The Gold Forward Offered Rate (GOFO) – Fever Chart of the LBMA
21 April 2011 16 Comments
GOFO is the interest we need to pay if we swap gold for US$, i.e. if we lend our gold and borrow currency, in this case US$, for the same fixed period. The LBMA publishes daily GOFO quotes for each of the periods of 1,2,3,6 and 12 months. These data are shown in Figure 1 above.
Whenever GOFO turns sharply lower or even negative, this indicates stress in the London gold market, i.e. that some of the participants are desperate to get their hands on gold on short notice and are prepared to pay a premium for borrowing gold against US$ collateral. GOFO is the (upside down) fever chart of the London gold market. As Figure 1 shows, the London gold market came down with a serious flu on several occasions: January 1993, November 1995, September 1999, May 2001, and November 2008. It caught a milder form of cold in September 1997, November 1997, March 1998 and September 1998.
In this article, we compile various pieces of information, including historical facts, articles in newspapers and market rumours in order to put the evolution of GOFO shown in Figure 1 into a broader context. We focus on November 1997, March 1998, September 1998, September 1999, May 2001, and October 2008. The article will appear in several parts. This is Part I.
The Gold Forward Offered Rate (GOFO)
GOFO is the interest paid if we lend gold and borrow US$ for the same period. If we split this swap into its two components, borrowing US$ and lending gold, we see that GOFO is approximately LIBOR minus the Gold Lease Rate (GLR). This assumes that the two pieces of counter-party risk compensate each other, namely the risk that the gold is not returned and the risk that the currency principal is not paid back. LIBOR is the London Inter Bank Offered Rate, the interest on an unsecured US$ loan, and GLR the Gold Lease Rate, the interest on an unsecured gold loan.
Under normal market conditions, GOFO is positive, i.e. the interest on US$ loans is higher than the GLR. This means that we need to pay interest if we lend gold and borrow US$. The gold basically serves as the collateral for borrowing US$ at a lower interest rate than we would have to pay for an unsecured loan.
Note that the transactions at the LBMA may be in unallocated accounts, i.e. they do not involve the delivery of physical gold, but rather credit or debit in an unallocated account.
Swapping (unallocated) gold for US$ has the following analogue in the purely physcial world of the commodities exchanges, for example at the New York Commodities Exchange (COMEX). The swap is (almost) equivalent to the following transaction in the spot and futures markets. We can synthesize the swap by selling physical gold for US$ in the spot market, i.e. for immediate delivery, and at the same time purchase a futures contract in order to buy back the same amount of gold at a future date. In the meantime, we can use (most of) the US$ proceeds of the gold sale as we please (while we need to post only a small amount of margin in order to hold the futures contract).
Under normal market conditions, GOFO is positive and the price of gold for delivery in the future is higher than the spot price. In the terminology of commodities trading, the gold market is in contango. In the following, we are interested in the situations in which GOFO turned negative and in which the gold market therefore was in backwardation, i.e. the gold for future delivery trades at a discount to the spot price.
Stress in the Gold Market
The following interpretation of gold backwardation rests on a few assumptions.
- The participants in the London bullion market act like banks. They borrow and lend both US$ paper currency and gold, but they are not primarily long or short gold versus US$. They do not primarily speculate and do not engage in proprietary trading. (This is why they are called bullion banks. They are institutions that lend and borrow gold and silver. Some of them have a commercial banking license, but some do not.)
- Their physical gold is fractionally reserved, i.e. they lend a certain multiple of the actual physical reserve by creating ledger entries for gold. These are the unallocated gold accounts.
- Because of the fractional reserve lending, the participants in the bullion market are prone to a classic run on the bank, i.e. a run on the physical bullion reserve. This happens when the depositors who hold gold in unallocated accounts wish to take possession of the physical bullion and request allocation, while the bank cannot recall their gold loans quickly enough in order to get hold of the physical bullion they have lent to somebody else. A previous article contains some sample balance sheets in order to illustrate how this works.
In the case of such a run on the bank, the bullion bank needs to borrow physical gold in the market for the period between the day on which the customer withdraws physical gold and the day on which an outstanding gold loan is repaid in physical gold.
One way of borrowing gold from the market is to swap US$ for gold, i.e. borrow gold and lend US$, in the LBMA forward market (In this swap, the borrower at first receives the gold as a credit to their unallocated account, but they can immediately request allocation, take possession of the physcial and deliver it to the customer who wanted to withdraw physical gold). The interest received for such a swap is GOFO. If, however, the bullion banks need to obtain large volumes of such swaps under pressure, this will drive down GOFO, possibly rendering it negative so that the bullion banks have to pay interest in order to borrow gold against a US$ collateral. This is a plausible explanation for the downward spikes in GOFO seen in Figure 1 above, some of which have sent the gold forward market into backwardation.
Another option is to borrow gold for US$ using the spot and futures markets by purchasing gold in the spot market and at the same time selling a futures contract, for example at the New York Commodities Exchange (COMEX). The signature of bullion banks borrowing gold from the market is then buying pressure in the market for physical gold plus an increasing short position in the futures market, possibly causing backwardation at the COMEX. Gold bugs often interpret this as outright market manipulation, but it may just be some bullion bank borrowing physical gold from the market for a fixed period of time.
Gold Backwardation Periods
Let us now look at some of the historical examples of gold backwardation at the LBMA. Since GOFO is approximately LIBOR minus GLR, the GOFO chart in Figure 1 roughly follows the trend of LIBOR. For example, the increase between 2004 and 2007 and the steep declines in January and July 2008 mainly show the behaviour of LIBOR. In addition to the general tendency of LIBOR, the GOFO chart contains some downward spikes when the GLR sharply increases. The reason for plotting GOFO rather than the GLR is that GOFO and LIBOR are measured while the GLR is derived from these, see a previous article.
1997 and 1998
Although GOFO did not drop into negative territory in 1997 and 1998, the events in fall 1997 and March 1998 can be viewed as a prelude to those of 1999 and provide some context. Figure 2 shows significant drops in short-term GOFO in 1997 and 1998.
What was happening? In order to understand the context, we take a look at the archive of the USA Gold Forum. Although the comprehensive archive does not date back beyond December 1998, some older messages by the user Another and their context have been archived. It is thought that Another was an insider from one of the Europen Central Banks, see, for example, the foreword to the archive. First, take a look at two messages in order to understand the context:
Date: Sat Mar 07 1998 13:08
ANOTHER (THOUGHTS!) ID#60253:
The Management of Gold, A Simple Tool for the 90s
The largest producers of gold were introduced to the use of large scale “forward contracts” by the Bullion Banks. Once the process started, good business required it to expand. Shareholders want maximum profits at all price levels and “forward deals” were good at any price of gold. Once hooked on “hedge profits” during the good times of a high gold price, the mines now “must have at all cost” “forward deals”, just to survive. Some say the mines will not forward sell at these, break even prices. However, the shareholders say it’s better to hedge now, for a lower price will bring doom! With the US$ price of gold holding at just above average break even levels, and the ensuing virtual bankruptcy of several well known companies, it appears that the mine owners are correct.
Understand, that many entities lend gold, but it is the CBs that started and do most of it. Their purpose was to create a “paper gold” market that would allow them to manage the “freely convertible” price of gold. The CB lends the gold to a bank that sells it on the open market. ( Usually, the gold is placed privately as it must go to the correct destination. ) Then the bank holds the money and draws interest as incremental payments are made to the mine for new gold delivered against the contract. Over the long period that a mine takes to produce and repay the gold, this money grows. To grasp the fact that the CBs had a plan, is to know that they lend the gold for only 1% or 2% while the proceeds set in a Bullion Bank and grow with interest for the benefit of the BB and the mine! And further, the lenders allow the return of the gold to be extended out for many years, as in “spot deferred”. The CBs allow public opinion to think of this as “typical government stupid”, it’s not!
Now that the gold price in US$ is around production cost, most mines must use “paper gold” to survive. The gold industry is coming under world bank domination, without signing away any sovereignty! Slowly, the CBs are gaining the ability to manage production and price with this simple tool.
“If they want new mine supply on the market, they roll over the contract to the BB. If they want new supply off the market, they allow the BB to pay for and take delivery of the gold and return it to the CB vault.” “Also, by offering ( or withholding ) vault gold from lease, they affect the lease rate and thereby control private lending as well”
Understand that the second sentence action is used because gold lending is done by many different entities. Many times a mine isn’t even involved. Sometimes, gold isn’t even involved, just paper. But, it’s still based on the gold price! The paper price, that is.
How would it be possible that the central banks (CBs) remote-control the operation of mining companies by leasing gold? In order to understand this, let us look at how central banks manage the supply of credit.
In a free market in absence of central banks, the interest for loans depends on the supply (savers offering to lend money) and on the demand (businesses bidding for loans). The interest rate discovered in the free market is the price of credit for which supply and demand match and at which the market is cleared, i.e. all supply that is offered at or below that price, is taken, and all demand that bids at or above that price, is met. We call this interest rate the market clearing price.
Now the central bank enters the game. They can offer additional credit in different ways.
First, they can target a quantity of credit. This means the central bank offers a fixed amount of credit and tries to maximize their profit, say, by auctioning the credit in order to achieve the highest possible interest rate. This way, the central bank becomes part of the market, and the interest rate discovered in the auction together with all the supply from savers and all the demand from businesses will be a market clearing price. The credit market is efficient.
Second, the central bank can target an interest rate. At least post-Volcker, the central banks have generally targeted interest rates below the market clearing price. At such an interest rate, demand for credit vastly exceeds supply. We say that under these conditions, credit is scarce. This term is not meant to suggest that someone has artificially limited the supply, but rather that the interest rate is so low that not all demand can be met. The central bank can now allot an arbitrary quantity of credit between zero and the natural demand minus the natural supply at that low interest rate. The remainder of the demand is not met. This way, the central bank directly controls the quantity of credit available (In this explanation, we have not mentioned the commercial banks. Usually, the central bank deals with the commercial banks who then multiply the available credit by their fractional reserve lending and deal with the actual borrowers). Note that in this situation, the interest rate is not a market clearing price. Therefore, the credit market is not efficient. (The central bank does not maximize the interest they receive – we cannot speak of maximizing profit here because creating credit does not cost them any `money’. Therefore, somebody else can pick up a free lunch. Prize question: Who is it?)
Of course, the central bank has to pay a price for being able to control the quantity of credit in this way. By setting the interest rate lower than the market clearing price, the credit volume grows faster than it would in an efficient market. Indeed, credit has typically been growing faster than GDP during all of this period. Second, this procedure encourages risk taking and extending loans for projects whose productivity is questionable. Think of all the consumer loans. When theoretical economists will study how the commercial banking system got into the 2007-2008 credit crisis, they will have to consider all these questions.
Bankers and other financiers will always hang themselves if provided with sufficient rope.
Andrew Smithers in Wall Street Revalued
Let us now get back to the gold market.
In the absence of central banks, the free market discovers the interest rate for gold loans, the GLR. If central banks enter the market and offer to lease gold for lease rates substantially lower than the market clearing GLR, they can control the volume of gold loans just in the same fashion as central banks control the volume of currency denominated credit. The central banks did not enter the market for gold swaps directly, but rather via some of the London bullion banks. The intention was apparently that the bullion banks would extend the cheap gold loans to mining companies. This would, indirectly, give the central banks control over the funding available to the relevant mining companies and thereby, indirectly, over their future production. Lending gold to a mining company entails little counter-party risk, simply because the mining company can pay back the loan by forward-selling a part of the gold that would be mined in the future.
There are some caveats, of course. First, the price of gold ought to be high enough to keep the mining companies profitable. A mining company that goes out of business cannot pay back the gold lent to them. Second, just as with currency denominated credit, if the central banks artificially lower the interest on gold loans, this (1) artificially increases the total volume of outstanding gold loans and (2) encourages risk taking. The following message indicates that both of these potential problems had become real.
Date: Sat Mar 07 1998 19:57
ANOTHER (THOUGHTS!) ID#60253:
The US$ price of gold is, now held between several extremes. In Nov. of last year, a understanding was reached that the “paper gold” market was out of control. The private market was using this medium in a way not intended. Gold was to be taken and held in the $320 to $360 range. This was good for all. As the down trend was not seen in price yet, it was known that trading was in place to drop the price. Perhaps, even below The BIS capital level! Several very large physical buys were made in the off market ( see Date: Sat Nov 29 1997 15:53 ANOTHER ( THOUGHTS! ) ID#60253:) as a warning. At that time the CBs started a slowdown in lending and sales. The market came close to a major resolution. In Jan. the BIS was close to a large open move, it would have caused a paper panic.
The user Another writes that the private market was using this medium in a way not intended. Apparently, the bullion banks had extended the cheap gold loans not only to mining companies (as intended) but also to other market participants (not intended). Of course, it was tempting to take the gold borrowed cheaply from the central banks and lend it to third parties at the (higher) market GLR. Technically, the contracts were probably swaps of gold for US$ currency, but if the interest on the currency component between central banks and bullion banks was the same as between bullion banks and third parties, this arbitrage opportunity was nevertheless present. Imagine one of the following happened:
- A bullion bank lends gold to the general market. This can be done by selling gold in the spot market and purchasing a forward contract in order to buy back the gold at some point in the future. In the meantime, the currency proceeds from the initial sale can be invested elsewhere. The gold for US$ swap for which the bullion bank has to pay interest as long as GOFO is positive, is thereby used as a source of cheap funding.
The new risk is that such a swap introduces the counter-party risk that is always present in the gold forward market into the central banks’ gold lease. The risk is that when the forward contract matures, the seller of the forward contract (who is perhaps not a mining company but some hedge fund and who was not required to post physical bullion as a collateral) will have to acquire the gold in the spot market. Even if the lease was only a ledger transaction, i.e. done in unallocated accounts, who knows how high the spot price is when the short seller will have to deliver, and whether they are able to deliver or perhaps default.
- Even worse, a speculator might borrow gold from a bullion bank and sell it in the spot market wihtout purchasing an offsetting forward contract. When the gold loan comes due, they would purchase bullion in the spot market and pay back the loan. This carry trade now has explicit price risk. As soon as the price of gold rises, the carry trader loses money.
Even if the above transactions involve only unallocated accounts (of course, central banks are most reluctant to remove the actual physical bullion from their vaults), another risk is that the third party who purchases the initially unallocated gold that was sold in the spot market, turns around, goes to the bullion bank and asks for allocation. If such moves deplete the physical reserve of the bullion banks, the bullion bank will have to approach the central bank and ask for allocation of the gold that was leased to the bullion bank. The central bank will try to avoid this at any cost.
The following message indicates that in November 1997, somebody indeed purchased (unallocated) gold in the spot market and asked for allocation:
Date: Mon Mar 09 1998 07:55
ANOTHER (THOUGHTS!) ID#60253:
The purchase of large physical stocks of gold in Nov. did send a message to the CBs. They did begin slowdown of sales/easing. The “management tool of gold in the 90s” ( see Date: Sat Mar 07 1998 13:08 ANOTHER ( THOUGHTS! ) ID#60253 ) is now to go into reverse! A large purchase, now, is sending another message, “bring gold back into $320 to $360 US$ range. We should see this in five to ten days. This will be a hard thing, as it may create a crush to cover. […]
This crush to cover was the drop in GOFO in March 1998. Figure 2 (above!) also shows a drop in GOFO in September 1997. The events in September 1997 might have been the reason for Another to enter the public discussion in the Kitco forum in October 1997, but we can only speculate about this.
Let us now turn to September 1998. Figure 3 shows the Gold Forward Offered Rate in 1998.
In September 1998, we see another drop in GOFO, albeit less serious than the one in March.
On September 23, 1998, the Hedge Fund Long Term Capital Management (LTCM) was about to fail. The Federal Reserve Bank of New York organized a bail-out of the fund by its major creditors. Why would the failure of LTCM as a consequence mainly of the debt default by Russia, cause a dent in GOFO? Was perhaps LTCM one of the hedge funds engaged in the gold carry trade? If yes, the closure of that carry trade in a hurry might be what is seen in the GOFO chart.
Some market rumours around the failure of LTCM have been reported by John Embry and Andrew Hepburn on page 29 of their article Not Free Not Fair – Long Term Manipulation of the Gold Price. According to Embry and Hepburn, LTCM’s attorney James G. Rickards maintained in a letter sent to attorneys for the Gold Anti-Trust Action Committee (GATA) that
None of LTCM, LTCP, nor their affiliates, has ever entered into any transaction involving the purchase or sale of gold, including without limitation, spot, forwards, options, futures, loans, borrowings, repurchases, coin or bullion, long or short, physical or derivative or in any other form whatsoever.
Nevertheless, rumours that LTCM was involved in gold carry trades have appeared persistently. According to Embry and Hepburn, in September 1999, TheStreet.com quoted Nesbitt Burns gold analyst Jeff Stanley as saying on a conference call:
We’ve learned Long Term Capital Management is short 400 tons
Anyway, as is evidenced by our GOFO chart, the failure of LTCM did affected the gold market, but it did so substantially less than the other events.
In Part 2 of this article, we will take a look at the events of September 1999 (Washington Agreement) and May 2001 (rumoured speculative attack and long-term reversal of the futures market). In case some of the insiders of the 1990s are reading this article, the author is always grateful for comments, opinions, and further details.