The recent decline in gold prices and the drain from physical ETFs have been interpreted by the media as signaling the end of the gold bull market. However, our analysis of the supply and demand dynamics underlying the gold market does not support this thesis.
For example, Non-Western Central Banks have been increasing their holdings of gold at a very rapid pace, going from 6,300 tonnes in Q1 2009 to more than 8,200 tonnes at the end of Q1 2013 (Figure 1a) while physical inventories are declining (Figure 1b) (or being raided, as we argued in the May 2013 Markets at a Glance)1 and physical demand from large (Figure 1c) and small (Figure 1d) scale buyers remains solid.
Source: World Gold Council, Bloomberg, Hong Kong Census and Statistics Department
Average premium calculated as the average premium for the following 1oz. coins, as reported by the Certified Coin Exchange (CCEX): American Eagle, Maple Leaf, Krugerrand, Philharmonic, Panda, Isle of Man and Kangaroo.
In previous articles we have argued that Western Central Banks have been filling the supply gap to satisfy the demand for physical gold.2 As shown in Figure 1a above, the official amount of gold held in the Western Central Banks and international institutions like the IMF has been steadily declining since 2000, only to stabilize at around 23,500 tonnes since 2008. Officially, this gold has primarily been sold by continental European Central Banks under what is known as the “Central Bank Gold Agreements” (CBGA) (also known as the Washington Agreement on Gold).3 Under this agreement (which expires after five years and has been repeatedly renewed since 1999), the “undersigned institutions will not enter the market as sellers, with the exception of “already decided sales” and “The signatories to this agreement have agreed not to expand their gold leasings and their use of gold futures and options over this period”. Sales under the CBGA are shown in Figure 2 below.
Source: World Gold Council
The two points referenced above are particularly interesting because gold leasing (or swaps) has been the primary instrument used by central banks and bullion dealers to suppress the price of gold over time (Alan Greenspan testified, on 24 July 1998, to the House of Representatives that “central banks stand ready to lease gold in increasing quantities should the price rise”).5
It is important to remember that bullion banks (members of the London Bullion Market Association, or LBMA) hold gold in their vaults for their clients.6 Most of those clients, according to the LBMA, deposit their gold (or purchase gold) through an LBMA bank, for example ScotiaMocatta, in what is called an unallocated account. “This is an account where specific bars are not set aside and the customer has a general entitlement to the metal. It is the most convenient, cheapest and most commonly used method of holding metal.”7 However, what the LBMA doesn’t say is that, just like regular fractional banking, the bullion bank does not need to keep the whole value of the gold deposit in gold, but only keeps a fraction of it in its vaults, hoping that not all depositors will request their gold at once. This creates a potential shortage of physical (and an increasing supply of paper) gold and is one reason why bullion banks sometimes need to lease gold from central banks. Leasing gold is analogous to a swap or a collateralized loan, where a Central Bank gets cash in exchange for gold as collateral, and pays an interest rate on the cash loan.
The gold leasing mechanism works in the following way (also shown in Schema 1 below):8
A Central Bank leases its gold to a bullion bank for a pre-specified period (say 1 month). In exchange, the Central Bank receives cash for the value of the gold and has to pay the Gold Forward Offered Rates (GOFO) to the bullion bank. Then, the Central Bank lends the cash on the market and receives LIBOR for 1 month, with net proceeds of LIBOR minus the GOFO, which is called the lease rate. If the lease rate is positive (and it usually is), then it is profitable for the Central Bank to lease its gold. A high lease rate increases the incentive for Central Banks to lease their gold.
The bullion bank, once it receives the gold from the Central Bank, sells it on the gold spot market and collects the cash (depressing the price in the process by increasing supply in the market). For the bullion bank, this transaction is cash flow neutral and pays a carry (the GOFO rate) (the bullion bank can also buy the gold forward one month to make this a risk free transaction, or hope the price of gold stays constant or declines when it’s time to buy it back). Thus, the GOFO rate is a measure of “how much the bullion bank desires physical gold”. If it is small (relative to LIBOR, which implies a large lease rate), the bullion bank wants gold. If it becomes negative, then it means the bullion bank is ready to pay (negative carry) the Central Bank for the privilege to lease its gold (presumably to deliver physical gold to clients that redeem physical gold from their unallocated accounts).
In theory, at the end of the month, the bullion bank buys the gold back from the market and gives it back to the Central Bank. If the bullion bank repurchases the gold from the market, there is no net effect on overall gold supply. We say in theory because, as highlighted above, the language used in the CBGA hints at something else.
It is our hypothesis that, in practice, the bullion banks do not purchase the gold back in the market at the end of the lease to give it back to the Central Banks. Instead, they only roll the transaction over with the Central Bank, resulting in gold IOUs (paper gold, referred to as “gold swapped or on loan” under Central Bank accounting jargon, in other words, a claim on gold that someone else holds) instead of real bullion in the Central Bank’s vaults. This can be seen in Figure 3 below. There is a clear negative relationship between the amount of gold leaving the vaults of the New York Federal Reserve Bank (other Central Bank’s official gold reserves) and the lease rate (how much carry a Central Bank owns by leasing out its gold). To us, this is a clear indication that Central Banks have been leasing out their physical gold against IOUs from their bullion bank partners.
Source: Federal Reserve Bulletin, Foreign Official Assets Held at Federal Reserve Banks, Earmarked Gold & LBMA. The 1-month lease rate is shown as an annual average.
Also, oddly enough, it seems from Figure 3 that the gold bull market started at about the same time (mid-2001/early 2002) as Central Banks and Bullion Banks stopped flooding the market with leased gold.
Another interesting observation is that the timing of official sales by European Central Banks and the International Monetary Fund (IMF) (Figure 2) do not really correspond with the outflows from the NY Fed’s vault (where most of the world’s Central Banks’ gold ex-US is stored; about 30% by our calculations). This is where the “already decided sales” concept referred to earlier comes into play.9
According to the IMF, they disposed of 403.3 tonnes of gold between 2009 and December 2010.10 However, the net outflows of gold out of the NY Fed’s vault were zero for those two years (and the NY Fed is the main vault for the IMF).11 If this rather large quantity of gold did not come from another vault, then it is plausible that it came out of the NY Fed’s vault, which experienced a net drawdown of 408 tonnes in 2007-2008, a full two years ahead of “official schedule”. Given this inconsistency, it is reasonable to believe that the IMF leased its gold reserves (in the manner explained above) to tame the gold market and rescue the bullion dealers, which probably got a lot of physical gold redemption requests they couldn’t meet during the financial crisis. Then, later on, they “sold” their paper gold to the dealers to net the IOUs and settled in cash.
A similar observation can be made of the European Central Banks’ CGBAs, which all happened well after all the gold outflows from the NY Fed’s vaults.
We are of the opinion that this is what an “already decided” sale is: a Central Bank leases gold to a bullion dealer, that dealer sells the gold (or delivers it to a client) but never pays back the Central Bank its physical gold. Then later on, to balance things out, the Central Banks declare official “sales” of gold, but all that changes hands at that point is paper gold and cash, the real gold is long gone.
Lessons for the current market
It is important to remember that the bullion dealers are the same banks that are deemed too big to fail by their respective governments. Thus, it is very unlikely that Central Banks would abstain from intervening à la Greenspan in order to save their bullion bank partners from a “bullion run” (analogous to a bank run). On the contrary, if the bullion dealers get in trouble because their reserves of physical gold are too small to match redemptions of physical (anecdotes) and risk a bullion run, Central Banks will use their firepower and “stand ready to lease gold in increasing quantities should the price rise” (Greenspan, 1998).
The thing is, it might not be that simple anymore… Since the beginning of the financial crisis, we have seen unprecedented demand for physical gold (see Figure 1a-d above) while at the same time, gold miners are shutting down mines and Central Banks have been relatively quiet in terms of official gold sales (Figure 2) (depressing supply). In fact, the announcement by the German Bundesbank (the second largest gold reserve in the world according to IMF data) that they would repatriate their gold from the NY Fed’s vaults can be seen as a sign that European banks are no longer as keen to lease (or swap) out their gold.12 Other very detailed documents released at the same time show that since 2008, the Bundesbank has not made use of gold leases or swaps.13 To us, this signals that Central Banks are less and less willing to participate in the gold leasing scheme.
Still, the price of gold has experienced a large decline over the past few months, only slightly recovering over the past 2 weeks or so. Given the strong physical demand, we think that this decline was engineered by a bullion bank that flooded the COMEX (paper market), only to then redeem physical gold from the various available sources at depressed prices (i.e. ETFs, see our discussion of this topic in the May 2013 Markets at a Glance). To further support our price manipulation hypothesis, we overlay the 1-month GOFO rate with days where the gold price suffered significant declines (more than 3%) in Figure 4. Unless it is the actual price drop that sparks all this increased demand, it seems counterintuitive that the gold price would decline precipitously before large declines in the GOFO rate, which implies increased demand for physical gold from bullion dealers.
It now seems that bullion banks are in desperate need of bullion, as evidenced by the increasingly negative GOFO rates we are seeing (Figure 4 below). Remember that a negative GOFO rate signifies that the bullion banks are ready to pay holders of physical to lease their gold, in this case for a month. Historically, negative GOFO rates have happened in very few occurrences. The last one was in November 2008 at the height of the financial crisis and after which gold rose 156% from through to peak. Before that, we saw negative GOFO rates in March of 2001 (about the start of the bull market) and September of 1999 (announcement of the first CBGA).
In our view, the bullion banks’ fractional gold deposit system is testing its limits. Too much paper gold exists for the amount of physical gold available. Demand from emerging markets, who do not settle for paper gold, has perturbed the status quo. Thus, our recommendation to investors is the following: empty unallocated gold accounts and redeem your gold in physical form (while you still can), invest in allocated, physically backed products (like the Sprott Physical Bullion Trusts) or in those that have access to physical gold in the ground (gold miners).
