions should be considered to signify an increase in investor preference for physical gold over paper GLD shares (which I think is the opposite of what the authors meant):

"Silly reporters. Instead of calling these “outflows” from the ETFs, it should be called what it is — a redemption of a basket of shares for physical gold by the Authorized Participants (e.g. bullion banks). Such share redemptions would actually be a bullish sign because it entails a reduction in the global supply of paper gold while at the same time signifying a preference by the redeeming party for having the metal over the ETF shares."

In the second sentence, they say that the price is "reset" each day to the price of gold, from which it fluctuates about 0.5% in either direction. The 10% thing simply means that each share represents roughly one-tenth of an ounce. I should mention, btw, that 0.5% is pretty darn good, and thanks to HFT, those fluctuations are often quite a bit tighter than 0.5%, and happen very quickly, as you can see in this chart:


The use of the phrase "reset each day" is a little confusing, perhaps intentionally so. Of course GLD stops trading at the end of each trading day, and starts up again at the beginning of the next. The opening price each day of course reflects any changes to the price of "gold" that happened overnight. Even if the market makers didn't "reset" the price, the market would, so I don't get the point of saying that. It tracks the price of gold, and it doesn't trade 24/7 like its underlying asset, so there are obviously gaps each night. I think using the word "reset" reflects their erroneous view of GLD as some sort of mutual fund or index-type product where the price actually does need to be reset each day by the operator due to some complex mix of underlying assets. In any case, it feels weird to use it for GLD.


Chart 2 – GLD Average Monthly Shares Outstanding vs. Average Monthly Closing Price from November 2010 through September 2014 (985 Trading Days – Almost 4 Years)




The chart above shows the plateau (phase 2 in my chart below) and the downslide (phase 3 in my chart).


The black arrow representing a little more than half of the time of the overall chart (552 trading days), covers the period of the most dramatic difference between the volatility in the price of "gold" (and GLD) and the stability (the plateau) in the inventory of GLD. This is the first of the periods that "exemplifies" "unexpected risk for investors in the GLD" that they want to complain about to the SEC.


Change in GLD Assets Over This 552 Trading Day Period

On April 15, 2013, the closing price of the GLD was $131.31. The last time the GLD was valued at this level was 552 trading days prior, on February 2, 2011 when the closing price was $130.45.



I just want to note that April 15, 2013 was the day "gold" dropped so much that Jim Sinclair emailed me. Also, I was just looking back at emails from that day, and another gold writer named Greg Canavan also emailed me this:

"Hi FOFOA,

Thought you might be interested in this. The way I view it, it confirms your GLD as BB reserve theory, although there’s probably other ways of looking at it too…

Since the start of the year, GLD inventory has lost 204.9 tonnes of gold. In April alone it’s lost 71.12 tonnes.

In the same timeframe, SLV has ADDED 366 tonnes.

I went back to 2008 in the last gold capitulation to see what the story was and from 1 Oct 2008 to 28 Nov 2008 GLD actually added around 3 tonnes.

You probably remember that during the 2008 storm, GOFO went negative briefly.

During the recent sell-off, GOFO has been pretty stable. I expected it to be very, very low…along the lines of 2008 given the price plunge.

So I think ‘maybe, instead of having GOFO plunge and give off the warning signal of tight physical supply/reluctance to lend (maintain the flow), GLD is the source.

Cheers, greg"



For most of the 552 days, the GLD was a relatively ‘hot’ ETF and the data suggests significant incoming investment into the GLD occurred.



By 'hot', they mean the trading volume was high. You can download the CSV (Excel) spreadsheet for GLD at this link and go look at those days. Trading volume is column I:

www.spdrgoldshares.com/assets/dynamic/GLD/GLD_US_archive_EN.csv

Of course "gold" was 'hot' during that period too, just look at the chart. Anything volatile is 'hot'. Just at a glance, it looks like the GLD daily volume stayed above ten million and got as high as tens of millions during that period, and at other times it can be well under ten million. So that's what they mean by 'hot'.


During the 552 trading days from February 2, 2011 through April 15, 2013, the reporting markets data showed more than 1 of every 2 GLD shares sold were the product of a short sale;



Here's comes the short sale pitch. Apparently they have data that shows or implies that more than half of the sales during that period were short sales, i.e., a sale before purchasing. And here's the wind up…


7.38 billion GLD shares were traded worth $1.17 trillion, but by April 15, 2013, the underlying physical gold value held by the ETF had declined by $784 million from the beginning of the period.



Ok, grab your napkin and a pen… 7.38 billion shares traded (by the way, there were only about 400 million shares even in existence, so they're just adding up the daily trading volumes (about 13 million per day) to come up with billions), and half the sellers were short sellers (me: HFT arbs), so I guess half of 7.38 billion shares was "incoming net investment"? Or should we divide by two to be safe? Half of 3.69 billion shares? So that would be 1.845 billion shares that should have been added… wait, let me grab another napkin…

Ok, each share is a tenth of an ounce, so divide by 10… 184.5 million ounces, or 5,738 tonnes should have been added? That would have meant a 550% increase in the size of GLD. Surely that's not what they're suggesting…


For the majority of this time, gold was ‘in-favor’ as an investment moving from $1,300 to $1,800 (in the first 150 days with flat or decreasing GLD shares outstanding). For the next 362 days the average price of gold remained above $1,600. It then took 40 trading days for gold to decline to $1,300 (February 18 through April 15, 2013), with the majority of the price decrease occurring within the final 3 days. Investment in physical gold and the GLD during this period had little economic influence on the GLD’s net purchasing of physical gold.



Just in case the absurd numbers achieved by adding up 552 days of daily volume and short selling wasn't enough to convince you that someone's being screwed by GLD not doing what we mistakenly believe it should be doing, consider that "gold" was 'in favor' (loosely defined as days gold was either going up or above $1,600) more than 90% of that time. Give me a break.

And I also have a bone to pick with the last sentence in that paragraph: "Investment in physical gold and the GLD during this period had little economic influence on the GLD’s net purchasing of physical gold."

GLD doesn't "purchase" physical gold. The process of adding physical gold goes like this. GLD is a Trust, which has allocated and unallocated gold accounts, as well as a dollar account, at HSBC bank in London. One of the APs asks the GLD Trust to create new shares. The minimum that can be created is 100,000 shares, and it must be in multiples of 100,000 shares. 100,000 shares currently represent about 9,507 ounces of physical gold (less than a third of a tonne). That's called a basket.

In order to create a basket (or baskets) of shares, the AP bank must transfer to one of the GLD Trust accounts at HSBC bank either unallocated gold credits, the dollar equivalent of the amount of gold at that time, or it could theoretically have some physical gold delivered to HSBC via armored truck. In any case, the transfer will be converted into unallocated credits in the GLD Trust's unallocated account, and then HSBC will allocate some of its unallocated gold to GLD within three days of the shares being created. Note that even in the unlikely case that physical gold is tendered in exchange for new shares, it won't necessarily be the same bars that get allocated to GLD.

Notice, also, that nowhere in this process is physical gold purchased by the GLD Trust. HSBC and most of the GLD APs are LBMA bullion banks. They don't purchase gold. They are the gold market, and gold flows through them, like art flows through a gallery. The slack in that flow is their reserves (remember, they are banks, with gold denominated books), and the physical that HSBC allocates to the GLD Trust comes out of those reserves. If there aren't enough reserves, I guess they borrow some from a CB storing at the BOE or something. What they don't do is go to the gold market and buy some. They are the gold market. It's a simple concept.


Table 23 shows the asset values and the total volume traded in the GLD during the 552-day period.

Table 23 – Asset Values and Volume Traded in the GLD from February 2, 2011 through April 15, 2013 (Two Years - 552 Trading Days)

Value

Asset Value February 2, 2011

$52,730,385,538

Asset Value April 15, 2013

$51,946,541,141

Highest Asset Value During Period (August 22, 2011)

$77,511,602,759

Lowest Asset Value During Period (April 15, 2013)

$51,946,541,141

Average Asset Value During Period

$66,500,114,530

Value of Shares Traded During Period

$1,170,756,995,770



Oh, here we go again with the vague insinuations of an absurd comparison between a ridiculously large aggregated volume number and the NAV at five different times during the period, which is more than an order of magnitude smaller.


By investing in the GLD, the investor pooled monies into an ETF that did not gain any new net holdings of physical gold (underlying assets of the GLD actually declined by $784 million) and investors only received the change in gold value (a GLD increase of $0.86 per share).



Hahahahahaha!!! "and investors only received the change in gold value" HAHAHAHAHA!!! Stop it, please, you're killing me! :D

Do the authors really believe that investors in GLD should receive more than the change in gold value? Did they even read the prospectus? HAHAHAHA! I wonder if anyone at the SEC laughed at this paragraph? :D

Just for the record, here's what the prospectus says about the objective of GLD:

"The investment objective of the Trust is for the Shares to reflect the performance of the price of gold bullion…"

There's nothing in there about investors gaining more bullion by investing in GLD. And here's what it says about when new shares will be created:

"The Trust creates and redeems the Shares from time to time…"

There's nothing in there about incoming net investment triggering the creation of new shares. There's nothing even about the reason for creations and redemptions, just that they happen "from time to time".


It appears that the pooled investor interests in the GLD were diluted with short sales that created large synthetic unreported short positions (producing excess leverage and undisclosed liabilities). The short sales were greater than half a trillion dollars when the average asset value was only $66.5 billion.



Ok, now we're finally getting to the real meat of their unsubstantiated belief, that short sales over 552 days add up to an absurd short position the way daily volume numbers add up to an absurd volume number. But as I said above, short sales only mean selling before purchasing. There's nothing but the authors' imagination to support such a ludicrous idea.

In fact, there is the opposite. Understand that GLD is almost always trading at either a slight premium or discount to the price of "gold", so it's fair to say that roughly half the time it's a premium, and the other half is a discount. Now, the HFT arbitrage trading computers will buy the lower of either GLD or XAU/USD and sell the other one simultaneously. Later, when the situation is reversed, they will unwind that trade. That means selling the one they bought, and buying back the one they sold short. That may be only minutes or hours later. So, you can imagine that half of the arb volume, whatever portion that is of the overall volume, is short sales that do not add up to a short position over time.


Simply put, despite investments, on February 2, 2011 and two years later on April 15, 2013, the ‘pot of gold’ owned by the GLD was the same pot of gold.



So what? That's the way it works. It reminds me of something I wrote back in 2011 in Who is Draining GLD?:

"Imagine GLD as a big lump of gold just sitting there in Town Square. The price of gold is "discovered" elsewhere and shares in this big lump just trade based on that elsewhere-discovered price."

Lump, pot, same difference. I mean, really, this nonsense is on the SEC website, bolded and underlined no less???


At the end of this 26-month period, an investor in the GLD did not have any more representative ownership of gold assets under management by its ETF operator than it started with and the investor asset value was diluted from high levels of short selling.



A) Why, again, should an investor in GLD have "representative ownership" of any more gold after 26 months than an investor in, say, physical gold itself? B) The dilution from short selling, again, is totally unsubstantiated, contrary to logic regarding the arbs that keep it tracking the price of "gold", and only in the authors' mistaken imagination. And C) does it bug you at all the way they say "the GLD" rather than just "GLD"? I guess that's the proper phrasing, but it sounds silly to me, so you won't see me doing it. ;D


These findings of static assets (inadequate net creations during positive investment periods) and sustained short selling have continued in multiple large ETFs we have reviewed.



Good lord. Think of all the time they must have spent on this.


GLD No Net Creation – Strong Positive Investor Sentiment in 2011

It takes net creation of new shares for the GLD to purchase its’ underlying assets of physical gold.



A) This is basically a tautology. B) As I explained above, GLD doesn't "purchase" its underlying asset. So what they're trying to say is that it takes net creations (more creations than redemptions at a given time) for the GLD inventory to increase. In other words, it takes net growth in GLD for GLD to grow. And C) "it's" is the contraction for "it is" or "it has", and "its" is the possessive, meaning "belonging to it". Its' is never correct. Ever.

Sorry, I couldn't resist. ;D


As shown in Chart 2, from February 1st through September 6, 2011 (151 trading days) the price of the GLD increased from $129.92 to $184.49 or by 42%, while there was essentially no net creation of ETF shares to purchase the underlying physical gold. There were 404 million shares outstanding on February 1st and on September 6th, there were 406.8 million shares outstanding; an increase of less than 1% (virtually no net change in the amount of gold in the pot).



This was the big run-up in gold, to its all-time high. You remember. It went from around $1,330 on February 1st up to $1,895 on September 6.

Yup, that's a 42% rise in the price of "gold" as well, so GLD did a fantastic job of tracking it. And the fact that there was virtually no change in inventory during this remarkable run should put to bed once and for all the idea that the arb is physical. A net total of 5 tonnes was added, actually less than half a percent.

There are various schools of thought on how the mythical physical arb might function, but this period alone should sufficiently debunk all of them. And the only alternative left is that creations and redemptions are a choice made by the banks, not forced upon them by investor sentiment or inflows. And if they're a choice (which they are), then "coat check" is the only view that makes sense.


During the 151-day period, 2.7 billion GLD shares traded worth $412 billion. Reporting markets/SROs data showed on average 52% of the shares were sold short, equating to approximately 1.4 billion shares sold short at a value of $215 billion. The only change in underlying asset value is attributable to the GLD price change (+$54.57 reflecting the price increase of gold), not because of new purchases of gold by the ETF operators/Authorized Participants. This data is shown in Table 24.



Again, for every buyer, there's a seller, even if it's the market-making bank. So, during that whole 42% run-up, there were as many sales as there were purchases. And the authors have data (it's not perfect data, but they believe it is reliable and I'm willing to accept it as reliable) that says roughly half of the sales during that period were short sales. And again, short selling means selling a share you didn't buy first. So why would there be so much short selling during a huge run-up?

Well, obviously there are going to be some gold bear traders calling the top all the way up, and shorting it all the way up. And GLD is an easy way to do that. There's also the arb mechanism I explained earlier. But in both cases, no one is building up a large short position. They are getting in and out quickly, or out and in as the case may be.

The authors' thesis seems to be that the GLD operators and APs should have been buying bullion and creating new shares hand over fist with all the "incoming net investment" money. Yet they weren't, so, the authors' think, someone must have been ripping off the investors. It must be either the GLD operators, the AP banks, or the short sellers diluting the pool with their synthetic supply and enormous short position.

But how can anyone say the GLD investors got ripped off when the price of their shares perfectly tracked the price of "gold", just like it's supposed to? The thing was working like clockwork.


Table 24 – GLD Data February 1, 2011 through September 6, 2011 (151 Trading Days)

2/1/2011

9/6/2011

Per Share NAV

$129.92

$184.49

Shares Outstanding

404,200,000

406,800,000

Total Assets Under Management

$52,513,965,320

$75,048,750,790

Volume

Value

February 1 - September 6, 2011

Between Dates

Traded Between Dates

Shares Executed (Consolidated Tape)

2,683,600,600

$412,372,915,050

Short Sales (Based on U.S. Reporting Exchanges Percent)

1,399,586,669

$214,968,361,274



Here we see that the authors added up all of the short sales during the 2011 run-up to not-so-subtly imply that an enormous short position exists. Can you imagine holding an enormous short position in gold or GLD from Feb. 1 to Sept. 6, 2011? Does that make any sense? Yeah, neither does Table 24.


Essentially, the increase in the total assets under management mirrors only the price increase of gold. Purchasers’ investment into the GLD did not generate net purchases of new gold assets. The data suggests products operating like this provide an efficient way to siphon investors’ monies through short sales while bypassing the purchasing or borrowing of assets. This is akin to selling fictitious financial instruments (thin air), which is illegal in the U.S.



Hahaha!!! See, it was the dirty shorts, I tells ya! :D


Large GLD Redemptions Caused Multiplying Losses of Physical Gold Assets – 2013

From March 1st through December 31, 2013 (212 trading days), the price of the GLD declined by $37.25 or 24%, aligning with the $378 drop in the price of gold. At the same time, there was a large amount of GLD redemptions that decreased the shares outstanding from 416 million to 266 million.



Ok, we're done with the period where GLD inventory changes didn't keep up with changes in the price of "gold", and now we're going to take a close look at a period when changes in GLD's inventory actually outpaced changes in the price. Get the picture? We're done with Phase 2, and we're moving on to Phase 3.



Due to the combined price decrease in the underlying asset (gold) and the net redemptions in the GLD, the value of assets under management dropped by $33 billion or 52% during the period. Simply put, the price of gold and the GLD declined by 24% and the value of the GLD assets fell by 52%.



This is like Captain Obvious stating the obvious, but with a furled brow and a sinister tone, so as to imply that something nefarious must be afoot.



This is very important. The 212-day period of net redemptions shows ETF underlying assets became unhinged from the underlying gold. Assets decreased by more than double the rate of gold's decline in price. Investors wishing to participate in the gold market would not buy the GLD if they knew that a price decline in gold could result in twice as much underlying asset decline for the GLD.

The summary of this period is shown in Table 25.



So they're saying that, during The Big Drain (Phase 3), GLD inventory changes "became unhinged" from POG changes. How is this possible if it was never hinged in the first place? Didn't they just get done showing us how it was unhinged in 2011? But this is very important, in a bold, underlined, and italicized kind of way.

Then they say that "gold" investors wouldn't choose GLD if they knew that the inventory could drain faster than the price of gold was declining. Well, this is one of those statements that I can attack from many different angles, yet may in fact be true, though not for the reasons the authors think.

First, they're wrongly assuming or implying a causal connection in the directionality, if not the speed, of the downward inventory and price action. There are times when the price went one direction and the inventory went the other.

Second, why should GLD investors care, as long as their shares are tracking the price of "gold"? They shouldn't care, so why do the authors say they would care?

It also seems a little silly to say that an investor would choose an asset that was going to decline in price, but not one that would be drained of parts not owned by the investor while it was declining in price. I know that's not exactly what they're saying, but when you invest in something, you expect it to go up, and you don't spend a lot of time thinking about what's going to happen to it while it goes down for 212 days, except maybe how quickly you'll sell the bastard.

Then again, if you read the authors' statement from a coat-check room perspective, it makes a little sense.


Table 25 – GLD NAV, Shares Outstanding and Assets Under Management March 31 through December 31, 2013 (212 Trading Days)

Date

Per Share NAV

Shares Outstanding

Total Assets Under Management

3/1/2013

$153.12

416,400,000

$63,758,194,679

12/31/13

$115.87

266,000,000

$30,822,044,650

Change

($37.25)

(150,400,000)

($32,936,150,030)

Percent Change

-24%

-36%

-52%



Column 1 is the price drop, column 2 is The Big Drain of 2013, and column 3 is the totally meaningless combination of columns 1 and 2, meant to vaguely imply that something nefarious must be afoot. What was afoot was, as Another once said, the Asians like to buy gold "low, lower and cheaper. They never run the price and they never run out of money." So maybe, just maybe, someone was buying physical gold (not GLD) even while the price was declining for 212 days. And maybe, just maybe, reserves in the LBMA were tight because of it, so they had to pull some out of the coat-check room. ;D



Chart 3 shows the decline in shares outstanding and the price of the GLD during the 212-day period.

Chart 3 – GLD Shares Outstanding vs. Per Share NAV March 1, 2013 through December 31, 2013 (212 Trading Days)




Yup, that's "phase 3". It looks much more dramatic and sinister in my chart. ;D



During the 212-day period, there were over 2.2 billion shares worth $298 billion traded for the GLD. The 51% short sale percentage on all reporting markets equates to approximately 1.1 billion shares sold short at a value of $151 billion.



Here we go again with the vague insinuation of something amiss by aggregating and comparing daily volume and short sales. Funny, though, how the short sale percentage is pretty consistent, regardless of what's happening with price or inventory. ;D


When the GLD assets under management declined at a significantly different rate than the price of the underlying gold, red flags should have been/were triggered at ETF operators and the GLD auditor that serious fundamental flaws and asset risks in the GLD exist that have not been properly disclosed to regulators and investors.



Which was it? Red flags should have been triggered? Or red flags were triggered? They were certainly triggered at my blog, but for different reasons. I sometimes wonder how much the "GLD operators" themselves understand that they are using GLD like a coat-check room for their gold reserves. Have you ever had one of those light-bulb moments where you heard an analogy or something that suddenly made you understand yourself, or your own behavior, a little better?

People often assume that if the coat-check view is the best way to understand GLD, then the people running it must be thinking in those same coat-check terms. But I don't think that's quite the way reality works.

The Big Drain of 2013 was remarkable to watch, and I do figure some red flags went up in the LBMA. But not for the reason the authors think, which is that GLD's inventory was improperly tracking the price of "gold". Instead, I figured that someone might have figured out that it not only meant that reserves were scarce, but that the drain itself, published daily and tweeted by BF, was telegraphing that scarcity to the world. That's not something the LBMA would ever want to do.


Again, A Period of No Net Creation for the GLD – 2014

On January 2, 2014 there were 264.8 million shares outstanding and on September 3, 2014 (169 trading days later) there were 264.1 million shares outstanding; virtually no net change. The underlying asset value (amount of gold holdings) remained flat at the reduced levels of over $30 billion less than when gold declined by 24% and the GLD gold holdings declined by 52% in the previous discussed period.



See, now this is interesting with the idea of that red flag in mind. Following The Big Drain, the draining just stopped, and remained basically flat for 169 trading days, before once again resuming.

Now, looking at this from a coat-check perspective, there are one or two different things that might have happened.

1. The slack in the flow might have started accumulating again rather than draining. Remember, you've got inflow and outflow, and the slack in that flow is the LBMA reserves. Very generally, think of the inflow as coming from western mines and recycling, and the outflow would be gold heading east, to China or wherever. So for the slack in the flow to start accumulating again, that would mean a shift in the differential between the inflow and the outflow. It's only a relative change, so we can't know for sure where it came from. But we do know that the POG did rise a little overall during the period in question, and we also know that mining production was a little higher in 2014 than in 2013.

2. In the absence of #1, or possibly even in addition to it, it is possible that someone took some sort of a drastic measure to stop the drain, like turning off the fire alarm to pretend the fire's out, or rigging the thermometer to pretend it's not hot. I don't know what that drastic measure would be, but I can think of a number of possibilities. One possibility could be to just borrow some gold for a while to stop having to drain GLD. If you cut off the fire alarm long enough, perhaps you can kill the fire narrative that's been circulating.


During this 169 days, over 1.1 billion GLD shares traded worth $143 billion with sales of the GLD averaging 62% short each day on reporting SROs/exchanges. Using the reporting markets percentage indicates approximately 704 million shares were sold short worth $88 billion, while there was again no net creation of shares. The summary of this period is shown in Table 26.



Here we go again with the shorts.

If you recall, we discussed shorts recently in How Gold is Different. Here's a little bit of it:

"Now, the ratio of spec shorts to paper claims outstanding… is known as the "short interest". In the real world, if the short interest is too high, the price is thought to be too low. So, contrary to the apparent bias I mentioned above, a short squeeze price spike is expected to be more likely when the short interest is higher than when it is lower. So how high is high?

Well, in the case of Volkswagen, the short interest was 13%. I'm no expert, but from what I've found in my research, 13% is verging on high. Investopedia implies that around 3% is normal, 25% is high, and a 40% short interest makes a company very vulnerable. I found this site called shortsqueeze.com where you can check a stock's short interest. Here is GLD's:

http://shortsqueeze.com/?symbol=GLD


It says the short interest in GLD is 8,853,600 shares. Total shares of GLD right now are 270.7 million, so the short interest is about 3.3%. Pretty normal.

Another metric is the "short interest ratio", which is the short interest as a percentage of the average daily trading volume. This is important because, during a short squeeze, the shorts are all trying to cover (buy) at once, so the short interest ratio is often stated as "days to cover". In GLD, the average daily trading volume is 6,362,100 shares, and the short interest is 8,853,600 shares, so the "days to cover" is 1.4 (8,853,600/6,362,100=1.4). (BTW, I'm only using GLD as a convenient (and perhaps ironic) example of anything and everything except gold. I could have used SLV which has a "days to cover" of 1.6, or even Google (1.9) or Facebook (1.3). Try it yourself!)

My point here is to give you an idea of the range of normal to very high in short interest for everything except gold. Under 5% is normal, and 40% is very high, and in days to cover, somewhere between 1 and 2 appears to be pretty normal."
Now, there's no reason to think the data from shortsqueeze.com is any less reliable than the data used by the authors of this paper. There may be reason to believe it's more reliable, but for now let's assume they're both equally reliable.

The authors are talking about half or more of the sales in the daily trading volume being short sales, 62% during this current period we're talking about, 2014. And I'm no expert in short selling, I simply have the internet and common sense, but it looks to me like the "short interest" on shortsqueeze.com is the existing short position on any given day. Please correct me if I'm wrong. And the short interest ratio or days to cover would be indicative of how long those short sales are, on average, being held.

Note that the short interest in all the things I checked, GLD, SLV, Google and Facebook, was more than the daily trading volume, but less than twice the daily trading volume. So if we assume 50% short sales, and an average short interest ratio of 1.5, then it would seem like the shorts are being closed out after three days on average. Again, please correct me if my impression is wrong.

In any case, it's only a matter of a few days on average, not the hundreds of days the authors are repeatedly adding up as if it implies an extraordinary short position. And GLD's short interest and days to cover appears normal, and comparable to everything else I've checked.

So I had this idea to check GLD on some random days in 2014 on shortsqueeze.com by using the Wayback Machine Internet Archive. Here's the link. Unfortunately, it only has snapshots taken in 2012, 2013 and 2017. So the best I could do was to compare the short interest ratio on random snapshot days during those years. Here's what I found:

1/20/12 – 1.3
4/20/12 – 1.0
10/31/13 – 2.3
1/13/17 – 1.4
6/23/17 – 1.3
8/5/17 – 1.2

As you can see, there's nothing much to see here, so where's the beef?


Table 26 – GLD NAV, Shares Outstanding and Assets Under Management January 2, 2014 through September 3, 2014 (169 Trading Days)

1/2/2014

9/3/2014

Per Share NAV

$118.14

$121.72

Shares Outstanding

264,800,000

264,100,000

Total Assets Under Management

$31,424,456,594

$32,146,670,696

Volume

Value

January 2 - September 3, 2014

Between Dates

Traded Between Dates

Shares Executed (Consolidated Tape)

1,145,395,200

$143,114,573,074

Short Sales (Based on SRO Reporting Markets Percent)

703,746,799

$87,997,877,090



That 703 million number is the sum total of shares sold short over 169 days, presented in the same chart as the total shares outstanding, which was 265 million. So 2½ times the total number of shares in existence were sold short over 169 days, the insinuation being that this must be unusual, extraordinary, and somehow sinister. But if we divide that number by 169 days, it comes out to about 4.1 million short sales per day. This is consistent with the GLD data from shortsqueeze.com, which was comparable to the other names I checked. So again, where's the beef? I.e., you have been served a big nothingburger.



Most investors believe ETFs perform like a type of mutual fund.56

56 In testimony before the Senate Subcommittee on Securities, Insurance and Investment in October 2011, Eileen Rominger, the director of the SEC’s Division of Investment Management explained the SEC’s understanding of physical ETFs; “ETFs offer investors an undivided interest in a pool of securities and other assets.” “Apart from the fact that ETFs trade intraday, most ETFs are similar to mutual funds in that they both translate investor purchases and sales in the fund (and changes in investor sentiment) into purchases and sales of underlying holdings.” Eileen Rominger, Director, Division of Investment Management, Testimony on Market Micro-Structure: An Examination of ETFs, October 19, 2011 http://www.sec.gov/news/testimony/2011/ts101911er.htm



Just because most people believe something or misunderstand how something works, doesn't make it so. Likewise, just because the director of the SEC’s Division of Investment Management misunderstands how something works, and says it in front of Congress, doesn't make it so.


For the GLD, the assets under management are deviating from what would be expected from a ‘mutual fund type investment’.





This is caused by the creation/redemption process implemented by the ETF operators and Authorized Participants, but not disclosed to regulators or investors.



Wrong, they do disclose the process in the prospectus. It says: "The Trust creates and redeems the Shares from time to time…"

If you want more detail on the process than that, you'll have to read my coat-check room posts. ;D


The data shows assets are not created despite incoming investment and synthetic shares increase the number of actual shares trading in the marketplace, which can exacerbate the downfall of the ETF assets under management during large redemption periods in stressed markets or over longer time periods, creating a slow insidious decline in asset value before the risks in the investment are discovered.



Another example of Captain Obvious stating the obvious in an Edwardian way with a run-on sentence, a furled brow and a sinister tone. The only thing I would probably change is that "asset value" should be "total net asset value" or NAV, or just inventory.



GLD Assets vs. Gold in 2015

When gold was at its high of approximately $1,895 per ounce in September 2011, the GLD was priced at $185. At the end of July 2015, the price of the GLD was fluctuating between $103 and $105. In both cases, whether it is the purchase of physical gold or the GLD, the price has declined since September 2011 by approximately the same percentage; 42%.



In other words, GLD did exactly what it was supposed to do, track the price of "gold", through all the various periods this paper detailed.


However, the investor in the GLD also suffered losses from the decline in physical gold assets held by the GLD. In September 2011, at gold’s high price, the GLD held approximately 39.6 million troy ounces of gold. By the end of July 2015, the GLD held approximately 21.8 million troy ounces. Not only did the GLD investor lose at the price level from gold’s decline, but the physical assets held by the GLD were almost cut in half.



Say what? How…? I mean… is this kind of sheer stupidity normal on the SEC website? How did the investor suffer losses from the decline in inventory from 2011 to 2015?



This has occurred because the process of creation has not been happening on a regular basis, but redemptions are being executed and the physical gold held by the GLD is being liquidated.





Since the 2008/2009 financial crisis, the GLD and other ETFs have diverged from their expected relationship with their underlying assets.



Wait, when did the 2008/2009 financial crisis come into this? Phase 2 actually started July 1st, 2010, so try again. And it wasn't consistent across all ETFs, or even similar ETFs. See my email from Greg Canavan above about SLV's opposite behavior. "Expected relationship" only means it diverged from the authors' own misunderstanding.


In this case, it has created a lose-lose for GLD investors, along with a potential collapse of GLD held assets.



Again, where's the second loss? According to the authors, GLD investors lost from the price of gold declining (since 2011), and from the GLD inventory declining. Someone call the cops!


Moreover, the excessive short selling indicates that there is massive over-leveraging of GLD shares sold that do not actually exist (there are multiple owners for each issued share of the GLD).



As I pointed out, it doesn't look like the short selling was excessive, so I'm not sure what they're comparing it to. Plus it was pretty consistent. And while it's true that there is some quantity of synthetic (not actually existing) shares out there, it doesn't appear to be massive. If shortsqueeze.com is correct, it's less than 3% of the total shares existing. So, by my calculations, there are 1.029 owners for each issued share. That's more than one, but not exactly "multiple". ;D


When these factors are taken into consideration, it appears that the value of the assets held by the GLD are seriously diluted and over-leveraged, which has created a potentially toxic ETF.





Again, these facts are not being disclosed by the ETF operators.



Maybe that's because they aren't facts.


The various metrics suggest that another $200 drop in the price of gold could cause another run in the GLD, which could leave the GLD with little assets.



Woohoo! We at the Speakeasy can't wait! :D


State Street has not addressed these changes in the makeup of the GLD.



What changes? These phase changes? You're right, they haven't addressed them, and never will! :D



Who would invest in the GLD if the above information was clearly disclosed?



Good question!


And that concludes your fisking for today. I hope you had as much fun reading it as I had writing it. Fisting is fun! :D

And hopefully you realized early on that this was more about being a new and different way of sharing my perspective on GLD than about debunking a two-year-old, stupid, anonymous submission to the SEC.

While this post isn't a full explanation of the coat-check room view like I've done in the past, I do want to reiterate a few important points about "coat-check". First, it only applies to GLD. It does not apply to other ETFs. It does not apply to other precious metals or commodity ETFs like SLV. And it does not even apply to other gold ETFs. Only GLD.

The reason is twofold: A) Because gold is different from everything else, even silver, and B) because GLD is the primary ETF of the LBMA. It's the same reason GLD is the biggest. It’s not because it was first. And it's not because it's the most popular among gold bugs. It simply has to do with how much gold the LBMA bullion banks had to check into the coat-check room.

Second, while Ari first explained "coat-check" to me back in January of 2011, about 6 months into Phase 2, this basic view of GLD has been around as long as GLD has been around. Here's an old reference I found, written in 2007, halfway through Phase 1:

"GLD is a passive investment trust. That is, it doesn’t do much. The underlying economic model is that of a warehouse in which gold periodically gets deposited in exchange for claim checks, and claim checks periodically get presented for gold."

The point being, this view works for all "phases", and wasn't a result of the dramatic phase transition that occurred on July 1, 2010.

Finally, buickgs455 asked the other day if this view stipulates that all reserves in the LBMA get deposited in GLD. The answer is no. As I explained above, the truth of this view doesn't even require that the participants understand it. They simply act in their own best interest, and they simply know what they see on their screens, without necessarily thinking too deeply about it.

DI had a similar question back in 2013, to which I responded in Another Angle on GLD 2. I had made the comment:

"Some will be getting more allocation and delivery demands while others may have excess reserves they want to check into GLD for whatever reason"

DI asked:

"Isn't the underlying reason to check into GLD an important part of the narrative?

Under the "pooled reserves" theory, would there not be an agreement in place to "check in" whenever a threshold amount of individual BB reserves is crossed, eg, 1 ton?

The underlying rationale for using pooled reserves would be that the system lasts until the last BB defaults, rather than until the first BB defaults, therefore, everyone must use the pool. No?"

My reply was that no, there's no requirement, mandate, agreement or stipulation needed. The banks simply have a natural incentive, the profit motive, to want to check their reserves into the coat check. Bullion banks don't view their gold reserves the same way Freegolders do. To bullion banks, they are a dead asset, a drag on their bottom line, and not even a positional bet on gold, because the bank's position is theoretically neutral or flat.

Banks need a minimal amount of reserves for clearing and redemptions, and they need access to more as needed, but from a money-making perspective, which is what they do, they want to carry as few reserves as possible. Banks don't want reserves, they need them. So as long as they can access them as needed, they prefer to carry as few as they possibly can, and beginning in late 2004, GLD gave them a place to check those reserves in and out as necessary, while someone else (the GLD investors) paid the carrying cost of securely storing those heavy reserves.

Additionally, I never assume that ALL of the reserves in the LBMA are put into GLD. First of all, there's a constant in/out flow, so some amount of gold is going to flow through without staying very long. GLD is like the cash you might park in a money market account or a CD. When it's draining fast, like in 2013, I do suspect that there's very little if any outside of GLD, but when it's staying put or rising, I assume some slack outside of GLD is more or less accumulating.

I was going to give you a list of "coat-check" posts for those who are new to this, but it turns out there really are a lot of them that touch on this subject. So here's a list in no particular order. I'm sure there are more, but this should get anyone started who wants to read more about the coat-check room view:

Coat Check 101
Milamber and Bron Face Off c.2013
Eight! Eight contains a summary of this series of posts which were written as the mystery was unfolding:

Plausible Explanations (for GLD’s Recent Behavior)
GLD Recap
GLD and the BOE
GLD and the WGC

Bullion Banking 201
Does High (Physical) Demand = Low (Paper) Price and Vice Versa?
Superhero
Another Angle on GLD
Another Angle on GLD 2
Open (Window?) Forum


In the beginning, the coat check room is empty. Also in the end.

Sincerely,
FOFOA